knowledge management – how KM used in your company and how knowledge transmitted
Document 4 page with 4 references. NO TITLE PAGE:
Describe how knowledge management used in your current company. Is there a Chief Information Officer? How is knowledge transmitted? Is knowledge hoarded or openly shared among workers.
Subject: Knowledge Management Systems and Processes
Reference:
1. “Why Knowledge Management?” by Antoine Tawa. link: https://www.youtube.com/watch?v=QWQp1EZP3eU
2. “Knowledge Management: An organization’s weapon of choice” link: https://www.youtube.com/watch?v=zUVZn44WT_c
3. Chapters 1 and 2 from book Knowledge Management Systems and Processes By Irma Becerra-Fernandez, Rajiv Sabherwal 2nd Edition
Knowledge
Management
This page intentionally left blank
Knowledge
Management
Systems and Processes
Second Edition
Irma Becerra-Fernandez and Rajiv Sabherwal
First published 2015
by Routledge
711 Third Avenue, New York, NY 10017
and by Routledge
2 Park Square, Milton Park, Abingdon, Oxon, OX14 4RN
Routledge is an imprint of the Taylor & Francis Group, an informa business
© 2015 Taylor & Francis
authors of this work has been asserted by them in accordance with sections 77
and 78 of the Copyright, Designs and Patents Act 1988.
All rights reserved. No part of this book may be reprinted or reproduced or utilised
in any form or by any electronic, mechanical, or other means, now known or
hereafter invented, including photocopying and recording, or in any information
storage or retrieval system, without permission in writing from the publishers.
Trademark notice: Product or corporate names may be trademarks or registered
to infringe.
Library of Congress Cataloging in Publication Data
Becerra-Fernandez, Irma, 1960–
Knowledge management : systems and processes / by Irma Becerra-Fernandez
and Rajiv Sabherwal. —
Second edition.
pages cm
Includes bibliographical references and index.
1. Knowledge management. 2. Information technology. I. Sabherwal, Rajiv. II.
Title.
HD30.2.B438 2015
658.4’038—dc23
ISBN: 978-0-7656-3915-8 (hbk)
ISBN: 978-1-315-71511-7 (ebk)
v
Contents
Preface and Acknowledgments xi
1. Introducing Knowledge Management 3
What Is Knowledge Management? 4
Forces Driving Knowledge Management 5
Knowledge Management Systems 8
Issues in Knowledge Management 9
Text Overview 10
Summary 12
Review 12
Application Exercises 12
Note 13
References 13
PART I. PRINCIPLES OF KNOWLEDGE MANAGEMENT
2. The Nature of Knowledge 17
What Is Knowledge? 17
Alternative Views of Knowledge 22
Different Types of Knowledge 24
Locations of Knowledge 32
Summary 35
Review 35
Application Exercises 36
References 37
3. Knowledge Management Foundations: Infrastructure, Mechanisms,
and Technologies 39
Knowledge Management 39
Knowledge Management Solutions and Foundations 41
Knowledge Management Infrastructure 43
Knowledge Management Mechanisms 50
Knowledge Management Technologies 50
vi CONTENTS
Management of Knowledge Management Foundations
(Infrastructure, Mechanisms, and Technologies) 53
Summary 54
Review 54
Application Exercises 55
Note 55
References 55
4. Knowledge Management Solutions: Processes and Systems 58
Knowledge Management Processes 58
Knowledge Management Systems 64
Managing Knowledge Management Solutions 69
Alignment Between Knowledge Management and Business
Strategy 70
Summary 70
Review 70
Application Exercises 72
References 72
5. Organizational Impacts of Knowledge Management 74
Impact on People 75
Impact on Processes 78
Impact on Products 82
Impact on Organizational Performance 84
Summary 87
Review 87
Application Exercises 88
Notes 88
References 88
PART II. KNOWLEDGE MANAGEMENT TECHNOLOGIES
AND SYSTEMS
6. Knowledge Application Systems: Systems that Utilize Knowledge 93
Technologies for Applying Knowledge 94
Developing Knowledge Application Systems 99
Types of Knowledge Application Systems 101
Case Studies 104
Limitations of Knowledge Application Systems 122
Summary 122
Review 123
Application Exercises 123
Notes 124
References 124
CONTENTS vii
7. Knowledge Capture Systems: Systems that Preserve and
Formalize Knowledge 127
What Are Knowledge Capture Systems? 127
Knowledge Management Mechanisms for Capturing Tacit Knowledge:
Using Organizational Stories 129
Techniques for Organizing and Using Stories in the Organization 132
Designing the Knowledge Capture System 133
Concept Maps 134
Context-based Reasoning 139
Knowledge Capture Systems Based on Context-based Reasoning 143
Barriers to the Use of Knowledge Capture Systems 145
Research Trends 146
Summary 150
Review 150
Application Exercises 150
Notes 151
References 151
8. Knowledge Sharing Systems: Systems that Organize and
Distribute Knowledge 154
What Are Knowledge Sharing Systems? 155
The Computer as a Medium for Sharing Knowledge 159
Designing the Knowledge Sharing System 160
Barriers to the Use of Knowledge Sharing Systems 161
Specific Types of Knowledge Sharing Systems 163
Lessons Learned Systems 165
Expertise Locator Knowledge Sharing Systems 170
The Role of Ontologies and Knowledge Taxonomies in the
Development of Expertise Locator Systems 171
Case Studies 176
Shortcomings of Knowledge Sharing Systems 187
Knowledge Management Systems that Share Tacit Knowledge 189
Summary 194
Review 194
Application Exercises 194
Notes 195
References 195
9. Knowledge Discovery Systems: Systems that Create Knowledge 198
Mechanisms to Discover Knowledge: Using Socialization to Create
New Tacit Knowledge 199
Technologies to Discover Knowledge: Using Data Mining to Create
New Explicit Knowledge 203
Designing the Knowledge Discovery System 209
Guidelines for Employing Data Mining Techniques 214
viii CONTENTS
Discovering Knowledge on the Web 221
Data Mining and Customer Relationship Management 225
Barriers to the Use of Knowledge Discovery Systems 227
Case Studies 229
Summary 237
Review 237
Application Exercises 237
Notes 238
References 238
PART III. MANAGEMENT AND THE FUTURE OF
KNOWLEDGE MANAGEMENT
10. Emergent Knowledge Management Practices 243
Web 2.0 243
Social Networking 247
Collaborative Content Creation via Wikis, Blogs, Mashups, and
Folksonomies 256
Open Source Development 258
Virtual Worlds 261
The Three Worlds of Information Technology: Does It Really Matter? 264
Summary 266
Review 266
Application Exercises 266
Notes 266
References 266
11. Factors Influencing Knowledge Management 269
A Contingency View of Knowledge Management 269
The Effects of Task Characteristics 272
The Effects of Knowledge Characteristics 274
The Effects of Organizational and Environmental Characteristics 276
Identification of Appropriate Knowledge Management Solutions 279
Illustrative Example 283
Summary 286
Review 287
Application Exercises 287
Note 288
References 288
12. Leadership and Assessment of Knowledge Management 290
Leadership of Knowledge Management 290
Importance of Knowledge Management Assessment 294
Types of Knowledge Management Assessment 295
Assessment of Knowledge Management Solutions 299
CONTENTS ix
Assessment of Knowledge 300
Assessment of Impacts 302
Conclusions About Knowledge Management Assessment 304
Summary 312
Review 311
Application Exercises 312
Notes 312
References 313
13. The Future of Knowledge Management 315
Using Knowledge Management as a Decision-Making Paradigm to
Address Wicked Problems 315
Promoting Knowledge Sharing While Protecting Intellectual Property 316
Involving Internal and External Knowledge Creators 320
Addressing Barriers to Knowledge Sharing and Creation 324
Concluding Remarks 327
Review 328
Application Exercises 328
Note 329
References 329
GLOSSARY 331
ABOUT THE AUTHORS 353
INDEX 355
This page intentionally left blank
xi
Preface
Knowledge Management: Systems and Processes is for students and managers who
seek detailed insights into contemporary knowledge management (KM). It explains
the concepts, theories, and technologies that provide the foundation for KM; the sys-
tems and structures that constitute KM solutions; and the processes for developing,
deploying, and evaluating these KM solutions. We hope this book will help readers
acquire the relevant suite of managerial, technical, and theoretical skills for managing
knowledge in the modern business environment.
The purpose of this book is to provide a thorough and informative perspective on
the emergent practices in knowledge management. Information technology has been,
and will continue to be, an important catalyst of this innovative field. Web-based
technologies including Web 2.0 and Web 3.0, artificial intelligence, expert systems,
analytics, and collaborative technologies continue to support and transform the field
of KM. However, these technologies would not be effective without the day-to-day
social aspects of organizations such as “water-cooler conversations,” brainstorming
retreats, and communities of practice. To further complicate matters, the current busi-
ness environment renders new skills obsolete in years or even months.
Knowledge management is defined in this book as doing what is needed to get the
most out of knowledge resources. KM is an increasingly important discipline that
promotes the discovery, capture, sharing, and application of the firm’s knowledge.
Indeed, we are witnessing a new era with advanced industrial economies being
revolutionized with the advent of the knowledge age and highly skilled knowledge-
based workers replacing industrial workers as the dominant labor group. Although
the benefits of KM may be obvious, it may not necessarily be so obvious to know
how to effectively manage this valuable resource. In this book, the discussion of KM
reflects the intimacy the authors have with this topic from a theoretical as well as a
practical standpoint and through their substantial and diverse experiences.
The book is divided into three parts:
Part I, Principles of Knowledge Management—This part provides a more detailed
discussion of the concepts of knowledge and knowledge management and describes
the key constituents of KM solutions including infrastructure, processes, systems,
tools, and technologies. The four types of KM processes are described and illustrated:
knowledge application, knowledge capture, knowledge sharing, and knowledge
discovery systems. The section also examines and provides examples of the ways in
which KM impacts contemporary organizations.
xii PREFACE
Part II, Knowledge Management Technologies and Systems—This section is
devoted to a discussion of the underlying technologies that enable KM systems
associated with the four types of KM processes. The four different types of KM
systems are described: knowledge discovery systems, knowledge capture systems,
knowledge sharing systems, and knowledge application systems. The mechanisms
and technologies to support these KM systems are discussed, and case studies related
to their implementation are presented.
Part III, Management and the Future of Knowledge Management—Some of the
issues related to management practices and the future of knowledge management are
presented here. The section describes how KM can benefit from emergent practices
and technologies, including social networks, communities of practice, wikis, and
blogs. It also examines the factors that affect KM and identifies the specific effects
of these factors. Moreover, the overall leadership and evaluation of KM are described
here. This section and the book conclude by examining aspects that are likely to be
important in the future of KM, including crowd sourcing or collective intelligence
and concerns related to privacy and confidentiality.
This book may be adapted in several different ways, depending on the course
and the students. It can be used as a one-semester course on KM for graduate MIS
students by covering selected topics from Parts I, II, and III. An instructor teaching a
course for engineering or computer science students may opt to concentrate on KM
technologies and systems by covering Chapters 1, 6, 7, 8, 9, and 10. Alternatively,
if the course is being taught to MBA students, a number of case studies could be as-
signed to complement the discussions presented in the book, and the discussion of
Chapters 6, 7, 8, and 9 could be emphasized less.
To complement the text and enhance the learning and pedagogical experience, we
provide the following support materials through the instructor’s Web site:
1. Solutions to the end-of-chapter problems.
2. PowerPoint slides for each chapter that describe the key concepts explained
in the text.
3. Sample syllabus and sample student projects.
4. List of relevant accompanying case studies.
5. References to KM software providers.
In addition, instructors adopting the book are encouraged to share with the authors
any relevant material that could be included on the Web site to reinforce and enhance
the students’ experience.
ACKNOWLEDGMENTS
We have so many people to acknowledge! First, we want to recognize our families who
were so supportive during the time we spent with our heads buried in our laptops.
We further thank those organizations that provided us with the fertile ground
to develop many of our ideas about KM: NASA-Kennedy Space Center, Goddard
Space Flight Center, Ames Research Center, Navy Center for Advanced Research in
PREFACE xiii
Artificial Intelligence, and the Institute for Human and Machine Cognition, among
others. We especially thank the individuals at these organizations who made it pos-
sible for us to formalize some of the concepts and techniques presented in this book.
We also thank all the authors that individually contributed to the many vignettes and
case studies presented throughout.
We also thank our administrators, who supported this effort, including President
Mark Rosenberg and Provost Doug Wartzok at Florida International University and
Dean Eli Jones at the University of Arkansas. Our sincere thanks are also directed
to Avelino Gonzalez of the University of Central Florida, who coauthored an earlier
edition of this book and selflessly contributed to some of the material contained in
this book.
We gratefully acknowledge the contributions of the students who collaborated in
the development of some of the KM systems and the material described here. We
are also grateful to Sayed A. Maleknia from University of Arkansas and John Glynn,
Soundarya Soundararajan, and Mouna Yerra from Florida International University
for their editorial assistance with the book.
Finally, we are deeply indebted to many individuals at M.E. Sharpe, Inc., who
enabled us to publish this book, especially the two individuals with whom we have
directly worked: our editor, Harry Briggs, and associate editor, Elizabeth Parker.
This page intentionally left blank
Knowledge
Management
This page intentionally left blank
3
1
Introducing Knowledge
Management
The scientific endeavor that culminated on July 20, 1969, with the first American
walking on the moon is considered one of the most significant accomplishments
in the history of humankind. What is especially noteworthy about this undertaking
is that when President John F. Kennedy issued the promise in 1961 that the United
States would land a man on the Moon and return him safely to Earth before the
end of that decade, most of the scientific and technological knowledge required
to take this “one small step for man, one giant leap for mankind” did not exist.
The necessary science and technology knowledge had to be discovered and de-
veloped in order to accomplish this extraordinary task. However, many of those
technological advances now have permanent presence in the landscape of our
lives, from cordless tools to cellular phones. These first missions to space carried
less computer power on board than what some of us typically lug around airports
on our portable computers. The computers on board Apollo 11, considered “state-
of-the-art” in the 1960s, had 4 KB of RAM, no disk drive, and a total of 74 KB of
auxiliary memory! From the knowledge management (KM) perspective, how did
they manage the extraordinary quantities of knowledge that had to be developed in
order to accomplish the task? The required knowledge about space travel, rocketry,
aerodynamics, systems, communications, biology, and many other disciplines had
to be developed and validated prior to being used in the space mission. From the
knowledge creation perspective, this was an extraordinarily successful endeavor.
On the other hand, a closer look reveals that attempts to elicit and capture the
knowledge resulting from these efforts may have been largely unsuccessful, and
some studies even suggest that NASA may have actually lost that knowledge.
In fact, in the words of Sylvia Fries, who was NASA’s chief historian between
1983 and 1990 and who interviewed 51 NASA engineers who had worked on the
Apollo program:
The 20th anniversary of the landing of an American on the surface of the Moon occasioned
many bittersweet reflections. Sweet was the celebration of the historic event itself. . . .
Bitter, for those same enthusiasts, was the knowledge that during the twenty intervening
years much of the national consensus that launched this country on its first lunar adventure
had evaporated . . . a generation of men and women who had defined their lives to a large
extent in terms of this nation’s epochal departure from Earth’s surface was taking its leave
of the program they had built. (Fries 1992)
4 CHAPTER 1
In this book, we hope to impart what we know about the important field of knowl-
edge management—what it is and how to implement it successfully with the tools
provided by the technological advances of our times. The book presents a balanced
discussion between theory and application of knowledge management to organizations.
The reader will find an overview of knowledge management theory and implementa-
tion, with a special emphasis on the technologies that underpin knowledge management
and how to successfully integrate those technologies. The book includes implementa-
tion details about both knowledge management mechanisms and technologies.
In this chapter, we first discuss what knowledge management is and what the forces
are that drive it. We also discuss organizational issues related to knowledge manage-
ment. Specifically, we introduce knowledge management systems and their roles in
the organization. Finally, we discuss how the rest of the book is organized.
WHAT IS KNOWLEDGE MANAGEMENT?
Knowledge management (KM) may simply be defined as doing what is needed to
get the most out of knowledge resources. Although KM can be applied to individuals,
it has recently attracted the attention of organizations. KM is viewed as an increas-
ingly important discipline that promotes the creation, sharing, and leveraging of the
corporation’s knowledge. Peter Drucker (1994), whom many consider the father of
KM, best defines the need for it:
Knowledge has become the key resource, for a nation’s military strength as well as for
its economic strength . . . is fundamentally different from the traditional key resources of
the economist—land, labor, and even capital . . . we need systematic work on the quality
of knowledge and the productivity of knowledge . . . the performance capacity, if not the
survival, of any organization in the knowledge society will come increasingly to depend on
those two factors. (pp. 66–69)
Thus, it can be argued that the most vital resource of today’s enterprise is the col-
lective knowledge residing in the minds of an organization’s employees, customers,
and vendors. Learning how to manage organizational knowledge has many benefits,
some of which are readily apparent, others not. These benefits may include leveraging
core business competencies, accelerating innovation and time-to-market, empower-
ing employees, innovating and delivering high-quality products, improving cycle
times and decision-making, strengthening organizational commitment, and building
sustainable competitive advantage (Davenport and Prusak 1998). In short, they make
the organization better suited to compete successfully in a much more demanding
environment. Organizations are increasingly valued for their intellectual capital. An
example of this fact is the widening gap between corporate balance sheets and inves-
tors’ estimation of corporate worth. It is said that knowledge-intensive companies
around the world are valued at three to eight times their financial capital. Consider
for example Microsoft Corporation, the highest-valued company in the world, with
a market capitalization that was estimated at around $302 billion as of January 2014.
Clearly, this figure represents more than Microsoft’s net worth in buildings, computers,
INTRODUCING KNOWLEDGE MANAGEMENT 5
and other physical assets. Microsoft’s valuation also represents an estimation of its
intellectual assets. This includes structural capital in the form of copyrights, customer
databases, and business-process software. Added to that is human capital in the form
of the knowledge that resides in the minds of all of Microsoft’s software developers,
researchers, academic collaborators, and business managers.
In general, KM focuses on organizing and making available important knowledge,
wherever and whenever it is needed. The traditional emphasis in KM has been on
knowledge that is recognized and already articulated in some form. This includes
knowledge about processes, procedures, intellectual property, documented best prac-
tices, forecasts, lessons learned, and solutions to recurring problems. Increasingly,
KM has also focused on managing important knowledge that may reside solely in
the minds of organizations’ experts.
Consider, for example, the knowledge of commercial pilots. Not only they are
expected to ensure the safety of passengers, but also keep their flights on time under
various weather conditions. They need to discover and establish the relevance of
all available information related to problems of flight, diagnose problems, identify
alternative action, and obtain the risk associated with each alternative within the
available time. The number of flight hours and years of flying experience have been
considered as indicators of a pilot’s level of expertise. This level of knowledge has
been obtained through many years of experience and successful decisions. With re-
tirement looming, how can an airline organization elicit and catalog this knowledge
so that new generations may benefit?
KM is also related to the concept of intellectual capital, which is considered by
many as the most valuable enterprise resource. An organization’s intellectual capital
refers to the sum of all its knowledge resources, which exist in aspects within or
outside the organization (Nahapiet and Ghoshal 1998). There are three types of intel-
lectual capital: human capital, or the knowledge, skills, and capabilities possessed
by individual employees; organizational capital, or the institutionalized knowledge
and codified experience residing in databases, manuals, culture, systems, structures,
and processes; and social capital, or the knowledge embedded in relationships and
interactions among individuals (Subramaniam and Youndt 2005). Recent study shows
that utilizing intellectual capital and knowledge management capabilities would lead
to innovation and firms’ performance improvement (Hsu and Sabherwal 2011).
FORCES DRIVING KNOWLEDGE MANAGEMENT
Today, organizations rely on their decision makers to make “mission critical”
decisions based on inputs from multiple domains. The ideal decisionmaker pos-
sesses a profound understanding of specific domains that influence the decision-
making process, coupled with the experience that allows her to act quickly and
decisively on the information. This profile of the ideal decisionmaker usually
corresponds to someone who has lengthy experience and insights gained from
years of observation. Although this profile does not mark a significant departure
from the past, the following four underlying trends are increasing the stakes in
the decision-making scenario:
6 CHAPTER 1
1. INCREASING DOMAIN COMPLEXITY
The complexity of the underlying knowledge domains is increasing. As a direct conse-
quence, the complexity of the knowledge required to complete a specific business pro-
cess task has increased as well. Intricacy of internal and external processes, increased
competition, and the rapid advancement of technology all contribute to increasing
domain complexity. For example, new product development no longer requires only
brainstorming sessions by the freethinking product designers of the organization,
but instead it requires the partnership of interorganizational teams representing vari-
ous functional subunits—from finance to marketing to engineering. Thus, we see an
increased emphasis from professional recruiters around the world seeking new job
applicants who not only possess excellent educational and professional qualifications,
but who also have outstanding communication and team-collaboration skills. These
skills will enable them to share their knowledge for the benefit of the organization.
2. ACCELERATING MARKET VOLATILITY
The pace of change, or volatility, within each market domain has increased rapidly in the
past decade. For example, market and environmental influences can result in overnight
changes in an organization. Corporate announcements of a missed financial quarterly
target could send a company’s capitalization, and perhaps that of a whole industry, in a
downward spiral. Stock prices on Wall Street have become increasingly volatile in the
past few years resulting in the phenomenon of day trading, where many nonfinancial
professionals make a living from taking advantage of the steep market fluctuations.
3. INTENSIFIED SPEED OF RESPONSIVENESS
The time required to take action based upon subtle changes within and across domains
is decreasing. The rapid advance in technology continually changes the decision-
making landscape, making it imperative that decisions be made and implemented
quickly, lest the window of opportunity closes. For example, in the past, the sales
process incorporated ample processing time, thus allowing the stakeholders a “comfort
zone” in the decision-making process. Typically in response to a customer request,
the sales representative would return to the office, discuss the opportunity with his
manager, draft a proposal, and mail the proposal to the client, who would then ac-
cept or reject the offer. The time required by the process would essentially provide
the stakeholders sufficient time to ponder the most adequate solution at each of the
decision points. Contrast yesterday’s sale process with today’s, like for example the
process required by many online bidding marketplaces thriving on the Web. Consider
the dilemma faced by a hotel manager that participates in an Internet auctioning mar-
ket of hotel rooms: “Should I book a $200 room for the bid offer of $80 and fill the
room or risk not accepting the bid hoping to get a walk-in customer that will pay the
$200?” Confronted with a decision to fill a room at a lower rate than what the hotel
typically advertises poses an important decision that the hotel manager must make
within minutes of a bid offer.
INTRODUCING KNOWLEDGE MANAGEMENT 7
4. EMPLOYEE TURNOVER
Organizations continue to face employee turnover due to voluntary (i.e., decided by
the employee, for example, due to opportunities for career advancement) as well as
involuntary (i.e., for reasons beyond the employee’s control, such as health-related
problems and termination of employment by the employer). Employee turnover is
especially important in tough economic conditions such as those faced in the 2008 to
2009 period, when several large companies laid off large numbers of employees. Such
employee turnover inevitably leads to the organization losing some of the knowledge
possessed by the departing individuals. Moreover, in some cases these individuals
might have knowledge that would be valuable to competitors. According to Kenny
(2007), “As staff leave, retraining is necessary. This strains company resources and
hinders growth. Replacing a full-time, private-sector worker costs, at a bare mini-
mum, 25 percent of his or her total annual compensation, estimates the Employment
Policy Foundation. Productivity nosedives, ultimately cutting into profitability.” In
case of turnover, new employees must be hired and trained. In addition to the cost
of training, there is considerable time required for a new employee to be effectively
productive.
So, what does this mean? Faced with increased complexity, market volatility, accel-
erated responsiveness, and employee turnover, today’s manager feels less adequate to
make the difficult decisions faced each day. In the decision-making scenario described
above, it is evident that knowledge can greatly assist the decisionmaker. In the past,
this knowledge resided mostly in the decisionmaker. The complications seen above
indicate that in modern organizations, the knowledge necessary to make good deci-
sions cannot possibly all reside with the decision maker, hence the need to provide
her with the requisite knowledge for making correct, timely decisions.
Perhaps nothing has made more evident the need for KM than the corporate down-
sizing trend at public and private organizations that marked the re-engineering era
of the 1990s, a well-known feature of the economic landscape of the late twentieth
century. The dominant driver of downsizing in most organizations is well understood:
Rapidly reduce costs in order to survive against competitors. Clearly, a negative side
effect of downsizing is the dissipation of the knowledge resources, resulting in de-
vitalized organizations. Some of the symptoms of such organizations are: decreased
morale, reduced commitment, inferior quality, lack of teamwork, lower productiv-
ity, and loss of innovative ability (Eisenberg 1997). The fact is, many individuals
who were laid off as a result of downsizing had performed significant tasks and had
acquired considerable and valuable skills over the years. Many companies are typi-
cally not prepared for downsizing, and few take any steps to prevent the escape of
knowledge that usually follows. To minimize the impact of downsizing, organizations
should first identify what skills and information resources will be needed to meet
mission-critical objectives. Therefore, effective methodologies, including tools and
techniques to capture vital knowledge, are essential for an organization to maintain
its competitive edge.
KM is important for organizations that continually face downsizing or a high
turnover percentage due to the nature of the industry. It is also important for all
8 CHAPTER 1
organizations since today’s decisionmaker faces the pressure to make better and
faster decisions in an environment characterized by a high domain complexity and
market volatility, even though she may in fact lack the experience typically expected
from a decisionmaker, and even though the outcome of those decisions could have
a considerable impact on the organization. In short, KM is important for everybody.
Box 1.1 illustrates this fact.
KNOWLEDGE MANAGEMENT SYSTEMS
Rapid changes in the field of KM have to a great extent resulted from the dramatic
progress we have witnessed in the field of information technology (IT). Information
technology facilitates sharing as well as accelerated growth of knowledge. IT allows
the movement of information at increasing speeds and efficiencies. For example, com-
puters capture data from measurements of natural phenomena, and then quickly ma-
nipulate the data to better understand the phenomena it represents. Increased computer
power at lower prices enables the measurement of increasingly complex processes,
which we possibly could only imagine before. According to Bradley (1997):
Today, knowledge is accumulating at an ever-increasing rate. It is estimated that knowledge
is currently doubling every 18 months and, of course, the pace is increasing. . . . Technology
facilitates the speed at which knowledge and ideas proliferate. (p. 54)
Thus, IT has provided the major impetus for enabling the implementation of KM
applications. Moreover, as learning has accrued over time in the area of social and
structural mechanisms, such as mentoring and retreats that enable effective knowledge
sharing, it has become possible to develop KM applications that best leverage these
improved mechanisms by deploying sophisticated technologies.
In this book, we therefore place significant focus on the applications that result
from the use of the latest technologies used to support KM mechanisms. Knowledge
management mechanisms are organizational or structural means used to promote KM.
The use of leading-edge information technologies (e.g., Web-based conferencing) to
support KM mechanisms in ways not earlier possible (e.g., interactive conversations
along with instantaneous exchange of voluminous documents among individuals
located at remote locations) enables dramatic improvement in KM. We call the ap-
plications resulting from such synergy between the latest technologies and social/
structural mechanisms knowledge management systems, as described in Chapters
6 through 9 of this book. Knowledge management systems utilize a variety of KM
mechanisms and technologies to support the knowledge management processes. Based
on observations on the KM systems implementations under way at many organiza-
tions, a framework emerges for classification of KM systems as:
1. Knowledge Application Systems (discussed in Chapter 6)
2. Knowledge Capture Systems (discussed in Chapter 7)
3. Knowledge Sharing Systems (discussed in Chapter 8)
4. Knowledge Discovery Systems (discussed in Chapter 9)
INTRODUCING KNOWLEDGE MANAGEMENT 9
Artificial intelligence and machine-learning technologies play an important role in
the processes of knowledge discovery, capture, sharing, and application, enabling the
development of KM systems. We provide a short introduction to these technologies
in each of these chapters. Because KM systems provide access to explicit company
knowledge, it is easy to learn from previous experiences. Experience management
is another recent term also related to knowledge management. Basically, experience
develops over time to coalesce into more general experience, which then combines
into general knowledge. Experiences captured over time can be managed by the use
of technology. We will discuss how intelligent technologies are used to manage ex-
periences as well as create new knowledge.
ISSUES IN KNOWLEDGE MANAGEMENT
In practice, given the uncertainty in today’s business environments and the reality of con-
tinuing layoffs, what could make employees feel compelled to participate in knowledge
management initiatives? Although many attempts have been made to launch KM initia-
tives, including the design and implementation of KM systems, not all KM implementa-
tions have been successful. In fact many KM systems implementations, for example of
lessons learned systems (discussed in Chapter 8), have fallen short of their promise. Many
KM systems implemented at organizations have failed to enable knowledge workers to
share their knowledge for the benefit of the organization. The case in point is that effective
KM is not about making a choice between “software vs. wetware, classroom vs. hands-
on, formal vs, informal, technical vs. social” (Stewart 2002). Effective KM uses all the
options available to motivated employees in order to put knowledge to work. Effective
KM depends on recognizing that all of these options basically need each other.
Box 1.1
Is Knowledge Management for Everybody?
John Smith owns an independent auto repair shop in Stillwater, Oklahoma, which he estab-
lished in 1985. Prior to opening his own shop, he had been repairing foreign cars as a mechanic
for the local Toyota dealership. In these days of increasing complexity in automobiles, he had to
learn about such new technologies as fuel injection, computer-controlled ignition, and multi-
valve and turbocharged engines. This has not been easy, but he managed to do it, and at the
same time created a successful business, one with an outstanding reputation. As his business
grew, he had to hire mechanics to help him with the workload. At first, training them was easy
since cars were simple. That has radically changed in the last ten years. He now finds himself
spending more time training and correcting the work of his mechanics instead of working on
cars himself, which is what he truly enjoys. To further complicate matters, his mechanics are so
well-trained that the local Toyota dealership is hiring them away from him for significant salary
increases. Being a small business he cannot afford to compete with them, so he finds himself
doing more and more training and correcting all the time. The turnover has now begun to affect
the quality of the work he turns over to his customers, increasing complaints and damaging
his hard-earned reputation. Basically, he has a knowledge problem. He has the knowledge
and needs to capture it in a way that it is easy to disseminate to his mechanics. He must find a
way to manage this knowledge in order to survive. How successful he is will dictate his future
survival in this business.
10 CHAPTER 1
One of the primary differences between traditional information systems and KM
systems is the active role that users of KM systems play on building the content of
such systems. Users of traditional information systems are typically not required to
actively contribute to building the content of such systems, an effort typically delegated
to the MIS department or to information systems consultants. Therefore, traditional
IS research has concentrated much of its efforts in understanding the factors leading
users to accepting, and thereby using, IT.1 As we will see later in Chapter 8, users of
lessons learned systems will not only utilize the system to find a lesson applicable to
a problem at hand but will typically also contribute lessons to the system database.
As a result, the successful implementation of KM systems requires that its users not
only effectively “use” such systems as in traditional information systems but that in
fact that they also “contribute” to the knowledge base of such systems. Therefore,
seeking to understand the factors that lead to the successful implementation of KM
systems is an important area of research that is still in its infancy.
Whereas technology has provided the impetus for managing knowledge, we now
know that effective KM initiatives are not only limited to a technological solu-
tion. An old adage states that effective KM is 80 percent related to organizational
culture and human factors and 20 percent related to technology. This means that
there is an important human component in KM. This finding addresses the fact that
knowledge is first created in the people’s minds. KM practices must first identify
ways to encourage and stimulate the ability of employees to develop new knowl-
edge. Second, KM methodologies and technologies must enable effective ways to
elicit, represent, organize, reuse, and renew this knowledge. Third, KM should not
distance itself from the knowledge owners but instead celebrate and recognize their
position as experts in the organization. This, in effect, is the essence of knowledge
management. More about the controversies surrounding KM will be presented in
Chapters 3, 5, and 13.
TEXT OVERVIEW
PART I. PRINCIPLES OF KNOWLEDGE MANAGEMENT
This section of the book includes the overview of knowledge management that we have
presented in this chapter, including the role that IT plays in KM and the relevance of
KM to modern organizations. Chapter 2 discusses the concept of knowledge in greater
detail and distinguishes it from data and information, summarizes the perspectives
commonly used to view knowledge, describes the ways of classifying knowledge,
and identifies some key characteristics of knowledge. Chapter 3 explains in greater
detail the concept of knowledge management. It also describes knowledge manage-
ment foundations, which are the broad organizational aspects that support KM in the
long-term and includes KM infrastructure, KM mechanisms, and KM technologies.
KM foundations support KM solutions. Chapter 4 describes and illustrates KM so-
lutions, which include two components: KM processes and KM systems. Chapter 5
describes the variety of ways in which KM can affect individuals and various aspects
of organizations.
INTRODUCING KNOWLEDGE MANAGEMENT 11
PART II: KNOWLEDGE MANAGEMENT TECHNOLOGIES AND SYSTEMS
This section of the book is devoted to a discussion of the underlying technologies
that enable the creation of knowledge management systems. Chapter 6 introduces
the reader to artificial intelligence (AI), its historical perspective, its relationship with
knowledge, and why it is an important aspect of knowledge management. This chap-
ter then discusses knowledge application systems, which refer to systems that utilize
knowledge and summarize the most relevant intelligent technologies that underpin
them, specifically rule-based expert systems and case-based reasoning. Case studies
of knowledge application systems are discussed. In Chapter 7 we introduce the reader
to knowledge capture systems, which refer to systems that elicit and preserve the
knowledge of experts so that it can be shared with others. Issues related to how to
design the knowledge capture system, including the use of intelligent technologies,
are discussed. In particular the role of RFID technologies in knowledge capture is
presented. Specific examples of knowledge capture systems are discussed. The chap-
ter also includes a discussion on mechanisms for knowledge capture and the use of
storytelling in organizations, and it concludes with a short discussion on research
trends on knowledge capture systems. In Chapter 8 we describe knowledge sharing
systems, which refer to systems that organize and distribute knowledge and comprise
the majority of the KM systems currently in place. This chapter also discusses the
Internet, the World Wide Web, and how they are used to facilitate communications.
Search techniques used in Web-based searches are also discussed. Design consider-
ations and special types of knowledge sharing systems are covered: lessons learned
systems and expertise locator systems. Case studies of knowledge sharing systems are
discussed based on the experience gained from their development. Finally, in Chapter
9 we introduce knowledge discovery systems, systems and technologies that create
knowledge. The chapter presents a description of knowledge discovery in databases
and data mining (DM), including both mechanisms and technologies to support the
discovery of knowledge. The material covers design considerations and the CRISP-
DM process. Two very relevant topics, DM and its relationship to discovering knowl-
edge on the Web and to customer resource management (CRM), are also presented
including the importance of “knowing” about your customer. Barriers to the use of
knowledge discovery are discussed. Case studies of knowledge discovery systems
are also presented. The chapter includes a discussion on mechanisms for knowledge
discovery and the use of socialization to catalyze innovation in organizations.
PART III: MANAGEMENT AND THE FUTURE OF KNOWLEDGE MANAGEMENT
This section of the book presents some of the issues related to management practices
and the future of knowledge management. Chapter 10 presents emergent KM practices
including a discussion of social networks and communities of practice, how they fa-
cilitate knowledge sharing, and how they benefit from communication technologies.
This chapter also incorporates a discussion of such emergent technologies as wikis,
blogs, and open source development and examines how they enable KM. Chapter 11
describes some of the factors influencing KM, including a discussion of the impact of
12 CHAPTER 1
the type of knowledge, the business strategy, and the industry environment on KM.
It also describes a methodology to prioritize implementation of KM solutions based
on knowledge, organizational, and industry characteristics. Chapter 12 presents a
mechanism for the evaluation and management of KM solutions in an organization.
It describes the reasons why such an assessment is needed as well as alternative ap-
proaches to conducting the evaluation. Finally, it discusses some overall approaches for
managing KM. Finally, Chapter 13 presents some issues on organizational leadership
and the future of KM. As KM becomes widely accepted in corporate organizations, it
will increasingly become critical for corporate managers to supply adequate leader-
ship for it as well as important safeguards for insuring the security and adequate use
of this knowledge. Also in this chapter, we present a discussion on the future of KM.
In the future, knowledge management systems are expected to help decisionmakers
make more humane decisions and enable them to deal with “wicked,” one-of-a-kind
problems. We anticipate a future where people and advanced technology will continue
to work together, enabling knowledge integration across diverse domains and with
considerably higher payoffs.
SUMMARY
In this chapter, you have learned about the following knowledge management issues
as they relate to the learning objectives:
1. A description of KM ranging from the system perspective to the organizational
perspective.
2. A discussion of the relevance of KM in today’s dynamic environments that
are augmented with increasing technological complexity.
3. Benefits and considerations about KM are presented, including an overview
of the nature of the KM projects currently in progress at public and private
organizations around the world.
4. Finally, IT plays an important role in KM. The enabling role of IT is discussed,
but the old adage of “KM is 80 percent organizational, and 20 percent about
IT” still holds today.
REVIEW
1. Describe knowledge management.
2. Discuss the forces driving knowledge management.
3. What are knowledge management systems? Enumerate the four types of KM
systems.
4. Describe some of the issues facing knowledge management.
APPLICATION EXERCISES
1. Identify an example of a knowledge management initiative that has been
undertaken in your organization. Has the initiative been successful? What are
INTRODUCING KNOWLEDGE MANAGEMENT 13
some of the issues, both technical and nontechnical, that were faced during
its implementation?
2. Design a knowledge management initiative to support your business needs.
3. Describe the nontechnical issues that you will face during its implementa-
tion.
4. Consider the four forces driving KM described in this chapter. Think of an-
other example that illustrates each of these forces.
NOTE
1. Much of the IS research has concentrated on the development of the technology acceptance
model (TAM; Davis 1989), which identifies two factors associated with user acceptance of information
technology to be Perceived Usefulness and Perceived Ease of Use.
REFERENCES
Becerra-Fernandez, I., and Sabherwal, R. 2005. Knowledge management at NASA-Kennedy Space
Center. International Journal of Knowledge and Learning, 1(1/2), 159–170.
Bradley, K. 1997. Intellectual capital and the new wealth of nations. Business Strategy Review, 8(1)
53–62.
Davenport, T.H., and Prusak, L. 1998. Working knowledge: How organizations manage what they know.
Boston: Harvard Business School Press.
Davis, F. 1989. Perceived usefulness, perceived ease of use, and user acceptance of information tech-
nology. MIS Quarterly, 13(3), 319–340.
Drucker, P. 1994. The age of social transformation. The Atlantic Monthly, 274(5), 53–70.
Eisenberg, H. 1997. Healing the wounds from reengineering and downsizing. Quality Progress,
May.
Fries, S. 1992. NASA Engineers and the Age of Apollo. Washington, DC (NASA SP-4104).
Hsu, I-C., and Sabherwal, R. 2011. From intellectual capital to firm performance: The mediating role
of knowledge management capabilities, IEEE Transactions on Engineering Management, 58(4),
626–642.
Kenny, B. 2007. The coming crisis in employee turnover. Forbes, April 25.
Nahapiet, J., and Ghoshal, S. 1998. Social capital, intellectual capital, and the organizational advantage.
Academy of Management Review, 23, 242–266.
Stewart, T. 2002. The case against knowledge management. Business 2.0 (February).
Subramaniam, M., and Youndt, M.A. 2005. The influence of intellectual capital on the types of innova-
tive capabilities. Academy of Management Journal, 48(3), 450–463.
This page intentionally left blank
PART I
PRINCIPLES OF
KNOWLEDGE MANAGEMENT
This page intentionally left blank
17
2
The Nature of Knowledge
In the previous chapter, we provided an introduction to the basic concepts of knowledge
management. This chapter takes the next step by explaining in detail what we mean
by knowledge. It also distinguishes knowledge from data and from information and
illustrates these three concepts using some examples. This chapter also summarizes
some of the perspectives commonly used to view knowledge, including both subjec-
tive and objective viewpoints. Moreover, it describes some of the ways to classify
knowledge and identifies some attributes that may be used to characterize different
types of knowledge. It also relates knowledge to the concept of intellectual capital
and its various dimensions. Finally, the chapter also explains the various reservoirs,
or locations, in which knowledge might reside.
WHAT IS KNOWLEDGE?
“Knowledge” is quite distinct from “data” and “information,” although the three terms
are sometimes used interchangeably. However, they are quite distinct in nature. In
this section, we define and illustrate these concepts and differentiate among them.
This discussion also leads to our definition of knowledge.
Data comprise facts, observations, or perceptions (which may or may not be cor-
rect). By itself, data represent raw numbers or assertions and may therefore be devoid
of context, meaning, or intent. Let us consider three examples of what is considered
to be data. We will then build upon these examples to examine the meaning of infor-
mation and knowledge.
Example 1: That a sales order at a restaurant included two large burgers and two
medium-sized vanilla milkshakes is an example of data.
Example 2: The observation that upon tossing a coin it landed heads also illustrates
data.
Example 3: The wind component (u and v) coordinates for a particular hurricane’s
trajectory, at specific instances of time, is likewise considered data.
Although devoid of context, meaning, or intent data can be easily captured, stored,
and communicated using electronic or other media.
Information is a subset of data, only including those data that possess context, rel-
evance, and purpose. Information typically involves the manipulation of raw data to
18 CHAPTER 2
obtain a more meaningful indication of trends or patterns in the data. Let us continue
with the three aforementioned examples:
Example 1: For the manager of the restaurant, the numbers indicating the daily sales (in
dollars, quantity, or percentage of daily sales) of burgers, vanilla milkshakes,
and other products are information. The manager can use such information
to make decisions regarding pricing and raw material purchases.
Example 2: Let us assume that the context of the coin toss is a betting situation where
John is offering to pay anyone $10 if the coin lands heads but take $8
if the coin lands tails. Susan is considering whether to take up John’s
bet, and she benefits from knowing that the last 100 times the coin was
tossed, it landed heads 40 times and tails on 60 occasions. The result
of each individual toss (head or tail) is data, but is not directly useful.
It is therefore data but not information. By contrast, that 40 heads and
60 tails resulted from the last 100 tosses are also data, but they can be
directly used to compute probabilities of heads and tails and hence to
make the decision. Therefore, they are also information for Susan.
Example 3: Based on the u and v components, hurricane software models may
be used to create a forecast of the hurricane trajectory. The hurricane
forecast is information.
As can be seen from these examples, whether certain facts are information or only
data depends on the individual who is using those facts. The facts about the daily sales
of burgers represent information for the store manager but only data for a customer.
If the restaurant is one out of a chain of 250 restaurants, these facts about daily sales
are also data for the CEO of the chain. Similarly, the facts about the coin toss are
simply data for an individual who is not interested in betting.
Knowledge has been distinguished from data and information in two different ways.
A more simplistic view considers knowledge as being at the highest level in a hierarchy
with information at the middle level and data at the lowest level. According to this view,
knowledge refers to information that enables action and decisions or information with
direction. Hence, knowledge is intrinsically similar to information and data, although it
is the richest and deepest of the three, and is consequently also the most valuable. Based
on this view, data refer to bare facts void of context, for example a telephone number.
Information is data in context, for example a phone book. Knowledge is information
that facilitates action, for example, individuals who are the domain experts within an
organization. An example of knowledge includes recognizing that a phone number
belongs to a good client who needs to be called once per week to get his orders.
Although this simplistic view of knowledge may not be completely inaccurate, we
feel it doesn’t fully explain the characteristics of knowledge. Instead, we use a more
complete perspective, according to which knowledge is intrinsically different from
information. Instead of considering knowledge as a richer or more detailed set of facts,
we define knowledge in an area as justified beliefs about relationships among concepts
relevant to that particular area. This definition has support in the literature (Nonaka
1994). Let us now consider how this definition works for the above examples.
THE NATURE OF KNOWLEDGE 19
Example 1: The daily sales of burgers can be used, along with other information (e.g.,
information on the quantity of bread in the inventory), to compute the
amount of bread to buy. The relationship between the quantity of bread
that should be ordered, the quantity of bread currently in the inventory,
and the daily sales of burgers (and other products that use bread) is an
example of knowledge. Understanding of this relationship (which could
conceivably be stated as a mathematical formula) helps to use the infor-
mation (on quantity of bread in the inventory and daily sales of burgers,
etc.) to compute the quantity of bread to be purchased. However, the
quantity of bread to be ordered should itself be considered information
and not knowledge. It is simply more valuable information.
Example 2: The information about 40 heads and 60 tails (out of 100 tosses) can
be used to compute the probability of heads (0.40) and tails (0.60).
The probabilities can then be used, along with information about the
returns associated with heads ($10 from Susan’s perspective) and tails
(–$8, again from Susan’s perspective), to compute the expected value
to Susan from participating in the bet. Both probabilities and expected
values are information, although more valuable information than the
facts that 40 tosses produced heads and 60 produced tails. Moreover,
expected value is more useful information than the probabilities; the
former can directly be used to make the decision, whereas the latter
requires computation of expected value.
The relationship between the probability of heads, the number of times
the coin lands heads, and the total number of tosses (i.e., that probability of
heads, or pH = nH/(nH + nT), assuming that the coin can only land heads
or tails), is an example of knowledge. It helps compute the probability
from the data on outcomes of tosses. The similar formula for probability
of tails is knowledge as well. In addition, the relationship between ex-
pected value (EV) and the probabilities (pH, pT) and returns (RH, RT)
for heads and tails (i.e., EV = pH*RH + pT*RT) is also knowledge. Using
these components of knowledge, probability of heads and tails can be
computed as 0.40 and 0.60, respectively. Then, the expected value for
Susan can be computed as 0.40*(+$10) + 0.60*(–$8) = –$0.80.
Example 3: The knowledge of a hurricane researcher is used to analyze the u and
v wind components as well as the hurricane forecast produced by the
different software models, to determine the probability that the hurricane
will follow a specific trajectory.
Thus, knowledge helps produce information from data or more valuable information
from less valuable information. In that sense, this information facilitates action such as
the decision of whether to bet or not. Based on the new generated information of the
expected value of the outcome as well as the relationship with other concepts, such as
Susan’s anticipation that the coin may be fair or not, knowledge enables Susan to decide
whether she can expect to win at the game. This aspect of the relationship between data and
information is depicted in Figure 2.1, which shows the relationship between data (which
20 CHAPTER 2
has zero or low value in making the decision) and information (which has greater value
than data, although different types of information might have differing values).
The above relationships between data, information, and knowledge are illustrated
using Example 2 in Figure 2.2. As may be seen from the figure, knowledge of how to
count helps convert data on coin tosses (each toss producing a head or tail, with the set
of 100 tosses producing 100 such observations, shown as H and T, respectively) into
information (number of heads and number of tosses). This information is more useful
than the raw data, but it does not directly help the decisionmaker (Susan) to decide on
whether to participate in the bet. Using knowledge of how to compute probabilities,
this information can be converted into more useful information—that is, the probabili-
ties of heads and tails. Moreover, combining the information about probabilities with
information about returns associated with heads and tails, it is possible to produce even
more information—that is, the expected value associated with participation in the bet.
In making this transition, knowledge of the formula for computing expected value from
probabilities and returns is utilized. Figure 2.2 illustrates how knowledge helps produce
information from data (e.g., probabilities based on outcomes of tosses of 60 heads and
40 tails) or more valuable information (expected value) from less valuable information
(e.g., probabilities and payoffs associated with heads and tails).
The above distinctions among data, information, and knowledge are consistent with
Nonaka and Takeuchi’s (1995) definition of knowledge as “a justified true belief.”
It is also consistent with Wiig’s (1999) view of knowledge as being fundamentally
different from data and information:
Knowledge consists of truths and beliefs, perspectives and concepts, judgments and expec-
tations, methodologies, and know-how. It is possessed by humans, agents, or other active
entities and is used to receive information and to recognize and identify; analyze, interpret,
and evaluate; synthesize and decide; plan, implement, monitor, and adapt—that is, to act
more or less intelligently. In other words, knowledge is used to determine what a specific
situation means and how to handle it.
Figure 2.3 depicts how knowledge, data, and information relate to information systems,
decisions, and events. As discussed, knowledge helps convert data into information. The
InformationData
VALUE
Zero Low Medium High Very High
Knowledge
Figure 2.1 Data, Information, and Knowledge
THE NATURE OF KNOWLEDGE 21
H T H T T
H H H T H
…
T T T H T
pH = 0.40
pT = 0.60
RH = +$10
RT = -$8
nH = 40
nT = 60
InformationData
VALUE
Zero Low Medium High Very High
Knowledge
EV = -$0.80
Counting
pH = nH/(nH+nT)
pT = nT/(nH+nT)
EV=pH RH+ pT RT
…
pH = nH/(nH+nT)
pT = nT/(nH+nT)
EV=pH RH+ pT RT
Figure 2.2 An Illustration of Data, Information, and Knowledge
Knowledge
InformationInformation
System
Decision
K
no
w
led
ge
Use of
information
Data
Events
Figure 2.3 Relating Data, Information, and Knowledge to Events
22 CHAPTER 2
knowledge could be stored in a manual or computer-based information system, which
receives data as input and produces information as output. Moreover, the use of information
to make the decision requires knowledge as well (e.g., in the context of the second example
above, the knowledge that expected value above zero generally suggests that the decision
is a good one). The decisions, as well as certain unrelated factors, lead to events, which
cause generation of further data. The events, the use of information, and the information
system might cause modifications in the knowledge itself. For example, in the context of
example 1 on ordering raw materials based on sales, information about changes in sup-
pliers (e.g., a merger of two suppliers) might cause changes in the perceived relationship
(i.e., knowledge) between the quantity on hand, the daily sales, and the quantity to be
ordered. Similarly, in example 2 on betting on the outcome of a coin toss, the individual’s
risk aversion, individual wealth, and so forth might cause changes in beliefs related to
whether expected value above zero justifies the decision to participate in the bet.
ALTERNATIVE VIEWS OF KNOWLEDGE
Knowledge can be viewed from a subjective or objective stance. The subjective
view represents knowledge using two possible perspectives: as a state of mind or
as a practice. On the other hand, the objective view represents knowledge in three
possible perspectives: as an object, as access to information, or as a capability. The
perspectives on knowledge are shown in Figure 2.4.
SUBJECTIVE VIEW OF KNOWLEDGE
According to the subjective view, reality is socially constructed through interactions
with individuals (Schultze 1999). Knowledge is viewed as an ongoing accomplishment
that continuously affects and is influenced by social practices (Boland and Tenkasi
1995). Consequently, knowledge cannot be placed at a single location, as it has no
existence independent of social practices and human experiences. According to the
subjective view, knowledge could be considered from two perspectives, either as a
state of mind or as practice.
Knowledge as State of Mind
This perspective considers knowledge as being a state of an individual’s mind. Organiza-
tional knowledge is viewed here as the beliefs of the individuals within the organization.
Moreover, to the extent the various individuals have differing experiences and backgrounds,
their beliefs and hence knowledge could differ from each other. Consequently, the focus
here is on enabling individuals to enhance their personal areas of knowledge so that they
can apply them to best pursue organizational goals (Alavi and Leidner 2001).
Knowledge as Practice
According to this perspective, knowledge is also considered as subjective but it is viewed
as being held by a group and not as being decomposable into elements possessed by
THE NATURE OF KNOWLEDGE 23
individuals. Thus, from this perspective, knowledge is “neither possessed by any one
agent, nor contained in any one repository” (Schultze 1999, p. 10). Moreover, knowl-
edge resides not in anyone’s head but in practice. Knowledge is comprised of beliefs,
consistent with our definition earlier, but the beliefs themselves are collective rather than
individual, and therefore are better reflected in organizational activities rather than in
the minds of the organization’s individuals. Viewed from this perspective, knowledge
is “inherently indeterminate and continually emerging” (Tsoukas 1996, p. 22).
OBJECTIVE VIEW OF KNOWLEDGE
The objective view is the diametrical opposite of the subjective stance. According to
the objective view, reality is independent of human perceptions and can be structured
in terms of a priori categories and concepts (Schultze 1999). Consequently, knowl-
edge can be located in the form of an object or a capability that can be discovered
or improved by human agents. The objective view considers knowledge from three
possible perspectives.
Knowledge as Objects
This perspective considers knowledge as something that can be stored, transferred, and
manipulated. Consistent with the definition of knowledge as a set of justified beliefs,
these knowledge objects (i.e., beliefs) can exist in a variety of locations. Moreover,
they can be of several different types, as discussed in the next section.
Knowledge as Access to Information
This perspective considers knowledge as the condition of access to information (Alavi
and Leidner 2001). Thus, knowledge is viewed here as something that enables access
and utilization of information. This perspective extends the above view of knowledge
as objects, emphasizing the accessibility of the knowledge objects.
Figure 2.4 Various Perspectives on Knowledge
Objective
View
Subjective
View
Perspectives on Knowledge
Knowledge as
Practice
Knowledge as
a State of Mind
Knowledge as
an Object
Knowledge as
Capability
Knowledge as
Access to
Information
24 CHAPTER 2
Knowledge as Capability
This perspective is consistent with the last two perspectives of knowledge as objects
or as access to information. However, this perspective differs in that the focus here is
on the way in which knowledge can be applied to influence action. This perspective
places emphasis on knowledge as a strategic capability that can potentially be applied
to seek a competitive advantage.
Thus, the five perspectives discussed above differ in their focus in viewing
knowledge, but they are all consistent in viewing knowledge as a set of beliefs about
relationships. The first perspective, knowledge as a state of mind, focuses on beliefs
within human minds; while the second perspective, knowledge as a practice, focuses
on beliefs implicit to actions or practice. In either case, the beliefs, and the knowl-
edge they comprise, are considered subjective. In contrast, the last three perspectives
(knowledge as objects, knowledge as access to information, and knowledge as a
capability) view knowledge as objective, focusing on beliefs as objects to be stored
and managed, as the condition of access to information, and as a capability that affects
action. We recognize all five perspectives as important, and consider them as simply
providing different ways of examining knowledge. However, in the remainder of the
book, we adopt a position that is more objective than subjective. This is due to the
desire to make this textbook useful for students and managers responsible for manag-
ing knowledge in their organizations; an objective view facilitates making practical
recommendations about how organizations should manage knowledge, whereas a
subjective view helps with understanding knowledge management but may be less
valuable in recommending actions for knowledge management.
We next examine the different forms of knowledge, which are clearly consistent
with the objective perspective of knowledge. However, an argument could also be
made that at least some types of knowledge discussed below (e.g., tacit) are not in-
consistent with a subjective view either.
DIFFERENT TYPES OF KNOWLEDGE
Knowledge has been classified and characterized in several different ways. For
example, knowledge has been categorized as individual, social, causal, conditional,
relational, and pragmatic (Alavi and Leidner 2001) and also as embodied, encoded,
and procedural (Venzin et al. 1998). In this section, we examine some of the more
important classifications of knowledge. It is important to understand the nature of
these various types of knowledge because different types of knowledge should be
managed differently, as discussed in detail in some of the later chapters.
PROCEDURAL OR DECLARATIVE KNOWLEDGE
The first distinction we examine is that between declarative knowledge (facts) and
procedural knowledge (how to ride a bicycle) (Kogut and Zander 1992; Singley
and Anderson 1989). Declarative knowledge (or substantive knowledge, as it is also
called) focuses on beliefs about relationships among variables. For example, all other
THE NATURE OF KNOWLEDGE 25
things being equal, a greater price charged for a product would cause some reduction
in its number of sales. Declarative knowledge can be stated in the form of proposi-
tions, expected correlations, or formulas relating concepts represented as variables.
For example, stating that the sum of the square of the sine of an angle and the square
of the cosine of the same angle would equal one is an example of declarative knowl-
edge. Similarly, identifying the specific product features a specific customer likes is
also an example of declarative knowledge.
Procedural knowledge, in contrast, focuses on beliefs relating sequences of steps
or actions to desired (or undesired) outcomes. An example of such procedural knowl-
edge is the set of justified beliefs about the procedure that should be followed in a
government organization in deciding on whom to award the contract for a particular
area (e.g., information system development).
Declarative knowledge may be characterized as “know what,” whereas procedural
knowledge may be viewed as “know-how.” To further understand the difference
between these two types of knowledge, let us consider the example of a hypothetical
automobile manufacturing firm. An instance of declarative knowledge in this context
is the set of justified beliefs about the effect that the quality of each component would
have on the final product. This could include the effect of quality on such features
as reliability, fuel consumption, deterioration over time, and quality of the ride of a
particular model. Such declarative knowledge, combined with information about the
set of components needed for each model and the prices of various alternatives for
each component, would help determine the specific components that should be used
in each model. An example of procedural knowledge in the same context would be
the set of beliefs about the process used to assemble a particular model of the car.
This could include such things as the steps in the engine assembly process, which
tasks can be performed in parallel, the amount of time that each step should take, the
amount of waiting time between successive steps, and so on.
TACIT OR EXPLICIT KNOWLEDGE
Another important classification of knowledge views it as tacit or explicit (Nonaka
1994; Polanyi 1966). Explicit knowledge typically refers to knowledge that has
been expressed into words and numbers. Such knowledge can be shared formally
and systematically in the form of data, specifications, manuals, drawings, audio-
and videotapes, computer programs, patents, and the like. For example, the basic
principles for stock market analysis contained in a book or manual are considered
explicit knowledge. This knowledge can be used by investors to make decisions about
buying or selling stocks. It should also be noted that although explicit knowledge
might resemble data or information in form, the distinction mentioned earlier in this
chapter is preserved; although explicated, the principles of stock market analysis are
justified beliefs about relationships rather than simple facts or observations. Also the
rules about how to process a travel reimbursement, which becomes embedded in an
enterprise resource planning system, is considered explicit knowledge.
In contrast, tacit knowledge includes insights, intuitions, and hunches. It is difficult
to express and formalize, and therefore difficult to share. Tacit knowledge is more
26 CHAPTER 2
likely to be personal and based on individual experiences and activities. For example,
through years of observing a particular industry, a stock market analyst might gain
knowledge that helps him make recommendations to investors in the stock market
regarding the likely short-term and long-term market trends for the stocks of firms
within that industry. Such knowledge would be considered tacit, unless the analyst
can verbalize it in the form of a document that others can use and learn from. Tacit
knowledge may also include expertise that is so specific that it may be too expensive
to make explicit; therefore, the organization chooses to let it reside with the expert.
As discussed above, explicit and tacit forms of knowledge are quite distinct. How-
ever, it is possible to convert explicit knowledge into tacit, as occurs, for example,
when an individual reads a book and learns from it, thereby converting the explicit
knowledge contained in the book into tacit knowledge in the individual’s mind.
Similarly, tacit knowledge can sometimes be converted into explicit knowledge, as
happens when an individual with considerable tacit knowledge about a topic writes
a book or manual formalizing that knowledge. These possibilities are discussed in
greater detail in the next chapter on knowledge management solutions.
GENERAL OR SPECIFIC KNOWLEDGE
The third classification of knowledge focuses on whether the knowledge is possessed
widely or narrowly (Sabherwal and Becerra-Fernandez 2005). General knowledge
is possessed by a large number of individuals and can be transferred easily across
individuals. For example, knowledge about the rules of baseball can be considered
general, especially among the spectators at a baseball park. One example of general
knowledge in this context is recognizing that when a baseball player takes the fourth
“ball,” he gets a walk; when he takes the third “strike,” he is out. It is general because
everyone with a basic understanding of baseball would possess this knowledge.
Unlike general knowledge, specific knowledge, or “idiosyncratic knowledge,”
is possessed by a very limited number of individuals, and is expensive to transfer
(Hayek 1945; Jensen and Meckling 1996; Sabherwal and Becerra-Fernandez 2005).
Consider the distinction between a professional coach and a typical fan watching a
baseball game. The coach has the knowledge needed to filter, from the chaos of the
game, the information required to evaluate and help players through advice such as
when to try to hit the ball, when to steal a base, and so on. For example, if Albert
Pujols is at bat, a slow man is on first, his team has two outs and is behind by one
run against a left-handed pitcher, Pujols should be allowed to swing away. Few fans
may have this knowledge, and so it is considered specific.
Specific knowledge can be of three types: technology-specific knowledge, context-
specific knowledge, and context-and-technology-specific knowledge. Technically
specific knowledge is deep knowledge about a specific area. It includes knowledge
about the tools and techniques that may be used to address problems in that area.
This kind of knowledge is often acquired as a part of some formal training and is then
augmented through experience in the field. Examples include the scientific knowledge
possessed by a physicist and the knowledge about computer hardware possessed by
a computer engineer. Within the engineering directorate at NASA-Kennedy Space
THE NATURE OF KNOWLEDGE 27
Center, the knowledge of project management techniques (such as PERT charts and
critical path analysis) is technology specific, as it pertains to project management in
general without being specific to NASA or Kennedy Space Center.
On the other hand, context-specific knowledge refers to the knowledge of particu-
lar circumstances of time and place in which work is to be performed (Hayek 1945;
O’Reilly and Pondy 1979; Sabherwal and Becerra-Fernandez 2005). Contextually
specific knowledge pertains to the organization and the organizational subunit within
which tasks are performed. For example, the detailed knowledge a design engineer
possesses about the idiosyncrasies of the particular design group in which she is work-
ing is contextually specific. Another example is a baseball catcher’s knowledge of the
team’s pitching staff. Contextually specific knowledge cannot be acquired through
formal training but instead must be obtained from within the specific context (such
as membership in the same design group or baseball team). Within the engineering
directorate at NASA-Kennedy Space Center, the knowledge of the mechanisms used
to patent and license NASA-developed technology for public use is context-specific,
because it depends primarily on the Kennedy Space Center’s context with minimal
effect of the particular technical discipline.
A third kind of specific knowledge, which may be called context-and-technol-
ogy-specific knowledge, is specific in terms of both the context and the technical
aspects. Context-and-technology-specific knowledge simultaneously involves both
rich scientific knowledge and an understanding of the particular context (Machlup
1980; Sabherwal and Becerra-Fernandez 2005). For example, knowledge of how
to decide on the stocks to acquire within an industry is context-and-technology-
specific; it blends an understanding of that industry’s dynamics as well as the tools
used to analyze stock performance. Similarly, in the engineering directorate at
NASA-Kennedy Space Center, the knowledge of how to plan and develop ground
and flight support systems is context-and-technology-specific because it depends on
both the design context of flight systems at Kennedy Space Center and principles
of engineering.
COMBINING THE CLASSIFICATIONS OF KNOWLEDGE
The above classifications of knowledge are independent. In other words, procedural
knowledge could be either tacit or explicit and either general or specific. Similarly,
declarative knowledge could be either tacit or explicit and either general or specific.
Combining the above three classifications and considering technically specific and
contextually specific knowledge as distinct, 12 (2 × 2 × 3) types of knowledge can
be identified as indicated and illustrated in Table 2.1.
KNOWLEDGE AND EXPERTISE
We define expertise as knowledge of higher quality. It addresses the degree of
knowledge. That is, one who possesses expertise is able to perform a task much
better that those who do not. This is specific knowledge at its best. The word
“expert” can be used to describe people possessing many different levels of skills
28 CHAPTER 2
or knowledge. A person can be an expert at a particular task irrespective of how
sophisticated that area of expertise is. For example, there are expert bus drivers
just as there are expert brain surgeons. Each of them excels in the performance of
tasks in their respective field.
Thus, the concept of expertise must be further classified for different types of domains.
The skill levels of experts from different domains should not be compared to each other.
All experts require roughly the same cognitive skills. The difference lies in the depth
of their expertise when compared to others from their own domains. For example, a
highly skilled bus driver has greater abilities than a novice driver, just as an expert brain
surgeon has greater skills than a surgical intern. Prior empirical research on expertise
indicates the importance of knowledge management: “It takes time to become an expert.
Even the most gifted performers need a minimum of ten years of intense training before
they win international competitions” (Ericsson et al. 2007, p. 18).
Expertise can be classified into three distinct categories. Expert systems have had
varying degrees of success when representing expertise from each of these categories.
These categories, discussed in the following subsections, are (1) associational (black
box), (2) motor skills, and (3) theoretical (deep) expertise.
Table 2.1
Illustrations of the Different Types of Knowledge
General Contextually Specific Technically Specific
Declarative
Explicit A book describing factors
to consider when deciding
whether to buy a company’s
stock. This may include
price to earnings ratio,
dividends.
A company document iden-
tifying the circumstances
under which a consultant
team’s manager should
consider replacing a team
member who is having prob-
lems with the project.
A manual describing the
factors to consider in
configuring a computer so
as to achieve performance
specifications.
Tacit Knowledge of the major
factors to consider when
deciding whether to buy a
company’s stock.
A human relations man-
ager’s knowledge of factors
to consider in motivating
an employee in a particular
company.
A technician’s knowledge
of symptoms to look for
in trying to repair a faulty
television set.
Procedural
Explicit A book describing steps to
take in deciding whether to
buy a company’s stock.
A company document
identifying the sequence of
actions a consultant team’s
manager should take when
requesting senior manage-
ment to replace a team
member having problems
with the project.
A manual describing how to
change the operating sys-
tem setting on a computer
so as to achieve desired
performance changes.
Tacit Basic knowledge of the
steps to take in deciding
whether to buy a company’s
stock.
A human relations man-
ager’s knowledge of steps
to take in motivating an
employee in a particular
company.
A technician’s knowledge
of the sequence of steps
to perform in repairing a
television set.
THE NATURE OF KNOWLEDGE 29
Associational Expertise
In most fields, it is usually desirable that experts have a detailed understanding of the
underlying theory within that field. But is this absolutely necessary? What about the
television repair technician considered an expert repairman but who does not under-
stand all of the complex internal workings of a transistor or a picture tube? He can
associate the observations of the performance of the device to specific causes purely
based on his experience. This individual may have expert-level associational under-
standings of these devices and may be able to fix almost any problem encountered.
However, if he encounters a new, previously unseen problem, he may not know how
to proceed because he does not understand the inner workings of the device.
Motor Skills Expertise
Motor skills expertise is predominantly physical rather than cognitive; therefore,
knowledge-based systems cannot easily emulate this type of expertise. Humans improve
these skills by repeated and coached practice. While some people have greater abilities
for these types of skills than others, real learning and expertise result from persistent
guided practice. For example, consider the tasks of riding a bicycle, hitting a baseball,
and downhill snow skiing. When you observe experts performing these activities, you
notice that their reactions seem spontaneous and automatic. These reactions result
from the experts’ continual and persistent and coached practice. For example, when a
skilled baseball player bats, he instinctively reacts to a curveball, adjusting his swing to
connect with the ball. This appropriate reaction results from encountering thousands of
curveballs over many years and the coaches’ recommendations on how to hit the ball
in a particular situation. A novice batter might recognize a curveball being thrown, but
due to a lack of practice reacts slower and consequently may strike out.
These processes do not involve conscious thinking per se. The batter merely reacts
instinctively and almost instantaneously to the inputs. In fact, many coaches maintain
that thinking in such situations degrades performance. Of course, some cognitive ac-
tivity is necessary—the batter must follow the track of the ball, recognize its motion
(curve, changeup, etc.), and make a decision on what to do (swing, let it go, etc.). The
issue, however, is that the result of the decision-making is manifested in very quick
physical actions and not in carefully pondered statements.
Theoretical (Deep) Expertise
Finding a solution to a technical problem often requires going beyond a superficial
understanding of the domain. We must apply creative ingenuity—ingenuity that is
based on our theoretical knowledge of the domain. This type of knowledge allows
experts to solve problems that have not been seen before and, therefore, cannot be
solved via associational expertise.
Such deeper, more theoretical knowledge is acquired through formal training and hands-
on problem-solving. Typically, engineers and scientists who have many years of formal
training possess this type of knowledge. Box 2.1 illustrates deep theoretical knowledge.
30 CHAPTER 2
SOME CONCLUDING REMARKS ON THE TYPES OF KNOWLEDGE
In addition to the above types of knowledge, some other classifications also deserve
mention. One of these classifications views knowledge as either simple or complex.
Whereas simple knowledge focuses on one basic area, complex knowledge draws
upon multiple distinct areas of expertise. Another classification focuses on the role
of knowledge within organizations. It divides knowledge into: support knowledge,
which relates to organizational infrastructure and facilitates day-to-day operations;
tactical knowledge, which pertains to the short-term positioning of the organization
relative to its markets, competitors, and suppliers; and strategic knowledge, which
pertains to the long-term positioning of the organization in terms of its corporate vi-
sion and strategies for achieving that vision.
Based in part on the above types of knowledge, a number of characteristics of
knowledge can be identified. One such characteristic is explicitness of knowledge,
which reflects the extent to which knowledge exists in an explicit form so that it can
be stored and transferred to others. As a characteristic of knowledge, explicitness
indicates that rather than simply classifying knowledge as either explicit or tacit, it
may be more appropriate to view explicitness as a continuous scale. Explicit and tacit
kinds of knowledge are at the two ends of the continuum, with explicit knowledge
being high in explicitness and tacit knowledge being low in this regard. Any specific
knowledge would then be somewhere along this continuum of explicitness.
Specific knowledge is directly related to the concept of knowledge specificity
(Choudhury and Sampler 1997). A high level of knowledge specificity implies that
the knowledge can be acquired and/or effectively used only by individuals possess-
ing certain prior knowledge (Jensen and Meckling 1996). Knowledge specificity
implies that the knowledge is possessed by a very limited number of individuals and
Box 2.1
Deep Theoretical Knowledge Enables Competitive Advantage
During the 1980s, two firms were involved in competition for a long-term (multiple decades)
and large (multibillion dollar) government contract for tactical missiles. Neither company had a
significant performance advantage over the other.
A scientist at one of the firms, who was not a member of the project team, broke the stalemate.
He had deep expertise in developing missiles due to over 20 years of experience in this area. He
was well regarded as a technical expert, and when he called a meeting of the major participants
in the project they all came. For several hours, he enchanted them with a comprehensive descrip-
tion of design changes that he had identified within a single week of committed effort. Making
no use of any kind of notes, he guided them through the reconfiguration of the entire missile. To
implement the extensive changes he suggested in hardware, wiring, and software, 400 individuals
would need to work full-time for a year and a half. However, the expert’s audience was convinced
that the redesign would produce tremendous competitive advantage. His proposal led to a frenzy
of activity and enabled his firm to win the contract. More than 20 years later, in 2004, the redesign
that this individual with deep expertise had created was still producing benefits.
Source: Compiled from Leonard and Swap 2004.
THE NATURE OF KNOWLEDGE 31
is expensive to transfer (Choudhury and Sampler 1997). Taking a step further, techni-
cally specific and contextually specific knowledge lead us to break down knowledge
specificity into contextual knowledge specificity and technical knowledge specific-
ity. Of course, contextually specific knowledge and technically specific knowledge
are high in contextual knowledge specificity and technical knowledge specificity,
respectively.
In addition, the distinction between simple and complex knowledge may be rep-
resented using complexity as a knowledge attribute. Similarly, the organizational
role of knowledge reflects the distinction among support, tactical, and strategic
knowledge.
An organization does not have only one of the above types of knowledge. Instead,
in any given organization, multiple different types of knowledge exist together. In
Box 2.2, we provide an example of how different types of knowledge exist together
within an organization.
Box 2.2
Different Types of Knowledge at Hill and Knowlton
Founded in 1927, Hill and Knowlton is a leading international communications consultancy
headquartered in New York, with 74 offices in 41 countries and an extensive associate net-
work. It is part of WPP Group Plc, which is one of the world’s largest communications services
groups and provides services to local, multinational, and global clients. Among other things, the
company is hired by organizations to manage their product launches, media relations, and com-
munication during crises.
In the late 1990s, turnover rates in certain practices in public relations, such as those related
to technology, increased from 15 percent to over 30 percent. The loss of talented individu-
als led to a leakage of important knowledge as well as information about specific projects. In
1988, in response to concerns by several key clients of the company, the Worldwide Advisory
Group (a summit of the company’s 200 managers) considered ways of addressing this issue of
knowledge leakage. This group identified three broad types of knowledge that were important to
the company. One of these was the company’s internal knowledge about its own products and
services. The second was external knowledge, such as economic forecasts and other related
research by outside experts. The third type of knowledge related to clients including budgets,
templates, and account activity.
Subsequently, Ted Graham was appointed as Hill and Knowlton’s worldwide director of
knowledge management. He concluded that while the company was performing well in terms of
capturing the structured knowledge such as case studies, proposals, and staff bios, it was not
doing so well in capturing unstructured knowledge such as knowledge embedded in speeches,
e-mail messages, and other information that had not been classified in any fashion. To deal
with this problematic situation, the advisory group decided to replace the current global Intranet
with “hK.net,” a “Web-based virtual workspace” serving the company’s offices across the world.
Based on Intraspect Software Inc.’s Salsa application and a password-protected Web site,
hK.net was designed to enable both the employees and clients to access internal and external
repositories of information and knowledge such as news about the company and the industry,
client-related budget information and e-mail archives, staff biographies, presentations, spread-
sheets, case studies, pictures, video clips, conference notes, research reports, and so on.
Clients as well as Hill and Knowlton executives appreciated hK.net because it reduced the time
spent in educating new members of project teams as well as training new employees.
Source: Compiled from Meister and Mark 2004, www.hillandknowlton.com/.
www.hillandknowlton.com/
32 CHAPTER 2
LOCATIONS OF KNOWLEDGE
Knowledge resides in several different locations or reservoirs, which are summarized
in Figure 2.5. They include people, including individuals and groups; artifacts, includ-
ing practices, technologies, and repositories; and organizational entities, including
organizational units, organizations, and interorganizational networks. These locations
of knowledge are discussed in the rest of this section.
KNOWLEDGE IN PEOPLE
A considerable component of knowledge is stored in people. It could be stored
either at the individual level or within a group or a collection of people (Felin and
Hesterly 2007).
Some knowledge is stored in individuals within organizations. For instance, in
professional service firms, such as consulting or law firms, considerable knowledge
resides within the minds of individual members of the firm (Argote and Ingram 2000;
Felin and Hesterly 2007). The knowledge stored in individuals is the reason several
companies continually seek ways to retain knowledge that might be lost because of
individuals retiring or otherwise leaving the organization.
In addition, considerable knowledge resides within groups because of the rela-
tionships among the members of the group (Felin and Hesterly 2007). When three
individuals have worked together for a long time, they instinctively know each other’s
strengths and weaknesses, understand the other’s approach, and recognize aspects that
need to be communicated and those that could be taken for granted (Skyrme 2000).
Consequently, groups form beliefs about what works well and what does not, and this
knowledge is over and above the knowledge residing in each individual member. In
other words, the collective knowledge is synergistic—greater than the sum of their
individual knowledge. Communities of practice that first develop as individuals in-
teract frequently with each other (physically or virtually) to discuss topics of mutual
interest, and they illustrate such embedding of knowledge within groups.
KNOWLEDGE IN ARTIFACTS
Over time, a significant amount of knowledge is stored in organizational artifacts
as well. Some knowledge is stored in practices, organizational routines, or sequen-
tial patterns of interaction. In this case, knowledge is embedded in procedures,
rules, and norms that are developed through experience over time and guide future
behavior (Levitt and March 1988). For example, fast-food franchises often store
knowledge about how to produce high-quality products in routines (Argote and
Ingram 2000).
Considerable knowledge is also often stored in technologies and systems. As discussed
earlier in this chapter, in addition to storing data, information technologies and computer-
based information systems can store knowledge about relationships. For example, a com-
puterized materials requirement planning system contains considerable knowledge about
relationships among demand patterns, lead times for orders, and reorder quantities.
THE NATURE OF KNOWLEDGE 33
Knowledge repositories represent a third way of storing knowledge in artifacts.
Knowledge repositories could either be paper based such as books, papers, and other
documents, or electronic. An example of a paper-based repository is a consultant’s
set of notes to herself about the kind of things the client might focus more on, when
examining the proposals submitted by the consultant firm’s and its competitors. On
the other hand, a Web site containing answers to frequently asked questions (FAQs)
about a product represents an electronic knowledge repository.
KNOWLEDGE IN ORGANIZATIONAL ENTITIES
Knowledge is also stored within organizational entities. These entities can be consid-
ered at three levels: organizational units (parts of the organization), an entire organi-
zation, and in interorganizational relationships (such as the relationship between an
organization and its customers).
Within an organizational unit, such as a department or an office, knowledge is
stored partly in the relationships among the members of the units. In other words, the
organizational unit represents a formal grouping of individuals, who come together
not because of common interests but rather, because of organizational structuring.
Over time, as individuals occupying certain roles in an organizational unit depart and
are replaced by others, the incumbents inherit some, but not all, of the knowledge
developed by their predecessors. This knowledge may have been acquired through the
systems, practices, and relationships within that unit. Moreover, contextually specific
knowledge is more likely to be related to the specific organizational unit.
An organization, such as a business unit or a corporation, also stores certain
knowledge, especially contextually specific knowledge. The norms, values, prac-
tices, and culture within the organization, and across its organizational units, contain
knowledge that is not stored within the mind of any one individual. The way in which
Figure 2.5 The Reservoirs of Knowledge
Organizational
Entities
People
Knowledge Reservoirs
Groups
Individuals Organizational Units
Interorganizational
Networks
Organizations
Artifacts
Practices RepositoriesTechnologies
Individuals
34 CHAPTER 2
the organization responds to environmental events is dependent, therefore, not only
upon the knowledge stored in individuals and organizational units but also in the
overall organizational knowledge that has developed through positive and negative
experiences over time.
Finally, knowledge is also stored in interorganizational relationships. As organi-
zations establish and consolidate relationships with customers and suppliers, they
draw upon knowledge embedded in those relationships. Customers who use the focal
organization’s products, and suppliers who provide the basic components from which
the products are made, often have considerable knowledge about the strengths and
weaknesses of those products. Consequently, organizations often learn from their
customers’ experience with products about how these can be improved. They can
also learn about new products that might be appealing to customers.
KNOWLEDGE IN COMMUNITIES OF PRACTICE
A community of practice (CoP) is an organic and self-organized group of individu-
als who are dispersed geographically or organizationally but communicate regularly
to discuss issues of mutual interest (Lave and Wenger 1991). Lave and Wegner ar-
gue that learning is not a process solely within an individual’s mind but is instead a
process that occurs through social interactions. This process of learning is facilitated
by discussions with colleagues and mentors or by observing how others apply the
knowledge and then try it themselves (Kimble, Hildreth, and Bourdon 2008). There
are many examples of communities of practice in real life, such as the community
of researchers at a scientific conference and investors at forums related to the stock
market. The members of a community of practice do not need to be co-located but
they necessarily interact to learn, share, and communicate their tacit and explicit
knowledge about shared interests.
Communities of practice have become associated with finding, sharing, transferring,
and archiving knowledge, as well as making explicit or tacit knowledge. Therefore,
communities of practice have significant value as they possess a rare source of tacit
knowledge.
KNOWLEDGE LOCATIONS AND FORMS OF INTELLECTUAL CAPITAL
An organization’s intellectual capital refers to the sum of all its knowledge resources,
which may be within or outside the organization (Stewart 1997; Subramaniam and
Youndt 2005; Youndt et al. 2004). Intellectual capital has been viewed as being of dif-
ferent types in terms of where the knowledge resides. Intellectual capital has recently
been classified into three types: human capital, or the knowledge, skills, and capabili-
ties possessed by individual employees; organizational capital, or the institutionalized
knowledge and codified experience residing in databases, manuals, culture, systems,
structures, and processes; and social capital, or the knowledge embedded in relation-
ships and interactions among individuals (Hsu and Sabherwal 2001; Subramaniam
and Youndt 2005; Youndt et al. 2004). These three types of intellectual capital relate
directly to the locations discussed above: Human capital relates to knowledge in
THE NATURE OF KNOWLEDGE 35
people, structural capital relates to knowledge in artifacts, and organizational capital
relates to knowledge in organizational entities.
SUMMARY
In this chapter, we have explained the nature of knowledge in considerable detail. Knowl-
edge is distinguished from data and information; highlighting that knowledge should
best be considered as fundamentally different from data and information rather than
considering data, information, and knowledge as being part of a hierarchy. We defined
knowledge in an area as justified beliefs about relationships among concepts relevant to
that particular area. Furthermore, we examined subjective and objective perspectives
for viewing knowledge, including perspectives that consider knowledge as a state of
mind, as practice, as an object, as access to information, and as a capability. We then
distinguished between procedural and declarative knowledge, between tacit and explicit
knowledge, and between general and specific knowledge. Some other ways of classify-
ing knowledge were also described. Based on the various classifications of knowledge,
we introduced knowledge characteristics such as tacitness, specificity, and so on. This
chapter also described the possible locations of knowledge including people, artifacts,
and organizational entities, and related these locations to different types of intellectual
capital. The next chapter builds on this one by explaining knowledge management and
describing the various aspects of KM infrastructure.
REVIEW
1. How do the terms data and knowledge differ? Describe each term with the
help of a similar example, elucidating the difference between the two.
2. Information contains data but not all data are information. Justify this
statement.
3. Explain why the same set of data can be considered as useful information by
some and useless by others. Further, could this useful information be termed
as knowledge? Why?
4. Describe the ways in which knowledge differs from data and information.
Justify your answer with a relevant diagram.
5. Explain the importance of knowledge in creation and utilization of information.
6. How does the subjective view of knowledge differ from the objective view?
Explain how knowledge can be viewed as a state of mind, as a practice, as
objects, as access to information and as capability.
7. What is the difference between knowledge characterized as know what and
know-how? In the above situations, how would you classify the knowledge
a computer programmer has?
8. Does a player in a card game use tacit or explicit knowledge? Why? Define
and explain the difference between the two.
9. What is general knowledge? How does it differ from specific knowledge?
Describe the types of specific knowledge with suitable examples.
10. What is expertise? Distinguish among the three types of expertise.
36 CHAPTER 2
11. What is a community of practice? What role does it play in knowledge
management?
12. Contrast the differences between knowledge in people and knowledge in ar-
tifacts. Describe the various repositories of knowledge within organizational
entities.
13. What is intellectual capital? What are the three types of intellectual capital,
and how do they relate to different knowledge locations?
APPLICATION EXERCISES
1. Consider five decisions you have made today. (They could be simple things
like taking a turn while driving or even choosing a soda at a convenience
store.) In each case determine the data, information, and/or knowledge that
were involved in the decision. Now consider how those decisions would
have been influenced by the lack of pre-existing data, information, or
knowledge.
2. You have recently invented a new product. Collect demographic data from a
sample population and determine how you would use this data and convert it
into information for marketing the product. Give an example about knowledge
that may be useful in converting the data into information.
3. Interview a manager in a manufacturing organization and one in a services-
based organization. Determine the contrasting views of knowledge between
the two due to the nature of their businesses.
4. Determine the various locations of knowledge within your organization (or
that of a friend/family member). Classify them appropriately. Now speculate
on the negative effects of not having one or more of those knowledge reposi-
tories and accordingly determine which repository is the most critical to the
organization. Which is the least?
5. Determine the various types of knowledge you used to read this chapter. You
should be able to state at least one of each type.
6. You are considering buying a new Ford Taurus. Gather tacit knowledge and
explicit knowledge on buying cars from various resources: for example, Ford
Web site (http://www.ford.com). List your findings and explain what source
of knowledge is important for your choice.
7. Suppose you desperately need technical advice on an Apple Inc. product. You
have several options. Four of them are: (a) Call Apple’s technical support; (b)
use Apple’s online customer support; (c) use Apple’s online discussion groups;
and (d) visit the genius bar at the nearest Apple store. Define your preferred
option and briefly explain your choice with the concepts of accessibility to
knowledge reservoir.
8. Wal-Mart Stores, Inc. (www.walmart.com) is said to be one of the leading
employers of older workers and considers seniors vital to its unique corporate
culture. Store managers are encouraged to recruit from senior citizen groups,
local AARP chapters, and churches. Analyze Wal-Mart’s above strategy in
terms of knowledge management.
http://www.ford.com
www.walmart.com
THE NATURE OF KNOWLEDGE 37
9. Use any organization with which you are familiar to answer this question. This
organization could be one where you currently work or one where you have
previously worked. For this organization, describe one example of knowledge
that would be classified as structural capital, one example of knowledge that
would be classified as organizational capital, and one example of knowledge
that would be classified as social capital.
REFERENCES
Alavi, M., and Leidner, D. 2001. Knowledge management and knowledge management systems:
Conceptual foundations and research issues. MIS Quarterly, 25(1), 107.
Argote, L., and Ingram, P. 2000. Knowledge transfer: Basis for competitive advantage in firms. Orga-
nizational Behavior and Human Decision Processes, 82(1) (May), 150–169.
Boland, R.J., and Tenkasi, R.V. 1995. Perspective making and perspective taking in communities of
knowing. Organization Science, 6(4), 350–72.
Choudhury, V., and Sampler, J. 1997. Information specificity and environmental scanning: An economic
perspective. MIS Quarterly, 21(1), 25–53.
Ericsson, K.A., Prietula, M.J., and Cokely, E.T. 2007. The making of an expert. Harvard Business
Review, July–August, 114–121.
Felin, T., and Hesterly, W.S. 2007. The knowledge-based view, nested heterogeneity, and new value
creation: Philosophical considerations on the locus of knowledge. Academy of Management Review,
32(1), 195–218.
Hayek, F.A. 1945. The use of knowledge in society. American Economic Review, XXXV(4) (Septem-
ber), 519–530.
Hsu, I-C., and Sabherwal, R. 2011. From intellectual capital to firm performance: The mediating role of knowl-
edge management capabilities. IEEE Transactions on Engineering Management, 58(4), 626–642.
Jensen, M.C., and Meckling, W.H. 1996. Specific and general knowledge, and organizational struc-
ture. In Knowledge management & organizational design, ed. P.S. Myers, 17–38. Newton, MA:
Butterworth-Heinemann.
Kimble, C., Hildreth, P., and Bourdon, I. (Eds.). 2008. Communities of practice: Creating learning
environments for educators. Charlotte, NC: Information Age.
Kogut, B., and Zander, U. 1992. Knowledge of the firm, combinative capabilities and the replication
of technology. Organization Science, 3(3), 383–397.
Lave, J., and Wenger, E. 1991. Situated learning: Legitimate peripheral participation. Cambridge,
UK: Cambridge University Press.
Leonard, D., and Swap, W. 2004. Deep smarts. Harvard Business Review, 82(9) (September), 88–97.
Levitt, B., and March, J.G. 1988. Organizational learning. Annual Review of Sociology, 14, 319–340.
Machlup, F. 1980. Knowledge: Its creation, distribution and economic significance, Volume 1. Princ-
eton, NJ: Princeton University Press.
Meister, D., and Mark, K. 2004. Hill & Knowlton: Knowledge management. Case #9B04E003. London,
ONT: Ivey Publishing.
Nonaka, I. 1994. A dynamic theory of organizational knowledge creation. Organization Science, 5(1),
14–37.
Nonaka, I., and Takeuchi, H. 1995. The knowledge creating company. New York: Oxford University
Press.
O’Reilly, C.A., and Pondy, L. 1979. Organizational communication. In Organizational behavior, ed.
S. Kerr, 119–150. Columbus, OH: John Wiley & Sons.
Polanyi, M. 1966. The tacit dimension. London: Routledge and Keoan.
Sabherwal, R., and Becerra-Fernandez, I. 2005. Integrating specific knowledge: Insights from Kennedy
Space Center. IEEE Transactions on Engineering Management, 52(3), 301–315.
38 CHAPTER 2
Schultze, U. 1999. Investigating the contradictions in knowledge management. In Information systems:
Current issues and future changes, ed. T.J. Larsen, L. Levine, and J.I. De Gross, 155–174. Laxenberg:
International Federation for Information Processing.
Singley, M., and Anderson, J. 1989. The transfer of cognitive skill. Cambridge, MA: Harvard Univer-
sity Press.
Skyrme, D.J. 2000. Developing a knowledge strategy: From management to leadership. In Knowledge
management: Classic and contemporary works, ed. D. Morey, M. Maybury, and B. Thuraisingham,
61–84. Cambridge, MA: The MIT Press.
Stewart, T.A. 1997. Intellectual capital. New York: Doubleday-Currency.
Subramaniam, M., and Youndt, M.A. 2005. The influence of intellectual capital on the types of innova-
tive capabilities. Academy of Management Journal, 48(3), 450–463.
Tsoukas, H. 1996. The firm as a distributed knowledge system: A constructionist approach. Strategic
Management Journal, 17 (Winter), 11–25.
Venzin, M., von Krogh, G., and Roos, J. 1998. Future research into knowledge management. In Know-
ing in firms: Understanding, managing and measuring knowledge, ed. G. von Krogh, J. Roos, and
D. Kleine. Thousand Oaks, CA: Sage.
Wiig, K. 1999. Introducing knowledge management into the enterprise. In Knowledge management
handbook, ed. Jay Liebowitz, 3-1 to 3-41. Boca Raton, FL: CRC Press.
Youndt, M.A., Subramaniam, M., and Snell, S.A. 2004. Intellectual capital profiles: An examination
of investments and returns. Journal of Management Studies, 41(2), 335–361.
39
3
Knowledge Management
Foundations: Infrastructure,
Mechanisms, and Technologies
In Chapter 2, we examined the nature of knowledge as well as its various forms
and locations. This chapter explains in greater detail the concept of knowledge
management. It also describes knowledge management solutions, which refer to the
variety of ways in which knowledge management can be facilitated. KM solutions
include two components: KM processes and KM systems. KM solutions depend on
three foundations: KM mechanisms, KM technologies, and KM infrastructure. This
chapter describes these three KM foundations. It is followed by a discussion of KM
solutions—that is, KM processes and KM solutions—in Chapter 4.
In this chapter we first discuss knowledge management, and then we describe
the five components of KM solutions. We subsequently describe and illustrate KM
mechanisms, KM technologies, and KM infrastructure followed by a brief discussion
of the management of these KM foundations and some concluding remarks.
KNOWLEDGE MANAGEMENT
Managing any resource may be defined as doing what is necessary to get the most
out of that resource. Therefore, at a very simple level, knowledge management may
be defined as doing what is needed to get the most out of knowledge resources. Let us
now consider this simple definition in some detail by providing a few elaborations.
First, it is important to stress that this definition can be applied at the individual
as well as organizational levels. Depending on the level, knowledge resources
might be those resources that are relevant to the decisions, goals, and strategies
of an individual or an organization. The “organization” may be a corporation, a
firm, a field office of a firm, a department within a corporation or firm, and so
forth. Moreover, the term knowledge resources refers not only to the knowledge
currently possessed by the individual or the organization but also to the knowledge
that can potentially be obtained (at some cost, if necessary) from other individu-
als or organizations.
Second, “get the most” reflects the impacts of knowledge management on the goal
achievement of the individual or the organization. Considering the impact knowledge
can have on individuals and organizations (as summarized in Chapter 1 and to be
discussed in greater detail in Chapter 4), the objective of knowledge management is
to enhance the extent to which knowledge facilitates the achievement of individual or
organizational goals. Furthermore, a cost/benefit assumption is implicit here. In other
40 CHAPTER 3
words, the objective is to enhance the impact of knowledge in a cost-effective fashion,
such that the benefits of knowledge management exceed the costs of doing so.
Finally, “the things needed” refers to a variety of possible activities involved in
knowledge management. These activities are broadly intended to: (a) discover new
knowledge, (b) capture existing knowledge, (c) share knowledge with others, or (d)
apply knowledge.
Based on these elaborations, a more detailed definition of knowledge management
can now be offered.
Knowledge management can be defined as performing the activities involved in discovering,
capturing, sharing, and applying knowledge so as to enhance, in a cost-effective fashion,
the impact of knowledge on the unit’s goal achievement.
Another important emergent technology related to knowledge management—
business intelligence (BI)—is sometimes used interchangeably with KM. Although
KM and BI are somewhat interrelated, they are quite distinct. BI focuses on providing
decisionmakers with valuable information and knowledge by utilizing a variety of
sources of data and structured and unstructured information (Sabherwal 2007, 2008),
via the discovery of the relationships that may exist between these sources of data
and information.
Unlike KM, which starts with information and knowledge as inputs, BI begins
with data and information as inputs. KM directly results in the discovery of new
knowledge, the conversion of knowledge from one form to another (i.e., from tacit
to explicit or vice versa), the sharing of knowledge, or the application of knowledge
while making a decision. In contrast, BI directly results in information (which is
presented in a friendly fashion, such as through dashboards) and newly created
knowledge or insights obtained by revealing previously unknown connections or
patterns within data and information. Thus, KM is by and large not directly con-
cerned with data (with the exception of knowledge discovery from data and infor-
mation using techniques such as data mining, which represents an area of overlap
between KM and BI). In contrast, data are critical to BI, which often depends on
activities like data warehousing and data mining. However, the results of BI can
be, and often are, useful inputs to KM.
KM incorporates knowledge capture, sharing, and application in addition to
discovery. On the other hand, BI focuses on data access, analysis, and presentation.
The connection between BI and knowledge is limited to knowledge creation (by
discovering patterns based on existing explicit data and information). Even in this
respect, BI focuses directly on discovery of explicit knowledge whereas KM con-
cerns discovery of both tacit and explicit knowledge. In other words, only explicit
knowledge can directly result from BI, whereas KM is concerned with activities that
produce both explicit and tacit knowledge.
KM involves using both social aspects as well as information technology, and
is sometimes viewed as being more social than technical. On the other hand, BI is
primarily technical in nature, and does not incorporate social mechanisms related to
knowledge discovery, such as meetings and brainstorming retreats.
KNOWLEDGE MANAGEMENT FOUNDATIONS 41
The above distinctions between knowledge management and business intelligence
are summarized in Table 3.1.
KNOWLEDGE MANAGEMENT SOLUTIONS AND FOUNDATIONS
Knowledge management depends on two broad aspects: KM solutions, which are
specific in nature; and KM foundations, which are broader and more long-term. KM
solutions refer to the ways in which specific aspects of KM (discovery, capture, shar-
ing, and application of knowledge) can be accomplished. KM solutions include KM
processes and KM systems. KM foundations are the broad organizational aspects
that support KM in the short- and long-term. They include KM infrastructure, KM
mechanisms, and KM technologies. Thus, KM solutions depend on KM foundations,
as shown in Figure 3.1. Next, we briefly explain the three components of KM founda-
tions and the two components of KM solutions.
KM infrastructure reflects the long-term foundation for knowledge management.
In an organizational context, KM infrastructure includes five major components
(e.g., organization culture and the organization’s information technology infra-
structure).
KM mechanisms are organizational or structural means used to promote knowledge
management. They may (or may not) involve the use of information technology, but
they do involve some kind of organizational arrangement or social or structural means
of facilitating KM. They depend on KM infrastructure and facilitate KM systems.
KM technologies are information technologies that can be used to facilitate
knowledge management. Thus, KM technologies are intrinsically no different from
information technologies, but they focus on knowledge management rather than in-
formation processing. KM technologies also support KM systems and benefit from
the KM infrastructure, especially the information technology infrastructure.
KM processes are the broad processes that help in discovering, capturing, sharing,
and applying knowledge. Box 3.1 illustrates the use of one of these processes—
knowledge sharing—at a health-care company. These four KM processes are supported
by KM systems and seven important types of KM subprocesses (e.g., exchange). KM
processes are described and illustrated in Chapter 4.
Table 3.1
A Comparison of Knowledge Management and Business Intelligence
Knowledge Management Business Intelligence
Intellectual Components Primary: Knowledge (Explicit and
Tacit)
Primary: Data
Secondary: Information, Data Secondary: Information, Explicit
Knowledge
Processes Knowledge Capture, Sharing,
Application, and Discovery
Data Access, Analysis, and
Presentation
Key Components Social Mechanisms and Information
Technology
Mainly Information Technologies
42 CHAPTER 3
Box 3.1
Performance Improvement in Healthcare Through Knowledge Sharing
To improve their development and submission of a capital plan to their budget holders, the
operations administration department of a nonprofit health care company used facilitated and
peer-assisted work sessions. In these work sessions, peers shared and transferred relevant
knowledge and experience with members of another team who were facing challenging issues.
This approach helped the team to draw upon the experience of a diverse group of peers from
across their regional parent organization.
The team consequently revised its plan such that it effectively addressed the latest shifts in
patients’ long-term care needs, while incorporating the key concerns of the company’s decision
makers. The end result was a revised capital plan that met the needs of the customers, provid-
ers, and budget holders. Moreover, participation in the process not only benefited the team that
needed help, but also the “visiting peers” that helped them.
Source: Compiled from Greenes Consulting 2014a, 2014b.
Figure 3.1 An Overview of Knowledge Management Solutions and Foundation
Systems
KM Mechanisms
KM Infrastructure
KM Processes
KM Solutions
KM FoundationKM Technologies
Systems
KM Processes
KM
KM systems are the integration of technologies and mechanisms that are developed
to support the above four KM processes. KM systems are described further in Chapter
4, and each of the four kinds of KM systems is then discussed in considerable detail
in Part II of the book.
Thus, KM infrastructure, which is at the organizational level, supports KM mecha-
nisms and technologies. KM mechanisms and technologies are used in KM systems,
with each KM system utilizing a combination of multiple mechanisms and multiple
technologies. Moreover, the same KM mechanism or technology could support multiple
KNOWLEDGE MANAGEMENT FOUNDATIONS 43
KM systems. KM systems enable KM processes, with a KM system focusing on one
specific KM process. Therefore, KM processes and KM systems are specific solutions for
KM needs whereas KM infrastructure, mechanisms, and technologies are broader: KM
mechanisms and technologies support multiple KM solutions, and the KM infrastructure
supports (through KM mechanisms and technologies) all KM solutions. However, over
time, KM infrastructure itself benefits from KM mechanisms and technologies as well
as KM processes, as shown by the curved arrows in Figure 3.1.
The remainder of this chapter describes the three components of the KM founda-
tion—that is, KM infrastructure, mechanisms, and technologies—in further detail.
The two aspects of KM solutions—KM processes and systems—are described in
Chapter 4.
KNOWLEDGE MANAGEMENT INFRASTRUCTURE
KM mechanisms and technologies rely on the KM infrastructure, which reflects
the long-term foundation for knowledge management. In an organizational context,
KM infrastructure includes five major components: organization culture, organi-
zation structure, information technology infrastructure, common knowledge, and
physical environment. These components are discussed in greater detail the next
five sections.
ORGANIZATION CULTURE
Organization culture reflects the norms and beliefs that guide the behavior of the
organization’s members. It is an important enabler of knowledge management in
organizations. Indeed, a survey of KM practices in U.S. companies (Dyer and Mc-
Donough 2001) indicated that the four most important challenges in knowledge
management are nontechnical in nature and include, in order of importance: (1) The
organization’s employees have no time for knowledge management; (2) the current
organization culture does not encourage knowledge sharing; (3) inadequate under-
standing of knowledge management and its benefits to the company; and (4) inability
to measure the financial benefits from knowledge management.
Though the second of the above challenges specifically mentions organization
culture, the first and third challenges are also directly dependent on organization
culture—a supporting organization culture helps motivate employees to understand
the benefits from knowledge management and also to find time for knowledge man-
agement. Indeed, getting people to participate in knowledge sharing is considered
the hardest part of KM. For example, individuals are usually reluctant to contribute
knowledge to repositories, as illustrated by the following comment by one knowledge
worker: “If I share my knowledge, others may take advantage of that. Will they do
the same for me?” (Standing and Benson 2000, p. 343). It is often believed that of
the organizations trying to implement KM, less than 10 percent have succeeded in
making it part of their culture (Koudsi 2000).
Attributes of an enabling organizational culture include understanding the value
of KM practices, management support for KM at all levels, incentives that reward
44 CHAPTER 3
knowledge sharing, and encouragement of interaction for the creation and sharing
of knowledge (Armbrecht et al. 2001). In contrast, cultures that stress individual
performance and hoarding of information within units encourage limited employee
interaction, and lack of an involved top management creates inhibited knowledge
sharing and retention. Moreover, people are often afraid of asking others if they know
the answer to a certain question, and especially if posting a question for the entire
company to see, fear it might reveal their ignorance (Koudsi 2000).
A case study of a baby food manufacturer revealed that built-in competition within
the corporate structure inhibited knowledge sharing practices that could have signifi-
cantly increased revenues. The performance of front-line salespeople was evaluated
comparing it to that of other salespeople. Because of this, a group of front-line sales-
people found a market niche in selling baby food to aging adults who could no longer
eat hard food, but they kept knowledge of their customer base to themselves and let
only their successful sales figures reveal their find. Because the company’s culture
bred competition among employees and offered incentives based on a curve, the firm
missed out not only on increased revenues across the organization but also through
additional sales in that niche market and also on potential product development to
better address the needs of this niche market (DeTienne and Jackson 2001).
In another case study, the CEO of a Web consulting company instituted several
measures to enhance the use of the company’s KM system (Koudsi 2000). He started
publicly recognizing people who stood out as strong knowledge contributors. He
also made usage of the KM system a part of everyone’s job description. He even
started paying employees to use this system. Each task on the KM system was as-
signed points. If a consultant placed his résumé in the system, he would receive one
point. If a consultant created a project record, she would receive five points. The
company’s knowledge manager acted as judge, deciding if entries deserved points.
The totals were tallied every three months, and the resulting score accounted for 10
percent of a consultant’s quarterly bonus. Before these metrics were introduced in
January 1999, only a third of the company’s employees were rated as good or better
in the usage of the KM system, but two months later, that usage had almost doubled
(Koudsi 2000).
Providing appropriate incentives is one way of building a culture that supports
knowledge sharing. Some companies (e.g., Shell Oil Company and Giant Eagle,
Inc.) provide informal recognition for individuals sharing knowledge by mentioning
their accomplishments in a newsletter, an e-mail, or during a meeting. Halliburton
Company uses a “most valuable player” program, acknowledging each month the
person who provides the best idea. The prestigious consulting firm Bain & Com-
pany provides its employees only with two annual awards, and one of them is for
the employee that best carried on the goals of knowledge management and innova-
tion. Companies often incorporate knowledge sharing within employees’ formal
job reviews; some companies make employees’ promotions and bonuses subject
to their sharing knowledge, while some other companies use it as one factor in the
overall evaluation process (Paul 2003). Box 3.2 illustrates the use of incentives
for knowledge sharing at Hill & Knowlton, Inc., a company we examined earlier
in Box 2.2 (see page 31).
KNOWLEDGE MANAGEMENT FOUNDATIONS 45
ORGANIZATION STRUCTURE
Knowledge management also depends to a considerable extent on the organization
structure. Several aspects of organization structure are relevant. First, the hierarchical
structure of the organization affects the people with whom each individual frequently
interacts, and to or from whom he is consequently likely to transfer knowledge.
Traditional reporting relationships influence the flow of data and information as well
as the nature of groups who make decisions together, and consequently affect the
sharing and creation of knowledge. By decentralizing or flattening their organiza-
tion structures, companies often seek to eliminate organizational layers, thereby
placing more responsibility with each individual and increasing the size of groups
reporting to each individual. Consequently, knowledge sharing is likely to occur
with a larger group of individuals in more decentralized organizations. In addition,
matrix structures and an emphasis on “leadership” rather than on “management”
also facilitates greater knowledge sharing primarily by cutting across traditional
departmental boundaries.
Second, organization structures can facilitate knowledge management through
communities of practice.1 For example, a tech-club at DaimlerChrysler included a
group of engineers who didn’t work in the same unit but met regularly, on their own
initiative, to discuss problems related to their area of expertise. Similarly, at Xerox
Corporation, a strategic community of IT professionals, involving frequent infor-
mal interactions among them, promotes knowledge sharing (Storck and Hill 2000).
Box 3.2
Incentives for Knowledge Sharing at Hill and Knowlton
Hill and Knowlton, which was founded in 1927, is a leading international communications con-
sultancy headquartered in New York with 74 offices in 41 countries and an extensive associate
network. It is part of one of the world’s largest communications services groups (WPP), and
provides services to local, multinational, and global clients. The company is hired by organiza-
tions to manage their product launches, media relations, and communication during crises.
Hill and Knowlton offered “Beenz,” which was a system of micropayments, each worth
$0.001, to encourage employees to contribute case studies and bios. Employees could redeem
Beenz online for CDs, books, and other items. For example, an employee could win a weekend
for two in a Caribbean villa for 110,000 Beenz. After the company offering Beenz shut down on
August 17, 2001, some offices started rewarding employees through gift certificates and pizza
parties.
Hill and Knowlton also offered bonuses to individuals managing departments that were active
in knowledge sharing. This was based on two criteria: whether the department made knowledge
contributions and whether the department extracted and used knowledge from another depart-
ment. Bonuses varied from one year to the next.
Hill and Knowlton also established a “best-seller” list to publicize the contributions that were
most frequently accessed. Users were encouraged to discuss their rank on the best-seller list
during conversations about advancement opportunities.
Source: Compiled from Meister and Mark 2004; http://www.hillandknowlton.com/.
http://www.hillandknowlton.com/
46 CHAPTER 3
Box 3.3 further illustrates communities of practice, using the example of Montgomery
Watson Harza.
Communities of practice provide access to a larger group of individuals than
possible within traditional departmental boundaries. Consequently, there are more
numerous potential helpers, and this increases the probability that at least one of
them will provide useful knowledge. Communities of practice also provide access to
external knowledge sources. An organization’s external stakeholders—for example,
customers, suppliers, and partners—provide a far greater knowledge reservoir than the
organization itself (Choo 1998). For instance, relationships with university research-
ers can help new biotechnology firms to maintain their innovativeness. Introduction
of communities of practice in a company is therefore expected to improve business
performance. For example, when Dell Inc. launched ITNinja, an online community of
practice for technical support personnel, the press release noted that the community
would help the IT community find answers to their software questions quickly so
they can focus on innovation and do more” (Dell Inc. 2012).
Although communities of practice are usually not part of a company’s formal
organization structure, company executives can facilitate them in several ways. For
example, they can legitimize them through support for participation in them. Moreover,
they can enhance the perceived value of participation in communities of practice by
seeking advice from them. They can also help communities of practice by providing
them with resources, such as money or connections to external experts and access to
Box 3.3
Communities of Practice at Montgomery Watson Harza
Montgomery Watson Harza1 (MWH) is a global engineering firm with over 3,600 specialists
spread across about 200 offices in 38 countries. From 1995 to 1999, its KM efforts had focused
primarily on information technologies and had encountered problems.
In 1999, it adopted a new KM approach characterized as “People First, Technology as Support.”
This approach led to a new name being given to the KM strategy: “KnowledgeNet.” KnowledgeNet
relied on formal and informal communities of practice, which would be supported by establishing
a global Intranet, called KNet. Each formal community—called a Knowledge Center and partially
funded by management—also established its theme as well as specific business objectives. Each
informal community, called a Knowledge Base, was locally driven and easier to set up. Whereas
Knowledge Centers were created from the top down to facilitate strategic initiatives, Knowledge
Bases were organic and started by practitioners themselves. Knowledge Bases could also be
networked together to represent a larger theme. Moreover, if a Knowledge Base became strategic
to the entire company, it could be converted into a Knowledge Center.
The implementation of this new strategy began in 2000. By early 2004, MWH had three Knowl-
edge Centers (a fourth center was being established) and 120 Knowledge Bases. MWH had about
6,000 employees at that time, and about 1,600 of them belonged to one of these formal or informal
communities. The new KM strategy based on communities of practice received positive responses
from top management, other employees, and clients. Top management perceived KnowledgeNet as
having provided the company with a competitive advantage in terms of obtaining client accounts.
Source: Compiled from Parise, Rolag, and Gulas 2004.
1See http://www.mwhglobal.com/.
http://www.mwhglobal.com/
KNOWLEDGE MANAGEMENT FOUNDATIONS 47
information technology that supports their virtual meetings and knowledge sharing
activities. Communities of practice benefit considerably from other emergent informa-
tion technologies, including blogs and social networking technologies.
Third, organization structures can facilitate knowledge management through
specialized structures and roles that specifically support knowledge management.
Three possibilities deserve special mention. First, some organizations appoint an
individual to the position of Chief Knowledge Officer and make this individual re-
sponsible for the organization’s KM efforts. Second, some organizations establish a
separate department for knowledge management, which is often headed by the Chief
Knowledge Officer. Finally, two traditional KM units—the R&D department and
the corporate library—also facilitate knowledge management, although they differ
in focus. Whereas the R&D department supports management of knowledge about
the latest, or future, developments, the corporate library supports business units by
facilitating knowledge sharing activities and serving as a repository of historical
information about the organization, its industry, and competitive environment. The
leadership of the KM function is further examined in Chapter 12.
INFORMATION TECHNOLOGY INFRASTRUCTURE
Knowledge management is also facilitated by the organization’s information technol-
ogy (IT) infrastructure. Although certain information technologies and systems are
directly developed to pursue knowledge management, the organization’s overall infor-
mation technology infrastructure, developed to support the organization’s information
systems needs, also facilitates knowledge management. The information technology
infrastructure includes data processing, storage, and communication technologies and
systems. It comprises the entire spectrum of the organization’s information systems,
including transaction processing systems and management information systems. It
consists of databases (DBs) and data warehouses, as well as enterprise resource
planning systems. One possible way of systematically viewing the IT infrastructure
is to consider the capabilities it provides in four important aspects: reach, depth, rich-
ness, and aggregation (Daft and Lengel 1986; Evans and Wurster 1999).
Reach pertains to access and connection and the efficiency of such access. Within
the context of a network, reach reflects the number and geographical locations of
the nodes that can be efficiently accessed. Keen (1991) also uses the term reach to
refer to the locations an IT platform is capable of linking, with the ideal being able
to connect to “anyone, anywhere.” Much of the power of the Internet is attributed
to its reach and the fact that most people can access it quite inexpensively. Reach is
enhanced not just by advances in hardware but also by progress in software. For in-
stance, standardization of cross-firm communication standards, and languages such as
XML, make it easier for firms to communicate with a wider array of trading partners,
including those with whom they do not have long-term relationships.
Depth, in contrast, focuses on the detail and amount of information that can be
effectively communicated over a medium. This dimension closely corresponds to the
aspects of bandwidth and customization included by Evans and Wurster (1999) in
their definition of richness. Communicating deep and detailed information requires
48 CHAPTER 3
high bandwidth. At the same time, it is the availability of deep and detailed informa-
tion about customers that enables customization. Recent technological progress—for
instance, in channel bandwidth—has enabled considerable improvement in depth.
Communication channels can be arranged along a continuum representing their
“relative richness” (Carlson and Zmud 1999). The richness of a medium is based on
its ability to: (a) provide multiple cues (e.g., body language, facial expression, tone
of voice) simultaneously; (b) provide quick feedback; (c) personalize messages; and
(d) use natural language to convey subtleties (Daft and Lengel 1986). Information
technology has traditionally been viewed as a lean communication medium. However,
given the progress in information technology, we are witnessing a significant increase
in its ability to support rich communication.
Finally, rapid advances in IT have significantly enhanced the ability to store and
quickly process information. This enables the aggregation of large volumes of infor-
mation drawn from multiple sources. For instance, data mining and data warehousing
together enable the synthesis of diverse information from multiple sources, potentially
to produce new insights. Enterprise resource planning systems (ERPs) also present
a natural platform for aggregating knowledge across different parts of an organiza-
tion. A senior IS executive at PricewaterhouseCoopers LLP, for example, remarks:
“We’re moving quite quickly on to an Intranet platform, and that’s giving us a greater
chance to integrate everything instead of saying to people, ‘use this database and that
database and another database.’ Now it all looks—and is—much more coordinated”
(Thomson 2000, p. 24).
To summarize, the above four IT capabilities enable knowledge management by
enhancing common knowledge or by facilitating the four KM processes. For example,
an expertise locator system (also called knowledge yellow pages or a people-finder
system) is a special type of knowledge repository that pinpoints individuals having
specific knowledge within the organization. These systems rely on the reach and depth
capabilities of IT by enabling individuals to contact remotely located experts and seek
detailed solutions to complicated problems. Another IS solution attempts to capture as
much of the knowledge in an individual’s head as possible and archive it in a search-
able database. This is primarily the aim of projects in artificial intelligence, which
capture the expert’s knowledge in systems based on various technologies, including
rule-based system and case-based reasoning, among others (Wong and Radcliffe 2000).
But the most sophisticated systems for eliciting and cataloging experts’ knowledge
in models that can easily be understood and applied by others in the organization
(see for example Ford et al. 1996) require strong knowledge engineering processes
to develop. Such sophisticated KM systems have typically not been advocated as
frequently for use in mainstream business environments, primarily because of the
high cost involved in the knowledge engineering effort.
COMMON KNOWLEDGE
Common knowledge (Grant 1996) represents another important component of the
infrastructure that enables knowledge management. It refers to the organization’s
cumulative experiences in comprehending a category of knowledge and activities
KNOWLEDGE MANAGEMENT FOUNDATIONS 49
and the organizing principles that support communication and coordination (Zander
and Kogut 1995). Common knowledge provides unity to the organization. It includes:
a common language and vocabulary, recognition of individual knowledge domains,
common cognitive schema, shared norms, and elements of specialized knowledge
that are common across individuals sharing knowledge (Grant 1996; Nahapiet and
Ghoshal 1998). The following comment by a senior executive at NASA-Kennedy
Space Center illustrates problems that might arise due to a lack of common knowledge
(Sabherwal and Becerra-Fernandez 2005, p. 302):
I used to consider myself a systems engineer. When I was in the shuttle program a systems
engineer was somebody who was an expert on a particular shuttle system. In the outside
world a systems engineer has a completely different definition. I was a technical expert on
the shuttle toilet. I called myself a systems engineer, my position description said I was a
systems engineer, but I could go over to the Payloads’ IT department and ask them what
their competency was, what their key effort was, and they would say “systems engineering.”
And I would say, “what system?” Because from my frame of reference, I was doing systems
engineering, from theirs, they were doing something called systems engineering.
Common knowledge helps enhance the value of an individual expert’s knowl-
edge by integrating it with the knowledge of others. However, because the common
knowledge is based on the above definition common only to an organization, this
increase in value is also specific to that particular organization and does not transfer
to its competitors. Thus, common knowledge supports knowledge transfer within the
organization but impedes the transfer (or leakage) of knowledge outside the organiza-
tion (Argote and Ingram 2000).
Two concepts related to common knowledge are absorptive capacity and shared
domain knowledge. Absorptive capacity is a firm’s ability to recognize the value
of new and external information, assimilate it, and apply it to commercial ends. It
is critical to a firm’s innovativeness (Cohen and Levinthal 1990). Moreover, the
development of absorptive capacity is highly dependent on previously acquired
knowledge and expertise. Consequently, greater common knowledge leads to greater
absorptive capacity. On the other hand, shared domain knowledge is one aspect of
common knowledge. More specifically, it is the ability of individuals from different
arenas, such as information technology and finance, to understand each other’s key
processes and recognize each other’s unique contribution and challenges. Reich and
Benbasat (2000) found that shared domain knowledge is one of the most important
influential factors affecting the long-run congruence of IT vision between IT and
business executives.
PHYSICAL ENVIRONMENT
The physical environment within the organization is often taken for granted, but
it is another important foundation upon which knowledge management rests. Key
aspects of the physical environment include the design of buildings and the sepa-
ration between them; the location, size, and type of offices; the type, number, and
nature of meeting rooms; and so on. Physical environment can foster knowledge
50 CHAPTER 3
management by providing opportunities for employees to meet and share ideas. Even
though knowledge sharing there is often not by design, coffee rooms, cafeterias,
water coolers, and hallways do provide venues where employees learn from and
share insights with each other. A 1998 study found that most employees thought
they gained most of their knowledge related to work from informal conversations
around watercoolers or over meals rather than from formal training or manuals
(Wensley 1998).
A number of organizations are creating spaces specifically designed to facilitate
this informal knowledge sharing. For example, the London Business School created
an attractive space between two major departments, which were earlier isolated,
to enhance knowledge sharing between them. Reuters News Service installed
kitchens on each floor to foster discussions. Moreover, a medium-sized firm in
the United States focused on careful management of office locations to facilitate
knowledge sharing (Stewart 2000). This company developed open-plan offices with
subtle arrangements to encourage what one senior executive calls knowledge ac-
cidents. Locations are arranged in this company so as to maximize the chances of
face-to-face interactions among people who might be able to help each other. For
example, an employee might walk down the hall so that she might meet someone
who knows the answer to her question, and she will meet such an individual not
due to chance but because a snack area is positioned where four project teams’
work areas intersect.
Table 3.2 summarizes the five dimensions of KM infrastructure, indicating the key
attributes related to each dimension.
KNOWLEDGE MANAGEMENT MECHANISMS
Knowledge management mechanisms are organizational or structural means used
to promote knowledge management. They enable KM systems, and they are them-
selves supported by the KM infrastructure. KM mechanisms may (or may not) utilize
technology, but they do involve some kind of organizational arrangement or social
or structural means of facilitating KM.
Examples of KM mechanisms include learning by doing, on-the-job training,
learning by observation, and face-to-face meetings. More long-term KM mecha-
nisms include the hiring of a Chief Knowledge Officer, cooperative projects across
departments, traditional hierarchical relationships, organizational policies, standards,
initiation process for new employees, and employee rotation across departments. To
illustrate KM mechanisms, we briefly examine in Box 3.4 the approach one company,
Viant Corporation, takes to manage knowledge.
Box 3.5 provides some additional examples of the use of KM mechanisms to fa-
cilitate knowledge management.
KNOWLEDGE MANAGEMENT TECHNOLOGIES
As mentioned earlier, KM technologies are information technologies that can be
used to facilitate knowledge management. Thus KM technologies are intrinsi-
KNOWLEDGE MANAGEMENT FOUNDATIONS 51
Table 3.2
A Summary of Knowledge Management Infrastructure
Dimensions of KM
Infrastructure Related Attributes
Organization Culture Understanding of the value of KM practices
Management support for KM at all levels
Incentives that reward knowledge sharing
Encouragement of interaction for the creation and sharing of knowledge
Organization Structure Hierarchical structure of the organization (decentralization, matrix
structures, emphasis on “leadership” rather than “management”)
Communities of practice
Specialized structures and roles (Chief Knowledge Officer, KM
department, traditional KM units)
Information Technology Reach
Infrastructure Depth
Richness
Aggregation
Common Knowledge Common language and vocabulary
Recognition of individual knowledge domains
Common cognitive schema
Shared norms
Elements of specialized knowledge that are common across individuals
Physical Infrastructure Design of buildings (offices, meeting rooms, hallways)
Spaces specifically designed to facilitate informal knowledge sharing
(coffee rooms, cafeterias, water coolers)
cally no different from information technologies, but they focus on knowledge
management rather than information processing. KM technologies also support
KM systems and benefit from the KM infrastructure, especially the information
technology infrastructure.
KM technologies constitute a key component of KM systems. Technologies that
support KM include artificial intelligence (AI) technologies including those used
for knowledge acquisition and case-based reasoning systems, electronic discussion
groups, computer-based simulations, databases, decision support systems, enterprise
resource planning systems, expert systems, management information systems, ex-
pertise locator systems, videoconferencing, and information repositories including
best practices databases and lessons learned systems. KM technologies also include
the emergent Web 2.0 technologies, such as wikis and blogs, which are discussed in
detail in Chapter 10.
Examples of the use of KM technologies include World Bank’s use of a combina-
tion of video interviews and hyperlinks to documents and reports to systematically
record the knowledge of employees that are close to retirement (Lesser and Prusak
52 CHAPTER 3
2001). Similarly, at BP plc, desktop videoconferencing has improved communica-
tion and enabled many problems at offshore oil fields to be solved without extensive
traveling (Skyrme 2000).
Box 3.6 provides one example of the use of knowledge management, and how it
helps deal with knowledge loss.
Box 3.4
Knowledge Management Mechanisms at Viant
Viant1 is a Boston-based consulting company specializing in helping clients build e-commerce
businesses. It considers knowledge management as a key objective of the processes through which
new employees are initiated into the organization and existing employees are rotated across func-
tions and locations as mechanisms for knowledge management. Viant makes excellent use of the
orientation process to provide newcomers with knowledge of key clients, some company-specific
skills, and the beginnings of an informal network. New employees begin their Viant career with three
weeks in Boston. On arrival employees receive their laptop, loaded with off-the-shelf and proprietary
software. Later that week they learn team skills and take a course in the company’s consulting
strategy and tools. For the next two weeks they switch back and forth between classroom work and
teams, participating in a mock consulting assignment. They bond, meet all the officers, listen to
corporate folklore, and party with the CEO. Employee rotation also plays an important role in knowl-
edge management at Viant. In fact, conventional reporting relationships do not work here. Because
people rotate in and out of assignments, consultants have no fixed relationship to a boss; instead,
senior managers act as “advocates” for a number of “advocados.” Performance reviews emphasize
the growth in the employee’s own skill level, and stock options recognize the knowledge they share.
Source: Stewart 2000.
1See http://www.viant.com.
Box 3.5
Knowledge Management Mechanisms from Three Organizations
At Phonak, Inc., a worldwide leader in digital hearing instruments, a series of events occur
throughout the year (every six weeks or so) enabling employees to get to know each other
through informal interactions including barbecues, company days out, and bicycle tours.
BP Amoco Chemical Company has benefited from retrospect meetings at the conclusion
of projects. Each retrospect meeting is facilitated by someone outside that project team and
focuses on the following questions: What was the goal of the project? What did we accomplish?
What were the major successes? Why? How can we repeat the successes? What were the
significant disappointments? Why? How can we avoid them in the future?
Katzenbach Partners uses light-hearted contests and events to facilitate knowledge manage-
ment. One example is “Stump Niko,” in which the managing director, who had the reputation that
he knew everything that was going on, would be asked a question about knowledge manage-
ment and the knowledge management system would then be asked the same question. The
objective was to demonstrate the potential of the knowledge management system.
Source: Compiled from Burgelman and Blumenstein 2007; Hoegl and Schulze 2005.
http://www.viant.com
KNOWLEDGE MANAGEMENT FOUNDATIONS 53
MANAGEMENT OF KNOWLEDGE MANAGEMENT FOUNDATIONS
(INFRASTRUCTURE, MECHANISMS, AND TECHNOLOGIES)
Knowledge management infrastructure, mechanisms, and technologies are the un-
derlying foundations for any organization’s KM solutions. KM infrastructure is of
fundamental importance with long-term implications and needs to be managed care-
fully, with close involvement from top executives. In any case, all components of
KM infrastructure (i.e., organization structure, organization culture, IT infrastructure,
common knowledge, and physical environment) affect not only KM but also all other
aspects of the organizational operations. Therefore, KM infrastructure does receive
attention from top management, although it is important that KM be explicitly con-
sidered in making decisions regarding these infrastructural aspects. In this regard, a
strong relationship between the leaders of the KM function (discussed in Chapter 12)
and the top executives of the organization plays an important role.
KM mechanisms and technologies work together and affect each other. KM
mechanisms depend on technology, although some mechanisms do so to a greater
extent than others. Moreover, improvement in KM technologies could, over time,
lead to changes (either improvements in, or in some cases, reduced emphasis on) in
KM mechanisms. In managing KM mechanisms and technologies, it is important to
recognize such interrelationships between mechanisms and technologies. Moreover,
it is important to achieve an appropriate balance between the use of technology and
social or structural mechanisms. Technological progress could lead to people focusing
Box 3.6
Knowledge Management Helps Address Knowledge Loss at Instituto Boliviano de
Comercio Exterior
Knowledge loss due to “brain drain” is a major challenge for most organizations. Recogniz-
ing the value of its in-house knowledge, the Instituto Boliviano de Comercio Exterior (IBCE)
documented its critical work procedures and processes. This enabled consistent delivery of its
core services to the Bolivian business community, despite employee rotation or turnover. More
specifically, IBCE took the following steps.
1. IBCE developed a large collection of video footage, screen recordings, and slideshows
describing the organization’s work processes. This repository helps retain most of IBCE’s
critical knowledge in-house.
2. IBCE started actively using an online collaboration tool that facilitates the sharing of training
materials, process descriptions, and other documents among its employees.
3. IBCE reduced traditional departmental boundaries, beginning with strategic and interactive
sessions on knowledge management, followed by an operational phase, adopting the new
tools and way of working. As a result, the mentality in the organization is gradually changing
from “isolated thinking” into synergies and collaborative approaches.
Since the teams observed the benefits of the above changes from the beginning, the
changes were received well and helped improve knowledge management across the
organization.
Source: Compiled from Plaisier and Boelaars 2012.
54 CHAPTER 3
too much on technology while ignoring structural and social aspects. On the other
hand, an organization with weak IT infrastructure may rely on social and structural
mechanisms while ignoring potentially valuable KM technologies.
Consequently, some organizations focus more on KM technologies, some focus
more on KM mechanisms, and some make a somewhat balanced use of KM technolo-
gies and mechanisms. For example, senior executives at Groupe Danone (Groupe
Danone 2009), which is a leading consumer-goods company with headquarters in
Paris and is known as Dannon in the United States (it is discussed in greater detail
later in Chapter 11, Box 11.1), believe that using IT to share knowledge would not
work as well for the company, and therefore rely primarily on social and structural
mechanisms (Edmondson et al. 2008).
Several other organizations switch from a focus on mechanisms to a focus on
technology, or vice versa. For example, Katzenbach Partners, LLC, relied almost
entirely on social and structural mechanisms to manage knowledge until 2005, but
then, in the light of its organizational growth, started focusing much more on KM
technologies, first using an Intranet and then using Web 2.0 technologies (which we
discuss in greater detail in Chapter 10). This is in contrast to Montgomery Watson
Harza (MWH Global, Inc.), which as discussed in Box 3.3 earlier in this chapter
(p. 46), focused primarily on information technologies from 1995 to 1999, and
then switched to a KM strategy based on “People First, Technology as Support”
(Parise et al. 2004).
SUMMARY
Building on the discussion of knowledge in Chapter 2, we have described the key
aspects of knowledge management in this chapter. We have provided a working
definition of knowledge management and discussed KM solutions as involving five
components: KM processes, KM systems, KM mechanisms, KM technologies, and
KM infrastructure. We have also discussed and illustrated three foundational com-
ponents—KM mechanisms, KM technologies, and KM infrastructure—and briefly
talked about how they could be managed. The next chapter examines the other key
aspects of KM solutions including KM systems, KM mechanisms and technologies,
and KM infrastructure.
REVIEW
1. What is knowledge management? What are its objectives?
2. What is business intelligence? How does knowledge management differ from
business intelligence?
3. Describe the ways to facilitate knowledge management and give suitable
examples.
4. Distinguish between KM foundation and KM solutions. What are the com-
ponents of KM foundation and KM solutions?
5. What is common knowledge? What does it include, and how does it support
knowledge management?
KNOWLEDGE MANAGEMENT FOUNDATIONS 55
6. What are absorptive capacity and shared domain knowledge? How do they
relate to common knowledge?
7. State the role of organizational culture in the development of a good knowl-
edge management infrastructure.
8. How can communities of practice affect business performance?
9. State the role of organizational structure in the development of a good knowl-
edge management infrastructure.
10. In what way does information technology infrastructure contribute to knowl-
edge management within an organization?
APPLICATION EXERCISES
1. Interview a manager of an organization where knowledge management prac-
tices have recently been implemented. Use the interview to study the nature
of the KM infrastructure and the ways in which its components are helping
or inhibiting those KM practices.
2. Consider an organization where you currently work, or are familiar with (either
through your own prior experience or through interactions with someone who
works there). What kind of mechanisms does this organization use to manage
knowledge? What are their effects?
3. Determine ways in which a local hospital would benefit from communities
of practice. Conduct interviews if necessary.
4. Consider a high school with which you are familiar. How can knowledge man-
agement at this high school benefit from information technologies? What kinds
of technologies does it currently use, and how could they be improved?
5. Interview at least three managers from local organizations that have recently
implemented knowledge management. Contrast the differences in organiza-
tion culture, structure, IT Infrastructure, common knowledge, and physical
environment within the organizations.
NOTE
1. For more information about communities of practice, see Chapter 2.
REFERENCES
Argote, L., and Ingram, P. 2000. Knowledge transfer: Basis for competitive advantage in firms. Orga-
nizational Behavior and Human Decision Processes, 82(1) (May), 150–169.
Armbrecht, F.M.R., Chapas, R.B., Chappelow, C.C., Farris, G.F., Friga, P.N., Hartz, C.A., McIlvaine,
M.E., Postle S.R., and Whitwell, G.E. 2001. Knowledge management in research and development.
Research Technology Management, 44(4), 28–48.
Burgelman, R.A., and Blumenstein, B. 2007. Knowledge management at Katzenbach Partners LLC.
Stanford Graduate School of Business, Case SM162.
Carlson, J.R., and Zmud, R.W. 1999. Channel expansion theory and the experiential nature of media
richness perceptions. Academy of Management Journal, 42(2), 153–170.
Choo, C.W. 1998. The knowing organization: How organizations use information to construct meaning,
create knowledge, and make decisions. New York: Oxford University Press.
56 CHAPTER 3
Cohen, W.M., and Levinthal, D.A. 1990. Absorptive capacity: A new perspective on learning and in-
novation. Administrative Science Quarterly, 35(1), 128–152.
Daft, R.L. and Lengel, R.H. 1986. Organization information requirements, media richness, and structural
design. Management Science, 32(5), 554–571.
Dell Inc. 2012. Dell launches ITNinja to provide best online community experience for IT administra-
tors. Press release, March 21. http://content.dell.com/us/en/corp/d/secure/2012-03-21-dell-itninja-
online-community (accessed February 1, 2014).
DeTienne, K.B., and Jackson, L.A. 2001. Knowledge management: Understanding theory and develop-
ing strategy. Competitiveness Review, 11(1), 1–11.
Dyer, G., and McDonough, B. 2001. The state of KM. Knowledge Management (May), 31–36.
Edmondson, A., Moingeon, B., Dessain, V., and Jensen, D. 2008. Global knowledge management at
Danone. Harvard Business School Publishing, Case 9–608–107, April 16.
Evans, P., and Wurster, T.S. 1999. Getting real about virtual commerce. Harvard Business Review,
77(6) (November-December), 85–94.
Fitzgerald, M. 2008. Why social computing aids knowledge management. CIO.com, June 13.
Ford, K.M., Coffey, J.W., Cañas, A.J., Andrews, E.J., and Turner, C.W. 1996. Diagnosis and explanation
by a nuclear cardiology expert system. International Journal of Expert Systems, 9, 499–506.
Grant, R.M. 1996. Toward a knowledge-based theory of the firm. Strategic Management Journal, 17,
109–122.
Greenes Consulting. 2014a. Health care. www.greenesconsulting.com/success-stories/health-care (ac-
cessed January 30, 2014).
———. 2014b. Team learning & performing. www.greenesconsulting.com/services/list-services/team-
learning-performing (accessed January 30, 2014).
Groupe Danone. 2009. www.danone.com/?lang=en.
Hoegl, M., and Schulze, A., 2005. How to support knowledge creation in new product development:
An investigation of knowledge management methods. European Management Journal, 23(3),
263–273.
Keen, P. 1991. Shaping the future: Business design through information technology. Boston: Harvard
Business School Press.
Koudsi, S. 2000. Actually, it is like brain surgery. Fortune, March 20.
Lave, J., and Wenger, E. 1991. Situated learning: Legitimate peripheral participation. Cambridge,
UK: Cambridge University Press.
Lesser, E., and Prusak, L. 2001. Preserving knowledge in an uncertain world. Sloan Management
Review (Fall), 101–102.
Meister, D., and Mark, K. 2004. Hill & Knowlton: Knowledge management. Case #9B04E003. London,
ONT: Ivey Publishing.
Nahapiet, J., and Ghoshal, S. 1998. Social capital, intellectual capital, and the organizational advantage.
Academy of Management Review, 23(2), 242–266.
Parise, S., Rolag, K., and Gulas, V. 2004. Montgomery Watson Harza and knowledge management.
Harvard Business School Publishing, Case BAB102, November 15.
Paul, L.G. 2003. Why three heads are better than one (How to create a know-it-all company). CIO
Magazine (December).
Plaisier, E., and R. Boelaars. 2012. Value your gold: Knowledge management in practice in Bolivia.
CBI Ministry of Foreign Affairs, June 15. www.cbi.eu/success-stories/value-your-gold-knowledge-
management-in-practice-in-bolivia/135806 (accessed January 31, 2014).
Reich, B.H., and Benbasat, I. 2000. Factors that influence the social dimension of alignment between
business and information technology objectives. MIS Quarterly, 24(1), 81–113.
Sabherwal, R. 2007. Succeeding with business intelligence: Some insights and recommendations.
Cutter Benchmark Review, 7(9), 5–15.
———. 2008. KM and BI: From mutual isolation to complementarity and synergy. Cutter Consortium
Executive Report, 8(8), 1–18.
http://content.dell.com/us/en/corp/d/secure/2012-03-21-dell-itninjaonline-community
http://content.dell.com/us/en/corp/d/secure/2012-03-21-dell-itninjaonline-community
CIO.com
www.greenesconsulting.com/success-stories/health-care
www.greenesconsulting.com/services/list-services/teamlearning-performing
www.greenesconsulting.com/services/list-services/teamlearning-performing
www.danone.com/?lang=en
www.cbi.eu/success-stories/value-your-gold-knowledgemanagement-in-practice-in-bolivia/135806
www.cbi.eu/success-stories/value-your-gold-knowledgemanagement-in-practice-in-bolivia/135806
KNOWLEDGE MANAGEMENT FOUNDATIONS 57
Sabherwal, R., and Becerra-Fernandez, I. 2005. Integrating specific knowledge: Insights from the Ken-
nedy Space Center. IEEE Transactions on Engineering Management, 52(3), 301–315.
Skyrme, D.J. 2000. Developing a knowledge strategy: From management to leadership. In Knowledge
management: Classic and contemporary works, ed. D. Morey, M. Maybury, and B. Thuraisingham,
61–84. Cambridge, MA: The MIT Press.
Standing, C., and Benson, S. 2000. Knowledge management in a competitive environment. In Deci-
sion support through knowledge management, ed. S.A. Carlsson, P. Brezillon, P. Humphreys, B.G.
Lundberg, A.M. McCosh, and V. Rajkovic, 336–348. Stockholm, Sweden: Department of Computing
Systems Sciences, University of Stockholm and Royal Institute of Technology.
Stewart, T.A. 2000. The house that knowledge built. Fortune, October 2.
Storck, J., and Hill, P. 2000. Knowledge diffusion through “strategic communities.” Sloan Manage-
ment Review, 41(2), 63–74.
Thomson, S. 2000. Focus: Keeping pace with knowledge. Information World Review, Issue 155 (Feb-
ruary), 23–24.
Wensley, A. 1998. The value of story telling. Knowledge and Process Management, 5(1), 1–2.
Wong, W., and Radcliffe, D. 2000. The tacit nature of design knowledge. Technology Analysis and
Strategic Management, 12(4), 493–512.
Zander, U., and Kogut, B. 1995. Knowledge and the speed of the transfer and imitation of organizational
capabilities: An empirical test. Organization Science, 6, 76–92.
58
4
Knowledge Management
Solutions: Processes and Systems
In Chapter 3, we provided an introductory discussion of knowledge management
solutions, which refer to the variety of ways in which knowledge management can be
facilitated. We indicated that KM solutions include KM processes and KM systems
and that KM solutions depend on KM foundations, which include KM mechanisms,
technologies, and infrastructure. We discussed KM foundations in detail in Chapter
3. This chapter provides a detailed discussion of KM solutions, including KM pro-
cesses and systems.
The next section describes and illustrates the various processes used to manage
knowledge including processes for applying knowledge, processes for capturing
knowledge, processes for sharing knowledge, and processes for creating knowledge.
In discussing these KM processes, we also examine the seven subprocesses that
facilitate them. The discussion of KM processes is followed by a discussion of KM
systems, followed by a discussion of the processes for managing KM processes and
systems, and then some concluding remarks.
KNOWLEDGE MANAGEMENT PROCESSES
We earlier defined knowledge management as performing the activities involved
in discovering, capturing, sharing, and applying knowledge so as to enhance,
in a cost-effective fashion, the impact of knowledge on the unit’s goal achieve-
ment. Thus, knowledge management relies on four main kinds of KM processes.
As shown in Figure 4.1, these include the processes through which knowledge
is discovered or captured. It also includes the processes through which this
knowledge is shared and applied. These four KM processes are supported by a
set of seven KM subprocesses, as shown in Figure 4.1, with one subprocess—
socialization—supporting two KM processes (discovery and sharing). Of the
seven KM subprocesses, four are based on Nonaka (1994). Focusing on the
ways in which knowledge is converted through the interaction between tacit
and explicit knowledge, Nonaka identified four ways of managing knowledge:
socialization, externalization, internalization, and combination. The other three
KM subprocesses—exchange, direction, and routines—are largely based on Grant
(1996) and Nahapiet and Ghoshal (1998).
KNOWLEDGE MANAGEMENT SOLUTIONS: PROCESSES AND SYSTEMS 59
KNOWLEDGE DISCOVERY
Knowledge discovery may be defined as the development of new tacit or explicit
knowledge from data and information or from the synthesis of prior knowledge. The
discovery of new explicit knowledge relies most directly on combination, whereas the
discovery of new tacit knowledge relies most directly on socialization. In either case,
new knowledge is discovered by synthesizing knowledge from two or more distinct
areas with explicit knowledge from two areas being synthesized through combina-
tion, and tacit knowledge from two areas being synthesized through socialization.
Combination and socialization are discussed now.
Combination
New explicit knowledge is discovered through combination, wherein the multiple
bodies of explicit knowledge (and/or data and/or information) are synthesized to
create new, more complex sets of explicit knowledge (Nonaka 1994). Through com-
munication, integration, and systemization of multiple streams of explicit knowledge,
new explicit knowledge is created—either incrementally or radically (Nahapiet and
Ghoshal 1998). Existing explicit knowledge, data, and information are reconfigured,
recategorized, and recontextualized to produce new explicit knowledge. For example,
when creating a new proposal to a client, explicit data, information, and knowledge em-
bedded in prior proposals may be combined into the new proposal. Also, data mining
techniques may be used to uncover new relationships amongst explicit data that may
be lead to create predictive or categorization models that create new knowledge.
Socialization
In the case of tacit knowledge, the integration of multiple streams for the creation
of new knowledge occurs through the mechanism of socialization (Nonaka 1994).
Capture
Externalization
Internalization
Sharing
Socialization
Exchange
Application
Direction
Routines
Discovery
Combination
Socialization
Figure 4.1 Knowledge Management Processes
60 CHAPTER 4
Socialization is the synthesis of tacit knowledge across individuals, usually through
joint activities rather than written or verbal instructions. For example, by transferring
ideas and images, apprenticeships help newcomers to see how others think. Davenport
and Prusak (1998) described how conversations at the watercooler helped knowledge
sharing among groups at IBM. In Box 4.1, we illustrate the knowledge discovery
process using the example of Xerox.
KNOWLEDGE CAPTURE
As we discussed in Chapter 2, knowledge can exist within people (individuals
or groups), artifacts (practices, technologies, or repositories), and organizational
entities (organizational units, organizations, interorganizational networks). More-
over, knowledge could be either explicit or tacit. It might sometimes reside within
an individual’s mind without that individual being able to recognize it and share
it with others. Similarly, knowledge might reside in an explicit form in a manual
but few people might be aware of it. It is important to obtain the tacit knowledge
from individuals’ minds as well as the explicit knowledge from the manual, such
that the knowledge can then be shared with others. This is the focus of knowledge
Box 4.1
Knowledge Discovery at Xerox
Julian Orr, who was earlier an anthropologist at Xerox’s Palo Alto Research Center (PARC),
studied the actions of customer service representatives who fix Xerox machines. One day, he
observed a representative working with an especially troublesome machine, which had been
recently installed but had never worked properly. Each time the machine failed, it generated a
different error message. Following the prescribed process for each particular message, such as
adjusting or replacing parts, failed to correct the overall problem. Moreover, the messages did
not make sense when considered together.
Frustrated with his inability to fix the troublesome machine, the representative called a spe-
cialist, but the specialist also failed to understand why the machine was behaving in this fash-
ion. Subsequently the representative and the specialist spent the afternoon cycling the machine
repeatedly, waiting for its crashes and recording its state when it crashed. While doing this, they
discussed other incidents of apparently similar problems. “The afternoon resembled a series
of alternating improvisational jazz solos, as each man took the lead, ran with it for a little while,
then handed it off to the other, this all against the bass-line continuo of the rumbling machine”
(Brown and Duguid 2000, p. 78).
During this process, the representative and the specialist gradually brought their different
ideas closer together toward a shared understanding of the machine. Finally, late in the day, ev-
erything clicked. The erratic behavior of the machine, the experiences of the representative and
the specialist, and the stories they both shared eventually formed a single, coherent account.
They were able to make sense of the machine and figure out how to fix it. Thus, by bringing
very different perspectives and experiences and then sharing them during their conversation—
with the problems encountered with the machine providing a common context—they were able
to create new knowledge and thereby solve the problem. Very soon, this new solution was
passed around for other technicians to use if they faced the same problem.
Source: Compiled from Brown and Duguid 2000.
KNOWLEDGE MANAGEMENT SOLUTIONS: PROCESSES AND SYSTEMS 61
capture, which may be defined as the process of retrieving either explicit or tacit
knowledge that resides within people, artifacts, or organizational entities. Also,
the knowledge being captured might reside outside the organizational boundaries
including consultants, competitors, customers, suppliers, and prior employers of
the organization’s new employees.
The knowledge capture process benefits most directly from two KM subprocesses—
externalization and internalization. Based on work by Nonaka (1994), externaliza-
tion and internalization help capture the tacit knowledge and explicit knowledge,
respectively.
Externalization involves converting tacit knowledge into explicit forms such as
words, concepts, visuals, or figurative language (e.g., metaphors, analogies, and nar-
ratives; Nonaka and Takeuchi 1995). It helps translate individuals’ tacit knowledge
into explicit forms that can be more easily understood by the rest of their group.
This is a difficult process because tacit knowledge is often difficult to articulate.
Nonaka (1994) suggested that externalization may be accomplished through the use
of metaphor—that is, understanding and experiencing one kind of thing in terms
of another. An example of externalization is a consultant team writing a document
that describes the lessons the team has learned about the client organization, client
executives, and approaches that work in such an assignment. This captures the tacit
knowledge acquired by the team members.
Internalization is the conversion of explicit knowledge into tacit knowledge. It
represents the traditional notion of learning. The explicit knowledge may be embodied
in action and practice so that the individual acquiring the knowledge can re-experience
what others have gone through. Alternatively, individuals could acquire tacit knowl-
edge in virtual situations, either vicariously by reading manuals or others’ stories or
experientially through simulations or experiments (Nonaka and Takeuchi 1995). An
example of internalization is a new software consultant reading a book on innovative
software development and learning from it. This learning helps the consultant, and
her organization, capture the knowledge contained in the book.
Box 4.2 provides an illustration of knowledge capture.
KNOWLEDGE SHARING
Knowledge sharing is the process through which explicit or tacit knowledge is com-
municated to other individuals. Three important clarifications are in order. First,
knowledge sharing means effective transfer, so that the recipient of knowledge can
understand it well enough to act on it (Jensen and Meckling 1996). Second, what
is shared is knowledge rather than recommendations based on the knowledge; the
former involves the recipient acquiring the shared knowledge as well as being able
to take action based on it, whereas the latter (which is direction, discussed in the next
section) simply involves utilization of knowledge without the recipient internalizing
the shared knowledge. Third, knowledge sharing may take place across individuals
as well as across groups, departments, or organizations (Alavi and Leidner 2001).
If knowledge exists at a location that is different from where it is needed, either
knowledge sharing or knowledge utilization without sharing (discussed in the next
62 CHAPTER 4
section) is necessary. Sharing knowledge is clearly an important process in enhanc-
ing organizational innovativeness and performance. This is reflected in the fact it
was one of the three business processes for which General Electric Company CEO
Jack Welch took personal responsibility (the others were allocation of resources and
development of people) (Stewart 2000).
Depending on whether explicit or tacit knowledge is being shared, exchange or
socialization processes are used. Socialization, which we have discussed above,
facilitates the sharing of tacit knowledge in cases in which new tacit knowledge is
being created as well as when new tacit knowledge is not being created. There is
no intrinsic difference between the socialization process when used for knowledge
discovery or knowledge sharing, although the way in which the process may be used
could be different. For example, when used to share knowledge, a face-to-face meet-
ing (a mechanism that facilitates socialization) could involve a question-and-answer
session between the sender and recipient of knowledge, whereas when used to cre-
ate knowledge a face-to-face meeting could take more the form of a debate or joint
problem-solving, as seen in Box 4.1.
Exchange, in contrast to socialization, focuses on the sharing of explicit knowl-
edge. It is used to communicate or transfer explicit knowledge among individuals,
groups, and organizations (Grant 1996). In its basic nature, the process of exchange
of explicit knowledge does not differ from the process through which information is
communicated. An example of exchange is a product design manual being transferred
by one employee to another, who can then use the explicit knowledge contained in
the manual. Exchanging a document could also be used to transfer information.
Box 4.3 provides an illustration of knowledge sharing.
KNOWLEDGE APPLICATION
Knowledge application is the process through which knowledge is utilized within the
organization to make decisions and perform tasks, thereby contributing to organiza-
Box 4.2
Knowledge Capture at Viant
Viant, the Boston-based company that we discussed in Box 3.4, uses a variety of means to
capture knowledge. It employs a number of simple but unavoidable forms. Before every project,
consultants are required to complete a quicksheet describing the knowledge they will need,
what aspects of knowledge can be leveraged from prior projects, and what they will need to
create along with the lessons they hope to learn that they can share with others later. A longer
report, a sunset review, is produced at a team meeting to document what worked and what did
not work well. Forgetting these reports is hard due to several reasons: “First, almost every docu-
ment ends up on Viant’s internal Web site, hot-linked every which way. Second, sunset reviews
are done with a facilitator who wasn’t on the team, which helps keep them honest. Third, every
six weeks Newell’s knowledge management group prepares, posts, and pushes a summary of
what’s been learned.”
Source: Stewart 2000.
KNOWLEDGE MANAGEMENT SOLUTIONS: PROCESSES AND SYSTEMS 63
tional performance. Of course, the process of knowledge application depends on the
available knowledge, and knowledge itself depends on the processes of knowledge
discovery, capture, and sharing, as shown in Figure 4.1. The better the processes
of knowledge discovery, capture, and sharing, the greater the likelihood that the
knowledge needed is available for effective application in decision-making and task
performance.
In applying knowledge, the party that makes use of it does not necessarily need to
comprehend it. All that is needed is that somehow the knowledge be used to guide
decisions and actions. Therefore, knowledge utilization benefits from two processes—
routines and direction—that do not involve the actual transfer or exchange of knowl-
edge between the concerned individuals but only the transfer of the recommendations
that is applicable in a specific context (Grant 1996).
Direction refers to the process through which the individual possessing the knowl-
edge directs the action of another individual without transferring to that individual the
knowledge underlying the direction. Direction involves the transfer of instructions
or decisions and not the transfer of the knowledge required to make those decisions,
and hence it has been labeled as knowledge substitution (Conner and Prahalad 1996).
This preserves the advantages of specialization and avoids the difficulties inherent
in the transfer of tacit knowledge. Direction is the process used when a production
worker calls an expert to ask her how to solve a particular problem with a machine
and then proceeds to solve the problem based on the instructions given by the expert.
He does this without himself acquiring the knowledge so that if a similar problem
reoccurs in the future, he would be unable to identify it as such and would therefore
be unable to solve it himself without calling an expert. Similarly a student taking
a test who asks his fellow classmate for the answer to a question gets a direction
(which of course could be wrong), and no knowledge is effectively shared between
the two, which means the next time the student faces that question, posed perhaps
Box 4.3
Knowledge Sharing at the Veteran’s Health Administration
Until 1997, the Veteran’s Health Administration (VHA) did not have any systematic mechanism
to enable its 219,000 employees to share their informal knowledge, innovations, and best
practices. To address this need and also to serve as a place where any VHA employee can ac-
cess knowledge capital of colleagues, the VHA Lessons Learned Project and its Web site, the
Virtual Learning Center (VLC), were initiated in 1997. The VHA indicates that a major reason
for initiating this project was a recognized need to transform the organization into a learning
organization. In 1999, the VLC became available on the Internet. The site now has international
participation from Korea, Canada, Spain, Pakistan, and elsewhere. By reducing red tape, cut-
ting across organizational silos, partnering and benchmarking with others, and establishing best
processing, the VHA is “saving countless hours of staff time by not having to reinvent the wheel
at its 173 medical centers, more than 600 clinics, 31 nursing home care units, 206 counseling
centers, and other federal and private healthcare institutions, Veterans Benefits and National
Cemetery offices.”
Source: Compiled from U.S. Department of Veterans Affairs, http://www.va.gov.
http://www.va.gov
64 CHAPTER 4
in a slightly different form, he will not be able to discern the right answer. Note the
difference between direction and socialization or exchange, where the knowledge is
actually transferred to the other person in either tacit form (socialization) or explicit
form (exchange).
Routines involve the utilization of knowledge embedded in procedures, rules, and
norms that guide future behavior. Routines economize on communication more than
directions as they are embedded in procedures or technologies. However, they take
time to develop, relying on “constant repetition” (Grant 1996). Routines could be
automated through the use of IT, such as in systems that provide help desk agents, field
engineers, consultants, and customer endusers with specific and automated answers
from a knowledge base (Sabherwal and Sabherwal, 2007). Similarly, an inventory
management system utilizes considerable knowledge about the relationship between
demand and supply, but neither the knowledge nor the directions are communicated
through individuals. Also, enterprise systems are coded with routines that describe
business process within industry segments.
Next, we examine KM systems that utilize KM mechanisms and technologies to
support the KM processes. In this discussion, we also identify the roles of several
specific KM technologies in enabling KM systems.
KNOWLEDGE MANAGEMENT SYSTEMS
Knowledge management systems are the integration of technologies and mechanisms
that are developed to support the four KM processes. Knowledge management systems
utilize a variety of KM mechanisms and technologies, discussed before, to support
the KM processes discussed in Chapter 3. Each KM system utilizes a combination of
multiple mechanisms and multiple technologies. Moreover, the same KM mechanism
or technology could, under differing circumstances, support multiple KM systems.
Depending on the KM process most directly supported, KM systems can be clas-
sified into four kinds, which are discussed in detail in Part II: knowledge application
systems (Chapter 6), knowledge capture systems (Chapter 7), knowledge sharing
systems (Chapter 8), and knowledge discovery systems (Chapter 9). Here we provide
a brief overview of these four kinds of systems and examine how they benefit from
KM mechanisms and technologies.
KNOWLEDGE DISCOVERY SYSTEMS
As discussed in Chapter 3, knowledge discovery systems support the process of
developing new tacit or explicit knowledge from data and information or from the
synthesis of prior knowledge. These systems support two KM subprocesses associ-
ated with knowledge discovery: combination, enabling the discovery of new explicit
knowledge; and socialization, enabling the discovery of new tacit knowledge.
Thus, mechanisms and technologies can support knowledge discovery systems by
facilitating combination and/or socialization. Mechanisms that facilitate combina-
tion include collaborative problem solving, joint decision-making, and collaborative
creation of documents. For example, at the senior-management level, new explicit
KNOWLEDGE MANAGEMENT SOLUTIONS: PROCESSES AND SYSTEMS 65
knowledge is created by sharing documents and information related to midrange
concepts (e.g., product concepts) augmented with grand concepts (e.g., corporate
vision) to produce new knowledge about both areas. This newly created knowledge
could be, for example, a better understanding of products and a corporate vision
(Nonaka and Takeuchi 1995). Mechanisms that facilitate socialization include ap-
prenticeships, employee rotation across areas, conferences, brainstorming retreats,
cooperative projects across departments, and initiation process for new employees.
For example, Honda Motor Company, Ltd., “set up ‘brainstorming camps’ (tama
dashi kai)—informal meetings for detailed discussions to solve difficult problems in
development projects” (Nonaka and Takeuchi 1995, p. 63).
Technologies facilitating combination include knowledge discovery systems (see
Chapter 9), databases, and Web-based access to data. According to Nonaka and Takeuchi
(1995), “reconfiguration of existing information through sorting, adding, combining,
and categorizing of explicit knowledge (as conducted in computer databases) can lead
to new knowledge” (p. 67). Repositories of information, best practice databases, and
lessons learned systems (see Chapter 8) also facilitate combination. Technologies can
also facilitate socialization, albeit to a lesser extent than they can facilitate combination.
Some of the technologies for facilitating socialization include videoconferencing and
electronic support for communities of practice (see Chapter 10).
KNOWLEDGE CAPTURE SYSTEMS
Knowledge capture systems support the process of retrieving either explicit or tacit
knowledge that resides within people, artifacts, or organizational entities. These
systems can help capture knowledge that resides within or outside organizational
boundaries including within consultants, competitors, customers, suppliers, and prior
employers of the organization’s new employees. Knowledge capture systems rely on
mechanisms and technologies that support externalization and internalization.
KM mechanisms can enable knowledge capture by facilitating externalization—that
is, the conversion of tacit knowledge into explicit form; or internalization—that is,
the conversion of explicit knowledge into tacit form. The development of models or
prototypes and the articulation of best practices or lessons learned are some examples
of mechanisms that enable externalization. Box 4.2, presented earlier, illustrates the
use of externalization to capture knowledge about projects in one organization.
Learning by doing, on-the-job training, learning by observation, and face-to-face
meetings are some of the mechanisms that facilitate internalization. For example, at
one firm, “the product divisions also frequently send their new-product development
people to the Answer Center to chat with the telephone operators or the 12 specialists,
thereby ‘re-experiencing’ their experiences” (Nonaka and Takeuchi 1995, p. 69).
Technologies can also support knowledge capture systems by facilitating external-
ization and internalization. Externalization through knowledge engineering, which
involves integrating knowledge into information systems to solve complex problems that
normally require considerable human expertise” (Feigenbaum and McCorduck 1983),
is necessary for the implementation of intelligent technologies such as expert systems,
case-based reasoning systems (see Chapter 6), and knowledge capture systems (see
66 CHAPTER 4
Chapter 7). Technologies that facilitate internalization include computer-based training
and communication technologies. Using such communication facilities, an individual
can internalize knowledge from a message or attachment thereof sent by another expert,
an AI-based knowledge capture system, or computer-based simulation.
KNOWLEDGE SHARING SYSTEMS
Knowledge sharing systems support the process through which explicit or tacit knowledge
is communicated to other individuals. They do so by supporting exchange (i.e., sharing of
explicit knowledge) and socialization (which promotes sharing of tacit knowledge).
Mechanisms and technologies that were discussed as supporting socialization also
play an important role in knowledge sharing systems. Discussion groups or chat groups
facilitate knowledge sharing by enabling an individual to explain her knowledge to
the rest of the group. In addition, knowledge sharing systems also utilize mechanisms
and technologies that facilitate exchange. Some of the mechanisms that facilitate ex-
change are memos, manuals, progress reports, letters, and presentations. Technologies
facilitating exchange include groupware and other team-collaboration mechanisms;
Web-based access to data and databases; and repositories of information, including
best practice databases, lessons learned systems, and expertise locator systems. Box
4.3 on the Veteran’s Health Administration (VHA), which was presented earlier,
provides one illustration of the importance of knowledge sharing.
KNOWLEDGE APPLICATION SYSTEMS
Knowledge application systems support the process through which some individuals
utilize knowledge possessed by other individuals without actually acquiring, or learn-
ing, that knowledge. Mechanisms and technologies support knowledge application
systems by facilitating routines and direction.
Mechanisms facilitating direction include traditional hierarchical relationships
in organizations, help desks, and support centers. On the other hand, mechanisms
supporting routines include organizational policies, work practices, organizational
procedures, and standards. In the case of both direction and routines, these mecha-
nisms may be either within an organization (e.g., organizational procedures) or across
organizations (e.g., industry best practices).
Technologies supporting direction include experts’ knowledge embedded in expert
systems (see Chapter 8) and decision-support systems, as well as troubleshooting sys-
tems based on the use of technologies like case-based reasoning. On the other hand,
some of the technologies that facilitate routines are expert systems (see Chapter 6),
enterprise resource planning systems, and traditional management information systems.
As mentioned for KM mechanisms, these technologies can also facilitate directions and
routines within or across organizations. These are discussed in detail in Chapter 6.
Box 4.4 provides an illustration of a knowledge application. Moreover, Box 4.5
provides an illustration of KM technologies.
Table 4.1 summarizes the discussion of KM processes and KM systems, and also
indicates some of the mechanisms and technologies that might facilitate them. As
KNOWLEDGE MANAGEMENT SOLUTIONS: PROCESSES AND SYSTEMS 67
Box 4.4
Automated Knowledge Application at DeepGreen Financial
Based in Cleveland, Ohio, DeepGreen Financial (which was acquired in March 2004 by Light-
year Capital, a New York-based private equity investment firm) has revolutionized the mortgage
industry by providing low-rate, home equity products that are easy to apply for and obtain over
the Internet. DeepGreen even offers to close the loan at the borrower’s home. DeepGreen’s
efficient and innovative technology has reduced the cost of loan production, which they pass on
to the consumer. An Internet-only home equity lender, DeepGreen originates loans in 47 states
and makes them available through its Web site and through partners such as LendingTree, LLC,
Priceline, and Costco Wholesale Corporation. DeepGreen originates home equity products at
five times the industry average in terms of dollars per employee.
Right from its start in August 2000, DeepGreen has been based on efficient knowledge utiliza-
tion. Right from the firm’s creation, the vision for it was to rely on automated decision technol-
ogy. DeepGreen created an Internet-based system that makes credit decisions within minutes
by selecting the customers with the best credit. Efficient knowledge utilization through routines
embedded as rules within an automated system, along with efficient use of online information, en-
abled only eight employees to process about 400 applications daily. Instead of competing on the
basis of interest rates, DeepGreen competed in terms of ease of application (a customer could
complete the application within five minutes) and by providing nearly instantaneous, unconditional
decisions without requiring the borrowers to provide the usual appraisals or paperwork upfront.
This quick decision is enabled through knowledge application and the computation of credit score
and property valuation using online data. In about 80 percent of the cases, a final decision is
provided to the customer within two minutes of the application being completed. Online Banking
Report named DeepGreen’s home equity lines of credit as the “Best of the Web.”
Source: Compiled from Davenport and Harris 2005; Harris and Brooks 2004; http://www.home-
equity-info.us/lenders_banks_deepgreen-bank.php; and http://www.deepgreenfinancial.com/.
Box 4.5
KM Technologies at Cisco
Cisco Systems Inc., utilizes Directory 3.0, which is its internal Facebook, in which the employee
listings are designed to identify the employee’s expertise area and promote collaboration. To
further promote knowledge sharing, it utilizes a variety of technologies including: Ciscopedia,
which is an internal document site; C-Vision, which is Cisco’s version of YouTube; and the Idea
Zone, which is a wiki for employees to post and discuss business ideas. Cisco has also been
developing a companywide social computing platform to enable knowledge creation and shar-
ing through strengthening of existing networks and facilitation of new connections (Fitzgerald
2008). According to Cisco’s VP, Communication and Collaboration IT: “We are always looking
for the applications that help people really have water-cooler talk, something that we thought
was impossible in a global business” (McGirt 2008).
Cisco’s CIO remarked in 2008 (Fitzgerald 2008): “CIOs need to consider issues of privacy,
data security, and the ability to scale across a global organization. It’s no good, if 15 different
business units develop 15 different online communities that can’t talk to each other.”
Sources: E. McGirt. “How Cisco’s CEO John Chambers Is Turning the Tech Giant Socialist,” Fast-
Company, December 1, 2008, http://www.fastcompany.com/1093654/how-ciscos-ceo-john- chambers-
turning-tech-giant-socialist. M. Fitzgerald, “Why Social Computing Aids Knowledge Management,”
CIO, June 13, 2008, http://www.cio.com/article/2435683/enterprise-software/why-social-computing-
aids- knowledge-management.html.
http://www.homeequity-info.us/lenders_banks_deepgreen-bank.php; and http://www.deepgreenfinancial.com/
http://www.homeequity-info.us/lenders_banks_deepgreen-bank.php; and http://www.deepgreenfinancial.com/
http://www.fastcompany.com/1093654/how-ciscos-ceo-john-chambersturning-tech-giant-socialist
http://www.fastcompany.com/1093654/how-ciscos-ceo-john-chambersturning-tech-giant-socialist
http://www.cio.com/article/2435683/enterprise-software/why-social-computingaids-knowledge-management.html
http://www.cio.com/article/2435683/enterprise-software/why-social-computingaids-knowledge-management.html
68
Ta
bl
e
4.
1
K
M
P
ro
ce
ss
es
a
n
d
S
ys
te
m
s,
a
n
d
A
ss
o
ci
at
ed
M
ec
h
an
is
m
s
an
d
T
ec
h
n
o
lo
g
ie
s
K
M
P
ro
ce
ss
es
K
M
S
ys
te
m
s
K
M
S
ub
pr
oc
es
se
s
Ill
us
tr
at
iv
e
K
M
M
ec
ha
ni
sm
s
Ill
us
tr
at
iv
e
K
M
T
ec
hn
ol
og
ie
s
K
no
w
le
dg
e
D
is
co
ve
ry
K
no
w
le
dg
e
D
is
co
ve
ry
S
ys
te
m
s
C
om
bi
na
tio
n
M
ee
tin
gs
, t
el
ep
ho
ne
c
on
ve
rs
at
io
ns
,
do
cu
m
en
ts
, a
nd
c
ol
la
bo
ra
tiv
e
cr
ea
tio
n
of
d
oc
um
en
ts
D
at
ab
as
es
, W
eb
-b
as
ed
a
cc
es
s
to
d
at
a,
d
at
a
m
in
in
g,
r
ep
os
ito
rie
s
of
in
fo
rm
at
io
n,
W
eb
p
or
ta
ls
, b
es
t
pr
ac
tic
es
, a
nd
le
ss
on
s
le
ar
ne
d
S
oc
ia
liz
at
io
n
E
m
pl
oy
ee
r
ot
at
io
n
ac
ro
ss
de
pa
rt
m
en
ts
, c
on
fe
re
nc
es
,
br
ai
ns
to
rm
in
g
re
tr
ea
ts
, c
oo
pe
ra
tiv
e
pr
oj
ec
ts
, i
ni
tia
tio
n
V
id
eo
-c
on
fe
re
nc
in
g,
e
le
ct
ro
ni
c
di
sc
us
si
on
g
ro
up
s,
a
nd
e
-m
ai
l
K
no
w
le
dg
e
C
ap
tu
re
K
no
w
le
dg
e
C
ap
tu
re
S
ys
te
m
s
E
xt
er
na
liz
at
io
n
M
od
el
s,
p
ro
to
ty
pe
s,
b
es
t p
ra
ct
ic
es
,
le
ss
on
s
le
ar
ne
d
E
xp
er
t s
ys
te
m
s,
c
ha
t g
ro
up
s,
b
es
t
pr
ac
tic
es
, a
nd
le
ss
on
s
le
ar
ne
d
da
ta
ba
se
s
In
te
rn
al
iz
at
io
n
Le
ar
ni
ng
b
y
do
in
g,
o
n-
th
e-
jo
b
tr
ai
ni
ng
, l
ea
rn
in
g
by
o
bs
er
va
tio
n,
an
d
fa
ce
-t
o-
fa
ce
m
ee
tin
gs
C
om
pu
te
r-
ba
se
d
co
m
m
un
ic
at
io
n,
A
I-
ba
se
d
kn
ow
le
dg
e
ac
qu
is
iti
on
, a
nd
co
m
pu
te
r-
ba
se
d
si
m
ul
at
io
ns
K
no
w
le
dg
e
S
ha
rin
g
K
no
w
le
dg
e
S
ha
rin
g
S
ys
te
m
s
S
oc
ia
liz
at
io
n
S
ee
a
bo
ve
S
ee
a
bo
ve
E
xc
ha
ng
e
M
em
os
, m
an
ua
ls
, l
et
te
rs
, a
nd
pr
es
en
ta
tio
ns
Te
am
c
ol
la
bo
ra
tio
n
to
ol
s,
W
eb
–
ba
se
d
ac
ce
ss
to
d
at
a,
d
at
ab
as
es
,
an
d
re
po
si
to
rie
s
of
in
fo
rm
at
io
n,
b
es
t
pr
ac
tic
es
d
at
ab
as
es
, l
es
so
ns
le
ar
ne
d
sy
st
em
s,
a
nd
e
xp
er
tis
e
lo
ca
to
r
sy
st
em
s
K
no
w
le
dg
e
A
pp
lic
at
io
n
K
no
w
le
dg
e
A
pp
lic
at
io
n
S
ys
te
m
s
D
ire
ct
io
n
Tr
ad
iti
on
al
h
ie
ra
rc
hi
ca
l r
el
at
io
ns
hi
ps
in
o
rg
an
iz
at
io
ns
, h
el
p
de
sk
s,
a
nd
su
pp
or
t c
en
te
rs
C
ap
tu
re
a
nd
tr
an
sf
er
o
f e
xp
er
ts
’
kn
ow
le
dg
e,
tr
ou
bl
es
ho
ot
in
g
sy
st
em
s,
an
d
ca
se
-b
as
ed
r
ea
so
ni
ng
s
ys
te
m
s;
de
ci
si
on
s
up
po
rt
s
ys
te
m
s
R
ou
tin
es
O
rg
an
iz
at
io
na
l p
ol
ic
ie
s,
w
or
k
pr
ac
tic
es
, a
nd
s
ta
nd
ar
ds
E
xp
er
t s
ys
te
m
s,
e
nt
er
pr
is
e
re
so
ur
ce
pl
an
ni
ng
s
ys
te
m
s,
a
nd
m
an
ag
em
en
t
in
fo
rm
at
io
n
sy
st
em
s
KNOWLEDGE MANAGEMENT SOLUTIONS: PROCESSES AND SYSTEMS 69
may be seen from this table, the same tool or technology can be used to support more
than one KM process.
MANAGING KNOWLEDGE MANAGEMENT SOLUTIONS
The management of KM systems will be discussed in Chapters 6 through 9, which will
examine each of the four types of KM systems in greater detail. Moreover, the selection
of KM processes and KM systems that would be most appropriate for the circumstances
will be discussed in Chapter 11. Finally, the overall leadership of the KM function will
be discussed in Chapter 12. Therefore, in this section, we focus on some overall recom-
mendations regarding the management of KM processes and systems.
First, organizations should use a combination of the four types of KM processes
and systems. Although different KM processes may be most appropriate in the light
of the organization’s business strategy, focusing exclusively on one type of KM pro-
cesses (and the corresponding type of KM systems) would be inappropriate because
they serve complementary objectives. More specifically, it is important to note the
following:
knowledge application could reduce knowledge creation, which often benefits
from individuals viewing the same problem from multiple different perspectives
and thereby leads to reduced effectiveness and innovation.
or from explicit form to tacit, and thereby facilitates knowledge sharing. However,
it might lead to reduced attention to knowledge creation. Moreover, knowledge
capture could lead to some knowledge being lost in the conversion process; not
all tacit knowledge is converted into explicit form during externalization, and
not all explicit knowledge is converted into tacit form during internalization.
much knowledge sharing could lead to knowledge leaking from the organization
and becoming available to competitors, and consequently reduce the benefits to
the focal organization.
knowledge discovery could lead to reduced efficiency. It is not always suitable to
create new knowledge, just as it may not always be appropriate to reuse existing
knowledge.
Second, each KM process could benefit from two different subprocesses, as depicted
in Figure 4.1. The subprocesses are mutually complementary, and should be used
depending on the circumstances as discussed in Chapter 11. For example, knowledge
sharing could occur through socialization or exchange. If knowledge being shared
is tacit in nature, socialization would be appropriate, whereas if knowledge being
shared is explicit in nature, exchange would be suitable. However, when individuals
need to share both tacit and explicit knowledge, the two subprocesses (socialization
and exchange) could be integrated together, such as in a face-to-face meeting (i.e.,
70 CHAPTER 4
using socialization to transfer tacit knowledge) where the participants are also sharing
printed reports containing explicit knowledge (i.e., using exchange to transfer explicit
knowledge). Overall, the seven KM subprocesses should be developed within a group
such that they can complement each other in an efficient fashion.
Third, each of the seven KM subprocesses of the KM processes depends on the
KM mechanisms and technologies, as discussed before. Moreover, the same mecha-
nism could be used to support multiple different subprocesses. Development and
acquisition of these mechanisms and technologies, respectively, should be done in
the light of the KM processes that would be most appropriate for the organizational
circumstances.
Finally, the KM processes and systems should be considered in the light of each other,
so that the organization builds a portfolio of mutually complementary KM processes
and systems over time. This requires involvement from senior executives, a long-term
KM strategy for the organization, and an understanding of the synergies as well as
common foundations (i.e., mechanisms and technologies that might support multiple
KM systems and processes) across the various KM systems and processes.
ALIGNMENT BETWEEN KNOWLEDGE MANAGEMENT AND
BUSINESS STRATEGY
Alignment between business strategy and knowledge management helps enhance
organizational performance (Kaplan and Norton 2004). Greater alignment between
a firm’s business strategy and its KM efforts indicates that these efforts are targeted
on areas that are critical to the firm’s success. For knowledge to become a source of
competitive advantage, firms need to match their learning and knowledge strategy
with their business strategy.
When a firm’s learning and knowledge strategy matches its business strategy, the
impact of knowledge and learning is positive. If this match is not achieved, knowledge
and learning may have no impact or even have a negative impact on performance
(Vera and Crossan 2004).
SUMMARY
Building on the discussion of knowledge management foundations in Chapter 3, we have
examined KM solutions, including KM processes and systems, in this chapter. Figure
4.2 provides a summary of the various aspects of knowledge management, indicating
the various aspects of KM processes (including the four overall processes as well as
the seven specific processes that support them), KM systems, KM mechanisms and
technologies, and KM infrastructure. The next chapter examines the value of knowledge
and KM solutions, highlighting their importance for organizational performance.
REVIEW
1. Give an example of one knowledge management mechanism that could be
used to facilitate each of the four knowledge management processes.
KNOWLEDGE MANAGEMENT SOLUTIONS: PROCESSES AND SYSTEMS 71
2. Give an example of one knowledge management technology that could be
used to facilitate each of the four knowledge management processes.
3. Briefly explain the four kinds of classifications for knowledge management
systems based on the process supported.
4. Distinguish between direction and routines. How do they relate to knowledge
substitution?
5. Socialization could be used for knowledge discovery as well as knowledge
sharing. Would the underlying process be any different depending on whether
it is being used for knowledge discovery or knowledge sharing?
6. How does knowledge management relate to business strategy?
7. What effect does alignment between knowledge management and business
strategy have on organizational performance?
8. Tacit knowledge could be transferred from one person to another in two
distinct ways. One possibility is to transfer it directly through socializa-
tion. The other possibility is to convert it into explicit form (through
externalization), then transfer it in explicit form to the recipient (through
exchange), who then converts it into tacit form (through internalization).
What are the pros and cons of each approach? If the purpose is to transfer
knowledge from one person to one other person, which approach would
you recommend? If the purpose is to transfer knowledge from one person
to 100 other individuals in different parts of the world, which approach
would you recommend? Why?
KM Processes
KM Mechanisms
KM Infrastructure
KM Technologies
ExternalizationCombination RoutinesSocialization Exchange DirectionInternalization
Knowledge
Capture Knowledge
Sharing
Knowledge
Application
Physical
Environment
Knowledge
Discovery
KM Systems
Knowledge
Discovery
Systems
Knowledge
Capture
Systems
Knowledge
Sharing
Systems
Knowledge
Application
Systems
Physical
Environment
Knowledge
Discovery
Systems
Knowledge
Capture
Systems
Knowledge
Sharing
Systems
Knowledge
Application
Systems
Organization
Culture
Organization
Culture
Organization
Structure
Organization
Structure
IT
Infrastructure
IT
Infrastructure
Common
Knowledge
Common
Knowledge
Analogies and metaphors
Brainstorming ret reats
On-the-job training
Face-to-face meetings
Apprenticeships
Employee rotation
Learning by observation
….
Analogies and metaphors
Brainstorming ret reats
On-the-job training
Face-to-face meetings
Apprenticeships
Employee rotation
Learning by observation
…. ….
Decision support systems
Web-based discussion groups
Repositories of best practices
Artificial intelligence systems
Case-based reasoning
Groupware
Web pages
…
Decision support systems
Web-based discussion groups
Repositories of best practices
Artificial intelligence systems
Case-based reasoning
Groupware
Web pages
……
Figure 4.2 A Detailed View of Knowledge Management Solutions
72 CHAPTER 4
APPLICATION EXERCISES
1. How would you, as a CEO of a manufacturing firm, facilitate the growth of
knowledge management practices within your organization?
2. How would you utilize knowledge discovery systems and knowledge capture
systems in an organization that is spread across the globe? Does geographic
distance hamper the utilization of these systems?
3. Suggest reasons why a knowledge sharing system could be established be-
tween rival organizations (e.g., Mastercard Inc. and Visa Inc.) for the mutual
benefit of both organizations.
4. Critique the following statement: “We have implemented several IT solutions:
expert systems, chat groups, and best practices/lessons learned databases.
These powerful solutions will surely induce our employees to internalize
knowledge.”
5. Consider the organization where you currently work or one with which you
are familiar (either through your own prior experience or through interactions
with someone who works there). What kind of knowledge management sys-
tems and processes does this organization use to manage knowledge? What
are their effects on this organization’s performance? In what order did the
organization develop these KM systems and processes, and why?
6. Interview at least three managers from local organizations that have recently
implemented a knowledge management system. How do these organizations
differ in terms of the KM systems they have developed? What reasons led
these organizations to develop these systems?
REFERENCES
Alavi, M., and Leidner, D. 2001. Knowledge management and knowledge management systems:
Conceptual foundations and research issues. MIS Quarterly, 25(1), 107–136.
Brown, J.S., and Duguid, P. 2000. Balancing act: How to capture knowledge without killing it. Harvard
Business Review (May–June), 73–80.
Conner, K.R., and Prahalad, C.K. 1996. A resource-based theory of the firm: Knowledge versus op-
portunism. Organization Science, 7(5), 477–501.
Davenport, T.H., and Harris, J.G. 2005. Automated decision making comes of age. Sloan Management
Review, 46(4) (Summer), 83–89.
Davenport, T., and Prusak, L. 1998. Working knowledge. Boston, MA: Harvard Business School
Press.
Feigenbaum, E., and McCorduck, P. 1983. The fifth generation. Reading, MA: Addison-Wesley.
Grant, R.M. 1996. Toward a knowledge-based theory of the firm. Strategic Management Journal, 17,
109–122.
Harris, J.G., and Brooks, J.D. 2004. In the mortgage industry, IT matters. Mortgage Banking, 65(3)
(December), 62–66.
Jensen, M.C., and Meckling, W.H. 1996. Specific and general knowledge, and organizational struc-
ture. In Knowledge management and organizational design, ed. P.S. Myers, 17–18. Newton, MA:
Butterworth-Heinemann.
Kaplan, R.S., and Norton, D.P. 2004. Measuring the strategic readiness of intangible assets. Harvard
Business Review (February), 52–63.
KNOWLEDGE MANAGEMENT SOLUTIONS: PROCESSES AND SYSTEMS 73
McKellar, H. 2001. The first annual KM World awards. The Fifth Annual KM World 2001 Conference
and Exposition, October 29–November 1. www.infotoday.com/kmw01/kmawards.htm.
Nahapiet, J., and Ghoshal, S. 1998. Social capital, intellectual capital, and the organizational advantage.
Academy of Management Review, 23(2), 242–266.
Nonaka, I. 1994. A dynamic theory of organizational knowledge creation. Organization Science, 5(1)
(February), 14–37.
Nonaka, I., and Takeuchi, H. 1995. The knowledge creating company: How Japanese companies create
the dynamics of innovation. New York: Oxford University Press.
Sabherwal, R., and Sabherwal, S. 2007. How do knowledge management announcements affect firm
value? A study of firms pursuing different business strategies. IEEE Transactions on Engineering
Management, 54(3) (August), 409–422.
Stewart, T.A. 2000. The house that knowledge built. Fortune, October 2.
Vera, D., and Crossan, M. 2004. Strategic leadership and organizational learning. Academy of Manage-
ment Review, 29(2), 222–240.
www.infotoday.com/kmw01/kmawards.htm
74
5
Organizational Impacts of
Knowledge Management
In the previous two chapters, we examined what we mean by knowledge manage-
ment and discussed KM foundations including KM infrastructure, mechanisms, and
technologies; and KM solutions including KM processes and solutions. In this chapter,
we examine the impacts of knowledge management. Consistent with our emphasis
on the use of KM in organizations, we focus our discussion on the impact of KM on
companies and other private or public organizations.
The importance of knowledge (and KM processes) is well recognized. According
to Benjamin Franklin, “An investment in knowledge pays the best interest” (NASA
2007). KM can impact organizations and organizational performance at several levels:
people, processes, products, and the overall organizational performance (Becerra-
Fernandez and Sabherwal 2008). KM processes can impact organizations at these
four levels in two main ways. First, KM processes can help create knowledge, which
can then contribute to improved performance of organizations along the above four
dimensions. Second, KM processes can directly cause improvements along these
four dimensions. These two ways in which KM processes can impact organizations
is summarized in Figure 5.1.
Figure 5.1 How Knowledge Management Impacts Organizations
Knowledge OrganizationKnowledge Management Knowledge OrganizationKnowledge Management
Figure 5.2 depicts the impacts of KM on the four dimensions mentioned above and
shows how the effect on one dimension can have an impact on another. The impact at
three of these dimensions—individuals, products, and the organization—was clearly
indicated in a joint survey by IDC1 and Knowledge Management Magazine in May
2001 (Dyer and McDonough 2001). This survey examined the status of knowledge
management practices in U.S. companies, and found three top reasons why U.S. firms
adopt knowledge management: (1) retaining expertise of employees, (2) enhancing
customers’ satisfaction with the company’s products, and (3) increasing profits or
revenues. We will examine these issues closely in the next four sections.
ORGANIZATIONAL IMPACTS OF KNOWLEDGE MANAGEMENT 75
IMPACT ON PEOPLE
Knowledge management can affect an organization’s employees in several ways.
First of all, it can facilitate their learning (from each other as well as from external
sources). This learning by individual employees allows the organization to be con-
stantly growing and changing in response to the market and the technology (Sabherwal
2008). Knowledge management also causes the employees to become more flexible
and enhances their job satisfaction. This is largely because of their enhanced ability
to learn about solutions to business problems that worked in the past, as well as those
solutions that did not work. These effects are now discussed.
IMPACT ON EMPLOYEE LEARNING
Knowledge management can help enhance the employee’s learning and exposure to
the latest knowledge in their fields. This can be accomplished in a variety of ways
including externalization and internalization, socialization, and communities of prac-
tice, which were all discussed in Chapter 3.
We earlier described externalization as the process of converting tacit knowledge
into explicit forms, and internalization as the conversion of explicit knowledge into
tacit knowledge (Nonaka and Takeuchi 1995). Externalization and internalization
work together in helping individuals learn. One possible example of externalization
is preparing a report on lessons learned from a project. In preparing the report, the
team members document, or externalize, the tacit knowledge they have acquired
during the project. Individuals embarking on later projects can then use this report
to acquire the knowledge gained by the earlier team. These individuals acquire tacit
knowledge through internalization—that is, by reading the explicit report and thereby
re-experiencing what others have gone through. Thus, an expert writing a book is
externalizing her knowledge in that area, and a student reading the book is acquiring
tacit knowledge from the knowledge explicated in the book.
Socialization also helps individuals acquire knowledge but usually through joint
activities such as meetings, informal conversations, and so on. One specific, but
important, way in which learning through socialization can be facilitated involves
Figure 5.2 Dimensions of Organizational Impacts of Knowledge Management
People ProductsProcesses
Organizational
Performance
Knowledge Management
People ProductsProcesses
Organizational
Performance
Knowledge Management
76 CHAPTER 5
the use of a community of practice, which we defined in Chapter 3 as an organic and
self-organized group of individuals who may be dispersed geographically or orga-
nizationally but communicate regularly to discuss issues of mutual interest. In Box
5.1, we describe how one organization was able to enable individual learning via the
implementation of communities of practice.
The experience of Xerox illustrates the way in which knowledge management can
enable the organization’s employees to learn from each other as well as from prior
experiences of former employees. It is also indicative of how such processes for
individual learning can lead to continued organizational success.
IMPACT ON EMPLOYEE ADAPTABILITY
When the knowledge management process at an organization encourages its em-
ployees to continually learn from each other, the employees are likely to possess the
information and knowledge needed to adapt whenever organizational circumstances
so require. Moreover, when they are aware of ongoing and potential future changes,
Box 5.1
Strategic Communities of Practice at Xerox
Xerox Corporation enabled individual learning through a strategic community of practice. Con-
sistent with our definition of a community of practice, the groups at Xerox included geographi-
cally distributed individuals from the headquarters as well as business units. However, these
groups were somewhat different from a traditional community of practice because they were not
voluntarily formed by the individuals themselves but instead were deliberately established by
the top management at Xerox with the goal of providing strategic benefits through knowledge
sharing. This is the reason Storck and Hill (2000) characterized them as “strategic” communi-
ties. One of these strategic communities, which had been tasked to help in the management of
technology infrastructure, consisted of a large group of information technology professionals
who provided leading-edge solutions, addressed unstructured problems, and stayed in touch
with the latest developments in hardware and software.
According to the group members surveyed by Storck and Hill, about two-thirds of the group’s
value resulted from face-to-face networking at the group’s meetings. This attention to knowledge
management by focusing on informal groups of employees has helped Xerox in its recent push
in global services. Jim Joyce, a senior executive at Xerox remarked: “It is about understanding
where knowledge is and how it is found. By working with human elements of this, there are real
things you can do to help people embrace the technology and incorporate it into the workflow”
(Moore 2001). Similarly, Tom Dolan, president, Xerox Global Services, recognized: “At the core
of Xerox’s heritage of innovation is a deep understanding of how people, processes and tech-
nology interact with each other in the creation of great work. As a result, our practical, results-
oriented, knowledge management solutions can help businesses streamline work processes,
enable better customer service, and grow revenue” (Business Wire 2002).
Xerox has continued the use of communities of practice, with about 15 learning communi-
ties, including more than 1,000 employees being launched in 2007 and 2008. According to Kent
Purvis, a managing principal with Xerox’s global services division: “We know there is a ground-
swell of knowledge among our managing principals, along all the lines of business. Now there is
a structure in place for sharing it” (Kranz 2008).
Source: Compiled from Business Wire 2002; Kranz 2008.
ORGANIZATIONAL IMPACTS OF KNOWLEDGE MANAGEMENT 77
they are less likely to be caught by surprise. Awareness of new ideas and involvement
in free-flowing discussions not only prepare them to respond to changes, but they also
make them more likely to accept change. Thus, knowledge management is likely to
engender greater adaptability among employees.
When Buckman Laboratories International, Inc., a privately owned U.S. specialty
chemicals firm with about 1,300 employees, was named “the 2000 Most Admired
Knowledge Enterprise,” its chairman, Bob Buckman, remarked that the company’s
knowledge management efforts were intended to continually expose its employees to
new ideas and enable them to learn from them (Business Wire 2000). He also empha-
sized that the employees were prepared for change as a result of being in touch with
the latest ideas and developments, and they consequently embraced change rather
than being afraid of it. The increased employee adaptability due to knowledge man-
agement enabled the company to become a very fast-changing organization around
the needs of its customers. Buckman Laboratories has subsequently won the Most
Admired Knowledge Enterprise award in 2001, 2003, 2004, 2005, and 2006, and
been nominated Best Practice Partner by American Productivity and Quality Center
(APQC) for their contributions to Leveraging Knowledge Across the Value Chain in
2006 (Buckman Laboratories International 2007).
IMPACT ON EMPLOYEE JOB SATISFACTION
Two benefits of knowledge management that accrue directly to individual employ-
ees have been discussed above: (a) They are able to learn better than employees
in firms that are lacking in KM, and (b) they are better prepared for change. These
impacts cause the employees to feel better because of the knowledge acquisition and
skill enhancement and also the impacts enhance their market value relative to other
organizations’ employees. A recent study found that in organizations having more
employees sharing knowledge with one another, turnover rates were reduced thereby
positively affecting revenue and profit (Bontis 2003). Indeed, exit interview data in
this study indicated that one of the major reasons many of the brightest knowledge
workers changed jobs was because “they felt their talent was not fully leveraged.” Of
course, it is possible to argue for the reverse causal direction; that is, more satisfied
employees are likely to be more willing to share knowledge. The causal direction of
the relationship between employee job satisfaction and knowledge sharing needs to
be researched further.
In addition, knowledge management also provides employees with solutions to
problems they face in case those same problems have been encountered earlier and
effectively addressed. This provision of tried-and-tested solutions (for example,
through the direction mechanism discussed in Chapter 3) amplifies employees’ ef-
fectiveness in performing their jobs. This helps keep those employees motivated, for
a successful employee would be highly motivated while an employee facing problems
in performing his job would likely be demotivated.
Thus, as a result of their increased knowledge, improved market value, and greater on-
the-job performance, knowledge management facilitates employees’ job satisfaction. In
addition, some approaches for knowledge management, such as mentoring and training,
78 CHAPTER 5
are also directly useful in motivating employees and therefore increasing employee job
satisfaction. Similarly, communities of practice provide the involved employees intimate
and socially validated control over their own work practices (Brown and Duguid 1991).
Figure 5.3 summarizes the above impacts knowledge management and knowledge
can have on employees of organizations.
Figure 5.3 How Knowledge Management Impacts People
Knowledge Employee Learning
Knowledge
Management
Employee
Adaptability
Employee
Job Satisfaction
Knowledge Employee Learning
Knowledge
Management
Employee
Adaptability
Employee
Job Satisfaction
IMPACT ON PROCESSES
Knowledge management also enables improvements in organizational processes such
as marketing, manufacturing, accounting, engineering, public relations, and so forth.
These impacts can be seen along three major dimensions: effectiveness, efficiency,
and degree of innovation of the processes. These three dimensions can be character-
ized as follows:
Effectiveness: Performing the most suitable processes and making the best pos-
sible decisions.
Efficiency: Performing the processes quickly and in a low-cost fashion.
Innovation: Performing the processes in a creative and novel fashion that im-
proves effectiveness and efficiency—or at least marketability.
Knowledge management can improve the above interrelated aspects of organiza-
tional processes through several means, including better knowledge being imparted
to individuals (through exchange, socialization, and so on) and the provision of work-
ORGANIZATIONAL IMPACTS OF KNOWLEDGE MANAGEMENT 79
able solutions (through directions and routines), for employees to solve the problems
faced in their tasks. The effects of KM on effectiveness, efficiency, and innovation
are discussed in more detail below.
IMPACT ON PROCESS EFFECTIVENESS
Knowledge management can enable organizations to become more effective by help-
ing them to select and perform the most appropriate processes. Effective knowledge
management enables the organization’s members to collect information needed to
monitor external events. This results in fewer surprises for the leaders of the organiza-
tion and consequently reduces the need to modify plans and settle for less effective
approaches. In contrast, poor knowledge management can result in mistakes by the
organization because they risk repeating past mistakes or not foreseeing otherwise
obvious problems. For example, Ford Motor Company and Firestone (now part of
Bridgestone Corporation) incurred numerous problems which may have been reduced
through greater knowledge sharing, either by exchanging explicit knowledge and
information or by using meetings (and other means of socialization) to share tacit
knowledge. These firms did possess the necessary information to warn them about
the mismatch of Ford Explorers and Firestone tires. However, the information was
not integrated across the two companies, which might have inhibited either company
from having the “full picture.” It is interesting to note that although Ford had a good
knowledge management process (the Best Practices Replication Process, discussed
later in this chapter), it was not used to manage the information and knowledge relat-
ing to the Ford Explorer and the Firestone tires, or identify the potential risk of the
tire’s tread peeling off, leading to tire disintegration with the likelihood of accident
in case the vehicle was then traveling at a high speed (Stewart 2000). The result was
significant loss in lives for their customers and unprecedented legal liability.
Knowledge management enables organizations to quickly adapt their processes
according to the current circumstances, thereby maintaining process effectiveness in
changing times. On the other hand, organizations lacking in knowledge management
find it difficult to maintain process effectiveness when faced with turnover of experi-
enced and new employees. An illustrative example is from a large firm that reorganized
its engineering department in 1996. This reorganization achieved a 75 percent reduc-
tion of the department’s workforce. An external vendor subsequently absorbed many
of the displaced engineers. However, like many organizations undergoing significant
downsizing, this company failed to institutionalize any mechanisms to capture the
knowledge of the employees that were leaving the department. A two-month review
of the results following the reorganization effort showed that several key quality in-
dicators were not met. This was a direct result of the loss of human knowledge with
the displacement of the workforce. One important reason for the lack of attention
to retaining knowledge is that the alternative approaches for capturing individual
knowledge (which were discussed in Chapter 3) are not well understood. We will
discuss some of these methods and technologies to capture knowledge in Chapter 7.
Box 5.2 illustrates how one particular organization was able to significantly improve
its processes in disaster management response through effective KM.
80 CHAPTER 5
IMPACT ON PROCESS EFFICIENCY
Managing knowledge effectively can also enable organizations to be more productive
and efficient. Upon exploring the “black box” of knowledge sharing within Toyota Motor
Corporation’s network, Dyer and Nobeoka (2000, p. 364) found that “Toyota’s ability
to effectively create and manage network-level knowledge sharing processes, at least
partially, explains the relative productivity advantages enjoyed by Toyota and its suppli-
ers.” Knowledge diffusion was found to occur more quickly within Toyota’s production
network than in competing automaker networks. This was because Toyota’s network
had solved three fundamental dilemmas with regard to knowledge sharing by devising
methods to: (1) motivate members to participate and openly share valuable knowledge
(while preventing undesirable spillovers to competitors); (2) prevent free riders—that is,
individuals who learn from others without helping others learn; and (3) reduce the costs
associated with finding and accessing different types of valuable knowledge.
Box 5.2
Knowledge Management at Tearfund
Tearfund1 is a large relief and development agency, based in the United Kingdom. It
regularly responds to natural and humanitarian disasters such as floods, hurricanes, ty-
phoons, famine, and displacement. Tearfund was introduced to knowledge management
by Paul Whiffen, who was earlier a knowledge management champion at British Petro-
leum (Milton 2004). Its knowledge management efforts were founded on the recognition
that learning from successes and failures during responses to disasters, both natural
and man-made, should improve responses to later ones. It has proved this by identifying,
consolidating, and then utilizing lessons learned in response to floods in Bangladesh, the
Orissa Cyclone in India, the Balkan crisis, and other disasters. Its knowledge manage-
ment efforts comprise of two main components. First, they utilize the learning opportu-
nities that arise during and after any major activity by involving key participants in the
activity to perform after-action reviews that describe lessons learned from the activity. In
each project, the key project members participate in a structured, facilitated process to
identify the key lessons learned and retrieve them again when they are next required.
Second, Tearfund creates communities of practice to connect people with similar roles,
issues, challenges, and knowledge needs. This enables Tearfund’s employees to share
their knowledge with its 350 United Kingdom and overseas partner organizations. Both
these steps rely on cultural change and use of technology.
Through these KM efforts, Tearfund has been consciously learning different disaster
responses, in each case identifying specific and actionable recommendations for future appli-
cation. The explicit and conscious sharing of these recommendations provides Tearfund with
the confidence and shared understanding needed to implement some of the lessons its many
individuals had learned. The outcome has been a more proactive and integrated response to
disasters that provides help to the beneficiaries more effectively. For example, Tearfund has
modified its processes so that someone would be in the field no later than 48 hours after a
disaster. It has also identified 300 specific and actionable recommendations. According to
Whiffen (2001), “success relies on not just identifying the lessons, but actually implementing
them the next time. It needs to be part of somebody ‘s job to make sure the learning happens
and lessons are embedded in the processes we follow next time there’s a disaster response.”
Source: Compiled from Milton 2004; Whiffen 2001; Wilson 2002.
1Visit http://www.tearfund.org for more information on this organization.
http://www.tearfund.org
ORGANIZATIONAL IMPACTS OF KNOWLEDGE MANAGEMENT 81
Another example of improved efficiency through knowledge management comes
from British Petroleum (Echikson 2001). A BP exploration geologist located off
the coast of Norway discovered a more efficient way of locating oil on the Atlantic
seabed in 1999. This improved method involved a change in the position of the drill
heads to better aim the equipment and thereby decrease the number of misses. The
employee posted a description of the new process on BP’s Intranet for everyone’s
benefit in the company. Within 24 hours, another engineer working on a BP well
near Trinidad found the posting and e-mailed the Norwegian employee requesting
necessary additional details. After a quick exchange of e-mail messages, the Ca-
ribbean team successfully saved five days of drilling and US$600,000. Of course,
in utilizing this knowledge, the employees of the Caribbean unit needed to either
trust their Norwegian colleagues or be able to somehow assess the reliability of
that knowledge. Issues of trust, knowledge ownership, and knowledge hoarding
are important and need to be examined in future research. This case study points
to a real instance where knowledge sharing and taking advantage of information
technology to quickly disseminate it resulted in a major cost savings to a company.
Overall, the use of knowledge management and Internet technologies enabled BP
to save US$300 million during the year 2001 while also enhancing innovation at
every step of its value chain.
IMPACT ON PROCESS INNOVATION
Organizations can increasingly rely on knowledge shared across individuals to
produce innovative solutions to problems as well as to develop more innovative
organizational processes. Knowledge management has been found to enable riskier
brainstorming (Storck and Hill 2000) and thereby enhance process innovation. In
this context, Nonaka’s (1998) concept of “ba”—which is equivalent to “place” in
English and refers to a shared space (physical, virtual, or mental) for emerging
relationships—is relevant. Unlike information, knowledge cannot be separated from
the context. In other words, knowledge is embedded in ba, and therefore a founda-
tion in ba is required to support the process of knowledge creation. J.P. Morgan
Chase & Co. recognized the impact knowledge can have on process innovation
when the following statement appeared in bold in their debut annual report: “The
power of intellectual capital is the ability to breed ideas that ignite value” (Stewart
2001, p. 192).
Buckman Laboratories, discussed earlier in this chapter, linked their research
and development personnel and technical specialists to their field-based marketing,
sales, and technical support staffs to insure that new products were developed with
the customers’ needs in mind and that customer needs were quickly and accurately
communicated to the product development group (Zack 1999). As a result, new
knowledge and insights were effectively exploited in the marketplace leading to better
products. In addition, the regular interactions with customers generated knowledge
to guide future developments.
Another example of the impact of KM on process innovation (and efficiency),
may be seen in the case of the Office of Special Projects, Veteran’s Health Ad-
82 CHAPTER 5
ministration (VHA), which was discussed in Box 4.3 in Chapter 4. The VHA
significantly enhanced innovation by reducing bureaucracy, breaking down
organizational barriers, benchmarking and partnering with others, and institu-
tionalizing best processes.
Through these process improvements, knowledge management also contributes to
the organization’s dynamic capabilities, which are viewed as identifiable and specific
organizational processes such as strategic decision-making and product development
that create value for organizations in dynamic environments (Eisenhardt and Martin
2000; Prieto and Easterby-Smith 2006). In one leading chemical company, knowl-
edge management, especially technical training, word-of-mouth transfer of market
knowledge, and informal exchanges between sales managers, seemed to facilitate
dynamic capabilities (Prieto and Easterby-Smith 2006).
Figure 5.4 summarizes the above impacts of knowledge management and knowl-
edge on organizational processes.
Figure 5.4 How Knowledge Management Impacts Organizational Processes
Knowledge
Knowledge
Management
Process Efficiency
roductivity improvement
ost savings
Process Innovation
mproved brainstorming
etter exploitation of new ideas
Process Effectiveness
ewer mistakes
daptation to changed circumstances
Knowledge
Knowledge
Management
Process Efficiency
roductivity improvement
ost savings
Process Innovation
mproved brainstorming
etter exploitation of new ideas
Process Effectiveness
ewer mistakes
daptation to changed circumstances
Process Efficiency
roductivity improvement
ost savings
Process Innovation
mproved brainstorming
etter exploitation of new ideas
Process Effectiveness
ewer mistakes
daptation to changed circumstances
Process Efficiency
roductivity improvement
ost savings
Process Innovation
mproved brainstorming
etter exploitation of new ideas
Process Effectiveness
ewer mistakes
daptation to changed circumstances
IMPACT ON PRODUCTS
Knowledge management also impacts the organization’s products. These impacts can
be seen in two respects: value-added products and knowledge-based products.
Whereas the impacts on the above dimensions come either through knowledge or
directly from KM, the impacts below arise primarily from knowledge created through
KM. This is depicted in Figure 5.5.
IMPACT ON VALUE-ADDED PRODUCTS
Knowledge management processes can help organizations offer new products or
improved products that provide a significant additional value as compared to earlier
products. One such example is Ford’s Best Practices Replication Process in manufac-
ORGANIZATIONAL IMPACTS OF KNOWLEDGE MANAGEMENT 83
turing. Every year Ford headquarters provides a “task” to managers, requiring them
to come up with a five percent, six percent, or seven percent improvement in key
measures—for example, improvements in throughput or energy use. Upon receiving
their task, the managers turn to the best practices database to seek knowledge about
prior successful efforts. Ford claims that its “best-practice replication” system, whose
use Ford tracks in meticulous detail, saved the company $245 million from 1996 to
1997 (Anthes 1998). Over a four-and-a-half year period from 1996 to 2000, more than
2,800 proven superior practices were shared across Ford’s manufacturing operations.
The documented value of the shared knowledge in 2000 was US$850 million, with
another $400 million of value anticipated from work in progress, for a total of $1.25
billion (Stewart 2000; Swarup 2005).
Value-added products also benefit from knowledge management due to the effect
the latter has on organizational process innovation. For example, innovative processes
resulting from knowledge management at Buckman Laboratories enables sales and
support staff to feed customer problems into their computer network in order to access
relevant expertise throughout the company and be able to develop innovative solutions
for the customers. Similarly, Steelcase Inc. uses information obtained through video
ethnography from its customers, the endusers of office furniture, to understand how
its products are used and then to redesign the products to make them more attractive
to customers (Skyrme 2000).
IMPACT ON KNOWLEDGE-BASED PRODUCTS
Knowledge management can also have a major impact on products that are inher-
ently knowledge based—for example, in consulting and software development in-
dustries. For instance, consultants at ICL2 can quickly access and combine the best
available knowledge and bid on proposals that would otherwise be too costly or too
time-consuming to put together. Indeed, in such industries, knowledge management
is necessary for mere survival.
Knowledge-based products can also sometimes play an important role in traditional
manufacturing firms. A classic example is Matsushita’s (now Panasonic Corporation)
development of an automatic breadmaking machine. In order to design the machine,
Matsushita sought a master baker, observed the master baker’s techniques, and then
incorporated them into the machine’s functionality (Nonaka and Takeuchi 1995).
Similarly, companies such as Sun Microsystems have enhanced the level of customer
service by placing solutions to customer problems in a shareable knowledge base.
Moreover, customers can download software patches from the Internet based on their
answers to an automated system that prompts customers with a series of questions
aimed at diagnosing the customer needs.
Figure 5.5 How Knowledge Management Impacts Products
Knowledge
alue-added products
nowledge-based products
Knowledge
Management
Knowledge
alue-added products
nowledge-based products
Knowledge
Management
84 CHAPTER 5
IMPACT ON ORGANIZATIONAL PERFORMANCE
In addition to potentially impacting people, products, and processes, knowledge
management may also affect the overall performance of the organization. Knowledge
management enhances the employees’ learning from each other and from external
sources, and helps to facilitate innovativeness, effectiveness, and efficiency of orga-
nizational processes. KM can also contribute to a firm by facilitating new knowledge-
based products or enabling improved products that provide significant additional
value. However, it would take time for a KM effort to produce these organizational
benefits (Sabherwal and Sabherwal 2007).
The Deutsche Bank put it all in a nutshell when it took out a big advertisement in
the Wall Street Journal (Stewart 2001, p. 192) that said: “Ideas are capital. The rest
is just money.” This advertisement reflects the belief that investments in knowledge
management should be viewed as capital investments. This investment may be capable
of producing long-term benefits to the entire organization rather than as assets that
provide value only at the present time.
Knowledge management can impact overall organizational performance either
directly or indirectly as discussed below.
The experience of a large company in benefiting from knowledge management is
described in Box 5.3.
DIRECT IMPACTS ON ORGANIZATIONAL PERFORMANCE
Direct impact of knowledge management on organizational performance occurs when
knowledge is used to create innovative products that generate revenue and profit or
when the knowledge management strategy is aligned with business strategy. Such
a direct impact concerns revenues and/or costs and can be explicitly linked to the
organization’s vision or strategy. Consequently, measuring direct impact is relatively
straightforward. It can be observed in terms of improvements in return on investment
(ROI). For example, one account director at British Telecom (BT Groups plc) indicated
that his sales team generated about US$1.5 million in new business based on briefings
from a new knowledge management system (Compton 2001). Similarly, speaking to
the Knowledge Management World Summit in San Francisco, California, on January
11, 1999, Kenneth T. Derr, the Chairman and CEO of Chevron Corporation stated:
Of all the initiatives we’ve undertaken at Chevron during the 1990s, few have been as impor-
tant or as rewarding as our efforts to build a learning organization by sharing and managing
knowledge throughout our company. In fact, I believe this priority was one of the keys to
reducing our operating costs by more than $2 billion per year—from about $9.4 billion to
$7.4 billion—over the last seven years.
INDIRECT IMPACTS ON ORGANIZATIONAL PERFORMANCE
Indirect impact of knowledge management on organizational performance comes about
through activities that are not directly linked to the organization’s vision, strategy,
ORGANIZATIONAL IMPACTS OF KNOWLEDGE MANAGEMENT 85
revenues, or costs. Such effects occur, for example, through the use of knowledge
management to demonstrate intellectual leadership within the industry, which, in
turn, might enhance customer loyalty. Alternatively, it could occur through the use of
knowledge to gain an advantageous negotiating position with respect to competitors
or partner organizations. Unlike direct impact, however, indirect impact cannot be
associated with transactions and, therefore, cannot be easily measured.
One example of indirect benefits is the use of knowledge management to achieve
economies of scale and scope. Before examining these effects, we briefly examine
what we mean by economies of scale and scope.
A company’s output is said to exhibit economy of scale if the average cost of
production per unit decreases with increase in output. Due to economy of scale, a
smaller firm has higher costs than those of larger firms, which makes it difficult to
compete with the larger firms in terms of price. Some of the reasons that result in
economies of scale include: large setup cost makes low-scale production uneconomic,
possibilities for specialization increase as production increases, and greater discounts
from suppliers are likely when production is large scale.
A company’s output is said to exhibit economy of scope when the total cost of that
same company producing two or more different products is less than the sum of the
costs that would be incurred if each product was produced separately by a different
company. Due to economy of scope, a firm producing multiple products has lower costs
than those of its competitors focusing on fewer products. Some of the reasons that result
in economies of scope include: incorporating new innovations into multiple products,
joint use of production facilities, and joint marketing or administration. Economy of
scope can also arise if the production of one good provides the other as a byproduct.
Knowledge management can contribute to economies of scale and scope by improv-
ing the organization’s ability to create and leverage knowledge related to products,
customers, and managerial resources across businesses. Product designs, components,
Box 5.3
Knowledge Management Helps a Premier Big Box Retailer
This premier office products company employs over 90,000 associates in more than two
dozen countries around the world and has four technical support centers spread across three
continents. It wanted to create a single repository of intellectual capital and support documents,
thereby achieving greater levels of operational efficiency and customer satisfaction, while also
lowering the number of issue escalations. This company benefited from the RightAnswers
Unified Knowledge Platform, which serves as the “central source of truth” for users, providing
answers to daily issues and questions, and thereby enabling associates to meet their goals and
objectives.
The use of this new platform improved the retailer’s performance in numerous ways. It raised
the First Call Resolution Rate by 15 percent, with about 20 percent of calls being positively
affected by off-the-shelf content and about 40 percent of calls directly benefiting from the knowl-
edge base. Average talk time decreased by 30 percent while the Level 1 training time reduced
by over 60 percent.
Source: Compiled from RightAnswers 2014.
86 CHAPTER 5
manufacturing processes, and expertise can be shared across businesses thereby reduc-
ing development and manufacturing costs, accelerating new product development, and
supporting quick response to new market opportunities. Similarly, shared knowledge of
customer preferences, needs, and buying behaviors can enable cross-selling of existing
products or development of new products. Finally, economies of scope also result from
the deployment of general marketing skills and sales forces across businesses. Although
economies of scale and scope could, and usually do, lead to improvements in return on
investments, the effect of knowledge management on scale and scope economies and
their subsequent effect on return on investments cannot be directly linked to specific
transactions and this is therefore considered as an “indirect” impact.
Another indirect impact of knowledge management is to provide a sustainable
competitive advantage. Knowledge can enable the organization to develop and exploit
other tangible and intangible resources better than the competitors can, even though the
resources themselves might not be unique. Knowledge, especially context-specific tacit
knowledge, tends to be unique and therefore difficult to imitate. Moreover, unlike most
traditional resources, it cannot easily be purchased in a ready-to-use form. To obtain
similar knowledge, the company’s competitors have to engage in similar experiences, but
obtaining knowledge through experience takes time. Therefore, competitors are limited
in the extent to which they can accelerate their learning through greater investment.
LeaseCo, an industrial garment and small equipment leasing company described by
Zack (1999), illustrates the use of knowledge management to gain a sustainable com-
petitive advantage. LeaseCo’s strategy involved occasionally bidding aggressively on
complex, novel, or unpredictable lease opportunities. These bidding, and subsequent nego-
tiation, experiences provided the company with unique and leverageable knowledge while
reducing the opportunity for competitors to gain that same knowledge. LeaseCo realized
two significant benefits over its competitors: first by investing in its strategic knowledge
platform and second by learning enough about the particular client to competitively and
profitably price leases for future opportunities with the same client. Sufficient mutual
learning occurred between LeaseCo and their client for the client to contract LeaseCo for
future leases without even going out for competitive bids. In essence, LeaseCo created
a sustainable (or renewable), knowledge-based barrier to competition.
Thus, sustainable competitive advantage may be generated through knowledge
management by allowing the organization to know more than its competitors about
Figure 5.6 How Knowledge Management Impacts Organizational Performance
Knowledge
ision
trategy
evenues
osts
Knowledge
Management
Organizational Performance
cale economies
cope economies
ustainable competitive advantage
Knowledge
ision
trategy
evenues
osts
Knowledge
Management
Organizational Performance
cale economies
cope economies
ustainable competitive advantage
ORGANIZATIONAL IMPACTS OF KNOWLEDGE MANAGEMENT 87
certain things. Competitors, on the other hand, would need considerable time to acquire
that same knowledge. Figure 5.6 summarizes the direct and indirect impacts KM and
knowledge can potentially have on organizational performance.
SUMMARY
In Table 5.1, we summarize the various impacts of knowledge management examined in
this chapter. The impact KM has on one level might lead to synergistic impacts on another
level as well. For example, employee learning facilitates impacts on processes as well as
on products. Thus, KM has the potential to produce several interrelated impacts on people,
products, processes, and organizations as we have described in this chapter.
Table 5.1
A Summary of Organizational Impacts of Knowledge Management
Levels of Impact Impacted Aspects
People
Processes
Products
Organizational Performance Direct Impacts
REVIEW
1. Briefly enumerate the ways in which knowledge can impact an organization.
2. State the importance of knowledge management with specific reference to its
impact on employee adaptability and job satisfaction.
3. Explain why poor knowledge management reduces the effectiveness of or-
ganizational processes.
4. What three dimensions are relevant for examining the impact of knowledge
management on business processes?
5. Identify ways in which knowledge management helps improve process ef-
fectiveness, efficiency, and innovation.
6. Describe how knowledge management can contribute to an organization’s
products. Illustrate using the example of Xerox.
7. How can we assess: (a) the direct impacts and (b) the indirect impacts of
knowledge management on organizational performance?
8. Knowledge management is an invaluable tool for the oil Industry. Justify this
statement with suitable examples.
88 CHAPTER 5
APPLICATION EXERCISES
1. Identify the possible ways in which knowledge management (or the lack
thereof) in your organization (it could be your academic institution or your
workplace) affects your learning and job satisfaction.
2. Identify the biggest positive impact on your organization (it could be your
academic institution or your workplace) due to the implementation of knowl-
edge management. Speculate on the possibilities if there were no knowledge
management practices in place.
3. Now identify the biggest negative impact on your organization (it could be
your academic institution or your workplace) due to improper/insufficient
knowledge management practices and suggest ways of improvement.
4. Interview a friend or a family member who works at a different organization
than you, and examine the overall effects of knowledge management on that
organization.
5. You are a CEO who considers implementing a knowledge management system
in your company. You have to decide one option out of two: (a) Our knowledge
management system can be accessed by customers, or (b) Our knowledge
management system cannot be accessed by customers. Describe your decision
and provide the reason in terms of organizational performance.
6. Critique the following analysis: Our investment on knowledge management
seems to be unsuccessful. The ROI decreased from 10 percent to 5 percent
at the year of system implementation. Since direct measure of organizational
performance decreased, we need to uninstall the knowledge management
system right away.
NOTES
1. IDC is one of the world’s leading providers of technology intelligence, in-
dustry analysis, market data, and strategic and tactical guidance to builders,
providers, and users of information technology. More information on it can
be obtained from http://www.idc.com.
2. Formed in 1968, ICL was bought by STC in 1984. Fujitsu bought an 80 percent
stake in ICL-UK from STC in 1990. In 2002, the consulting arm of ICL-UK
was merged with DMR Consulting, and its service division became Fujitsu
Services.
REFERENCES
Anthes, G. 1998. Defending knowledge. ComputerWorld, 32(7) (February 16), 41–42.
Becerra-Fernandez, I., and Sabherwal, R. 2008. Individual, group, and organizational learning. In
Knowledge management: An evolutionary view, vol. 12, Advances in Management Information
Systems, ed. I. Becerra-Fernandez and D. Leidner, 13–39. Armonk, NY: M.E. Sharpe.
Bontis, N. 2003. HR’s role in knowledge management. Canadian HR Reporter, 16(5) (March
10), G8.
http://www.idc.com
ORGANIZATIONAL IMPACTS OF KNOWLEDGE MANAGEMENT 89
Brown, J.S., and Duguid, P. 1991. Organizational learning and communities-of-practice: Toward a
unified view of working, learning, and innovation. Organization Science, 2(1), 40–57.
Buckman Laboratories International. 2007. Awards and recognitions for Buckman Laboratories. Knowl-
edge nurture. www.knowledge-nurture.com/recognitions.html (accessed February 12, 2009).
Business Wire. 2000. 2000 most admired knowledge enterprises announced. Business Wire, June 5.
———. 2002. Xerox ranked as one of North America’s most admired knowledge enterprises: Winning
practices available to customers through Xerox Global Services. Business Wire, May 6.
Compton, J. 2001. Dial K for knowledge. CIO, June 15.
Dyer, G., and McDonough, B. 2001. The state of KM. Knowledge Management, May, 21–36.
Dyer, J.H., and Nobeoka, K. 2000. Creating and managing a high-performance knowledge-sharing
network: The Toyota case. Strategic Management Journal, 23(3), 345–367.
Echikson, W. 2001. When oil gets connected. Business Week, December 3.
Eisenhardt, K.M., and Martin, J.K. 2000. Dynamic capabilities: What are they? Strategic Management
Journal, 21, 1005–1121.
King, J. 2001. Shell strikes knowledge gold. ComputerWorld, July 16.
Kranz, G. 2008. At Xerox, learning is a community activity. Workforce Management Online (December).
www.workforce.com/section/11/feature/26/05/23/index.html (accessed February 12, 2009).
Milton, N. 2004. Knowledge management in the aid and development sector: A case study in imple-
mentation at Tearfund. In Performance through learning: Knowledge management in practice, ed.
C. Gorelick, N. Milton, and K. April, 143–161. Boston: Elsevier Butterworth-Heinemann.
Moore, C. 2001. Xerox makes global services push. InfoWorld, November 19. www.itworldcanada.
com/a/Daily-News/de565b27-f77d-4fd5-82ff-e861f4bbcfb9.html.
National Aeronautics and Space Administration (NASA). 2007. Quotes related to knowledge manage-
ment or collaboration. http://km.nasa.gov/whatis/KM_Quotes.html (accessed February 12, 2009).
Nonaka, I. 1998. The concept of ba: Building a foundation for knowledge creation. California Manage-
ment Review, 40(3) (Spring), 40–54.
Nonaka, I., and Takeuchi, H. 1995. The knowledge creating company: How Japanese companies create
the dynamics of innovation. New York: Oxford University Press.
Prieto, I.M., and Easterby-Smith, M. 2006. Dynamic capabilities and the role of organizational knowl-
edge: An exploration. European Journal of Information Systems, 15, 500–510.
RightAnswers. 2014. Success story: Premier big box retailer. www.rightanswers.com/resources/case-
studies/client-stories/big-box-retailer/ (accessed January 31, 2014).
Sabherwal, R., and Sabherwal S. 2007. How do knowledge management announcements affect firm
value? A study of firms pursuing different business strategies. IEEE Transactions on Engineering
Management, 54(3), 409–422.
Sabherwal, R. 2008. KM and BI: From mutual isolation to complementarity and synergy. Cutter Con-
sortium Executive Report, 8(8), 1–18.
Skyrme, D.J. 2000. Developing a knowledge strategy: From management to leadership. In Knowledge
management: Classic and contemporary works, ed. D. Morey, M. Maybury, and B. Thuraisingham,
61–84. Cambridge, MA: The MIT Press.
Stewart, T.A. 2000. Knowledge worth $1.25 billion. Fortune, November 27, 302–303.
———. 2001. Intellectual capital: Ten years later, how far we’ve come. Fortune, May 28, 192–193.
Storck, J., and Hill, P. 2000. Knowledge diffusion through strategic communities. Sloan Management
Review, 41(2), 63–74.
Swarup, S. 2005. Applying KM to improve quality. InsideKnowledge, October 10.
Whiffen, P. 2001. Seizing learning opportunities at Tearfund. Knowledge Management Review, No-
vember/December.
Wilson, J. (Ed.). 2002. Knowledge management review: The practitioner’s guide to knowledge manage-
ment. Chicago: Melcrum Publishing Limited.
Zack, M.H. 1999. Developing a knowledge strategy. California Management Review, 41(3) (Spring),
125–145.
www.knowledge-nurture.com/recognitions.html
www.workforce.com/section/11/feature/26/05/23/index.html
www.itworldcanada.com/a/Daily-News/de565b27-f77d-4fd5-82ff-e861f4bbcfb9.html.
www.itworldcanada.com/a/Daily-News/de565b27-f77d-4fd5-82ff-e861f4bbcfb9.html.
www.rightanswers.com/resources/casestudies/client-stories/big-box-retailer/
www.rightanswers.com/resources/casestudies/client-stories/big-box-retailer/
http://km.nasa.gov/whatis/KM_Quotes.html
This page intentionally left blank
PART II
KNOWLEDGE MANAGEMENT
TECHNOLOGIES AND SYSTEMS
This page intentionally left blank
93
6
Knowledge Application Systems:
Systems that Utilize Knowledge
In the last chapter, we discussed the organizational impacts of knowledge management.
In this chapter, we describe knowledge application systems, how they are developed,
and relate experiences of how organizations have implemented such systems. As we
discussed in Chapter 4, knowledge application systems support the process through
which individuals utilize the knowledge possessed by other individuals without actu-
ally acquiring, or learning, that knowledge. Both mechanisms and technologies can
support knowledge application systems by facilitating the knowledge management
processes of routines and direction. Knowledge application systems are typically en-
abled by intelligent technologies. In this chapter, we introduce the reader to artificial
intelligence (AI), its historical perspective, its relationship with knowledge, and why
it is an important aspect of knowledge management. We also summarize the most
relevant intelligent technologies that underlie most KM systems, from rule-based
expert systems, to case-based reasoning (CBR), to traditional management informa-
tion systems. Moreover, we discuss different types of knowledge application systems:
expert systems, help desk systems, and fault diagnosis systems. The case studies in
this chapter narrate the implementation of knowledge application systems. Each is
based on different intelligent technologies and designed to accomplish different goals:
provide advice, enhance fault detection, and facilitate creative reasoning. Finally,
limitations of knowledge application systems are discussed.
You may recall, from Chapter 4, that knowledge application depends on direction and
routines. Mechanisms facilitating direction include hierarchical relationships, help desks,
and support centers; whereas mechanisms facilitating routines include organizational
policies, work practices, and standards. Technologies supporting direction and routines
include expert systems, decision support, advisor systems, fault diagnosis (or trouble-
shooting) systems, and help desk systems. These technologies may support direction,
as in the case of a field service technician seeking to troubleshoot a particular product;
or may support routines, as in the case of a customer service representative who may
need to identify alternative product delivery mechanisms while preparing the shipment
of an order. Moreover, mechanisms and technologies can facilitate knowledge applica-
tion through direction and routines either within or across organizations.
For a quick overview of what knowledge application systems are and how they
are used, let us look at a brief case study of how NEC Corporation redefined the way
the organization is able to apply their collective experience in order to better produce
high-quality software—see Box 6.1.
94 CHAPTER 6
TECHNOLOGIES FOR APPLYING KNOWLEDGE
ARTIFICIAL INTELLIGENCE
In this section we begin by describing the historical perspective of AI, the area of
computer science that deals with the design and development of computer systems
that exhibit human-like cognitive capabilities. Artificial Intelligence (AI) refers to
enabling computers to perform tasks that resemble human thinking ability. Much
like KM and human intelligence, AI is associated with knowledge. Definitions for AI
range from: systems that act like humans, systems that think like humans, systems that
think rationally, to systems that act rationally (Russell and Norvig 2002). Systems that
act like humans refer to those that pass the Turing Test, which refers to a computer
passing a test by a human interrogator, who cannot tell whether the responses came
from a person or not. Systems that think like humans refer to a computer program
whose input to output behavior matches those of humans, for example when solving
problems, like playing chess or performing a medical diagnosis. Systems that think
rationally refer to those that follow a specific logic to solve a problem. Finally, systems
that act rationally refer to those computer agents that are expected to have specific
Box 6.1
Applying Organizational Experiences to Produce Quality Software
NEC is a leading global company that manufactures cutting edge products for the broadband
networking and mobile Internet market. In 1981, NEC recognized the need to extend their qual-
ity control (QC) activity to the domain of software development. In order to accomplish this goal,
the company established a company-wide corporate structure to assist employees in apply-
ing the principles of software quality control (SWQC). QC activities typically resulted in a case
report that outlined the problem analysis, its possible root cause, the corrective actions taken,
and the results of the corrective actions.
By 1991, the company had collected over 25,000 such cases in an effort to apply the pro-
ductivity improvements across the organization. Initially the case reports were stored in a book
and later in a searchable database, but people found it difficult to search and apply the QC
cases. NEC then decided to implement the software quality control advisor (SQUAD),1 based
on case-based reasoning methodology, to improve user access and application of the reported
QC cases. The cases in SQUAD were nominated through a review committee that reviews each
case and selects the best cases. Cases are selected on the basis of the quality of the analysis,
significance of the results, and how generalizable the problem is.
Adequate incentives were established to encourage employee participation. Initially about
3,000 cases were submitted each year, and later new submissions decreased to about 1,000
cases a year. The significant drop in the rate of new cases submission came about because
most typical cases were already reported in the system. By 1994, the system represented
about 24,000 cases and served over 150,000 users.
Some of the success factors that marked the development of SQUAD included low develop-
ment cost, since its development only required four person-months. Furthermore, the develop-
ment of SQUAD supported incremental modifications, since it allowed for cases to incremen-
tally be included in the case database. By 1991, it was estimated that SQUAD had already paid
off to the organization over 100 million dollars per year.
1For further details about SQUAD, refer to Cheetham and Watson 2005; Kitano and Shimazu 1996.
KNOWLEDGE APPLICATION SYSTEMS: SYSTEMS THAT UTILIZE KNOWLEDGE 95
characteristics that enable them to operate autonomously within their environments,
and even adapt to change in the face of uncertainty.
Typically computers perform repetitive, logical tasks extremely well, such as
complex arithmetic calculations and database retrieval and storage. One common
characteristic of these conventional computer tasks is their algorithmic nature,
which means that they engage precise and logically designed instructions to result
in a distinct correct output. Humans, by contrast, excel at solving problems using
symbols to which a specific meaning can be attached, such as understanding the
meaning of a poem. Artificial intelligence deals with the manipulation of these
symbols. Therefore, our own definition of AI (Becerra-Fernandez et al. 2004) in
more specific terms is:
The science that provides computers with the ability to represent and manipulate symbols
so they can be used to solve problems not easily solved through algorithmic models.
Modern AI systems are based on the understanding that intelligence and knowl-
edge are tightly intertwined. Knowledge is associated with the cognitive symbols we
manipulate, while human intelligence refers to our ability to learn and communicate
in order to solve problems. However, when we judge a student’s performance in
class or decide whom to hire, we generally focus on how much they know, not their
intelligence. People are born with a certain degree of intelligence, which they use to
learn and thus acquire new knowledge. Some AI systems (also known as knowledge-
based systems or knowledge application systems) try to imitate the problem-solving
capabilities of skillful problem-solvers in a particular domain. Intelligent systems
offer us technologies to manage knowledge—that is, to apply, capture, share, and
discover it.
The idea of creating computers that resemble the human intelligent abilities can
be traced to the 1950s, when scientists predicted the development of such machines
within a decade. Although those same scientists may have underestimated the com-
plexities of the human mind, AI research made significant strides. The term “artifi-
cial intelligence” was coined by John McCarthy during a workshop he organized at
Dartmouth College in 1956, where he convened the four pioneers of the field: John
McCarthy, Marvin Minsky, Allan Newell, and Herbert Simon.
AI research first focused on games and natural language translation. In the area
of gaming, scientists developed numerous chess programs including Greenblatt’s
“Mac Hack” and Slate and Atkin’s “Chess 4.5” (Hsu et al. 1990). Back in 1997, an
AI program named Big Blue defeated Boris Kasparov, the reigning world champion
in chess, in a widely publicized match. On the other hand, early efforts in machine
translation of natural languages were not nearly as successful. Another seminal devel-
opment in AI was the development of General Problem Solver (GPS) by Simon and
Newell (Newell and Simon 1963). The importance of GPS was that it demonstrated
the computer’s ability to solve some problems by searching for an answer in a solu-
tion space, which represented a new trend for AI.
One of the areas in AI that has witnessed the greatest popularity is knowledge-
based systems, which we refer to here as knowledge application systems. Knowledge
96 CHAPTER 6
application systems are the topic of this chapter, and they basically apply knowledge
to solve specific problems. Other areas of research within AI include natural language
understanding, classification, diagnostics, design, machine learning, planning and
scheduling, robotics, and computer vision. Next we describe the two most relevant
intelligent technologies that underpin the development of knowledge application
systems: rule-based expert systems and case-based reasoning.
RULE-BASED SYSTEMS
Traditionally, the development of knowledge-based systems had been based on the
use of rules or models to represent the domain knowledge. The development of such
systems requires the collaboration of a subject matter expert with a knowledge engi-
neer, the latter being responsible for the elicitation and representation of the expert’s
knowledge. We will see two examples of rule-based expert systems when we present
cases on Westinghouse Electric Corporation’s GenAID and the SBIR/STTR Online
Advisor later in this chapter.
The process of developing knowledge application systems requires eliciting the
knowledge from the expert and representing it a form that is usable by computers.
This process is called knowledge engineering. Knowledge engineers typically build
knowledge application systems by first interviewing in detail the domain expert and
representing the knowledge more commonly in a set of heuristics, or rules-of-thumb.
Experts develop these rules-of-thumb over years of practical experience at solving
problems. In order for the computer to understand these rules-of-thumb, we represent
them as production rules or IF-THEN statements. For example: IF the number of
employees is less than 500, THEN the firm is a small business is one of the rules that
the SOS Advisor checks to ensure the firm is eligible for the SBIR/STTR program.
Rules are the most commonly used knowledge representation paradigm, perhaps due
to their intuitive implementation. The IF portion is the condition (also premise or
antecedent), which tests the truth-value of a set of assertions. If the statement is true,
the THEN part of the rule (also action, conclusion, or consequence) is also inferred
as a fact.
In addition to rules, other paradigms to represent knowledge include frames,
predicates, associative networks, and objects. Rule-based systems have posed
some disadvantages. One is that in many circumstances, the number of rules that
may be needed to properly represent the domain may be quite large. For example,
the GenAID system that we describe in the first case study below consisted of
about 10,000 rules when it was first deployed. Although later developments of
GenAID may have condensed the number of rules by about 3,000, it was still
considered a large system. Expert systems with such a large number of rules offer
many disadvantages, namely (1) difficulty in coding, verifying, validating, and
maintaining the rules; and (2) reduction in the efficiency of the inference engine
executing the rules. As an alternative, we consider the use of cases as a method
to represent knowledge. For more details on rule-based systems refer to Chapter
8 of the book Knowledge Management: Challenges, Solutions, and Technologies
(Becerra-Fernandez et al. 2004).
KNOWLEDGE APPLICATION SYSTEMS: SYSTEMS THAT UTILIZE KNOWLEDGE 97
CASE-BASED REASONING SYSTEMS
Although the rules approach to knowledge representation has produced many examples
of successful knowledge application systems, many of these systems are increasingly
based on the implementation of case-based reasoning methodology.
Case-based reasoning (CBR) is an artificial intelligence technique designed to
mimic human problem solving. CBR is based on Schank’s (1982) model of dynamic
memory. Its goal is to mimic the way humans solve problems. When faced with a
new problem, humans search their memories for past problems resembling the current
problem and adapt the prior solution to “fit” the current problem. CBR is a method
of analogical reasoning that utilizes old cases or experiences in an effort to solve
problems, critique solutions, explain anomalous situations, or interpret situations
(Aamodt and Plaza 1994; Kolodner 1991, 1993; Leake 1996; Watson 2003). A typical
case-based knowledge application system will consist of the following processes:
1. Search the case library for similar cases. This implies utilizing a search engine
that examines only the appropriate cases and not the entire case library, as it
may be quite large.
2. Select and retrieve the most similar case(s). New problems are solved by
first retrieving previously experienced cases. This implies having a means
to compare each examined case to the current problem, quantifying their
similarity, and somehow ranking them in decreasing order of similarity.
3. Adapt the solution for the most similar case. If the current problem and
the most similar case are not similar enough, then the solution may have
to be adapted to fit the needs of the current problem. The new problem
will be solved with the aid of an old solution that has been adapted to the
new problem.
4. Apply the generated solution and obtain feedback. Once a solution or clas-
sification is generated by the system, it must be applied to the problem. Its
effect on the problem is fed back to the CBR system for classification of its
solution (as success or failure).
5. Add the newly solved problem to the case library. The new experience is
likely to be useful in future problem solving. This step requires identifying
if the new case is worth adding to the library and placing it in the appropriate
location in the case library.
There are several advantages to using CBR over rules or models for developing
knowledge application systems. These advantages come to light when the relation-
ship between the case attributes and the solution or outcome is not understood well
enough to represent in rules. Alternatively, CBR systems are advantageous when the
ratio of cases that are “exceptions to the rule” is high, as rule-based systems become
impractical in such applications. CBR is especially useful in such situations because
it incorporates the solution of a newly entered case. It is in such situations that meth-
ods for adaptation are used, providing the user with steps to combine and derive a
solution from the collection of retrieved solutions.
98 CHAPTER 6
There are several variants of CBR, such as exemplar-based reasoning, instance-
based reasoning, and analogy-based reasoning. These different variations of CBR are
described below (Aamodt and Plaza 1994; Leake 1996):
1. Exemplar-based reasoning—These systems seek to solve problems through
classification, that is, finding the right class for the unclassified exemplar.
Essentially the class of the most similar past case then becomes the solution
to the classification problem, and the set of classes are the possible solutions
to the problem (Kibler and Aha 1987).
2. Instance-based reasoning—These systems require a large number of instances
(or cases) that are typically simple; that is, they’re defined by a small set of
attribute vectors. The major focus of study of these systems is automated
learning, requiring no user involvement (Aha et a1. 1991).
3. Analogy-based reasoning—These systems are typically used to solve new
problems based on past cases from a different domain (Aamodt and Plaza
1994; Veloso and Carbonnell 1993). Analogy-based reasoning focuses on
case reuse, also called the mapping problem, which is finding a way to map
the solution of the analogue case to the present problem.
CBR, rules, and models are not the only type of intelligent technology underpin-
ning the development of knowledge application systems. Other important technolo-
gies used to develop knowledge application systems are worth mentioning—namely,
constraint-based reasoning, model-based reasoning, and diagrammatic reasoning.
Constraint-based reasoning is an artificial intelligence technique that uses essen-
tially “what cannot be done” to guide the process of finding a solution (Tsang 1994,
p. 148). This technique is useful in naturally constrained tasks such as planning and
scheduling. For example, to schedule a meeting all the individuals that need to attend
must be available at the same time, otherwise the “availability constraint” will be
violated. Model-based reasoning (MBR) is an intelligent reasoning technique that
uses a model of an engineered system to simulate its normal behavior (Magnani et al.
1999, p. 148). The simulated operation is compared with the behavior of a real system
and noted discrepancies can lead to a diagnosis; for example, a hurricane model can
be designed and implemented to predict a hurricane’s trajectory, given the set of cur-
rent weather conditions such as wind speed, presence of a cold front, temperature,
and so forth. Finally, diagrammatic reasoning is an artificial intelligence technique
that aims to understand concepts and ideas using diagrams that represent knowledge
(Chandrasekaran et al. 1993; Glasgow et al. 1995, p. 148). These technologies are
radically different from rule-based systems or CBR systems and have very specific
application areas.
In summary, rule-based systems and case-based reasoning, as well as constraint-
based reasoning, model-based reasoning, and diagrammatic reasoning are all tech-
nologies used to develop knowledge application systems. The applicability of each
technology is dictated primarily by the characteristics of the domain as described
above. Table 6.1 summarizes the technologies to develop knowledge application
systems and the characteristics of the domain that define their applicability. The
KNOWLEDGE APPLICATION SYSTEMS: SYSTEMS THAT UTILIZE KNOWLEDGE 99
next sections describe specific types of knowledge application systems based on the
aforementioned technologies.
As we previously mentioned, it has become increasingly clear that the most popular
technique for the implementation of knowledge application systems in businesses
today is case-based reasoning. The reasons why CBR is more commonly used in the
development of such systems include the fact that CBR implementations are, at least
on the surface, more intuitive. In addition, CBR implementations take advantage of
explicit knowledge that may already exist in the organization, for example in problem
reports. We will see two examples of CBR systems: National Semiconductor Corpo-
ration’s Total Recall and NASA’s Out-of-Family-Disposition prototype later in this
chapter. In the next section, we describe how to implement knowledge applications
systems. For the reasons mentioned here, we will assume that the underpinning tech-
nology for the knowledge application system will be CBR, although the methodology
applies to any of the aforementioned technologies.
DEVELOPING KNOWLEDGE APPLICATION SYSTEMS
Here we describe how to build a knowledge application system. We make extensive
use of examples and boxes in order to enhance the learning experience. The next
section discusses the different types of knowledge application systems, and specific
examples are presented in subsequent sections.
The effective implementation of the knowledge application system requires a care-
fully thought-out methodology. The Case-Method Cycle (Kitano 1993; Kitano and
Shimazu 1996) is a methodology that describes an iterative approach to effectively
Table 6.1
Technologies for Knowledge Application Systems
Technology Domain Characteristics
Rule-based Systems Applicable when the domain knowledge can be defined by a
manageable set of rules or heuristics.
Case-based Reasoning Applicable in weak-theory domains, that is, where an expert
either doesn’t exist or does not fully understand the domain. Also
applicable if the experience base spans an entire organization,
rather than a single individual.
Constraint-based Reasoning Applicable in domains that are defined by constraints, or what
cannot be done.
Model-based Reasoning (MBR) Applicable when designing a system based on the description of
the internal workings of an engineered system. This knowledge
is typically available from design specifications, drawings, and
books, and can be used to recognize and diagnose its abnormal
operation.
Diagrammatic Reasoning Applicable when the domain is best represented by diagrams and
imagery, such as when solving geometric problems.
100 CHAPTER 6
develop CBR and knowledge application systems in general. The Case-Method Cycle
describes the following six processes:
1. System development process—This process is based on standard software
engineering approaches, and its goal is to develop a knowledge application
system that will store new cases and retrieve relevant cases.
2. Case library development process—The goal of this process is to develop and
maintain a large-scale case library that will adequately support the domain
in question.
3. System operation process—This process is based on standard software engi-
neering and relational database management procedures. Its goal is to define
the installation, deployment, and user support of the knowledge application
system.
4. Database mining process—This process uses rule-inferencing techniques
and statistical analysis to analyze the case library. This step could help infer
new relationships between the data, which could be articulated to enhance
the knowledge application system.
5. Management process—This process describes how the project task force will
be formed and what organizational support will be provided to the project.
6. Knowledge transfer process—This process describes the incentive systems
that will be implemented to encourage user acceptance and support of the
knowledge application system. This step will ensure that users will feel com-
pelled to augment the case library with new cases.
In terms of actually developing the case library (step 2 above), the process can also
be described in terms of the following subprocesses (Kitano and Shimazu 1996):
1. Case Collection—This process entails the collection of seed cases, which
provide an initial view of the application. For example, for the SQUAD sys-
tem described in Box 6.1, the developers started with 100 seed cases. These
seed cases were used to define a format for the collection of future cases and
for the design of the database structure. Seed cases typically do not follow a
predefined structure, while the subsequent collection of cases will follow the
defined format. The number of seed cases may vary according to the applica-
tion, as we will see in this chapter, and may even be generated artificially by
creating permutations of the cases available, as discussed in the case study
later in the chapter.
2. Attribute-Value Extraction and Hierarchy Formation—This step is essential
for indexing and organizing the case library. The goal of this phase is to ex-
tract the attributes that define the case representation and indexing. This phase
will seek to create a list of attributes that define each case, a list of values
for each attribute, and a possible grouping of such attributes. In addition, the
relationships among the attributes must also be defined. After the hierarchy
is defined, the relative importance of each attribute is determined. This deci-
sion is typically reflective of the implementation domain. This phase results
KNOWLEDGE APPLICATION SYSTEMS: SYSTEMS THAT UTILIZE KNOWLEDGE 101
in a concept hierarchy created for each attribute, assigned with similarities
between values. Also, this step will require mapping a hierarchy into a rela-
tional database or flat case library.
3. Feedback—This phase will provide necessary feedback to those supplying
the cases to the CBR system, so the quality of the cases can be improved.
The use of the Case Method Cycle has been shown to result in significant reduction
in system development workload and costs (Kitano and Shimazu 1996). For example,
use of the Case method during the development of SQUAD resulted in a savings of
six person-months from the expected development time for the entire system. Fur-
thermore, the workload required for the system maintenance was reduced to less than
10 percent of the initial workload.
Knowledge application systems not only apply a solution to a similar problem but
can also serve as a framework for creative reasoning (Leake 1996). For example,
analogy-based reasoning could provide the initial ideas in solving new problems.
Case memories can provide humans with the experience base they may lack. Faced
with a problem, experts may recall experiences from the case library, and perform the
adaptation and evaluation of the solutions that is sometimes relegated to the knowl-
edge application systems. This is the emphasis of the SQUAD system presented in
Box 6.1.
Knowledge application systems have enabled the implementation of decision
support systems to support design tasks in diverse domains such as architecture,
engineering, and lesson planning (Domeshek and Kolodner 1991, 1992, 1993; Grif-
fith and Domeshek 1996). These decision-support systems, also called case-based
design aids or CBDAs, help human designers by making available a broad range of
commentated designs. CBDAs can serve to illustrate critical design issues, explain
design guidelines, and provide suggestions or warnings regarding specific design
solutions. One of the critical components in the development of such systems is the
supporting indexing system used to perform the relevant case search.
Finally, case libraries can serve to accumulate organizational experiences and
can often be viewed as a corporate memory. For example, the case library for a help
desk system could be considered a corporate memory of organizational experiences
related to customer support. The same thing can be said of a rule-base supporting an
expert system. For more details on a case-based reasoning systems, refer to Chapter
9 of the book Knowledge Management: Challenges, Solutions, and Technologies
(Becerra-Fernandez et al. 2004).
TYPES OF KNOWLEDGE APPLICATION SYSTEMS
Recall that knowledge application systems include advisor systems, fault diagnosis
or troubleshooting systems, expert systems, help desk systems, and decision-support
systems in general.
One area where knowledge application systems are specifically important is in
the implementation of help desk technologies. For example, Compaq Computer
Corporation implemented a help desk support technology named SMART (Acorn
102 CHAPTER 6
and Walden 1992), to assist help desk employees track calls and resolve customer
service problems. Compaq’s SMART system was developed to support its Customer
Service Department when handling user calls through its toll-free number. SMART
was an integrated call-tracking and problem-solving system, supported by hundreds
of cases that help resolve diagnostic problems resulting from the use of Compaq
products (Allen 1994). The philosophy behind the design of SMART was to develop
a system that minimized the learning curve for help desk employees in diagnosing
and resolving user’s problems. Historical cases similar to what the customer currently
faced were automatically retrieved by the system from the case library. The customer
service representative then used that solution to help customers solve the problem at
hand. SMART developers reported an increase from 50 percent to 87 percent of the
problems that could be resolved directly by the first level of customer support. The
implementation of SMART at Compaq paid for itself in one year with the productiv-
ity improvements it brought to the company.
The earliest help desk system was Lockheed’s CLAVIER, a system for laying
out composite parts prior to baking in an industrial convection oven. The CLAVIER
system became the classic CBR system that demonstrated CBR could be applied
to solve a problem where no explicit decision model existed and that CBR systems
can learn by acquiring new cases and therefore improve their performance with time
(Watson and Marir 1994).
Fault diagnosis is increasingly becoming a major emphasis for the development
of knowledge applications systems, as we will discuss below. Fault diagnosis has
been one of the main focuses of intelligent systems implementation (Davis 1984;
de Kleer 1976; Genesereth 1984; Patton et al. 2000, p. 154). One of the earliest
successful implementations of knowledge application systems for the diagnosis and
recovery of faults in large multistation machine tools was CABER at Lockheed Martin
Corporation (Mark et al. 1996). Although these milling machines are equipped with
self-diagnostic capabilities, typically they resolved only 20 percent to 40 percent of
the systems faults. The expectation for the CABER system was that it had to help
identify how the equipment experienced the fault and how to safely exit the faulted
state. Typically, an equipment fault results in a call to the field-service engineer. For
the creation of the case library that supports this system, Lockheed counted on over
10,000 records collected by the field service engineers. CABER augmented the self-
diagnostic capabilities of the milling machine, which provided junior field-service
engineers with the necessary tools to resolve the fault and reduce machine downtime.
In addition, Compaq also developed a fault diagnosis system early on for its Page-
Marq printer line known as QuickSource (Nguyen et al. 1993). A case base of over
500 diagnostic cases supported the QuickSource knowledge applications system.
This system, designed to run in a Windows environment, was shipped with printers
to enable customers to do their own diagnosis.
Finally, another prominent CBR system was FormTool, the system developed at
General Electric (GE) in order to determine the correct formulas to color plastics ac-
cording to the customers’ specifications (Cheetham 2005). What is difficult about the
process of determining the colorants and levels to be added to the plastics is that the
possible number of colorants is very large, the amounts of each colorant also need to
KNOWLEDGE APPLICATION SYSTEMS: SYSTEMS THAT UTILIZE KNOWLEDGE 103
be determined, there’s no exact algorithm for predicting the color produced by a set of
colorants, a given formula can appear different due to lighting conditions, and different
base plastics have different starting colors. FormTools was originally developed in
1994 and continued to be used ten years later, saving GE millions of dollars in pro-
ductivity and colorant costs. The benefits from this development included improved
color matcher productivity, pigment cost reduction, global color consistency, improved
color-match speed, and served as the basis of development for other tools, such as
a tool to control the color produced by the manufacturing line. Also, the technology
developed in this knowledge application system was used to create an online color-
selection for GE’s customers, and a customer innovation center developed around the
system’s software. Other successful CBR implementations at GE include automation
of an appliance call center via the Support the Customer (STC) tool, which helps call
takers solve customers’ problems by suggesting questions that could help with the
problem diagnosis (Cheetham and Goebel 2007). This implementation improved the
accuracy of the diagnostic process and the speed of resolution, ultimately improving
customer satisfaction. The STC knowledge application system also was in use for
around ten years, saving the company more than $50 million.
The design of Verdande Technology’s (2014) CBR-driven Edge platform is based
on the principle that similar problems have similar solutions. The Edge platform is
used to identify, capture, and analyze data patterns in real time, predicting future
problems based on past events while offering solutions based on prior actions. The
company has launched vertical applications for the oil and gas, financial services,
and health-care industries.
VIDUR is a CBR-based advisory system developed by the agricultural experts
from the Central Agricultural University in India and funded by the Centre for
Development of Advanced Computing in Mumbai (C-DAC 2014), for the purpose
of supporting farmers in Manipur state with certain aspects of the farming process.
Currently, the system can be used for paddy variety selection and weed control. With
this advisory system, farmers now have a virtual expert available as and when help
may be needed.
jCOLIBRI (GAIA 2014) is a reference platform for developing CBR applica-
tions. It defines the architecture for designing CBR systems and provides a refer-
ence for the implementation of the components required to build the applications,
in a way that is extensive and reusable across different domains and CBR families.
The first version of jCOLIBRI was released in 2005. The main scope is to support
the research community, but the applications can also be deployed in commercial
applications.
Collaborative Agile Knowledge Engine (CAKE) is a collaboration support architec-
ture that provides unified access to knowledge within organizations and uses CBR to
distribute this knowledge. The system integrates adaptive workflows and collaborative
patterns among agents. CAKE has been used to build various prototypical applications
in diverse areas, which include the fire services, medicine, agile software development,
e-government, and geographical information systems. The CAKE system consists of
four main components: the agile workflow engine, a knowledge engine, a common
storage layer, and a browser-based user interface (Bergmann et al. 2006).
104 CHAPTER 6
In the following five sections, we discuss the development and implementation
details for five knowledge application systems. The first system, GenAID, is one of
the earliest diagnostic knowledge application systems. GenAID is based on the use
of rules and is still operational today. Later, we describe the development of SOS
Advisor, a Web-based expert system built using a set of rules. The reason heuristics
were used for the implementation of this system is that a small number of rules can
define the domain—that is, defining the eligibility potential for companies interested in
applying for a specific federal program. Following that, we describe the development
of a knowledge application system based on CBR technology, which was designed
with the goal of reusing the solutions to software quality problems, as these problems
recur throughout the organization. The system deployed at Darty, described in the
following section is also based on CBR technology and it’s in use at call centers to
help resolve problems at tier support and minimize the deployment of technicians to
the field. CLAIM, the CBR-based system developed at GE Healthcare presented later
in this chapter, is improving the process of healthcare services reimbursement. The
knowledge application system described is somewhat different in the sense that it’s
designed to assist in the solution of new problems as they occur, by identifying similar
problems that may have happened in the past and their corresponding solutions.
We begin by describing the development of GenAID.
CASE STUDIES
GENAID: A KNOWLEDGE APPLICATION SYSTEM FOR EARLY FAULT
DETECTION AT WESTINGHOUSE
By the year 1990, there were over 3,000 AI-based systems in use around the world for
a variety of purposes including Ace (telephone cable maintenance advisory system),
XCON (computer configuration system), Dispatcher (printed-wire, board assembly,
work-dispatching system), APES (electronic design), CDS (configuration-dependent
part sourcing), National Dispatcher (transportation sourcing and routing), XFL (floor
layout assistance), XSEL (sales assistance), Compass (network management), Cooker
(food-processing control), ESP (facility analysis), Ocean (computer configuration),
Opgen (process planning), Trinity Mills Scheduler (scheduling), VT (elevator con-
figuration), CDS (flexible manufacturing system cell control), GenAID (generator
diagnosis), Intellect (natural language database interfacing), Mudman (drilling mud
analysis), and Telestream (telemarketing assistance) (Fox 1990).
In the early 1980s, Westinghouse Electric Corporation, a manufacturer of large
power generation equipment (now Siemens Power Generation), started the develop-
ment of Process Diagnosis System (PDS), also known as GenAID1 (Gonzalez et al.
1986). The goal of GenAID was to enable the early detection of abnormal operating
conditions of their turbine generators, which could cause them to eventually malfunc-
tion. For Westinghouse’s customers, electrical power utilities that operate generation
plants, a plant outage could represent costs that range from $60,000 to $250,000 per
day, depending on the size and the type of the plant. A generator malfunction could
cause unplanned outages to repair the broken unit that could last up to six months.
KNOWLEDGE APPLICATION SYSTEMS: SYSTEMS THAT UTILIZE KNOWLEDGE 105
Clearly, anticipating a generator outage when it’s a minor fault and before it’s com-
pletely inoperative and becomes a major incident can result in corrective actions
that could reduce the magnitude of the problem and thus reduce the downtime from
months to perhaps days.
GenAID was one of the first real-time, sensor-based diagnosis systems. Before the
development of GenAID, sensor data from the power generation equipment which is
transmitted to the data acquisition system would be periodically inspected by a technician
who traveled to the site. Typically power plant personnel at the site lacked the expertise
to analyze the sensor measurements. The goal behind GenAID was to continuously
analyze the generator sensor data, therefore anticipating major destructive incidents.
The early attempts at performing this analysis date from the 1970s, when West-
inghouse developed a computer-based system that used probabilistic analysis. The
system performed well for a limited number of malfunctions, but posed serious limi-
tations which forced the research attention to shift to intelligent systems. In 1980, a
collaboration between Westinghouse’s Research and Development Laboratories in
Pittsburgh and the Robotics Institute at Carnegie-Mellon University resulted in the
development of the Process Diagnostic System (PDS). The combination of PDS and
the diagnostic knowledge was called GenAID, which stands for Generator Artificial
Intelligence Diagnostics. Originally, GenAID was located in Orlando, Florida, where
it processed data on a semicontinuous basis from each of the power plant sites across
the United States to produce a diagnosis of potential malfunctions in real time. GenAID
would later on be colocated with each power plant generator to produce a malfunction
diagnosis for each power plant operator.
GenAID went into production use in 1985, and in 1986 it was recognized as one of
the top 100 engineering achievements of the year. By 1990, about 14 generators were
connected to the system. Since the time of its development, the PDS shell was modified
to run on general-purpose personal computers and its user interface was completely
revamped. The system contains in the neighborhood of 2,500 rules. GenAID has suc-
cessfully diagnosed numerous malfunctions that otherwise could have resulted in serious
outages. As a result, Siemens extended this concept to other pieces of equipment, includ-
ing steam turbines and gas turbines. GenAID is a prototypical knowledge application
system. Its development required the elicitation of important knowledge possessed by
human experts, capturing this knowledge electronically in a knowledge base, and the
ability to apply this knowledge in a way that multiplied manifold the original utility of
that knowledge by placing it in an automatic monitoring and diagnostic system.
To this date, the GenAID system is still sold as a site-based monitor for power plants.
In the late 1990s, Siemens began a research and development project to expand the
system to cover other power generation equipment. The new expanded product is simi-
lar in features but is run remotely in the Siemens Power Diagnostics Center, known as
the Siemens Power and Process Automation D3000 (SPPA-D3000 Diagnostic Suite),
which monitors nearly 300 power plants worldwide. The new version covers not only
the generator but the gas turbine, steam turbine, and balance of plant equipment as well
and has saved Siemens’ customers millions of dollars in the last decade.2
Next we describe the development of SOS Advisor, a simple system based on a
small set of heuristics.
106 CHAPTER 6
THE SBIR/STTR ONLINE SYSTEM (SOS) ADVISOR: A WEB-BASED EXPERT
SYSTEM TO PROFILE ORGANIZATIONS
The SBIR/STTR Online System (SOS) Advisor, a Web-based expert system, was
developed to assist potential applicants to the Small Business Innovation Research
(SBIR) and Small Business Technology Transfer Research (STTR) programs. Es-
tablished by Congress in 1982, the SBIR and STTR programs help federal agencies
develop innovative technologies by providing competitive research contracts to U.S.-
owned small business companies with fewer than 500 employees. These programs
also help by providing seed capital to increase private sector commercialization of
innovations resulting from federal research and development (see, for example, NASA
2008 and USDOD 2009). The goal of the SOS Advisor was to optimize the time re-
quired to examine the potential eligibility for companies seeking SBIR/STTR funding
by prompting users through an interactive questionnaire that was used to evaluate the
company’s potential eligibility to be a grant recipient. The user only needed to click
on Yes or No to answer the 10 questions that frame the eligibility criteria.
Once the user submitted the registration information to the Web-based system,
the SOS Advisor Questionnaire page was launched. The questionnaire consists of
10 questions used to determine the eligibility of the company. The profile questions
are listed in Table 6.2, to which users could respond by selecting the radio button
next to the Yes, No, or Not Sure. Answering Not Sure will prompt users for more
information, necessary in order to define the potential candidate’s eligibility for
funding. In order to match the SBIR winners’ profile, users were expected to answer
according to the responses specified in Table 6.2. Question 6 was for information
purposes only, since it does not constitute a necessary criterion for eligibility. Each
question had a Tip icon that allowed the user to obtain additional information related
to the corresponding question through the use of a pop-up window. In this man-
ner, users could learn about SBIR/STTR requirements and the reasoning for each
question. The one-page questionnaire format allowed users to spend minimum time
when answering the profile questions. Furthermore, the user had the opportunity
to see at once all the questions and answers in order to review and modify the an-
swers before submission. The suggestions field provided users with the option of
providing feedback to the development team. Figure 6.1 describes the architecture
of the SOS Advisor.
The user information provided and the corresponding answers to the question-
naire were stored in the SOS Advisor database and evaluated automatically by the
system. Using a set of rules that evaluate the user responses, the system identified if
the user profile matched the profile of an SBIR/STTR candidate. SOS Advisor then
would automatically send an e-mail to the user with the results of the evaluation.
At the same time, if the user profiles match was positive the system automatically
notified via e-mail the corresponding agency program personnel, with the user
point-of-contact.
The rules used to evaluate the user profiles were developed using a scripting code.
The scripts were used to evaluate the answers given and based on predefined rules
to generate the user profiles. Based on the user’s answers to the questionnaire, the
KNOWLEDGE APPLICATION SYSTEMS: SYSTEMS THAT UTILIZE KNOWLEDGE 107
SOS Advisor was able to determine whether the user profile indeed matched that of
an SBIR/STTR recipient and provided sufficient information to educate potential
applicants about related funding opportunities.
In summary, the SOS Advisor is an example of a knowledge application system
that was used to identify those companies whose profiles matched that of an SBIR/
STTR candidate and, therefore, helped focus the resources of federally funded
assistance programs. The system prompts users to answer a set of questions that
describe if the company meets the stipulated criteria defined for companies in-
terested in the SBIR/STTR program. The system then used a set of heuristics or
rules to quickly examine each company’s qualifications rather than attempting to
transfer the knowledge about the program requirements to each of the companies
interested in applying for the program (although that option is available through
launching the Tip section). For those companies with matching profiles, the SOS
Advisor automatically sent the company’s contact information to a federal agency
employee who then identified appropriate assistance resources available for the
benefit of the inquiring company. In this way, the SOS Advisor helped minimize
distractions to the federal program representative caused by casual cyber surfers,
since it would only forward the information for the companies that matched the
qualification criteria. In addition, a by-product of this effort was the creation of a
database with point-of-contact information for each company that completed the
survey, which could be used to generate mailings and announcements of upcoming
SBIR/STTR informative events.
Table 6.2
SBIR/STTR Profile Framing Questions
Question
SBIR winners’
profile
1. I would like to know if your company is independently owned and operated. Yes
2. Is this company located in the United States? Yes
3. Is this company owned by at least 51% U.S. citizens or permanent U.S. residents? Yes
4. Regarding your company size, does it have less than 500 employees? Yes
5. What about your proposed innovation? Has it been patented or does it have any
patents pending?
No
6. Could it be patented, copyrighted, or otherwise protected? Don’t care
7. Are you planning on using SBIR/STTR funding to conduct any of the following: No
a. Systems studies
b. Market research
c. Commercial development of existing products or proven concepts
d. Studies
e. Laboratory evaluations
f. Modifications of existing products without innovative changes
8. Does your technology area align with any of the following research areas of
interest to NASA?
Yes
9. Is there a likelihood of your proposed technology having a commercial
application?
Yes
10. Has your firm been paid or is currently being paid for equivalent work by any
agency of the federal government?
No
108
F
ig
ur
e
6.
1
S
O
S
A
d
vi
so
r
A
rc
h
it
ec
tu
re
In
te
rn
etC
a
n
d
id
a
te
R
es
ea
rc
h
ar
ea
s
o
f
in
te
re
s
t
In
te
rn
et
S
O
S
A
dv
is
or
N
A
SA
S
O
S
A
d
vi
so
r
S
er
ve
r
S
O
S
A
dv
is
or
E
m
ai
l
S
ys
te
m
N
A
S
A
p
ro
vi
de
s
fe
ed
ba
ck
to
p
os
si
bl
e
C
an
di
da
te
r
eg
ar
di
ng
S
B
IR
/S
TT
R
Pr
og
ra
m
ht
tp
://
w
w
w
.s
bi
r.
fiu
.e
du
S
O
S
A
dv
is
or
Q
ue
st
io
nn
ai
re
S
O
S
A
dv
is
or
p
ro
vi
de
s
em
ai
l w
ith
fe
ed
ba
ck
S
O
S
A
dv
is
or
p
ro
vi
de
s
N
A
S
A
w
ith
c
om
pa
ny
’s
p
oi
nt
o
f c
on
ta
ct
In
te
rn
etC
a
n
d
id
a
te
In
te
rn
et
In
te
rn
etC
a
n
d
id
a
te
R
es
ea
rc
h
ar
ea
s
o
f
in
te
re
s
t
In
te
rn
et
S
O
S
A
dv
is
or
N
A
SA
R
es
ea
rc
h
ar
ea
s
o
f
in
te
re
s
t
In
te
rn
et
S
O
S
A
dv
is
or
N
A
SA
S
O
S
A
d
vi
so
r
S
er
ve
r
S
O
S
A
dv
is
or
E
m
ai
l
S
ys
te
m
N
A
S
A
p
ro
vi
de
s
fe
ed
ba
ck
to
p
os
si
bl
e
C
an
di
da
te
r
eg
ar
di
ng
S
B
IR
/S
TT
R
Pr
og
ra
m
ht
tp
://
w
w
w
.s
bi
r.
fiu
.e
du
S
O
S
A
dv
is
or
Q
ue
st
io
nn
ai
re
S
O
S
A
dv
is
or
p
ro
vi
de
s
em
ai
l w
ith
fe
ed
ba
ck
S
O
S
A
dv
is
or
p
ro
vi
de
s
N
A
S
A
w
ith
c
om
pa
ny
’s
p
oi
nt
o
f c
on
ta
ct
S
O
S
A
dv
is
or
http://www.sbir.fiu.edu
KNOWLEDGE APPLICATION SYSTEMS: SYSTEMS THAT UTILIZE KNOWLEDGE 109
The key importance of SOS Advisor is that it enabled federal agency personnel to
apply the knowledge about qualification requirements for the SBIR/STTTR program
without tying up the time of the program representative. Prior to the development
of the SOS Advisor, a federal employee provided this information and performed
the initial assessment of companies interested in the program. The federal agency
representative repeatedly used her tacit knowledge base to perform the analysis. The
SOS Advisor helped to apply this knowledge, freeing up the employee to use her
time so she can provide more personalized advice to those companies that meet the
program’s stipulated criteria.
Now we will discuss how knowledge technologies can be used to reuse organiza-
tional knowledge about software quality.
PRODUCT QUALITY ANALYSIS FOR NATIONAL SEMICONDUCTOR
National Semiconductor3 was established in 1959 in Santa Clara, California. Since that
time, with manufacturing sites around the world, the company has been a leader in
the semiconductor industry. National Semiconductor had annual sales of $1.42 billion
in fiscal 2010 and approximately 5,800 employees worldwide with a net income of
$325.8 billion. National Semiconductor has set the pace for revolutionary electronics
technologies. A leader in the development of thousands of patents and analog products,
this company’s achievements range from the design and manufacture of early discrete
transistors to the introduction of sophisticated integrated circuit product lines.
Although National’s product quality record is extraordinary, today’s environment
requires semiconductor component deliveries with zero defects, enabling manufac-
turers to achieve lower costs and just-in-time manufacturing schemes. Therefore,
a rare failure is a cause of immediate concern for both National and the customer, as
it is imperative to quickly determine and take corrective action; in particular, if the
failure indicates that a manufacturing process is moving out of statistical control. Na-
tional’s customers demand rapid and complete failure analysis, as well as the adoption
of corrective actions that will ensure the accurate identification and solution for the
root cause of the failure. The advanced technology and high degree of complexity
in today’s semiconductors make this analysis a major challenge. For this purpose,
National depends on its Worldwide Quality Network, a centralized manufacturing
quality assurance (QA) group. This group consists of engineers who use the Product
Quality Analysis (PQA) process to focus on root cause determination and finding
solutions to each of these failures.
In order to support the engineers involved in the PQA and other quality-related
business processes, an in-house team at National developed the Advanced Quality and
Reliability Information System (AQUARIS) in 1995, a tracking system that provided
a searchable repository of PQAs. Some of the limitations of AQUARIS related to its
inefficient ability to query similar past failures. The workflow associated with the
PQA process is quite elaborate and can involve many analysis steps to determine the
true causes of a device’s failure or, in some cases, just a cursory analysis revealing
that there is no problem with the part at all. In any event, these steps taken to analyze
parts are carefully and methodically accomplished and the interim and final results
110 CHAPTER 6
are captured and stored in the AQUARIS system. Many times, engineers engaged in
the various stages of analysis make “hunch” decisions based on prior experiences or
anecdotal information, which can significantly shorten the analysis cycle.
By 1999, it was recognized that AQUARIS did not provide an effective means of
recalling information that could prevent unnecessary work from being performed,
while at the same time promoting learning from prior failures. Engineers would typi-
cally spend hours attempting to search on AQUARIS for similar past PQAs that they
distantly remembered, based on some similarity to their current analysis. Typically
since their search centered on retrieving recalled PQAs, it only focused on those with
which they were previously involved and didn’t include the work of others. Soon
National recognized the need to better collect these experiences in a way that could
be adequately applied by others, because written reports collected in AQUARIS did
not provide an efficient means to extract and apply this knowledge when needed.
To respond to this challenge, National developed a knowledge application system
based on the use of CBR technology. The development team adapted and expanded
the back-end relational database that had driven AQUARIS to provide integration
with the CBR system. The overall application, titled Total Recall, can be viewed as
consisting of four components and the Web client.
Figure 6.2 illustrates the Total Recall System Architecture. Users typically entered
PQAs into the Application Server, and the users’ workflow is illustrated with the dark
arrows labeled as User operation in the figure. The Total Recall Database was used
to collect the results from the testing performed in the different PQA activities. Data
from the Total Recall Database were used to create the CBR case library. Note that
not all PQAs produce new cases for the case library. A nomination process to the case
library administrator was used to denote potential cases for the case library, illustrated
by the arrows labeled as Admin. only in the figure.
The case library stores the experience gained from the PQA process as a collec-
tion of cases. During an active analysis, the Total Recall system relayed queries to
the CBR server by gathering information entered up to that point. The CBR server
responded to the query with an ordered set of cases sorted by declining similarity.
The footprint number identified the cases in the CBR server, and the Total Recall
application performed an additional search to translate these footprint numbers into
the original PQA and device serial numbers that were more meaningful to the user.
This information then allowed the engineers to retrieve, online, the corresponding
PQA reports. Engineers could then study the reports identified as similar to the case
at hand and decide if the failure mechanism and corrective actions described for these
earlier failures applied to their current situation. The engineer made the final decision
to adopt or adapt these findings.
Each basic component of Total Recall is described as follows:
1. Application Server: The main server for the Total Recall application. This
server performed data manipulation and user presentation. This component
is the result of the system development process, described on page 100.
2. Total Recall Database: Maintained all the information related to the testing
results of the PQA process.
KNOWLEDGE APPLICATION SYSTEMS: SYSTEMS THAT UTILIZE KNOWLEDGE 111
3. Case Library: A separate database containing CBR representation of cases,
including mapping information that related the case footprint numbers to
specific devices analyzed during PQA processing. This component is the
result of the case library development process, which is described later.
4. CBR Server: The final case library and CBR engine.
A CBR query was made from the Application Server to the CBR engine. The
CBR engine responded with a set of footprint numbers that represented the set of
cases that were similar to the case in the query. Other devices on other PQAs may
have failed in a similar way. Rather than treat these as separate cases, they were
considered reference cases. The same footprint number was also used to identify
these reference cases.
As mentioned earlier, when an engineer completed a new PQA analysis, she decided
if she would nominate the PQA for possible inclusion in the case base. The Total
Recall System at this point attempted to take advantage of her current knowledge of
the failure by allowing a thorough refinement of the case description. At this point
the case-base administrator performed the final evaluation of the nominated case by
searching the case library to identify if similar cases already existed in the library.
In general, it’s preferred not to have numerous similar cases in the case library in
order to provide users with manageable search results. If multiple cases reflecting the
same type of failure and analysis were to be included in the case library, one needed
to be designated as the footprint case while the others were designated as reference
Figure 6.2 Total Recall System Architecture
Application Server Total Recall
Database
CBR ServerCBR Database
Web Client
User operation
Application Server Total Recall
Database
CBR ServerCBR Database
Web Client
User operation
Admin. only
112 CHAPTER 6
cases. Figure 6.3 presents the details of how the CBR database is populated. The case
library administrator can make the decision to treat newly nominated cases as a new
footprint or as a reference case to an existing case. Also, this decision may be left to
an engineering technical review board.
One of the most time-consuming tasks required for the implementation of Total
Recall was the initial population of the case library. A subset of PQAs from AQUARIS
was evaluated, and the corresponding test data had to be cleansed and augmented prior
to manually representing them as cases in the new system. This task also presented
significant cultural challenges, since it required the involvement of failure analysis
engineers to review each potential case at a technical level. This task represented sig-
nificant additional work requirement from this group, so only limited success could be
claimed. In addition, lack of adequate CBR training and commitment from the users
could also be attributed to the low level of support from this group. The initial case
library represented approximately 200 cases, which was barely enough to perform the
initial testing. The size of the case library was expected to grow substantially during
the implementation of Total Recall. This was not expected to impact the application
adversely, given the system’s structured architecture.
In terms of its user interface, Total Recall mimicked much of the AQUARIS work-
flow, with Web-based functionality and options required to capture new information
such as activities and the explicit, related observations. Testing included the analysis
CBR Admin.
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Admin Data
CBR Admin.
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Admin Data
CBR Admin.
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Admin Data
CBR Admin.
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Admin Data
CBR Admin.
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Admin Data
CBR Cases
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Footprint
No.
CBR Cases
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Footprint
No.
CBR Cases
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Footprint
No.
Footprint no.
PQA #
S/N
Ref. flag
Footprint no.
PQA #
S/N
Ref. flag
Footprint no.
PQA #
S/N
Ref. flag
Footprint no.
PQA #
S/N
Ref. flag
Response to
CBR Query
Loaded to
CBR Server
Subset of
nominated
cases
CBR Admin.
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Admin Data
CBR Admin.
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Admin Data
CBR Admin.
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Admin Data
CBR Admin.
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Admin Data
CBR Admin.
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Admin Data
CBR Admin.
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Admin Data
CBR Admin.
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Admin Data
CBR Admin.
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Admin Data
CBR Admin.
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Admin Data
CBR Admin.
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Admin Data
CBR Admin.
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Admin Data
CBR Admin.
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Admin Data
CBR Admin.
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Admin Data
CBR Admin.
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Admin Data
CBR Admin.
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Admin Data
CBR Cases
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Footprint
No.
CBR Cases
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Footprint
No.
CBR Cases
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Footprint
No.
CBR Cases
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Footprint
No.
CBR Cases
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Footprint
No.
CBR Cases
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Footprint
No.
CBR Cases
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Footprint
No.
CBR Cases
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Footprint
No.
CBR Cases
________________
________________
________________
________________
________________
________________
________________
________________
________________
CBR Data
Footprint
No.
Footprint no.
PQA #
S/N
Ref. flag
Footprint no.
PQA #
S/N
Ref. flag
Footprint no.
PQA #
S/N
Ref. flag
Footprint no.
PQA #
S/N
Ref. flag
Footprint no.
PQA #
S/N
Ref. flag
Footprint no.
PQA #
S/N
Ref. flag
Footprint no.
PQA #
S/N
Ref. flag
Footprint no.
PQA #
S/N
Ref. flag
Response to
CBR Query
Loaded to
CBR Server
Subset of
nominated
cases
Figure 6.3 CBR Database Detail
Source: Courtesy of National Semiconductor.
KNOWLEDGE APPLICATION SYSTEMS: SYSTEMS THAT UTILIZE KNOWLEDGE 113
of prior solved cases through Total Recall. Engineers familiar with these cases then
correlated these results with prior results. The testing process confirmed that the results
of using Total Recall were consistent with the prior cases’ solution.
Although the team at National was faced with the evaluation of the usefulness of
Total Recall, the benefits it offered prompted other developments at the company.
Systems like Total Recall are not intended to eliminate the analysis engineer. The Total
Recall system acts as a cognitive prosthesis for the engineer, who is able to make a
faster and perhaps more accurate prognosis of the failure case on hand.
The key importance of Total Recall is that it enables the application of knowledge
gained from completed failure analyses performed throughout the worldwide quality
organization. Prior to the development of Total Recall, only some of this information
was kept in the AQUARIS database. The AQUARIS database wasn’t a useful plat-
form for knowledge application, since it was hard to identify and apply prior relevant
knowledge. The Total Recall system helps to apply knowledge resulting from the
software-enabled quality process. This knowledge can be applied to prevent unneces-
sary work from being performed while promoting learning from prior failures. For
more details on the Total Recall system, refer to Watson (2003).
In the next section, we describe the implementation of a CBR-based call center
application at Darty, the European retailer of electronic products.
DARTY IMPROVES CUSTOMER SATISFACTION THROUGH EFFECTIVE CALL CENTER
PROBLEM RESOLUTION
Darty, a consumer electronic retailer, was established in 1957 and since then has
become one of the largest retailers in Europe, with operations originally in France,
but now extending to Turkey, Italy, and Switzerland. Darty currently has revenues of
over $2.9 billion, over 209 stores, over 11,000 employees, a fleet of over 400 trucks,
and 1,000 field technicians.4 Darty counts on seven call centers that handle over 4
million calls a year concerning 1.3 million home repairs. The company was facing
increasing productivity losses as first-level agents lacked the required technical back-
ground to resolve the majority of the call center calls, 95 percent of which resulted in
calls that escalated to the next level of more experienced and expensive technicians.
In order to improve the efficiency of the level 1 call center technicians, the retailer
implemented a knowledge application system to support call center operations. The
system was developed over a three-month period and initially covered approximately
1,500 cases, or problem-resolution pairs. Three months into the implementation of the
knowledge application solution, the number of calls that required escalation to level
2 went down to 20 percent, and more products were being fixed remotely without the
need to dispatch a technician to the trouble site.
The user interface for the knowledge application system implemented at Darty
allowed users to access the system via a combination of search methods as depicted
in Figure 6.4. Users can query the system by entering any of the following:
1. Free text search—such as “washing machine has a leak during the rinse cycle.”
2. Guided search—when the system poses a set of questions for the user in order
to further refine the search such as “are there any error lights on the panel?”
114 CHAPTER 6
3. Expert search—where all possible questions are presented to the user, who
in this case is likely to be a more experienced technician.
4. Browse—enables users to review all the problem/solutions pairs for a par-
ticular product or model.
The Domain Model layer enables to interpret and translate the user’s request into
the common vocabulary of the knowledge base. The Retrieve layer is built on a CBR
engine, but in addition the question engine helps refine the resulting set of solutions
if too many are presented to the user at this level. Finally the Manage layer enables
to build reports as well as support the administrative modules. Figure 6.5 describes
in details the search operations of the Darty knowledge application system.
Figure 6.6 presents the user interface that was used for the Darty knowledge appli-
cation system, including the sensory search engine that in this case is used to explain
the problem that the user entered as “laundry white marks” in the wash, which the
system allows the user to select the picture that best depicts the discolored material.
Note that in April 2010 Kaidara, Inc. was acquired by PTC Servigistics, Inc. and is
part of its PTC Service Lifecycle Management (SLM) solutions.
We have seen how knowledge application systems can help resolve recurring
problems at Darty. The next section describes in detail an innovative application of
CBR: to improve the process of claim reimbursement in healthcare.
AUTOMATING IDENTIFICATION OF ATTACHMENTS FOR HEALTHCARE CLAIMS
In 1992, healthcare costs constituted 14 percent of all domestic spending in the United
States. In 2012, this rose to 22 percent (Zuckerman, 2013).5 Thirty percent of these
costs ($660 billion) were spent processing claims for reimbursement of healthcare
services. Efficient processing of these claims could greatly reduce the overall cost of
Figure 6.4 Technology Architecture for the Darty Knowledge Application System
Free Text
Search
Guided
Search
Expert
Search
Browse
Search
Domain Model (Ontology)
Structured Knowledge Base (Solutions)
Retrieval Engine—
Case-based Reasoning (CBR)
Questioning Engine—
Dynamic Induction (DI)
Technologies
Interact
(UI)
Interpret
Retrieve
& Refine
Manage
F
u
n
ct
io
n
s
Free Text
Search
Guided
Search
Expert
Search
Browse
Search
Domain Model (Ontology)
Structured Knowledge Base (Solutions)
Retrieval Engine—
Case-based Reasoning (CBR)
Questioning Engine—
Dynamic Induction (DI)
Free Text
Search
Guided
Search
Expert
Search
Browse
Search
Domain Model (Ontology)
Structured Knowledge Base (Solutions)
Retrieval Engine—
Case-based Reasoning (CBR)
Questioning Engine—
Dynamic Induction (DI)
Technologies
Interact
(UI)
Interpret
Retrieve
& Refine
Manage
F
u
n
ct
io
n
s
KNOWLEDGE APPLICATION SYSTEMS: SYSTEMS THAT UTILIZE KNOWLEDGE 115
Figure 6.5 Darty Knowledge Application System Search Example
Interact
(UI)
Interpret
Retrieve
& Refine
Manage
F
u
n
ct
io
n
s
“w/m has a leek
during rinse cycle”
Are there lights on?
A – Filter
B – Dispenser
Dispenser
KB
Appliance* = Washer
Symptom** = Leak
Cycle = Rinse
There are 29 results
that match your
criteria
Determine the best
question to refine/
narrow the results
Appliance = Washer
Symptom = Leak
Cycle = Rinse
Error Light = Dispenser
There are 2 best results
***** Service Flash 8973
***** Clean Dispenser
and 4 other results
*Synonym
**Misspelling
Interact
(UI)
Interpret
Retrieve
& Refine
Manage
F
u
n
ct
io
n
s
“w/m has a leek**
during rinse cycle”
Are there lights on?
A – Filter
B – Dispenser
Dispenser
KB
Appliance* = Washer
Symptom** = Leak
Cycle = Rinse
There are 29 results
that match your
criteria
Determine the best
question to refine/
narrow the results
Appliance = Washer
Symptom = Leak
Cycle = Rinse
Error Light = Dispenser
There are 2 best results
***** Service Flash 8973
***** Clean Dispenser
and 4 other results
*Synonym
**Misspelling
Figure 6.6 Darty Knowledge Application System User Interface
kaldara
advisor
SEMCH* AppllJTKf
f iw Tsit Sesrh Guided Search Bro^e Solutions
E
Apokarte Modd Sokm» M**O
Frfte T*Kt Sfrtrcll Solution. 6*»OJution.
1: «E.Y:0U5. • – ; – • COMPiKL f tOluTtl
Laundry / to *d I^UJCS
; L sundry is dam.i^d
Deter-gent m ‘̂k
: Discoidied
VHW mm values-
I . ,:HM1H\ •! tM. •!
” idundrv «vhiTe una
. S * F V : 3 L £ *F£Tft>
Ctck on th* *TI*9*5 dtavt to fltt th* U(«f
Uw Msnuilt
S*l*Ct: Honf i l :,,:,:- t ),tTM
UoriJfV h** d«t*rotr>t markj du* to too much tfattfotnt b*»a u«>d
Laundry K wonkJed after d ŷmq dm to bwm left m dryer Too tonq.
m i d r y ti j j f r ua* j Ju* to b***ch.
Applunce mafces nase cw damaogd laondry due ft’ forenyi Qb]ect-5.
LJU–if’;Y Î Cjmjg^d duy to foreign ‘iutet^rK-g 1 lQ3d.
Laundry a. r>ot elaarnd dm to ov^rfifcto th* dao
SEARCH SEARCH BY
Languages Logout (Edward Extenal-User))
Next
PLEASE CLAREYFY BY WHITE MARKS, DO YOU MEAH
RESTART SKIP NEXT
116 CHAPTER 6
healthcare. General Electric provides automation tools and outsourcing services for
processing claims. This case study describes a project aimed at making those tools
and services more efficient.
There is a shift under way in the healthcare industry: a focus on quality and results.
As such, health insurance providers (or “payers”) want more information to determine
the appropriateness of care as well as documentation on outcomes, setting up a new
shift in paying for performance versus paying by procedure. This results in providers
being required to send more and more clinical documentation to payers as part of
the claim process. A claim that is not sent with the documentation deemed necessary
by a payer is typically denied. This stretches out the reimbursement process, as the
healthcare provider needs to then locate the right information to be included and then
has to resubmit the claim over again. This added delay can cost the healthcare provider
in lost payments and interest. Unfortunately, it is hard to know at claim submission
time what additional documentation the payer requires. There are over 500 payers in
the United States. Each payer has different requirements for what clinical documenta-
tion needs to be submitted with a claim. There are over 30,000 documented medical
procedures and over 20,000 unique diagnoses, all of which can be associated with
documentation needs by a payer. The attachments requested by one payer may not be
the same documentation that was requested by another. This makes for an astronomi-
cal amount of knowledge that both healthcare and related software providers need to
be aware of when submitting claims.
To make matters more complex, the relevant knowledge constantly changes.
Payers’ requirements for additional documentation are not static but are continually
evolving as new procedures become widespread or experimental procedures become
accepted. Many payers do not publish their policies on attachment requirements,
leaving healthcare providers to determine the corresponding rules by trial and error.
For those payers that do publish their policies, the material may be in nonstandard
formats. Even within a single payer’s documentation it may be hard to locate the re-
quired information, resulting in a significant manual effort for the provider to unveil
the relevant knowledge for each payer.
The goal for the Clinical Artificial Intelligence Manager (CLAIM) project was to
automate the identification of attachments for healthcare claims by creating tools that
can maintain and use the knowledge of the payers’ “needs for attachments.” Then
that knowledge can be used to determine if additional documentation is required for
a specific claim prior to submission to a payer. GE Healthcare is already in the busi-
ness of facilitating the submission of medical claims and has a database with over
300 million past claims and remittances (responses from the payer as to whether or
not the claim is rejected, approved, or partially approved). Selected data from this
database were used as a case base for a case-based reasoning (CBR) system. For this
research, in order to preserve patient confidentiality, the data used were scrubbed to
remove any information that could identify any specific person. In the CBR system
a new claim is compared to similar past claims to determine if an attachment was
needed. The claim is similar to a past claim if it is for the same payer, procedure,
diagnosis, and procedure modifiers. Additional factors can make the claim more or
less similar. The set of similar claims is then analyzed to determine if an attachment
KNOWLEDGE APPLICATION SYSTEMS: SYSTEMS THAT UTILIZE KNOWLEDGE 117
is needed. Because of variability within payers, policy changes, and other factors,
these similar past claims will often not all have required the same attachment. When
discrepancies arise between similar claims in CLAIM, the system uses a confidence
algorithm to determine which attachments are needed. The algorithm takes informa-
tion such as the date of the claim, degree of similarity match, and claim quality in
performing this calculation. After the new claim is paid or rejected it is added to the
database for use on future claims.
Since there’s a great deal of similarity in claims that are submitted, it is necessary
to create a single prototypical case that represents the entire set of similar cases,
called a protocol. Payers could have their own set of protocols. An automated learn-
ing algorithm was created to look through historical claims and remittance data on
a periodic basis, searching for cases where a claim was denied and the denial reason
was that it required “additional documentation not submitted.” From this set of cases,
the learning algorithm clusters like claims/remittances based on common character-
istics. Then the number of remittances denied for “needing attachment” are grouped
and tallied for each unique value found for that field. For each of the field values,
count of denials for “needing attachment” can then be compared against all other
claims/remittances with that same field value that were not denied. If the number
of denials versus acceptances is significant, then the value can be used to define a
protocol. CLAIM identifies if a protocol with similar criteria already exists, and if
not it creates a new protocol for that field/value. The field and value are captured as
a “condition” of the protocol.
The learning algorithm is as automated as possible. However, there still ex-
ists the benefit of having a person in the loop to review and approve a newly
discovered protocol before making it available for use when processing claims.
This review and approval process has been implemented as part of a prototype
graphical user interface (GUI) tool that allows users to manually approve,
create, and maintain protocols. Once protocols are approved they are stored in a
repository where they can be accessed by algorithms for processing claims and
predicting additional documentation needs. A high-level picture of this toolset is
shown in Figure 6.7.
The graphical user interface can also be used to create protocols that are based on
documented or tacit knowledge. A person, acting as the protocol manager can browse
or search for protocols, select a protocol, and view the details of a protocol. They
can also modify existing protocols to add new conditions, references to additional
information and so forth, or delete protocols that no longer are necessary. In this way
the repository is always kept up to date.
Pilot studies of the CLAIM system using a subset of these new claims showed
the accuracy in selecting the correct attachments for an initial claim to be over 90
percent. This was better than what existed prior because many providers would
just send claims without the proper attachments knowing that when the claim was
rejected by the payer, information would be provided on the documentation that
was needed.
Next, we see how knowledge application systems can assist the problem-solving
process, even when these are new problems.
118 CHAPTER 6
OUT-OF-FAMILY DISPOSITION SYSTEM FOR SHUTTLE PROCESSING
The Shuttle Processing Directorate of the Kennedy Space Center provided preflight,
launch, landing, and recovery services for KSC. Within the directorate, the Shuttle
Vehicle Engineering department was responsible for the engineering management and
technical direction of preflight, launch, landing, and recovery activities for all Space
Shuttle vehicles and integration of payloads. An important function of this group
was to perform the out-of-family disposition (OFD) process, which dealt with any
operation or performance outside the expected range or that had not been previously
experienced. These anomalies were described as out-of-family in the sense that they
were new anomalies and differentiated from in-family anomalies that had previously
occurred. In the OFD process, new problems (which we’ll reference as cases) were
referenced, solved, and documented. As in problem-solving, drawing analogies to
similar prior cases helps to solve new problems. Therefore, this process lent itself
to the adoption of knowledge application technologies and to documenting these
anomalies in a way that made the solutions to these problems available to the rest of
the organization. As more unfamiliar cases were documented within the knowledge
application system, the case database grew and became more comprehensive.
In order to build the OFD prototype, a sample set of twelve OFD Problem Reports
(PR) were collected, each describing an anomaly identified during the processing of
P
R
O
V
I
D
E
R
S
P
A
Y
E
R
S
New
Prediction
Tool
New Learning &
Maintenance Tool
P
re
di
ct
io
n
P
ro
ce
ss
in
g
A
ut
om
at
ed
L
ea
rn
in
g
New Protocols Claim & Response data
Claim
Historical records of
denied and paid claims
Claim with
Automated selection
of attachments
LEGEND:
Existing Process
NEW TECHNOLOGY
Payment
P
R
O
V
I
D
E
R
S
P
A
Y
E
R
S
New
Prediction
Tool
New Learning &
Maintenance Tool
P
re
di
ct
io
n
P
ro
ce
ss
in
g
A
ut
om
at
ed
L
ea
rn
in
g Claim & Response data
Claim
Historical records of
denied and paid claims
Claim with
Automated selection
of attachments
LEGEND:
Existing Process
NEW TECHNOLOGY
Payment
Figure 6.7 Process Flow for the CLAIM Knowledge Application System
KNOWLEDGE APPLICATION SYSTEMS: SYSTEMS THAT UTILIZE KNOWLEDGE 119
the Space Shuttle, together with the anomaly resolution. The OFD PRs comprised
a cover page with 36 entries that described details for the OFD anomaly. Part of the
report also detailed the description and requirement of the troubleshooting plan,
damaged parts’ specifications, and alternative replacement parts. PRs were typically
from 10 to 70 pages in length and did not follow a prescribed format. The final pages
of the report included the most reasonable rationale, which detailed the most likely
reason for the failure, as well as the most reasonable repair plan and justification.
Also, the PR included engineering orders for replacement parts as well as related
part specifications. Finally, the OFD PR included a problem summary and conclu-
sion. Each of the twelve OFD PRs were used to build a case in the case library. The
twelve OFD PRs were very different from one another. The steps in the creation of
the case library were:
1. Identify and establish a set of categories or clusters to stratify them through
analysis of their similarities and differences. In our example, we found the PRs
grouped into four categories. This resulted in a 4th order distribution tree (see Figure
6.8). The most appropriate problem categories identified were Computer, Electrical,
Mechanical, and Materials. Given that this was only a prototype based on a total of
only 12 PRs, there was only an average of two to three cases corresponding to each
problem category.
2. Analyze each PR to identify a case title, a description, a set of characterizing ques-
tions and answers, and a resulting action. Building the case library required combining
information from the sections of the PR, since the reporting format of the PR was dif-
ferent from the way cases were stored in the case library. For example, for each case in
the library, the description of the action taken was deduced from a combination of the
PR sections describing the most reasonable rationale, summary, and conclusion.
3. Develop a set of descriptive questions for each case, which needed to be in
natural-language format. Some CBR software packages require that the set of de-
scriptive questions for each case must be normalized to ensure that the similarity
function will work properly. The objective of the similarity function is to identify
from the case database those cases that are similar to the case under analysis by the
end user. Consider for example the Electrical category in Figure 6.8. Normalization
means that cases 4, 7, and 11 should be described by a similar number of
those catalogued in different clusters) can be defined with fewer questions because
retrieval conflicts are less likely to occur.
4. Just as an expert draws upon her wide experience in order to infer solutions to
new problems, case-based systems work best when the case library is large enough to
be representative of the total set of possible anomalies. As such, in order to develop a
working prototype, the application developers were compelled to add permutations of
the OFD Problem Reports to the case library, so that the additional cases improved the
system’s ability of finding a relevant solution. With a total of 12 cases, permutations of
these original reports were developed to allow the library to represent a larger subset
of possible anomalies. These permutations were created through the definition of varia-
tions for each
120 CHAPTER 6
Referring back to Figure 6.4, each PR corresponds to case numbers 1 through 12. In
this example, the case corresponding to the Data Drop Out PR appears as case 2, and
the permutations corresponding to this case appear as 2.3, 2.4, and 2.6. The diagnosing
solution for case 2 is found after answering yes or no to the set questions that accurately
describe the Data Drop Out problem. The permutations for case 2 correspond to a dif-
fering answer to the questions that essentially describe the case. This process of adding
permuted cases resulted in a total set of 34 cases in the case library.
5. Following the development of the case library, the case library must be validated
to ensure the proper execution of the application. The validation process requires that
none of the following conditions exist in the case library, which essentially diminishes
the accuracy of the application:
a. Disjunctions (i.e., otherwise identical cases with separate solutions): Disjunc-
tive cases must be combined into a single case.
b. Internal disjunctions: These are characterized by situations in which a single
case in a cluster contains multiple questions that are not answered in any other
case in the same cluster. To resolve internal disjunctions, combine these ques-
tions into a single question with multiple answers, which allows the system
to match a conversation’s query containing either answer and also reduces
the number of questions in the cases.
c. Subsumed cases: These are characterized by one case being a logical spe-
cialization of another and having the same solution. In these circumstances,
eliminate the more specific case.
d. Validation: some case-authoring tools provides the ability to validate the
case library through an automated testing functionality. This functionality
allows for verification if the retrieval precision for the case library is ac-
ceptable.
Figure 6.8 Distribution Order Tree for the OFD Problem Reports
Prompt Problem
Computer Electrical Mechanical Materials
Data drop
out
FCMS
GMT
Discrep.
PCM3
shows up
in PCM2
Micro-
switch
malfunction
Un-
explained
power
drops
Helium ISO
valves
Back up
HGDS
Seal port
dynat.
Stress
corrosion
cracking
Catch bot.
relief valve
Debris
detected in
stiffener
ring
Cracked
A8U panel
2
2.3
2.4
2.6
12 4
4.1
7 11
11.1
11.2
11.3
11.4
11.5
11.6
1
1.1
1.2
1.3
3 5 6
6.1
10
10.1
10.2
10.3
8
8.2
8.3
9
9.1
9.2
Prompt Problem
Computer Electrical Mechanical Materials
Data drop
out
FCMS
GMT
Discrep.
Data drop
out
FCMS
GMT
Discrep.
PCM3
shows up
in PCM2
Micro-
switch
malfunction
Un-
explained
power
drops
PCM3
shows up
in PCM2
Micro-
switch
malfunction
Un-
explained
power
drops
Helium ISO
valves
Back up
HGDS
Seal port
dynat.
Stress
corrosion
cracking
Catch bot.
relief valve
Helium ISO
valves
Back up
HGDS
Seal port
dynat.
Stress
corrosion
cracking
Catch bot.
relief valve
Debris
detected in
stiffener
ring
Cracked
A8U panel
Debris
detected in
stiffener
ring
Cracked
A8U panel
2
2.3
2.4
2.6
2
2.3
2.4
2.6
12 4
4.1
4
4.1
7 11
11.1
11.2
11.3
11.4
11.5
11.6
11
11.1
11.2
11.3
11.4
11.5
11.6
1
1.1
1.2
1.3
1
1.1
1.2
1.3
3 5 6
6.1
6
6.1
10
10.1
10.2
10.3
10
10.1
10.2
10.3
8
8.2
8.3
8
8.2
8.3
9
9.1
9.2
9
9.1
9.2
KNOWLEDGE APPLICATION SYSTEMS: SYSTEMS THAT UTILIZE KNOWLEDGE 121
The user inputs the topic of trouble in the description box. After this is done, the software
will output all relevant cases in a ranked order inside the Ranked Cases box. Cases are
given a rank score according to their relevance to the description topic. The application
also outputs all relevant questions that are to be answered by the user in order to further
constrain the search to the most relevant topic. These questions are also ranked according
to their relevance to the topic. Once the user has answered the questions, a single most
relevant case is identified. Figure 6.9 presents the screen with the dialogue box which
shows the output of the search, the title, and description of the most relevant case (the
highest ranking case) and the steps and actions needed to solve the case. More information
about this case study can be found in Becerra-Fernandez and Aha (1998).
The key importance of the OFD system is that it enables one to apply the knowl-
edge gained through solving prior problems when solving new anomalies experienced
during the Shuttle Processing process. The OFD system helps to apply knowledge
to prevent unnecessary work from being performed, while promoting learning from
prior failures. Prior to the development of the OFD system, this information remained
in the tacit knowledge base of the engineer in charge of the process. Since NASA
enjoys the advantage of having a relatively stable workforce, engineers used their
own knowledge base to identify similar cases that they have solved in the past. But as
downsizing and requirements continue to be part of the federal government landscape,
systems like the OFD will be essential as a platform for knowledge application in
order to identify and apply prior relevant knowledge.
In the next section, we discuss how rule-based systems can be instrumental to the
design of troubleshooting systems that have stood the test of time.
Figure 6.9 Search Results
NaCoDAE • CONVtnsI H
Library Edi View Trial Cire Experiment
Start Tr;EJ New Query Step End Questions End Trial
Description Tile damage
Dialogue
Number of questions asked: 2
Retrieved Case Description:
case2
Title: Tile is slumped, C3 0.3″ x 0.15″
Actions:
acbon2 FWD2-25-5993?Repair Process A
, r k . . j mifm o*: ^ao.-jTO-.^.-t).-,,,.,,™ TS
Ranked Questions
“Yes” Question #3; Can You notice a contour change in the tile?
“Yes” Question #9: Was the tiling exposed to exceeding roughness conditions?
Ranked Cases
100 case2 Tie is slumped, C3 0.3″ x 0.15″
50 case3 Tile sidewall/ 1ML damage S3 / F 8 2 6″ x 0 35″ x 0.06″
I’.i-.n JAVA NdCoDAE CONVER.. 4:53 PM
122 CHAPTER 6
LIMITATIONS OF KNOWLEDGE APPLICATION SYSTEMS
There are some practical limitations to the development of knowledge application systems.
These relate to the fact that most of these systems are developed to serve a task-specific
domain problem and are typically not integrated with the organization’s enterprise systems.
Other limitations also exist—for example, for knowledge application systems based on
CBR technologies, the following limitations apply (Kitano and Shimazu 1996):
1. Security: Cases may include sensitive information. Knowledge application sys-
tems must consider the incorporation of security measures, including access control
according to the user’s organizational role. If knowledge application systems do not
incorporate security measures, systems may not realize their maximum value.
2. Scalability: Knowledge application systems must represent a large enough
number of cases so that the majority of the new experiences are represented in the
case-based system. This means the knowledge application system must reach satura-
tion prior to its deployment. Reaching system saturation means that most typical cases
would have already been reported in the system. The number of cases necessary to
reach the saturation point changes according to the domain. For SQUAD, discussed
earlier, reaching this point required the inclusion of about 3,000 cases each year, a
number that later was reduced to 1,000 per year. The more complex the domain, the
higher the importance of keeping the growth of the case base viable. Clearly, the
continual growth of the case library will also require the use of complex indexing
schemes, which may result in decreased system stability.
3. Speed: As the size of the case library grows to a more comprehensive repre-
sentation of real environments, computing and searching costs will also increase.
Therefore, developers of knowledge application systems must consider the use of
complex indexing schemes that will guarantee acceptable case-retrieval times and
performance levels.
In addition, knowledge application systems may not be able to solve all the prob-
lems that they encounter. In particular, diagnosing problems may be increasingly
difficult in complex environments, as evidenced by the Space Shuttle Columbia
tragedy. Diagnosing everything that went wrong or may go wrong in such envi-
ronments may not be possible with systems like the ones already described. New
technologies will need to be developed in order to prevent incidents in complex
engineering environments.
Some rule-based systems could suffer from other limitations, namely the lack of
scalability. Other technologies offer a different set of limitations. But in essence, the
benefits that the implementation of knowledge application systems brings to the or-
ganization outweigh their limitations, and they will continue to provide competitive
advantages to those organizations that continue to implement them.
SUMMARY
In this chapter we discussed what knowledge application systems are, along with
design considerations and specific types of intelligent technologies that enable such
KNOWLEDGE APPLICATION SYSTEMS: SYSTEMS THAT UTILIZE KNOWLEDGE 123
systems. The Case-Method Cycle, a methodology to effectively develop knowledge
application systems is presented. Also the chapter discusses different types of knowl-
edge application systems: expert systems, help desk systems, and fault diagnosis
systems. Six case studies that describe the implementation of knowledge application
systems are presented, each based on different intelligent technologies and designed
to accomplish different goals: provide advice, recognize fault detection, and spur
creative reasoning. The first system uses rules to troubleshoot electrical generators
in real time. The second system is based on rules to advise potential applicants to the
SBIR/STTR program if they meet the program’s criteria. The third system, based
on CBR technology, helps engineers diagnose faulty chips. The fourth system, also
based on CBR technology, helps improve customer service at Darty. The fifth system,
also CBR-based, is improving the reimbursement of claims for healthcare services.
Finally, the sixth system helps NASA engineers find solutions to new problems faced
while processing the Shuttle, assuming that new problems could be related or be a
combination of old problems. Finally, limitations of knowledge application systems
are discussed.
REVIEW
1. What are some of the intelligent technologies that provide the foundation for
the creation of knowledge application systems?
2. Describe in your own words when you should use rules as opposed to CBR
when developing a knowledge application system.
3. Describe the four steps in the CBR process.
4. Describe the steps and the importance of the Case-Method Cycle.
5. Explain the case library development process.
6. What are some of the limitations of knowledge application systems?
APPLICATION EXERCISES
1. Identify examples of knowledge application systems in use in your orga-
nization. What are some of the intelligent technologies that enable those
systems?
2. Describe five knowledge application scenarios that could be supported via
CBR systems and explain why. For example, one such scenario is a system
that will identify the most likely resolution of a court case based on the out-
come of prior legal cases.
3. Design a knowledge application system to support your business needs.
Describe the type of system and the foundation technologies that you would
use to develop such a system.
4. Design the system architecture for the system described in question 2
above.
5. Identify three recent examples in the literature of knowledge application
systems.
124 CHAPTER 6
NOTES
1. We acknowledge Avelino Gonzalez of the University of Central Florida for this case study.
2. We acknowledge Monica Wood, Sheila Oliva, and Carolyn Joiner all of Siemens for contribut-
ing the current status of the GenAID system. For more information on SPPA-D300 consult Siemens
2014.
3. We acknowledge National Semiconductor, now part of Texas Instrument’s Analog business, in
particular Art Hamilton, Mike Glynn, Mike Meltzer, and Amir Razavi, for their support in creating
this section.
4. We acknowledge Glenn Gardner and Greg Leary of Kaidara Software, Inc. for contributing this
material. Note that on April 2010, Kaidara Software, Inc. was acquired by Servigistics, Inc. (www.
servigistics.com).
5. We acknowledge William Cheetham, Bernhard Scholz, and Deborah Belcher of General Electric
Research for this box.
REFERENCES
Aamodt, A., and Plaza, E. 1994. Case-based reasoning: Foundational issues, methodological variations,
and system approaches. AI Communications, 7(1), 39–52.
Acorn, T., and Walden, S. 1992. SMART: Support management automated reasoning technology for
Compaq customer service. In Proceedings of the 4th Innovative Applications of Artificial Intelligence
Conference, San Jose, California, July 12–16, 1992.
Aha, D., Kibbler, D., and Albert, M. 1991. Instance-based learning algorithms. Machine Learning, 6(1).
Allen, B. 1994. Case-based reasoning: Business applications. Communications of the ACM, 37(3),
40–42.
Becerra-Fernandez, I. and Aha, D. 1998. Case-based problem solving for knowledge management
systems. In Proceedings of the 12th Annual International Florida Artificial Intelligence Research
Symposium (FLAIRS): Knowledge Management Track, Orlando, Florida, May 1999.
Becerra-Fernandez, I., Gonzalez, A., and Sabherwal, R. 2004. Knowledge management: Challenges,
solutions, and technologies. Upper Saddle River, NJ: Prentice Hall.
Bergmann, R., Freßmann, A., Maximini, K., Maximini, R., and Sauer, T. 2006. Case-based support for
collaborative business. In Advances in case-based reasoning, ed. T.R. Roth-Berghofer, M.H. Göker,
and H.A. Güvenir, 519–533. Berlin: Springer.
C-DAC Mumbai. 2014. Center for Development of Advance Computing Mumbai. www.cdacmumbai.
in/ (accessed January 5, 2014).
Chandrasekaran, B., Narayanan, H., and Iwasaki, Y. 1993. Reasoning with diagrammatic representa-
tions. Artificial Intelligence Magazine, 14(2), 49–56.
Cheetham, W. 2005. Tenth anniversary of plastics color matching. Artificial Intelligence Magazine,
26(3), 51–61.
Cheetham, W., and Goebel, K. 2007. Appliance call center: A successful mixed-initiative case study.
Artificial Intelligence Magazine, 28(2), 89–100.
Cheetham, W., and Watson, I. 2005. Fielded applications of case-based reasoning. The Knowledge
Engineering Review, 20(3), 321–323.
Davis, R. 1984. Diagnostic reasoning based on structure and behavior. Artificial Intelligence, 24(1–3),
347–410.
de Kleer, J. 1976. Local methods for localizing faults in electronic circuits. Memo 394, MIT Artificial
Intelligence Laboratory, Cambridge, MA, 154.
Domeshek, E. and Kolodner, J. 1991. Toward a case-based aid for architecture; toward a case-based
aid for conceptual design. International Journal of Expert Systems, 4(2), 201–220.
———. 1992. A case-based design aid for architecture. In Artificial intelligence in design 92, ed. J.S.
Gero, 487–516. Norwell, MA: Kluwer.
www.servigistics.com
www.servigistics.com
www.cdacmumbai.in/
www.cdacmumbai.in/
KNOWLEDGE APPLICATION SYSTEMS: SYSTEMS THAT UTILIZE KNOWLEDGE 125
———. 1993. Using the points of large cases. Artificial Intelligence for Engineering Design, Analysis,
and Manufacturing, 7(2), 87–96.
Fox, M. 1990. AI and expert system myths, legends, and facts. Intelligent Systems and Their Applica-
tions, 5(1), 8–20.
Genesereth, M. 1984. The use of design descriptions in automated diagnosis. Artificial Intelligence,
24(1), 411–436.
Glasgow, J., Narayanan, H., and Chadrasekaran, B. (Eds.). 1995. Diagrammatic reasoning: Cognitive
and computational perspectives. Cambridge, MA: MIT Press.
Gonzalez, A.J., Osborne, R.L., Kemper, C., and Lowenfeld, S. 1986. On-line diagnosis of turbine
generators using artificial intelligence. IEEE Trans. on Energy Conversion, EC-1(2), June,
68–74.
Griffith, A. and Domeshek, E. 1996. Indexing evaluations of buildings to aid conceptual design. In
Case-based reasoning experiences, lessons, and future directions, ed. D. Leake, 68–80. Menlo
Park, CA: AAAI Press.
Group of Artificial Intelligence Applications (GAIA). 2014. jCOLIBRI. http://gaia.fdi.ucm.es/research/
colibri/jcolibri.
Hsu, F., Anantharaman, T., Campbell, M., and Nowatzyk, A. 1990. A grandmaster chess machine.
Scientific American, 263(4) (October), 44–50.
Kibler, D. and Aha, D. 1987. Learning representative exemplars of concepts: An initial study. In Pro-
ceedings of the Fourth International Workshop on Machine Learning, UC-Irvine, June 24–29.
Kitano, K. 1993. Challenges for massive parallelism. In Proceeding of the 13th Annual Conference on
Artificial Intelligence (IJCAI-93), Chabery, France, 813–834.
Kitano, H., and Shimazu, H. 1996. The experience-sharing architecture: A case study in corporate-wide
case-based software quality control. In Case-based reasoning: experiences, lessons, and future
directions, ed. D. Leake, 235–268. Menlo Park, CA: AAAI Press.
Kolodner, J. 1991. Improving human decision making through case-based decision aiding. AI Maga-
zine, 12(2), 52–68.
———. 1993. Case-based reasoning. San Mateo, CA: Morgan Kaufmann.
Leake, D. 1996. CBR in context: The present and future. In Case-based reasoning: Experiences, les-
sons, and future directions, ed. D. Leake, 3–30. Menlo Park, CA: AAAI Press.
Magnani, L., Nersessian, N.J., and Thagard, P. (Eds.). 1999. Model-based reasoning in scientific dis-
covery. New York: Kluwer Academic Press.
Mark, W., Simoudis, E., and Hinkle, D. 1996. CBR: Expectations and results. In Case-based reason-
ing: Experiences, lessons, and future directions, ed. D. Leake, 269–294. Menlo Park, CA: AAAI
Press.
National Aeronautics and Space Administration (NASA). 2008. SBIR and STTR solutions. http://sbir.
nasa.gov/SBIR/solicit.htm.
Newell, A., and Simon, H.A. 1963. GPS, a program that simulates human thought. In Computers and
thought, ed. E.A. Fiegenbaum and J. Feldman, 279–296. New York: McGraw-Hill.
Ngyen, T., Czerwishki, M., and Lee, D. 1993. COMPAQ QuickSource: Providing the consumer with
the power of artificial intelligence. In Proceedings of the 5th Innovative Applications of Artificial
Intelligence Conference, Washington, D.C., July 11–15.
Patton, R.J., Frank, P.M., and Clark, R.N. (Eds.). 2000. Issues of fault diagnosis for dynamic systems.
Berlin: Springer-Verlag.
Russell, S., and Norvig, P. 2002. Artificial intelligence. Upper Saddle River, NJ: Pearson Education.
Schank, R. 1982. Dynamic memory: A theory of learning in computers and people. New York: Cam-
bridge University Press.
Siemens. 2014. SPPA-D3000 Diagnostic Suite. www.energy.siemens.com/hq/en/automation/power-
generation/diagnosis-monitoring-protection/sppa-d3000-diagnostic-suite.htm (accessed January
5, 2014).
Tsang, E. 1994. Foundations of constraint satisfaction. London: Academic Press.
http://gaia.fdi.ucm.es/research/colibri/jcolibri
http://gaia.fdi.ucm.es/research/colibri/jcolibri
http://sbir.nasa.gov/SBIR/solicit.htm
http://sbir.nasa.gov/SBIR/solicit.htm
www.energy.siemens.com/hq/en/automation/powergeneration/diagnosis-monitoring-protection/sppa-d3000-diagnostic-suite.htm
www.energy.siemens.com/hq/en/automation/powergeneration/diagnosis-monitoring-protection/sppa-d3000-diagnostic-suite.htm
126 CHAPTER 6
U.S. Department of Defense. 2009. Small Business Innovation Research and Small Business Technol-
ogy Transfer. www.acq.osd.mil/osbp/sbir/.
Velloso, M., and Carbonnell, J. 1993. Derivational analogy in PRODIGY. Machine Learning, 10(3),
249–278.
Verdande Technology. 2014. www.verdandetechnology.com/.
Watson, I. 2003. Applying case-based reasoning techniques for enterprise systems. San Francisco:
Morgan Kauffmann.
Watson, I., and Marir, F. 1994. Case-based reasoning: A review. Knowledge Engineering Review, 9(4),
327–354.
Zuckerman, M.B. 2013. The high cost of staying well. US News & World Report, October 22.
www.acq.osd.mil/osbp/sbir/
www.verdandetechnology.com/
127
7
Knowledge Capture Systems:
Systems that Preserve and
Formalize Knowledge
In the previous chapter, we discussed knowledge application systems. In this chapter,
we discuss what knowledge capture systems are about and how they serve to elicit
and store organizational and individual knowledge. Knowledge capture systems are
designed to help elicit and store knowledge, both tacit and explicit. Knowledge can
be captured using mechanisms or technologies so that the captured knowledge can
then be shared and used by others. Perhaps the earliest mechanisms for knowledge
capture date to the anthropological use of stories—the earliest form of art, education,
and entertainment. Storytelling is the mechanism by which early civilizations passed
on their values and their wisdom from one generation to the next.
In this chapter, we first discuss issues about organizational storytelling and how
this mechanism can support knowledge capture. We then discuss how technol-
ogy can enable the knowledge capture process. We also describe issues related
to how to design the knowledge capture system, including the use of intelligent
technologies in support of this process. In particular, the role of RFID technologies
in knowledge capture is discussed. We discuss two types of knowledge capture
systems: one that serves best to support educational settings; and a second system
that serves best to capture tactical knowledge. Recall from Chapter 2 that tactical
knowledge is defined as knowledge that pertains to the short-term positioning of
the organization.
For a quick overview of how organizations can utilize strategic stories, let us look
at a brief case study and how 3M Corporation uses stories to embody their innova-
tive culture in Box 7.1.
WHAT ARE KNOWLEDGE CAPTURE SYSTEMS?
As discussed in Chapter 4, knowledge capture systems support the process of eliciting
either explicit or tacit knowledge that may reside in people, artifacts, or organizational
entities. These systems can help capture knowledge existing either within or outside
organizational boundaries, among employees, consultants, competitors, customers,
suppliers, and even prior employers of the organization’s new employees. Knowledge
capture systems rely on mechanisms and technologies that support externalization and
internalization. Both mechanisms and technologies can support knowledge capture
systems by facilitating the knowledge management processes of externalization and
internalization.
128 CHAPTER 7
You may recall from Chapter 4 that knowledge capture mechanisms facilitate exter-
nalization (i.e., the conversion of tacit knowledge into explicit form) or internalization
(i.e., the conversion of explicit knowledge into tacit form). The development of mod-
els or prototypes, and the articulation of stories are some examples of mechanisms
that enable externalization. Learning by observation and face-to-face meetings are
Box 7.1
Using Stories to Build Effective Business Plans at 3M
Few companies rival 3M’s 100 record years of innovation. From the invention of sandpaper in
1904 to the invention of masking tape in 1925 and Post-it Notes in 1980, 3M’s culture is noted
by its use of stories. Stories are part of 3M’s sales representatives’ training, award ceremonies,
and in short a “habit-of-mind.” At 3M, the power of stories is recognized as a means to “see
ourselves and our business operations in complex, multidimensional forms—that we’re able to
discover opportunities for strategic change. Stories give us ways to form ideas about winning”
(Shaw et al. 1998, p. 41.)
Recently, recognition about the power of stories reached 3M’s boardroom. Traditionally at 3M,
business plans were presented through bulleted lists. Cognitive psychologists have proven that
lists are ineffective learning artifacts since item recognition decreases with the length of the list
(Sternberg 1975), and typically only items at the beginning or end of the list are remembered
(Tulving 1983). As a contrast, a good story can better represent a business plan, since it in-
cludes a definition of the relationships, a sequence of events, and a subsequent priority among
the items which in turn causes the strategic plan to be remembered. Therefore, stories are cur-
rently used as the basic building blocks for business plans at 3M.
Shaw et al. (1998) define an effective business plan to be a lot like a good story, and ap-
propriately illustrate this with a narrative example. The strategic business plan must first set the
stage or define the current situation. For example:
Global Fleet Graphics (a 3M division that makes durable graphic-marking systems for
buildings, signs, and vehicles) was facing increasing demand from customers at the
same time that they were experiencing eroding market share due to diminishing patent
advantages and competitors’ low-cost strategies.
Next, the strategic story must introduce the dramatic conflict. Continuing with the same
example:
The 3M division had to effect a quantum change in the production system that enabled
the quick and competitive delivery of products. The solution included the development of
innovative technologies that enabled this group’s product offerings to differentiate from its
competitors. In addition, sales and marketing skills had to appropriately match the new
strategy.
Finally, the strategic narrative must reach a resolution. In other words, it must summarize
how the organization will win through effectively drawing upon the diverse technological skills
required to transform the business.
Studies at 3M show that the adequate use of a narrative business plan results in an improved
understanding of the requirements for the plan to succeed. In addition, using a narrative strat-
egy spurs excitement among 3M employees as well as generating commitment about the plan.
As summarized by Shaw et al. (1998):
When people can locate themselves in the story, their sense of commitment and involve-
ment is enhanced. By conveying a powerful impression of the process of winning, narra-
tive plans can motivate and mobilize an entire organization.
KNOWLEDGE CAPTURE SYSTEMS: SYSTEMS THAT PRESERVE AND FORMALIZE KNOWLEDGE 129
some of the mechanisms that facilitate internalization. Technologies can also support
knowledge capture by facilitating externalization and internalization. Externalization
through knowledge engineering, described earlier in Chapter 6, is necessary for the
implementation of intelligent technologies such as expert systems and case-based
reasoning systems, also described in Chapter 6. Technologies that facilitate internal-
ization include computer-based communication and computer-based simulations. For
example, an individual can use communication facilities to internalize knowledge
from a message sent by another expert or an AI-based knowledge-acquisition system.
Furthermore, computer-based simulations can also support individual learning. Both
knowledge capture mechanisms and technologies can facilitate externalization and
internalization within or across organizations.
KNOWLEDGE MANAGEMENT MECHANISMS FOR CAPTURING TACIT
KNOWLEDGE: USING ORGANIZATIONAL STORIES
The importance of using metaphors and stories as a mechanism for capturing and
transferring tacit knowledge is increasingly drawing the attention of organizations.
For example as illustrated in Box 7.1, 3M currently uses stories as part of its busi-
ness planning to set the stage, introduce dramatic conflict, reach a resolution to
the challenges the company is facing, and generate excitement and commitment
from all the members of the organization (Shaw et al. 1998). Storytelling at 3M
has taken center stage, as it seeks to develop a culture of problem preventers rather
than “an eleventh-hour problem-solver.” In order to reinforce this paradigm shift,
3M leaders turn to telling pet stories that describe “what not to do” (Clark 2004).
Storytelling is now considered of strategic importance as organizations recognize
the need to develop their company’s next generation leaders and has been recog-
nized as one of the most effective means to develop high-potential managers in a
firm (Ready 2002).
Stories are considered to play a significant role in organizations characterized by
a strong need for collaboration. Organizational stories are defined as
a detailed narrative of past management actions, employee interactions, or other intra- or
extra-organizational events that are communicated informally within organizations. (Swap
et al. 2001)
Organizational stories typically include a plot, major characters, an outcome,
and an implied moral. Stories originate within the organization and typically reflect
organizational norms, values, and culture. Because stories make information more
vivid, engaging, entertaining, and easily related to personal experience and because
of the rich contextual details encoded in stories, they are the ideal mechanism to
capture tacit knowledge (Swap et al. 2001). Stories have been observed to be use-
ful to capture and communicate organizational managerial systems (how things are
done), norms, and values.
Dave Snowden (1999), a long-time proponent of storytelling at IBM, identifies the
following set of guidelines for organizational storytelling:
130 CHAPTER 7
1. Stimulate the natural telling and writing of stories.
2. Stories must be rooted in anecdotal material reflective of the community in
question.
3. Stories should not represent idealized behavior.
4. An organizational program to support storytelling should not depend on ex-
ternal experts for its sustenance.
5. Organizational stories are about achieving a purpose, not entertainment.
6. Be cautious of overgeneralizing and forgetting the particulars. What has
worked in one organization may not necessarily work in others.
7. Adhere to the highest ethical standards and rules.
According to Phoel (2006), the following eight steps to successful storytelling will
help work magic in the organization:
1. Have a clear purpose.
2. Identify an example of successful change.
3. Tell the truth.
4. Say who, what, when.
5. Trim detail.
6. Underscore the cost of failure.
7. End on a positive note.
8. Invite your audience to dream.
But Phoel also emphasizes that to tell the story right, it is not just what you say but
how you say it that will determine its success. In this regard, stories should speak as
one person to another, they should present the truth as one sees it, should seem spon-
taneous but must be rehearsed, and one should relive the story as one tells it. Perhaps
the one theme that all storytelling experts agree on is the “crucial importance of truth
as an attribute of both the powerful story and the effective storyteller” (Guber 2007).
In fact, according to Guber, there are four kinds of truth in each effective story:
1. Truth to the teller—the storyteller must be congruent to her story.
2. Truth to the audience—the story must fulfill the listeners’ expectations by
understanding what the listeners know about, meeting their emotional needs,
and telling the story in an interactive fashion.
3. Truth to the moment—since great storytellers prepare obsessively and never
tell a story the same way twice.
4. Truth to the mission—since great storytellers are devoted to the cause, which
is embodied in the story, capturing and expressing the values that she believes
in and wants others to adopt as their own.
Other important considerations in the design of an effective organizational story-
telling program include (Post 2002):
1. People must agree with the idea that this could be an effective means of cap-
turing and transferring tacit organizational knowledge.
KNOWLEDGE CAPTURE SYSTEMS: SYSTEMS THAT PRESERVE AND FORMALIZE KNOWLEDGE 131
2. Identify people in the organization willing to share how they learned from
others about how to do their jobs.
3. Metaphors are a way to confront difficult organizational issues.
4. Stories can only transfer knowledge if the listener is interested in learning
from them.
In fact, one of the strengths of stories is that they are clearly episodic in nature, which
means related to events directly experienced. To the extent that the storyteller is able to
provide a sufficiently vivid account for the listener to vicariously experience the story,
many features of the story will be encoded in the listener’s memory and later available
for retrieval (Swap et al. 2001). In fact, the emphasis on the use of case studies at most
business schools is related to the effectiveness of stories as a pedagogical tool. Much
like case studies, Steve Denning (2000), who is best known for his efforts to implement
communities of practice and storytelling at the World Bank, describes the importance
of springboard stories. Springboard stories enable a leap in understanding by the audi-
ence in order to grasp how an organization may change by visualizing from a story in
one context what is involved in large-scale organizational transformations. Springboard
stories are told from the perspective of a protagonist who was in a predicament, which
may resemble the predicament currently faced by the organization. As an example of
a springboard story, consider the story used by Denning to convince his colleagues at
the World Bank about the importance of knowledge management:
In June of last year, a health worker in a tiny town in Zambia went to the Web site of the
Centers for Disease Control and got an answer to a question about the treatment of malaria.
Remember that this was in Zambia, one of the poorest countries in the world, and it was in a
tiny place six hundred kilometers from the capital city. But the most striking thing about this
picture, at least for us, is that the World Bank isn’t in it. Despite our know-how on all kinds of
poverty-related issues, that knowledge isn’t available to the millions of people who could use
it. Imagine if it were. Think what an organization we would become. (Denning 2005, p. 4)
An interesting question is the role storytelling plays with respect to analytical think-
ing. Denning (2000) supports the argument that storytelling supplements analytical
thinking by enabling us to imagine new perspectives and new worlds. He sees story-
telling as ideally suited to communicating change and stimulating innovation, because
abstract analysis is easier to understand when seen through the lens of a well-chosen
story and can of course be used to make explicit the implications of a story.
Finally, Denning (2000) describes the organizational areas where storytelling can
be effective, including:
1. Igniting action in knowledge-era organizations: Storytelling can help manag-
ers and employees actively think about the implications of change and the
opportunities for the future of their organization. Listeners actively understand
what it would be like if things were done a different way, re-creating the idea
of change as an exciting and living opportunity for growth.
2. Bridging the knowing-doing gap: This view proposes that storytelling can
exploit the interactive nature of communication by encouraging the listener
132 CHAPTER 7
to imagine the story and to live it vicariously as a participant. The listener
perceives and acts on the story as part of their identity.
3. Capturing tacit knowledge: Probably this line of reasoning is best captured
in Denning’s (2000) words: “Storytelling provides a vehicle for conveying
tacit knowledge, drawing on the deep-flowing streams of meaning, and of
patterns of primal narratives of which the listeners are barely aware, and so
catalyzes visions of a different and renewed future.”
4. To embody and transfer knowledge: A simple story can communicate a com-
plex multidimensional idea by actively involving the listeners in the creation
of the idea in the context of their own organization.
5. To foster innovation: Innovation is triggered by the inter-relatedness of ideas.
Storytelling enables to easily absorb and relate knowledge, the same spark
that triggers innovation.
6. Launching and nurturing communities: In many large organizations, the
formation of communities of practice enables the grouping of professionals
who come together voluntarily to share similar interests and learn from each
other. These communities of practice may be known under different names:
thematic groups (World Bank), learning communities or learning networks
(Hewlett-Packard Company), best practice teams (Chevron), and family
groups (Xerox). Denning (2000) explains how a storytelling program provides
a natural methodology for nurturing communities and integrating them to the
organization’s strategy and structure because:
a. Storytelling builds trust—enabling knowledge seekers in a community to
learn from knowledge providers through the sharing of candid dialogue.
b. Storytelling unlocks passion—because it enables the members of the
community to commit “passionately” to a common purpose, being the
engineering design of a new artifact, or sharing the discovery of a new
medical remedy.
c. Storytelling is nonhierarchical—because storytelling is collaborative, with
the members of the community pooling resources to jointly create the story.
7. Enhancing technology: Most people agree that e-mail has made increasing
demands in our lives, resulting in the expectation that we’re available 24/7
to answer electronic requests that span from office memos to virtual garbage
mail. Communities of practice and storytelling can enable us to interact with
our neighbors and remain connected when we want to, providing us with
“tranquility yet connectedness.”
8. Individual growth: The world of storytelling is one that proposes avoiding
adversarial contests and win-win for all sides: the knowledge seeker and the
knowledge-provider.
TECHNIQUES FOR ORGANIZING AND USING STORIES IN THE
ORGANIZATION
The power of narratives or stories as a knowledge capture mechanism in an organi-
zation lies in the fact that narratives capture the knowledge content as well as its
KNOWLEDGE CAPTURE SYSTEMS: SYSTEMS THAT PRESERVE AND FORMALIZE KNOWLEDGE 133
context and the social networks that define the way “things are done around here.” In
order to capture organizational knowledge through narratives, it is best to encourage
storytelling in a work context. In addition to the knowledge-elicitation techniques
described in Chapter 6, here we present knowledge-elicitation techniques pertaining
specifically to stories.
One technique described by Snowden (2000) for narrative knowledge capture
is anthropological observation, or the use of naïve interviewers, citing an example
where they used a group of school children to understand the knowledge flows in an
organization. The children were naïve, therefore they asked innocent and unexpected
questions which caused the subjects to naturally volunteer their anecdotes. They were
also curious, which resulted in a higher level of knowledge elicitation.
He also describes a second technique, storytelling circles, formed by groups having
a certain degree of coherence and identity such as a common experience in a project.
Story circles are best recorded in video. Certain methods can be used for eliciting
anecdotes such as:
1. Dit-spinning—or fish tales—represents human tendencies to escalate or better
the stories shared previously.
2. Alternative histories—are fictional anecdotes which could have different
turning points, based for example on a particular project’s outcome.
3. Shifting character or context—are fictional anecdotes where the characters
may be shifted to study the new perspective of the story.
4. Indirect stories—allow disclosing the story with respect to fictional characters,
so that any character similarities with the real-life character are considered
to be pure coincidence.
5. Metaphor—provides a common reference for the group to a commonly known
story, cartoon, or movie.
Once a number of stories has been elicited and captured, the next problem is how
to store the narratives so people will find them. Narrative databases can be indexed by
the theme of the story, by the stakeholders of the story, or by archetypal characters.
The theme could be, for example, innovative stories. The stakeholders could be the
scientists, the marketing group, or the customers. The archetypal characters represent
well-known characters that represent a virtue, for example, the good father archetype
represented by Bill Cosby in his TV role as Dr. Cliff Huxtable.
DESIGNING THE KNOWLEDGE CAPTURE SYSTEM
Typically the documentation available in organizations is the result of applying
expertise rather than expertise itself. For example, a radiologist interpreting high-
precision functional images of the heart will have the results of his diagnosis captured
in a document, but the reasoning process by which he reaches the diagnosis is not
usually captured. In addition, consider the process of engineering for complex sys-
tems. Traditional methods for documenting and representing the engineered designs
include creating engineering drawings, specifications, and computer-aided design
134 CHAPTER 7
(CAD) systems. But often the decisions leading to the design choices including
the assumptions, constraints, and considerations, are not captured. Capturing these
decisions is not only important but may lead to a more useful representation of the
design, specifically when designing complex systems in an environment character-
ized by high uncertainty.
Knowledge-elicitation techniques have been studied and used extensively in AI
for the development of expert systems (Chapter 6). The purpose of these techniques
is to assist the knowledge-elicitation process based on interview sessions between a
knowledge engineer and the domain expert, with the goal of jointly constructing an
expertise model. Although computers may understand the resulting expertise mod-
els, these models may not directly meet the objective of capturing and preserving
the expert’s knowledge so it can be transferred to others, or in other words, so others
can learn from it.
Next, we discuss how technology can facilitate capturing the knowledge of experts.
We will describe two such systems based on different methodologies and intelligent
technologies. The first system is based on the use of concept maps as a knowledge-
modeling tool. The second system is based on the use of context-based reasoning
(CxBR) to simulate human behavior. Each of these systems is best suited for certain
specific situations. For example, the use of concept maps may be best suited to capture
the knowledge of experts when supporting educational settings. On the other hand,
CxBR is best suited to capture the tactical knowledge of experts, which requires as-
sessment of the situation, selecting a plan of action, and acting on the plan. Both of
these knowledge capture systems can then be used to construct simulation models
of human behavior.
CONCEPT MAPS
KNOWLEDGE REPRESENTATION THROUGH THE USE OF CONCEPT MAPS
One type of knowledge capture system that we describe in this chapter is based on
the use of concept maps as a knowledge-modeling tool. Concept maps, developed
by Novak (Novak 1998; Novak and Cañas 2008; Novak and Gowin 1984), aim to
represent knowledge through concepts enclosed in circles or boxes of some types,
which are related via connecting lines or propositions. Concepts are perceived regu-
larities in events or objects that are designated by a label.
In the simplest form, a concept map contains just two concepts connected by a link-
ing word to form a single proposition, also called a semantic unit or unit of meaning.
For example, Figure 7.1 is a concept map that describes the structure of concept maps.
Based on the concept map represented in Figure 7.1, the two concepts—concept maps
and organized knowledge—are linked together to form the proposition: “Concept
maps represent organized knowledge.” Additional propositions expand the meaning
of concept maps, such as “Concepts are hierarchically structured.”
In a concept map, the vertical axis expresses a hierarchical framework for organiz-
ing the concepts. More general, inclusive concepts are found at the top of the map
with progressively more specific, less inclusive concepts arranged below them. These
135
F
ig
ur
e
7.
1
C
o
n
ce
p
t
M
ap
a
b
o
u
t
C
o
n
ce
p
t
M
ap
s
So
ur
ce
:
N
ov
ak
a
nd
C
añ
as
2
00
8.
I C
on
ce
pt
M
ap
s
re
pr
es
en
t
C
on
ce
pt
s
ar
e
ar
e
Pe
rc
ei
ve
d
R
eg
ul
ar
iti
es
L
ab
el
ed
m
be
gi
n
w
ith
E
ve
nt
s
(H
ap
pe
ni
ng
s)
O
bj
ec
ts
(T
hi
ng
s)
O
rg
an
iz
ed
K
no
w
le
dg
e
Pr
op
os
iti
on
s
in
cl
u
d
e
| S
ym
bo
ls
|
| W
or
ds
|
| C
re
at
iv
ity
be
gi
ns
w
ith
C
on
te
xt
D
ep
en
de
nt
A
ss
oc
ia
te
d
Fe
el
in
gs
or
A
ff
ec
t
Ef
fe
ct
iv
e
T
ea
ch
in
g
sh
aw
E
ff
ec
tiv
e
L
ea
rn
in
g
ne
ed
ed
to
s
ee
In
te
rr
el
at
io
ns
hi
ps
co
m
bi
ne
to
f
o
rm
a
re
‘
w
it
h
ai
ds
m
ay
b
e
ne
ce
ss
ar
y
fo
r
in
cl
ud
esis
In
fa
nt
s
is
C
ro
ss
lin
ks
H
ir
ea
ic
hi
ca
ll
y
St
ru
ct
ur
ed
be
tw
ee
n D
iff
er
en
t
M
ap
S
eg
m
en
t
136 CHAPTER 7
maps emphasize the most general concepts by linking them to supporting ideas with
propositions. Concept maps represent meaningful relationships between concepts in
the form of propositions. In addition relationships between concepts (propositions)
in different domains of the concept map are defined as cross-links. These cross-links
help to visualize how different knowledge domains are related to each other.
Sometimes the difference between concept maps and semantic networks could
be a source of confusion. Semantic networks, also called associative networks, are
typically represented as a directed graph connecting the nodes (representing concepts)
to show a relationship or association between them. This type of associative network
can be useful in describing, for example, traffic flow in that the connections between
concepts indicate direction, but directed graphs do not connect concepts through
propositions. Furthermore, in a directed graph there’s no assumption about the pro-
gression of generality to more specific concepts as the nodes are traversed from the
top of the network. The same holds true for associative networks in general.
Concept maps were developed based on Ausubel’s (1963) learning psychology
theory. Ausabel’s cognitive psychology research provides us with the understanding
that learning takes place through the assimilation of new concepts and propositions
into existing concept frameworks by the learner. Ausubel’s studies uncovered the
conditions for meaningful learning to include (1) a clear presentation of the material,
(2) the learner’s relevant prior knowledge, and (3) the learner’s motivation to integrate
new meanings into their prior knowledge. Concept maps can be useful in meeting
the conditions for learning by identifying concepts prior to instruction, building new
concept frameworks, and integrating concept maps through cross-links.
In educational settings, concept-mapping techniques have been applied to many
fields of knowledge. Their rich expressive power derives from each map’s ability to
allow its creator the use of a virtually unlimited set of linking words to show how
meanings have been developed. Consequently, maps having similar concepts can vary
from one context to another. Also, concept maps may be used to measure a particu-
lar person’s knowledge about a given topic in a specific context. Concept maps can
help formalize and capture an expert’s domain knowledge in an easy to understand
representation of an expert’s domain knowledge. Figure 7.2 shows a segment of a
concept map from the domain of nuclear cardiology.
KNOWLEDGE CAPTURE SYSTEMS BASED ON CONCEPT MAPS
The goal of CmapTools (Cañas et al. 2004),1 a concept map-based browser, is to
capture the knowledge of experts. The navigation problem, an important concern in
hypermedia systems, is alleviated by the use of concept maps, which serve as guides
in the traversing of logical linkages among clusters of related objects. The Cmap-
Tools extend the use of concept maps beyond knowledge representation to serve as
the browsing interface to a domain of knowledge.
Figure 7.3 shows the concept map-based browser as the interface for the expla-
nation subsystem of a nuclear cardiology expert system (Cañas et al. 1997, 2001,
2003, 2004; Cañas and Novak 2005; Ford et a1. 1996). Each of the concept nodes
represents an abstraction for a specific cardiology pathology, which is fully described
137
F
ig
ur
e
7.
2
A
S
eg
m
en
t
o
f
a
C
o
n
ce
p
t
M
ap
f
ro
m
t
h
e
D
o
m
ai
n
o
f
N
u
cl
ea
r
C
ar
d
io
lo
g
y
So
ur
ce
:
Fo
rd
e
t a
l.
19
96
.
P
ro
gn
os
tic
s!
ly
S
ig
ni
fic
an
t
m
ay
be
E
je
ct
io
n
: F
ra
ct
io
n
N
or
m
al
LV
F
un
ct
io
n
us
ua
lly
m
ea
ns
N
or
m
al
W
al
l
M
ot
io
n
w
hi
ch
ex
hi
bi
ts
A
bn
or
m
al
G
lo
ba
l
F
un
ct
io
n
w
ith
50
%
‘L
O
W
m
ay
in
di
ca
te
•r
i\
e
n
be
lo
u
m
ig
ht
b
e
D
ia
gn
os
tic
al
ly
S
ig
ni
fic
an
tisr
io
c
ca
n
be
1
‘H
IG
H
‘
/o
ft
e
n
M
te
te
s
w
he
n
N
or
m
al
G
lo
ba
l
F
un
ct
io
n
G
en
er
al
iz
ed
A
bn
or
m
al
iti
es
m
an
ife
st
as
‘B
U
L
L
5
E
Y
E
S
‘
‘D
IF
F
U
S
E
B
LU
E
1
er
e
bo
th
si
gn
s
of
C
ar
di
om
yo
pa
th
y
m
ay
in
d
ie
*!
–
O
O
C
U
f
w
ith
d
sp
ta
ys
R
ad
io
nu
cl
id
e
V
en
lr
ic
ul
og
ra
m
Js
p
ta
ys
A
re
pr
es
en
ta
tio
n
of
H
ea
rt
W
a
ll
M
ot
io
n
C
*J
I
b
e
‘
D
iff
er
en
tia
tio
n
A
bi
lit
y
be
tw
ee
n
m
ay
d
ec
re
as
e
w
ith
S
ev
er
e
Le
ft
V
en
tr
ic
le
D
ys
fu
nc
tio
n
“I
C
E
•
C
R
E
A
M
C
O
N
E
a
re
;
g
.-
i:
>.
o
f
V
al
ve
D
is
ea
se
w
he
n
sm
al
l
w
ha
n in
fe
rio
rd
is
pl
ay
s
if
?.
st
at
is
tic
al
ve
lo
ci
ty
of
in
te
rn
s
of
di
sp
la
y
bo
lu
s
ej
ec
tio
n
co
un
ts
R
at
e
Im
ag
es
N
or
m
al
S
fl-
“c
an
b
e
ca
n
be
be
tw
ee
n
S
ys
to
lic
D
ec
re
as
e
D
ia
st
ol
ic
I
nc
re
as
e
ir
~
S
eg
rn
en
ta
l
A
bn
on
na
lit
ie
s
pn
od
w
ce
s
m
an
ife
st
m
an
ife
st
?5
m
an
ife
st
as
‘B
A
LL
E
R
IN
A
F
O
O
T
‘
w
su
sJ
Iy
in
d
ita
ie
s
M
itr
al
V
al
ve
P
ro
la
ps
e
Is
ch
em
ia
ar
e
bo
th
si
gn
s
of
‘A
S
S
Y
M
E
T
R
I
Z
O
N
E
S
‘
‘B
LU
E
F
IN
G
E
R
S
138 CHAPTER 7
by the icons at the concept node. For the cardiologist, the image result of a Nuclear
Medicine Radionuclide Ventriculogram2 scan resembling a picture of “asymmetric
blue fingers” (later depicted in Figure 7.4) is a sign of myocardial ischemia or chronic
heart failure. An image resembling a “ballerina foot” is usually a representation of
a mitral valve prolapse. Clearly the first patient will quickly need to be rushed to a
hospital for emergency surgery, while the second may be given medication and a diet
to relieve him of his symptoms.
The icons below the concept nodes provide access to auxiliary information that helps
explain the concepts in the form of pictures, images, audio-video clips, text, Internet
links, or other concept maps related to the topic. These linked media resources and
concept maps can be located anywhere accessible via the Internet (Cañas et al. 2001).
The browser provides a window showing the hierarchical ordering of maps, highlights
the current location of the user in the hierarchy, and permits movement to any other
map by clicking on the desired map in the hierarchy. This concept map-based interface
provides a unique way of organizing and browsing knowledge about any domain.
CmapTools provides a practical application of the idea of utilizing concept maps
to capture and formalize knowledge resulting in context-rich knowledge representa-
Figure 7.3 Segment of a Concept Map from the Domain of Nuclear Cardiology,
Represented Using CmapTools
Source: Ford et al. 1996.
• . . I
, . . , • .
lr.-h,jn-r^..-.;S,-‘
^;.;jlLn;.r
IT^t>
• 4 0 * 1 –
EtKbcriudide
VenU kukfran ‘
“*s*j».
ri*hrtuJ
iKtHtfiMnh
.• it r i n 11 >L
CkUl
FunCBCn
mil •
• W l
V i –
•r,.
i. .
tKt
Function
• * i
NamMJWai:
r…
V . |
NomJLV
. . rivi’-^i’1:”
nft*-
:
” ^
tlbrTOtf 4f
bcdLv^jblkm
• i
Malta *.
Concept Map
J4TTMACES”
n .
INTQ^RUt’J JJJ
d»riHdv
• i- : I H I ,
FWJSFLTOAWALL
ta h
LCS^CFZCWK*
lhc*i
f;M-
.•>•.: iiii-::^:.’
NOTUMnffiTO
HUJ^.1MMH!R1’1R
I….
!•
FTJ1 -AlTT) Ffl Ai ,hJfFMTFT
YELLCV/REGlCN
• L L . [ ‘^••’.-l-l.-^
* • • • .
• . – • *
rr+jjhTW
;•-;;>•! ^ ” :
HH]
vi :.ij:.’ji.i Kv.;
IN – BIUU i ‘• Dn
•..I. ii
b>T.v;.-.i.-L-
CAROCMVCfATHV
• . – . . .
nftutei
INFERTR
LATTRAL
A H C A L
• v».L^
M.^t –
. . . ! –
GENEKALEE3
,.-.
Atriwnwnf,
li *m
K>:::-
• f •
. – : , .
5SLUE:-TNL-liBi:
• ” I , I I . .’Ij- Jix-
_ECHEM1A_
“”ECFgT IMAGES
REGIONALLY
Top Level Concept Map
First Pass
Diagnotically
Significan
File Edit Actions Display Help
KNOWLEDGE CAPTURE SYSTEMS: SYSTEMS THAT PRESERVE AND FORMALIZE KNOWLEDGE 139
tion models that can be viewed and shared through the Internet. CmapTools takes
advantage of the richness provided by multimedia, providing an effective platform
for aspiring students to learn from subject-matter experts.
Concept maps provide an effective methodology to organize and structure the
concepts representing the expert’s domain knowledge. During the knowledge cap-
ture process, the knowledge engineer and domain expert interact to collaboratively
construct a shared conceptual model of the domain which eventually becomes the
concept map for the multimedia system. Users later browse this conceptual model
through the CmapTools. Browsing enables the learner to implicitly gain the expert’s
view of the domain. In general, this model for knowledge representation provides a
broad view of the domain as understood by that particular domain expert.
Links in concept maps are explicitly labeled arcs, and usually connect two concepts
to form a concept-link-concept relation that may be read as a simple proposition.
CmapTools users learn about the domain by clicking on the small icons depicted at
the nodes in the concept map and directly navigate to other contexts (or subcontexts)
through hyperlinks where other concepts are described. Figure 7.4 shows some of the
different media windows opened from the windows in Figure 7.3.
Another advantage of using concept maps for knowledge representation is that,
because of their hierarchical organization, concept maps can easily scale to large
quantities of information. This particular characteristic can then enable the easy
integration of domain concepts together.
CmapTools has been shown to facilitate virtual collaboration and the creation
of concept maps at a distance (Cañas et al. 2003), which are stored on public serv-
ers that can be accessed via the Internet. Concept maps are the ideal mechanism
to make explicit and capture ideas so they can later be shared to collaboratively
create new knowledge. Studies also show that the use of concept maps has helped
improve areas of knowledge such as reading comprehension and language (Liu,
Chen, and Chang 2010).
In summary, concept-mapping tools like CmapTools can be an effective way to
capture and represent the knowledge of domain experts in representation models that
can later be used by potential students of the domain (Cañas and Novak 2005; Cañas,
Reiska and Novak 2009; Correia, Cicuto,and Aguiar 2014). Practically speaking, the
knowledge representation models illustrated in the aforementioned Figures 7.2, 7.3,
and 7.4 could be used by students in the field of cardiology to effectively learn the
practical aspects of the domain from one of the best experts in the field.
CONTEXT-BASED REASONING
KNOWLEDGE REPRESENTATION THROUGH THE USE OF CONTEXT-BASED REASONING
Recall from Chapter 2 that tactical knowledge is defined as pertaining to the short-term
positioning of the organization relative to its markets, competitors, and suppliers; and
is contrasted to strategic knowledge, which pertains to the long-term positioning of the
organization in terms of its corporate vision and strategies for achieving that vision. In the
context of this example tactical knowledge refers to the human ability that enables domain
140 CHAPTER 7
experts to assess the situation at hand (therefore short-term) among a myriad of inputs,
select a plan that best addresses the current situation, and execute that plan (Gonzalez and
Ahlers 1998; Thorndike and Wescort 1984). Consider the following scenario:
The commanding officer of the submarine is generally bombarded with a multitude of inputs
when performing his job. He receives audio inputs such as engine noise, electronic noise, and
conversations with others around him. He likewise receives visual inputs such as the radar and
sonar screens, possibly the periscope and so forth, and tactile inputs such as vibrations of the
submarine. He is able to cognitively handle these inputs rather easily when they are all in the
normal expected range. However, if one of these should deviate from normal, such as abnormal
noise and vibrations, the officer will immediately focus only on these inputs in order to recognize
the present situation as, for instance, a potential grounding, collision, or engine malfunction. All
other inputs being received, meanwhile, are generally ignored during the crisis.
Alternatively, consider an example more relevant to our daily lives:
The daily routine drive to and from work is marked by a myriad of inputs while performing
the task. A Dad driving to work with his children receives audio inputs such as the noise
from babies, siblings vying for attention, pop music, their spouse’s conversation, and who
Figure 7.4 The Explanation Subsystem Based on the Concept Map
Source: Ford et al. 1996.
File nu t Hi tiuns Display
rnnrrpt Map
1:19PM
• f c j j . •
FtatFas
Kadkjnudide
“”Ventriculogsm””
I > ̂ ^jiLtd:.-.IK;
PiDgioaicalljr’
5^iifitsnt
RATE IMAGES
“1
j . .:.,
mailto:elockhart@business.queensu.ca
KNOWLEDGE DISCOVERY SYSTEMS: SYSTEMS THAT CREATE KNOWLEDGE 205
perhaps the recent proliferation of e-commerce applications providing reams of hard
data ready for analysis presents us with an excellent opportunity to make profitable
use of these techniques. The increasing availability of computing power and inte-
grated DM software tools, which are easier than ever to use, have contributed to the
increasing popularity of DM applications to businesses. Many success stories have
been published in the literature describing how data mining techniques have been
used to create new knowledge. Here we briefly describe some of the more mature
and/or specifically relevant applications of data mining to knowledge management
for business.
Over the last decade, data mining techniques have been applied across business
problems.1 Examples of such applications are as follows:
1. Marketing—Predictive DM techniques, like artificial neural networks
(ANN), have been used for target marketing including market segmenta-
tion. This allows the marketing departments using this approach to segment
customers according to basic demographic characteristics such as gender,
age group, and so forth, as well as their purchasing patterns. They have also
been used to improve direct marketing campaigns through an understand-
ing of which customers are likely to respond to new products based on their
previous consumer behavior.
2. Retail—DM methods have likewise been used for sales forecasting. These
take into consideration multiple market variables, such as customer profiling
based on their purchasing habits. Techniques like market basket analysis
also help uncover which products are likely to be purchased together. Box
9.6 describes how Amazon takes advantage of this technique to improve the
user’s shopping experience and increase their sales.
Figure 9.1 IDEATECTS Problem Solving Model
206 CHAPTER 9
3. Banking—Trading and financial forecasting have also proven to be excellent ap-
plications for DM techniques. These are used to determine derivative securities
pricing, futures price forecasting, and stock performance. Inferential DM tech-
niques have also been successful in developing scoring systems to identify credit
risk and fraud. An area of recent interest is attempting to model the relationships
among corporate strategy, financial health, and corporate performance.
4. Insurance—DM techniques have been used for segmenting customer groups
to determine premium pricing and to predict claim frequencies. Clustering
techniques have also been applied to detecting claim fraud and to aid in cus-
tomer retention.
5. Telecommunications—Predictive DM techniques, like artificial neural net-
works, have been used mostly to attempt to reduce churn, that is, to predict
when customers will attrition to a competitor. In addition, predictive tech-
niques like neural networks can be used to predict the conditions that may
cause a customer to return. Finally, market basket analysis has been used to
identify which telecommunication products customers are likely to purchase
together.
6. Operations management—Neural networks have been used for planning and
scheduling, project management, and quality control.
Diagnosis is a fertile ground for mining knowledge. Diagnostic examples typically
abound in large companies with many installed systems and a wide network of service
representatives. The incidents are typically documented well, and often in a highly
structured form. Mining the incident database for common aspects in the behavior
of particularly troublesome devices can be useful in predicting when they are likely
to fail. Having this knowledge, the devices can be preventatively maintained in the
short-range and designed or manufactured in a way to avoid the problem altogether
Box 9.6
Amazon Making Use of Market Basket Analysis
The popular online retailer Amazon.com takes advantage of market basket analysis techniques
to provide its customers with high-quality recommendations of products that they’re most likely
to buy. How does it work? Amazon uses advanced algorithms to analyze its customer’s activ-
ity once they log onto their Web site. The company analyzes what items the customer views,
searches for, purchases and rates, as well as the shopping cart activity. For example, Amazon
keeps track of any items added and removed from the shopping carts as well as those left un-
purchased. By doing so, the company is able to optimize the shopping experience and custom-
ize each Web page as if it were personally tailored for the customer.
After processing all of this data with its algorithms, Amazon can combine this informa-
tion from each customer’s current and past shopping sessions with that of other customers.
The company’s patented algorithm, item-to-item collaborative filtering, matches each of the
user’s purchased and rated items to similar items, then combines those similar items into a
recommendation list. By leveraging this data, Amazon is able to present its customers with rec-
ommendations that are most likely to be of interest and are most likely to lead to a sale (Linden
et al. 2003).
KNOWLEDGE DISCOVERY SYSTEMS: SYSTEMS THAT CREATE KNOWLEDGE 207
in the long term. Witten (2000) mentions a specific example where diagnostic rules
were mined from 600 documented faults in rotating machinery (e.g., motors, genera-
tors) and compared to the same rules elicited from a diagnostic expert. It was found
that the learned rules provided slightly better performance than the ones elicited from
the expert.
In the electric utility business, neural networks have been used routinely to predict
the energy consumption load in power systems. The load on a power system depends
mostly on the weather. In hot weather areas, air conditioners during the summer rep-
resent the biggest load. In cold regions, it is the heating load in the winters. Knowing
the weather forecast and how that maps to the expected load can help forecast the
load for the next 24, 48, and 72 hours, and thereby place the appropriate generating
capacity in readiness to provide the required energy. This is particularly important
because efficient power stations cannot be turned on and off within minutes if the
load is greater than expected. Nuclear power stations (the most efficient) take several
days or weeks to place online from a cold state. Coal- or oil-fired stations (the next
most efficient) take the better part of one day to do the same. Although other types of
generating equipment can be turned quickly on and off, they are highly inefficient and
costly to operate. Therefore, utilities greatly benefit if they can bring their efficient
units online in anticipation of energy load increases, yet running them unnecessarily
can also be expensive.
All major electric utilities have entire departments expressly dedicated to this
load-forecasting function. The expected temperature is the most influential factor.
However, other attributes such as the day of the week, the humidity, and wind speed
have some influence as well. Data mining in this context consists of training neural
networks to predict the energy load in a certain area for a specified period of time.
This is considered supervised training. The relations are embedded in the weights
computed by the training algorithm, typically the back-propagation algorithm. By
mining a database containing actual recorded data on ambient temperatures, wind
speed, humidity, day of the week (among others), and the actual power consumed
per hour, the network can be trained. Then, the forecast values can be fed the same
attributes and it can predict the load on a per hour basis for 24, 48, and 72 hours.
Positive results in this arena have led the Electric Power Research Institute to offer
neural network-based tools to perform this specific function.
Witten (2000) describes a use of data mining for credit applications. In this project,
a credit institution undertook a project in data mining to learn the characteristics of
borrowers who defaulted on their loans in order to better identify those customers who
are likely to default on their loans. Using 1,000 examples and 20 attributes, a set of
rules was mined from the data, which resulted in a 66 percent successful prediction
rate. Back in the mid-1990s, 95 percent of the top banks in the United States were
already utilizing data mining techniques (Smith and Gupta 2000). For example, Bank
of Montreal was facing increased competition and the need to target-sell to its large
customer base. Earlier telemarketing attempts had proven unsuccessful; therefore the
bank embarked on an attempt to develop a knowledge discovery system to determine
a customer’s likelihood of purchasing new products. As a result, the bank could seg-
ment its customers for more targeted product marketing campaigns (Stevens 2001).
208 CHAPTER 9
A more recent example by Li, Wang, and Yen (2012) describes a use of data mining
for identifying fraudulent bank accounts. In this project, the authors collected transac-
tion data from 10,216 accounts from a bank in Taiwan. They analyzed metrics such
as transaction type, transaction time, online banking activity, and transaction amount
and found that the system had an effectiveness of about 65 percent.
Nevertheless, the most common and useful applications are in product marketing
and sales and in business operations. Every time someone purchases a product, a sales
record is kept. Often, these records contain demographic information on the buyer
and other times not. In any case, obtaining a personal profile of the purchasers of the
product can serve to better direct the product to this cross-section of consumers or
expand its appeal to other cross sections not currently purchasing the product. This is
true for not only hard products but also for services such as mobile services, Internet
service providers, banking and financial services, and others.
For example, Proflowers is a Web-based flower retailer. Flowers perish quickly;
therefore, Proflowers must level its inventory as the day progresses in order to
adequately serve its customers. Proflowers has achieved better management of its
customer traffic via inventory optimization that downplays the better-selling products
on their Web-storefront while highlighting the slower-selling ones. Based on their
analysis of Web purchases, Proflowers is able to change their Web site throughout
the day and therefore effectively attract attention to lower-selling items through their
Web site (Stevens 2001). Proflowers has also been innovative in their use of Customer
Relationship Management (CRM) software. By connecting growers to customers
via the Proflowers Web site and eliminating the middleman, Proflowers is able to
guarantee fresh flowers to its customers. Their CRM software is highly scalable and
flexible because it can communicate easily with its suppliers’ systems to ensure fast
and accurate order processing (ProFlowers 2012).
Another example is eBags, a Web-based retailer of suitcases, wallets, and related
products. Through the use of Web content mining, the company is able to determine
which Web pages result in higher customer purchases. This information is used to
adequately determine how Web content can drive the sales process. Finally eBags
uses the results from their Web content mining to help them personalize their retailing
Web pages on the fly, based on customer’s buying preferences and even geographic
location. For example, capturing the Web visitor’s zip code could be used to infer how
affluent the online shopper is. If she comes from an affluent neighborhood, the Web
site may feature designer items. If the online shopper comes from a zip code marked
by a large number of apartments, discounted offers would be made prominent in the
user’s view of the Web store (Stevens 2001).
Data mining techniques have also been used in areas as diverse as facilitating the
classification of a country’s investing risk based on a variety of factors and identify-
ing the factors associated with a country’s competitiveness (Becerra-Fernandez et al.
2002). For a quick overview of what knowledge discovery systems are and how they
are used, let us look at Box 9.7, which shows how Britain’s Safeway supermarkets
(now Morrisons Supermarkets) have used data mining techniques to recommend
products to shoppers and thus increase sales. In Box 9.8, we describe the role that
data mining can play in detecting money laundering and terrorist financing.
KNOWLEDGE DISCOVERY SYSTEMS: SYSTEMS THAT CREATE KNOWLEDGE 209
The KDD process is viewed as both an interactive and iterative process that turns
data into information and information into business knowledge. Next, we discuss the
steps in the KDD process.
DESIGNING THE KNOWLEDGE DISCOVERY SYSTEM
Discovering knowledge can be different things for different organizations. Some organi-
zations have large databases, while others may have small ones. The problems faced by
the users of data mining systems may also be quite different. Therefore, the developers of
DM software face a difficult process when attempting to build tools that are considered
generalizable across the entire spectrum of applications and corporate cultures. Early
efforts to apply data mining in business operations faced the need to learn, primarily
via trial and error, how to develop an effective approach to DM. In fact, as early adopt-
ers of DM observed an exploding interest in the application of techniques, the need to
Box 9.7
Filling up the Grocery Cart at Morrisons Supermarkets
Morrisons Supermarkets (previously Safeway Stores) is one of Britain’s leading food grocers,
employing around 90,000 staff and owning more than 500 stores across the country. Since
the penetration of personal computers (PCs) in Britain is lower than in the United States, IBM
developed an application for Britain’s Morrisons Supermarkets that enabled customers to
prepare their shopping lists on a personal digital assistant (PDA) and transmit their order to the
store for subsequent pickup without having to walk the aisles of the market. Shoppers quickly
jumped on the convenience, which removes the marketer from the opportunity to suggest via
attractive displays the spontaneous purchase of additional products that invariably fill up the
shopping cart.
In order to provide for a means to suggest the purchase of additional products, the supermarket
turned to the use of data mining as a means to recommend additional purchases to its clients. The
idea of personalizing the recommendations was based on the prior successful implementations
of such systems, which work by filtering a set of items (for example, grocery products) through a
personal profile. This filtering may be content-based, which recommends items based on what a
person has liked in the past. Alternatively, the filtering may be collaborative, which recommends
items that other people, similar to the one at hand, have liked in the past (Lawrence et al. 2001).
Market shoppers construct a shopping list via the PDA and e-mail it to the grocer’s server. Shop-
pers select products from lists residing in the personal catalog, the store’s recommendations, and
special promotions. Customers are clustered based on their prior purchasing behavior, and a list of
most popular products is generated to represent the preferred product purchases for the customers
in each cluster. The recommender system then ranks this list of products, according to computed
affinities with each customer, to produce a list of 10 to 20 of the highest-ranking products. When
customers synchronize with the store server they are presented with the recommended list, which in
fact contains products that were not previously purchased by the customer.
Results demonstrated that using the recommender resulted in 25 percent of the orders including
something from the recommendation list and a 1.8 percent increase in revenue. The study demon-
strates that data mining can help improve the understanding of customer preferences and thereby
boost the business revenue. For more information on this study, see Lawrence et al. (2001).
Other developments at Morrisons include testing of new in-store shopping cart technology to
flash personalized ads to customers while they shop (Gilbert 2002). As customers walk down
the store aisles, the shopping cart screen flashes promotions and coupons based on prior
purchasing patterns which remain stored in their loyalty club cards.
210 CHAPTER 9
Box 9.8
Using Data Mining to Detect Money Laundering and Terrorist Financing
In the past, U.S. intelligence agencies have prevented money laundering and terrorist financ-
ing via focused attention on transactions in the financial service sectors, such as banks and
other financial service institutions. But a more significant and largely overlooked mechanism for
money laundering is via abnormal international trade pricing. Overvaluing imports or undervalu-
ing exports is perhaps the oldest technique used to elude the government’s attention when
laundering money across international borders. Paying a higher value for an imported product
means the money is shifted to the foreign exporter who could be an operative of a terrorist
organization. Similarly, undervaluing exports, preferred by terrorists and money launderers
because they avoid the use of financial institutions, involves purchasing products at market
price for cash and then exporting to a colluding importer at below market prices who resells the
goods for their true value. All of these activities can translate to customs fraud, income tax eva-
sion, and money laundering (Zdanowicz 2004, 2009).
A data mining study of the 2001 U.S. import and export transactions produced by the U.S.
Department of Commerce reported suspicious prices that translated to overvalued imports and
undervalued exports to the tune of US$156.22 billion in 2001. Money laundered from the United
States to countries appearing in the U.S. State Department Al Qaeda watch list was estimated
to be around $4.27 billion that same year (Zdanowicz 2004, 2009). In this respect, data mining
is without a doubt critical in order to win the war against terrorism.
develop a standard process model for KDD became apparent. This standard should be
well-reasoned, nonproprietary, and freely available to all DM practitioners.
In 1999, a consortium of vendors and early adopters of DM applications for busi-
ness operations—consisting of Daimler-Chrysler (then Daimler-Benz AG, Germany),
NCR Systems Engineering Copenhagen (Denmark), SPSS/Integral Solutions Ltd.
(England), and OHRA Verzegeringen en Bank Groep B.V. (The Netherlands)—
developed a set of specifications called Cross-Industry Standard Process for Data
Mining (CRISP-DM) (Brachman and Anand 1996; Chapman et al. 2000; Edelstein
1999). CRISP-DM is an industry consortium that developed an industry-neutral and
tool-neutral process for data mining. CRISP-DM defines a hierarchical process model
that defines the basic steps of data mining for knowledge discovery as follows.
BUSINESS UNDERSTANDING
The first requirement for knowledge discovery is to understand the business prob-
lem. In other words, to obtain the highest benefit from data mining, there must be
a clear statement of the business objectives. For example, a business goal could be
“to increase the response rate of direct mail marketing.” An economic justification
based on the return of investment of a more effective direct mail marketing may be
necessary to justify the expense of the data mining study. This step also involves an
assessment of the current situation, for example:
the current response rate to direct mail is 1 percent. Results of the study showed that using
35 percent of the current sample population for direct mail (the one that is likely to buy the
product), a marketing campaign could reach 80 percent of the prospective buyers.
KNOWLEDGE DISCOVERY SYSTEMS: SYSTEMS THAT CREATE KNOWLEDGE 211
In other words, the majority of the people in a marketing campaign who receive a
target mail do not purchase the product. This example illustrates how you could ef-
fectively isolate 80 percent of the prospective buyers by mailing only to 35 percent of
the customers in a sample marketing campaign database. Identifying the most likely
prospective buyers from the sample and targeting the direct mail to those customers
could save the organization significant costs, mainly those associated with mailing a
piece to 65 percent of the customers who are the least likely to buy the new product
offering. The maximum profit occurs from mailing to the 35 percent of the custom-
ers that are most likely to buy the new product. Finally, this step also includes the
specification of a project plan for the DM study.
DATA UNDERSTANDING
One of the most important tenets in data engineering is “know thy data.” Knowing the
data well can permit the designers to tailor the algorithm or tools used for data mining to
their specific problem. This maximizes the chances for success as well as the efficiency
and effectiveness of the knowledge discovery system. This step, together with prepara-
tion and modeling, consumes most of the resources required for the study. In fact, data
understanding and preparation may take from 50 percent to 80 percent of the time and
effort required for the entire knowledge discovery process. Typically, data collection for
the data mining project requires the creation of a database, although a spreadsheet may
be just as adequate. Data mining doesn’t require data collection in a data warehouse and
in the case the organization is equipped with a data warehouse, it’s best not to attempt to
manipulate the data warehouse directly for the purpose of the DM study. Furthermore,
the structure of the data warehouse may not lend itself for the type of data manipulation
required. Finally, the construction of a data warehouse that integrates data from multiple
sources into a single database is typically a huge endeavor that could extend a number of
years and cost millions of dollars (Gray and Watson 1998). Most data mining tools enable
the input data to take many possible formats, and the data transformation is transparent to
the user. The steps required for the data understanding process are as follows.
1. Data Collection
This step defines the data sources for the study, including the use of external public
data (e.g., real estate tax folio) and proprietary databases (e.g., contact information
for businesses in a particular zip code). The data collection report typically includes
the following: a description of the data source, data owner, who (organization and
person) maintains the data, cost (if purchased), storage format and structure, size
(e.g., in records, rows, etc.), physical storage characteristics, security requirements,
restrictions on use, and privacy requirements.
2. Data Description
This step describes the contents of each file or table. Some of the important items in
this report are number of fields (columns) and percent of records missing. Also for
212 CHAPTER 9
each field or column: data type, definition, description, source, unit of measure, num-
ber of unique values, list and range of values. Also some other valuable specifics are
about how the data were collected and the time frame when the data were collected.
Finally, in the case of relational databases, it is important to know which attributes
are the primary or foreign keys.
3. Data Quality and Verification
In general, good models require good data; therefore, the data must be correct and
consistent. This step determines whether any data can be eliminated because of ir-
relevance or lack of quality. In addition, many data mining packages allow specifying
which columns in a table will be ignored (for the same reasons) during the modeling
phase. Furthermore, missing data can cause significant problems. Some data mining
algorithms (e.g., C5.0) can handle the missing data problem by automatically mas-
saging the data and using surrogates for the missing data points. Other algorithms
may be sensitive to missing values. In that case, one approach would be to discard
the data sample if some of the attributes or fields are missing which could cause a
substantial loss of data. A better approach is to calculate a substitute value for the
missing values. Substitute values could consist of the mode, median, or mean of the
attribute variable depending on the data type.
4. Exploratory Analysis of the Data
Techniques such as visualization and online analytical processing (OLAP) enable
preliminary data analysis. This step is necessary to develop a hypothesis of the prob-
lem to be studied and to identify the fields that are likely to be the best predictors. In
addition, some values may need to be derived from the raw data, for example factors
such as per capita income may be a more relevant factor to the model than the factor
income.
DATA PREPARATION
The steps for this task are as follows.
1. Selection
This step requires the selection of the predictor variables and the sample set. Selecting
the predictor variables is necessary because typically data mining algorithms don’t
work well if all the variables (fields or database columns) are considered as potential
predictors. In essence, that’s why data mining requires an understanding of the domain
and the potential variables influencing the outcome in question. As a rule-of-thumb,
the number of predictors (columns) must be smaller than the number of samples
(rows) in the data set. In fact, the number of simple observations should be at least
10 to 25 times the number of predictors. As the number of predictors increases, the
computational requirement to build the model also increases. Selecting the sample
KNOWLEDGE DISCOVERY SYSTEMS: SYSTEMS THAT CREATE KNOWLEDGE 213
set is necessary because when the data set is large, a sample of the data set can be
selected to represent the complete data set. In selecting the sample, attention must
be paid to the constraints imposed by sampling theory in order for the sample to be
representative of the complete data set.
2. Construction and Transformation of Variables
Often, new variables must be constructed to build effective models. Examples include
ratios and combination of various fields. Furthermore, some algorithms, like market
basket analysis, may require data to be transformed to categorical format (integer)
when in fact the raw data exist in continuous form. This may require transformations
that group values in ranges like low, medium, and high.
3. Data Integration
The data set for the data mining study may reside on multiple databases, which
would need to be consolidated into one database. Data consolidation may require
redefinition of some of the data fields to allow for consistency. For example, differ-
ent databases may relate to the same customer with different names; for instance,
one database may refer to the National Aeronautics and Space Administration while
other database fields may just use NASA. These incompatibilities must be reconciled
prior to data integration.
4. Formatting
This step involves the reordering and reformatting of the data fields as required by
the DM model.
MODEL BUILDING AND VALIDATION
Building an accurate model is a trial-and-error process. The process often requires the
data mining specialist to iteratively try several options until the best model emerges.
Furthermore, different algorithms could be tried with the same data set and the results
then compared to see which model yields the best results. For example, both neural
network and rule induction algorithms could be applied to the same data set to develop
a predictive model. The results from each algorithm could be compared for accuracy
in their respective predictive quality. Following the model development, the models
must be evaluated or validated. In constructing a model, a subset of the data is usually
set aside for validation purposes. This means that the validation data set is not used
to develop or train the model but to calculate the accuracy of predictive qualities of
the model. The most popular validation technique is n-fold cross-validation, specifi-
cally ten-fold validation. The ten-fold cross validation divides the population of the
validation data set into ten approximately equal-sized data sets and then uses each of
the ten holdout sets a single time to evaluate the models developed with the remaining
nine training sets. For each of the ten models (the last model includes using the whole
214 CHAPTER 9
data set) the accuracy is determined, and the overall model accuracy is determined
as the average of each of the model samples.
EVALUATION AND INTERPRETATION
Once the model is determined, the validation data set is fed through the model.
Because the outcome for this data set is known, the predicted results are compared
with the actual results in the validation data set. This comparison yields the accuracy
of the model. As a rule-of-thumb, a model accuracy of around 50 percent would be
insignificant because that would be the same accuracy as for a random occurrence.
DEPLOYMENT
This step involves implementing the “live” model within an organization to aid the
decision-making process. A valid model must also make sense in the real world, and
a pilot implementation is always warranted prior to deployment. Also, following
implementation it’s important to continue to monitor how well the model predicts
the outcomes and the benefits that this brings to the organization. For example, a
clustering model may be deployed to identify fraudulent Medicare claims. When
the model identifies potential instances of fraud, and these instances are validated as
indeed fraudulent, the savings to the organization from the deployment of the model
should be captured. These early successes will then act as champions and will result in
continued implementation of knowledge discovery models within the organization.
Figure 9.2 summarizes the CRISP-DM process methodology. Figure 9.3 illustrates
the iterative nature of the CRISP-DM process.
CRISP-DM is only one of the institutions that have ongoing efforts towards stream-
lining the KDD process. Another similar consortium is the Data Mining Group
(DMG), an independent, vendor led group, which develops data mining standards,
such as the Predictive Model Markup Language (PMML).
In general, the goal that these standards pursue is to facilitate the planning, documen-
tation, and communication in data mining projects and to serve as a common reference
framework for the DM industry. Many of these standards were developed based on
practical experience resulting from the implementation of DM projects. In fact, the
purpose of these standards is to help people involved in the process to communicate.
The next section describes in detail step four in the KDD process, which is building
and validating the data mining predictive model.
GUIDELINES FOR EMPLOYING DATA MINING TECHNIQUES
Once the goal of the data mining system is understood (step 1 of the CRISP-DM
process) and the data have been collected (step 2) and prepared (step 3), the next step
involves building and validating the data mining model (step 4). In terms of defining
the adequate data mining techniques to be used, the nature of the data will play the
deciding role as to which technique is most appropriate. Input variables (also called
predictors) and output variables (also called outcomes) could be continuous or discrete
215
F
ig
ur
e
9.
2
C
R
IS
P
-D
M
D
at
a
M
in
in
g
P
ro
ce
ss
M
et
h
o
d
o
lo
g
y
In
iti
al
D
at
a
C
ol
le
ct
io
n
In
iti
al
d
at
a
co
lle
ct
io
n
re
po
rt
D
at
a
D
es
cr
ip
tio
n
D
at
a
de
sc
rip
tio
n
re
po
rt
D
at
a
Q
ua
lit
y
V
er
ifi
ca
tio
n
D
at
a
qu
al
ity
r
ep
or
t
E
xp
lo
ra
to
ry
A
na
ly
si
s
E
xp
lo
ra
to
ry
a
na
ly
si
s
re
po
rt
D
at
a
U
nd
er
st
an
di
ng
D
at
a
se
t
D
at
a
se
t d
es
cr
ip
tio
n
S
el
ec
tio
n
R
at
io
na
le
fo
r
in
cl
us
io
n/
ex
cl
us
io
n
C
le
an
in
g
D
at
a
cl
ea
ni
ng
r
ep
or
t
C
on
st
ru
ct
io
n
D
er
iv
ed
v
ar
ia
bl
es
G
en
er
at
ed
r
ec
or
ds
ra
ns
fo
rm
at
io
n
In
te
gr
at
io
n
M
er
gi
ng
A
gg
re
ga
tio
n
F
or
m
at
tin
g
R
ea
rr
an
gi
ng
a
ttr
ib
ut
es
R
eo
rd
er
in
g
re
co
rd
s
W
ith
in
-v
al
ue
r
ef
or
m
at
tin
g
D
at
a
P
re
pa
ra
tio
n
G
en
er
at
e
Te
st
D
es
ig
n
es
t d
es
ig
n
B
ui
ld
M
od
el
ar
am
et
er
s
et
tin
gs
od
el
s
M
od
el
E
va
lu
at
io
n
od
el
d
es
cr
ip
tio
n
A
ss
es
sm
en
t
M
od
el
lin
g
E
va
lu
at
e
R
es
ul
ts
A
pp
ro
ve
d
m
od
el
s
A
ss
es
sm
en
t o
f d
at
a
m
in
in
g
re
su
lts
w
.r.
t.
bu
si
ne
ss
s
uc
ce
ss
c
rit
er
ia
R
ev
ie
w
P
ro
ce
ss
ev
ie
w
o
f p
ro
ce
ss
D
et
er
m
in
e
N
ex
t S
te
ps
Li
st
o
f p
os
si
bl
e
ac
tio
ns
ec
is
io
nE
va
lu
at
io
n
P
la
n
D
ep
lo
ym
en
t
ep
lo
ym
en
t p
la
n
P
ro
du
ce
F
in
al
R
ep
or
t
in
al
r
ep
or
t
in
al
p
re
se
nt
at
io
n
P
la
n
M
on
ito
rin
g
an
d
M
ai
nt
en
an
ce
ai
nt
en
an
ce
p
la
n
R
ev
ie
w
P
ro
je
ct
E
xp
er
ie
nc
e
do
cu
m
en
ta
tio
n
D
ep
lo
ym
en
t
D
et
er
m
in
e
B
us
in
es
s
O
bj
ec
tiv
es
B
ac
kg
ro
un
d
B
us
in
es
s
ob
je
ct
iv
es
B
us
in
es
s
su
cc
es
s
cr
ite
ria
S
itu
at
io
n
A
ss
es
sm
en
t
nv
en
to
ry
o
f r
es
ou
rc
es
R
eq
ui
re
m
en
ts
A
ss
um
pt
io
ns
C
on
st
ra
in
ts
R
is
ks
a
nd
c
on
tin
ge
nc
ie
s
er
m
in
ol
og
y
C
os
ts
a
nd
b
en
ef
its
D
et
er
m
in
e
D
at
a
M
in
in
g
G
oa
l
D
at
a
m
in
in
g
go
al
s
D
at
a
m
in
in
g
su
cc
es
s
cr
ite
ria
P
ro
du
ce
P
ro
je
ct
P
la
n
P
ro
je
ct
p
la
n
B
us
in
es
s
U
nd
er
st
an
di
ng
–
P
la
n
D
ep
lo
ym
en
t
ep
lo
ym
en
t p
la
n
P
ro
du
ce
F
in
al
R
ep
or
t
in
al
r
ep
or
t
in
al
p
re
se
nt
at
io
n
P
la
n
M
on
ito
rin
g
an
d
M
ai
nt
en
an
ce
ai
nt
en
an
ce
p
la
n
R
ev
ie
w
P
ro
je
ct
E
xp
er
ie
nc
e
do
cu
m
en
ta
tio
n
D
ep
lo
ym
en
t
P
la
n
D
ep
lo
ym
en
t
ep
lo
ym
en
t p
la
n
P
ro
du
ce
F
in
al
R
ep
or
t
in
al
r
ep
or
t
in
al
p
re
se
nt
at
io
n
P
la
n
M
on
ito
rin
g
an
d
M
ai
nt
en
an
ce
ai
nt
en
an
ce
p
la
n
R
ev
ie
w
P
ro
je
ct
E
xp
er
ie
nc
e
do
cu
m
en
ta
tio
n
D
ep
lo
ym
en
t
216 CHAPTER 9
(also called categorical). In addition, the data could also be textual, which command
a different set of data mining techniques.
In general the first step in defining the data mining technique to be used involves
defining if the study is of a predictive nature, meaning there is an outcome in mind.
For example, to build a model to predict credit risk for customers seeking a loan
from a bank, a data set must exist that includes for each customer their correspond-
ing characteristics (such as credit score, salary, years of education, etc.), which will
serve as the predictors or inputs to the model as well as the outcome variable credit
risk. In this example, there is an outcome in mind: credit risk. This is also called
inferential techniques or supervised learning. In unsupervised learning, there is not
a previously known outcome in mind, and we describe these methods later as the
descriptive techniques that appear in Table 9.4 on page 221.
Data mining techniques include both statistical as well as nonstatistical techniques.
Statistical techniques are known as traditional data mining methods including regres-
sion, logistic regression, and multivariate methods. Nonstatistical techniques, also
known as intelligent techniques or data-adaptive methods, include memory-based
reasoning, decision trees, and neural networks.
Table 9.1 summarizes the different inferential statistic techniques and their appli-
cability pertaining to the characteristics of the input and output variables. Inferential
Figure 9.3 The Iterative Nature of the KDD Process
Source: SPSS 2000.
Deployment
Evaluation
DATA
Data
preparation
Modelling
Business
understanding
Data
understanding
Deployment
Evaluation
DATADATA
Data
preparation
Modelling
Business
understanding
Data
understanding
<
KNOWLEDGE DISCOVERY SYSTEMS: SYSTEMS THAT CREATE KNOWLEDGE 217
statistical techniques are differentiated from descriptive statistics. Inferential statistics
are used to generalize from data and thus develop models that generalize from the
observations. Descriptive statistics (which appear in Table 9.4) are used to find pat-
terns or define classes of similar objects in the data collected.
The key difference between using statistical techniques and nonstatistical
techniques is that the statistical techniques require a hypothesis to be specified
beforehand. In addition, statistical techniques often are subject to stringent assump-
tions such as normality of the sample data, uncorrelated error, or homogeneity of
variance. In particular when the number of explanatory variables is large, model
specification and selection are increasingly difficult, making it harder to work with
statistical techniques. However, statistical techniques provide for a more rigorous
test of hypotheses. Table 9.2 summarizes the predictive nonstatistical techniques
Table 9.1
Summary of Applicability of Inferential Statistical Techniques
Goal
Input
Variables
(Predictors)
Output Variables
(Outcomes)
Statistical
Technique Examples (SPSS, 2000)
Find linear
combination of
predictors that
best separate the
population
Continuous Discrete Discriminant
Analysis
remain or leave (churners or not)
respond to a new product or offer
medical procedures
Predict the
probability of
outcome being
in a particular
category
Continuous Discrete Logistic and
Multinomial
Regression
customer will buy
to fail
Output is a linear
combination of
input variables
Continuous Continuous Linear
Regression dollars from a new customer
callers to an 800 number
based on patient characteristics
and medical condition
For experiments
and repeated
measures of the
same sample
Most inputs
must be
discrete
Continuous Analysis of
Variance
(ANOVA)
factors are likely to cause
cancer
To predict future
events whose
history has been
collected at
regular intervals
Continuous Continuous Time Series
Analysis past sales records
218
Ta
bl
e
9.
2
S
u
m
m
ar
y
o
f
A
p
p
lic
ab
ili
ty
o
f
N
o
n
in
fe
re
n
ti
al
P
re
d
ic
ti
ve
T
ec
h
n
iq
u
es
G
oa
l
In
pu
t (
P
re
di
ct
or
) V
ar
ia
bl
es
O
ut
pu
t (
O
ut
co
m
e)
V
ar
ia
bl
es
S
ta
tis
tic
al
T
ec
hn
iq
ue
E
xa
m
pl
es
(
S
P
S
S
, 2
00
0)
P
re
di
ct
o
ut
co
m
e
ba
se
d
on
va
lu
es
o
f n
ea
re
st
n
ei
gh
bo
rs
C
on
tin
uo
us
, D
is
cr
et
e,
a
nd
Te
xt
C
on
tin
uo
us
o
r
D
is
cr
et
e
M
em
or
y-
ba
se
d
R
ea
so
ni
ng
(M
B
R
)
ou
tc
om
es
P
re
di
ct
b
y
sp
lit
tin
g
da
ta
in
to
su
bg
ro
up
s
(b
ra
nc
he
s)
C
on
tin
uo
us
o
r
D
is
cr
et
e
(D
iff
er
en
t t
ec
hn
iq
ue
s
us
ed
b
as
ed
o
n
da
ta
ch
ar
ac
te
ris
tic
s)
C
on
tin
uo
us
o
r
D
is
cr
et
e
(D
iff
er
en
t t
ec
hn
iq
ue
s
us
ed
b
as
ed
o
n
da
ta
ch
ar
ac
te
ris
tic
s)
D
ec
is
io
n
Tr
ee
s
w
ill
le
av
e
P
re
di
ct
o
ut
co
m
e
in
c
om
pl
ex
no
nl
in
ea
r
en
vi
ro
nm
en
ts
C
on
tin
uo
us
o
r
D
is
cr
et
e
C
on
tin
uo
us
o
r
D
is
cr
et
e
N
eu
ra
l N
et
w
or
ks
KNOWLEDGE DISCOVERY SYSTEMS: SYSTEMS THAT CREATE KNOWLEDGE 219
and their applicability pertaining to the characteristics of the input and output vari-
ables. Memory-based reasoning (MBR) is a DM technique that looks for the nearest
neighbors of known data samples and combines their values to assign classification
or prediction values for new data samples. It is very similar to Case-based Reason-
ing as described in Chapter 6. Like CBR, MBR uses a distance function to find
the nearest element to a new data sample, and a combination function to combine
the values at the nearest neighbors to make a prediction. For more information on
MBR refer to Berry and Linoff (2011).
Among the techniques described in Table 9.2, decision trees (or rule induction
methods) are used to predict the outcome by splitting data into subgroups. Different
decision tree and rule induction methods are applicable depending on the character-
istics of the data. Table 9.3 summarizes the various methods.
Table 9.4 summarizes the different descriptive techniques, including both as-
sociation and clustering methods, and their applicability pertaining to the charac-
teristics of the input variables. Note that for all these techniques, the outcome or
output variable is not defined. Market basket or association analysis can include
the use of two techniques: Apriori is an association rule algorithm that requires the
input fields to be discrete. Apriori is generally faster to train than Generalized Rule
Induction (GRI). Apriori allows only the specification of logical (or dichotomies)
for the input variables, such as (True, False) or (1,0) to indicate the presence (or
absence) of the item in the market basket. Generalized rule induction is an associa-
tion rule algorithm, capable of producing rules that describe associations between
attributes to a symbolic target, and is capable of using continuous or logical data
as its input.
Typically, several methods could be applied to any problem with similar results.
The knowledge discovery process is an iterative process as we will describe in the
next section. Once the model has been developed, the results must be evaluated.
The potential for errors and their consequences must be carefully considered when
performing a data mining exercise. An intelligent computer system can make two
types of errors when trying to solve a problem. To simplify the argument, let’s assume
that the possible solutions can only be “yes” or “no.” One instance of this could be
in a medical diagnosis of a serious disease, such as cancer. The two possible errors
are: (1) an indication of “no” when the true answer is “yes,” and (2) an indication of
“yes” when the true answer is “no.” The former is called the user’s risk, while the
latter is called the developer’s risk. Depending on the application of the system, one
type of error may be tolerable while the other may not be. For example, in the case
of medical diagnosis such as cancer, a false positive (“no” when the answer is truly
“yes”) can be very costly in terms of the seriousness of the error. On the other hand,
a system designed to select a type of wine for a meal may tolerate such an error quite
acceptably.
The cost of errors must be carefully evaluated when the model is examined. For
example, Table 9.5 presents the results of a study to predict the diagnosis of patients
with heart disease based on a set of input variables. In the table, the columns represent
predicted values for the diagnostic and the rows represent actual values for diagnostics
of patients undergoing a heart disease examination. Actual values are coded in the
220 CHAPTER 9
Table 9.3
Summary of Applicability of Decision Tree Techniques
Goal
Input
(Predictor)
Variables
Output
(Outcome)
Variables
Statistical
Technique Examples (SPSS, 2000)
Predict by splitting
data into more than
two subgroups
(branches)
Continuous,
Discrete, or
Ordinal
Discrete Chi-square
Automatic
Interaction
Detection
(CHAID)
combinations of predictors yield
the highest probability of a sale
causing product defects in
manufacturing
Predict by splitting
data into more than
two subgroups
(branches)
Continuous Discrete C5.0
are considered a “good” risk
associated with a country’s
investment risk
Predict by splitting
data into binary
subgroups
(branches)
Continuous Continuous Classification
and Regression
Trees (CART)
associated with a country’s
competitiveness
are predictors of increased
customer profitability
Predict by splitting
data into binary
subgroups
(branches)
Continuous Discrete Quick,
Unbiased,
Efficient,
Statistical Tree
(QUEST)
care after heart surgery
cells, with percentages coded in parentheses along the actual values. In this table,
the predictions made along the quadrant (Actual No Disease/Predicted No Disease)
represent patients that were correctly predicted as being healthy. That means that 118
patients (or 72 percent of a total of 164 patients) were diagnosed with No Disease
and they were indeed healthy. On the other hand, looking at the quadrant (Actual
Presence of Disease/Predicted Presence of Disease) 96 patients (or 69.1 percent
of a total of 139 patients) were diagnosed with Presence of v and they were indeed
sick. So for the patients in these two quadrants the classification algorithm correctly
predicted their heart disease diagnosis. But the patients whose diagnosis falls off this
diagonal were incorrectly classified. In this example, 46 patients (or 28 percent of a
total of 164 patients) were diagnosed with the disease when in fact they were healthy.
Furthermore, 43 patients (or 30.9 percent of a total of 139 patients) were incorrectly
diagnosed with no disease when in fact they were sick.
In summary, 70.6 percent of the patients in this example were correctly classi-
fied with the prediction algorithm. Note that in this example, the cost of incorrectly
giving a patient a sound bill of health, when in fact she is sick, is considered much
higher than incorrectly predicting the patient to be sick, when in fact he is healthy.
The former may cause the patient to die without the proper care, while the latter will
give the patient a jolt but further tests are likely to exonerate him.
KNOWLEDGE DISCOVERY SYSTEMS: SYSTEMS THAT CREATE KNOWLEDGE 221
DISCOVERING KNOWLEDGE ON THE WEB
Business organizations can profit greatly from mining the Web. The business need
for Web DM is clear:
Companies venturing into e-commerce have a dream. By analyzing the tracks people make
through their Web site, they’ll be able to optimize its design to maximize sales. Information
about customers and their purchasing habits will let companies initiate e-mail campaigns and
other activities that result in sales. Good models of customers’ preferences, needs, desires,
and behaviors will let companies simulate the personal relationship that businesses and their
clientele had in the good old days. (Edelstein 2001)
Table 9.4
Summary of Applicability of Clustering and Association Techniques
Goal
Input
(Predictor)
Variables
Output
(Outcome)
Variables
Statistical
Technique Examples (SPSS 2000)
Find large groups of
cases in large data
files that are similar
on a small set of input
characteristics
Continuous or
Discrete
No outcome
variable
K-means Cluster
Analysis marketing
claims
To create large
cluster memberships
Kohonen Neural
Networks segments based on demo-
graphics and buying patterns
Create small set
associations and look
for patterns between
many categories
Logical No outcome
variable
Market Basket
or Association
Analysis with
Apriori
are likely to be purchased
together
students are likely to take
together
Create small set
associations and look
for patterns between
many categories
Logical or
Numeric
No outcome
variable
Market Basket
or Association
Analysis with
GRI
students are likely to take
together
To create linkages
between sets of items
to display complex
relationships
Continuous or
Discrete
No outcome
variable
Link Analysis
between a network of
physicians and their
prescriptions
Table 9.5
Classification Table Results
Heart Disease Diagnostic Predicted No Disease Predicted Presence of Disease
Actual No Disease 118 (72%) 46 (28%)
Actual Presence of Disease 43 (30.9%) 96 (69.1%)
Source: SPSS 2000.
222 CHAPTER 9
Web-based companies are expecting to discover all this knowledge in the logs main-
tained by their Web servers. The expectation is that a customer’s path through the data
may enable companies to customize their Web pages, increase the average purchase
amount per customer visit to the site, and in a nutshell increased profitability.
Certainly, e-business provides a fertile ground for learning market trends as well
as what the competitors are up to. Therefore, learning to mine the Web can lead to
a tremendous amount of new knowledge. Web pages and documents found on the
Web can provide important information at a minimal cost to develop or maintain.
Text mining refers to automatically “reading” large documents (called corpora) of
text written in natural language and being able to derive knowledge from the process.
Web mining is “Web crawling with on-line text mining” (Zanasi 2000). Zanasi re-
ports about Online Analyst, a system that can mine the Web to provide competitive
intelligence—a term that indicates knowledge leading to competitive advantages for
a business organization. This system provides the user with an intelligent agent that
surfs the Web in an intelligent fashion, and reads and quickly analyzes documents
that are retrievable online. This system has the advantage that it can review many
more documents than a human analyst can, even working 24 hours per day. Some of
the documents may be well hidden (unintentionally or otherwise), and often times
the relevant information can be found deeply buried within one document. Zanasi
does not describe the techniques behind Online Analyst, probably to protect its own
secrets. The system was developed by IBM-Bologna in Italy and is used as a tool
for consulting.
Unfortunately, the information and data in the Web are unstructured. This can
lead to difficulties when mining the Web. Conventional data mining techniques de-
scribed earlier are not all applicable to Web mining, because by their nature they are
limited to highly structured data. There are several differences between traditional
data mining and Web mining. One significant difference is that Web mining requires
linguistic analysis or natural language processing (NLP) abilities. It is estimated that
80 percent of the world’s online content is based on text (Chen 2001). Web mining
requires techniques from both information retrieval and artificial intelligence domains.
Therefore, Web text mining techniques are rather different from the DM techniques
described previously.
Web pages are indexed by the words they contain. Gerald Salton (1989) is generally
considered the father of information retrieval (IR). IR indexing techniques consist
of calculating the function term frequency inverse document frequency (TFIDF).
The function consists of the product of a term frequency and its inverse document
frequency, which depends on the frequency of occurrence of a specific keyword-term
in the text and the number of documents it appears in. The term frequency (TF) refers
to how frequently a term occurs in the text, which represents the importance of the
term. The inverse document frequency (IDF) increases the significance of terms that
appear in fewer documents, while downplaying terms that occur in many documents.
TFIDF then highlights terms that are frequently used in one document but infrequently
used across the collection of documents. The net effect is that terms like cryogenics
which may occur frequently in a scientist’s Web page, but infrequently across the
whole domain of Web pages in an organization, will result in a good indexing term.
KNOWLEDGE DISCOVERY SYSTEMS: SYSTEMS THAT CREATE KNOWLEDGE 223
WEB MINING TECHNIQUES
Web mining techniques can be classified into four main layers (Chen 2001).
1. Linguistic Analysis/NLP
Linguistic Analysis/NLP is used to identify key concept descriptors (the who, what,
when, or where), which are embedded in the textual documents. In NLP the unit of
analysis is the word. These functions can be combined with other linguistic tech-
niques such as stemming, morphological analysis, Boolean, proximity, range, and
fuzzy search. For example, a stemming algorithm is used to remove the suffix of a
word. Stoplists are used to eliminate words that are not good concept descriptions,
such as prepositions (e.g., and, but, etc.). Linguistic techniques can be combined with
statistical techniques, for example to represent grammatically correct sentences. Also
semantic analysis is used to represent meaning in stories and sentences.
2. Statistical and Co-occurrence Analysis
Statistical and co-occurrence analysis is similar to the TFIDF function mentioned
before. For example, link analysis is used to create conceptual associations and au-
tomatic thesauri for keyword concepts. Also similarity functions are used to compute
co-occurrence probabilities between concept pairs.
3. Statistical and Neural Networks Clustering and Categorization
Like those discussed previously in “Designing the Knowledge Discovery System”
section, statistical and neural networks clustering and categorization are used to group
similar documents together as well as communities into domain categories. Kohonen
NN techniques work well for large-scale Web text mining tasks and its results can be
graphically visualized and intuitive.
4. Visualization and Human Computer Interfaces
Visualization and human computer interfaces (HCI) can reveal conceptual as-
sociations, which can be represented in various dimensions (one-, two-, and three-
dimensional views). Furthermore, interaction techniques, such as zooming, can be
incorporated to infer new knowledge.
USES FOR WEB DATA MINING
There are three types of uses for Web data mining. They are as follows.
1. Web Structure Mining
Web structure mining refers to examining how the Web documents are struc-
tured and attempts to discover the model underlying the link structures of the Web.
224 CHAPTER 9
Intra-page structure mining evaluates the arrangement of the various HTML or
XML tags within a page; inter-page structure refers to hyperlinks connecting one
page to another. Web structure mining can be useful to categorize Web pages, and to
generate relationships and similarities among Web sites (Jackson 2002).
2. Web Usage Mining
Web usage mining, also known as clickstream analysis, involves the identification
of patterns in user navigation through Web pages in a domain. Web usage mining tries
to discover knowledge about the Web surfer’s behaviors through analysis of his/her
interactions with the Web site including the mouse clicks, user queries, and transac-
tions. Web usage mining includes three main tasks: preprocessing, pattern discovery,
and pattern analysis (Jackson 2002):
a. Preprocessing—converts usage, content, and structure from different data
sources into data sets ready for pattern discovery. This step is the most chal-
lenging in the data mining process, since it may involve data collection from
multiple servers (including proxy servers), cleansing of extraneous informa-
tion, and using data collected by cookies for identification purposes.
b. Pattern analysis—This step takes advantage of visualization and Online
Analytical Processing (OLAP) techniques, like the ones discussed earlier,
to aid understanding of the data, notice unusual values, and identify possible
relationships between the variables.
c. Pattern discovery—based on the different DM techniques discussed earlier
except that certain variations may be considered. For example in a market
basket analysis of items purchased through a Web storefront, the click-order
for the items added to the shopping cart may be significant, which is not typi-
cally studied in brick-and-mortar settings.
3. Web Content Mining
Web content mining is used to discover what a Web page is about and how to un-
cover new knowledge from it. Web content data include what is used to create the
Web page including the text, images, audio, video, hyperlinks, and metadata. Web
content mining is based on text mining and IR techniques, which consist of the orga-
nization of large amounts of textual data for most efficient retrieval—an important
consideration in handling text documents. IR techniques have become increasingly
important, as the amount of semistructured as well as unstructured textual data pres-
ent in organizations has increased dramatically. IR techniques provide a method to
efficiently access these large amounts of information.
Mining Web data is by all means a challenging task, but the rewards can be great
including aiding the development of a more personalized relationship with the virtual
customer, improving the virtual storefront selling process, and increasing Web site
revenues.
KNOWLEDGE DISCOVERY SYSTEMS: SYSTEMS THAT CREATE KNOWLEDGE 225
DATA MINING AND CUSTOMER RELATIONSHIP MANAGEMENT
Customer relationship management (CRM) is the mechanisms and technologies
used to manage the interactions between a company and its customers. Database
marketers were the early adopters of CRM software, in order to automate the pro-
cess of customer interaction. CRM implementations can be characterized as being
operational and/or analytical. Operational CRM includes sales force automation and
call centers. Most global companies have implemented such systems. The goal of
operational CRM is to provide a single view and point of contact for each customer.
On the other hand, analytical CRM uses data mining techniques to uncover customer
intelligence that serves to better understand and serve the customer.
In particular, the financial services, retailing, and telecommunications industry
among others, facing increasing competitive markets, have turned to analytical CRM
in order to (Schwenk 2002):
a. Integrate the customer viewpoint across all touchpoints—since many CRM
solutions combine infrastructure components such as enterprise application
integration (EAI) technology and data warehouses, as well as OLAP and
data mining. The CRM promise is to build an integrated view of the customer,
to understand the customer touchpoints, and resulting customer intelligence
that will enable organizations to better recognize and service the needs of the
customer.
b. Respond to customer demands in “Web time”—because the Web has changed
the dynamics of decision making, and competitive environments require
organizations to react to increasingly complex customer requests at faster
speeds. Also the analysis and interpretation of Web data can be used to enhance
and personalize customer offerings. Analysis of Web data can uncover new
knowledge about customer behavior and preferences, which can be used to
improve Web site design and content.
c. Derive more value from CRM investments—since data mining analysis can
be used to perform market segmentation studies that determine what custom-
ers could be targeted for certain products, to narrowcast (send out target e-
mails) customers, and to perform other related studies such as market basket
analysis.
For example Redecard, a company that captures and transmits MasterCard, Din-
ers Club, and other credit and debit card transactions in Brazil, uses CRM to analyze
transaction and customer data. The company performs market segmentation analysis
to determine which customers to target for certain products (Lamont 2002). Also
Soriana, a Mexican grocery retailer, uses the market basket analysis capability of its
CRM product to study promotion effectiveness and the impact of price changes on
purchasing behavior (Lamont 2002).
The first step in the CRM process involves identifying customer market segments
with the potential of yielding the highest profit. This step requires sifting through
large amounts of data in order to find the “gold nuggets”—the mining promise. CRM
226 CHAPTER 9
software automates the DM process to find predictors of purchasing behaviors. In
addition, CRM technology will typically integrate the solution of the DM study into
campaign management software used to manage the targeted marketing campaign.
The goal of campaign management software is to effectively manage the planning,
execution, assessment, and refinement of myriad marketing campaigns at an organi-
zation. Campaign management software is used to manage and monitor a company’s
communications with its customers, including direct mail, telemarketing, customer
service, point of sale, and Web interactions.
In CRM applications, the data mining prediction models are used to calculate a
score, which is a numeric value assigned to each record in the database to indicate
the probability that the customer represented by that record will behave in a specific
manner. For example, when using DM to predict customer attrition or the likelihood
that the customer will leave, a high score represents a high probability that the cus-
tomer will indeed leave. The set of scores is then used to target customers for specific
marketing campaigns.
Perhaps one of the best known innovative implementation of CRM is Harrah’s
Entertainment, one of the most recognized brand names in the casino entertainment
industry. Harrah’s comprehensive data warehouse and CRM implementation enabled
them to keep track of millions of customers’ activities allowing them to market more
effectively, thus increasing the attraction and retention of targeted customers. Armed
with deep knowledge about their customers’ preferences, they customized their cus-
tomer rewards program and were able to target promotions based on individual prefer-
ences. As an example, Harrah’s provided hotel vouchers to their out-of-town guests
and free show tickets to day-trip visiting customers. Harrah’s CRM implementation
won them national recognition in addition to an increase of their share of the gaming
budget from 36 percent to 42 percent between 1999 and 2002 as well as 110 percent
increase in their earnings per share between 1999 and 2002 (Lee et al. 2003). More
recently, Hilton Hotels launched their “Customers Really Matter” (also abbreviated as
CRM) strategy aimed at improving service delivery and consistency across the Hilton
brand. The CRM strategy was viewed as “a way to use technology to give you the
power to solidify relationships with our best customers” (Applegate et al. 2008) via the
consolidation of far-flung customer data to produce comprehensive arrival reports of
their top customers. This strategy aimed to achieve a “holistic view” of the customer
and thus improve their experience at every one of their customer touchpoints and
recognize them properly. The online gambling industry has also made an entry into the
CRM business, as it seeks to predict churn in order to understand customer retention.
Churn prediction refers to the process of identifying gamblers with a high probability
of leaving based on their prior behavior (Coussement and DeBock 2013).
Consider the following scenario: The result of a DM study at a large national bank
revealed that many of its customers only take advantage of the checking account
services it provides. A typical customer at this institution would deposit their check,
quickly moving the funds once they became available to mutual funds accounts and
other service providers outside the bank. Using the integrated capabilities for campaign
management, the software automatically triggers a direct marketing piece for those
customers with a sizable deposit to encourage them to keep their money at the bank.
KNOWLEDGE DISCOVERY SYSTEMS: SYSTEMS THAT CREATE KNOWLEDGE 227
DM and campaign management software can work together to sharpen the focus of
prospects, therefore increasing marketing response and effectiveness. For more details
about the relationship between DM and CRM, please refer to Berson et al. (2000).
The CRM market is expected to continue to expand in the foreseeable future. Even
in times of economic slowdown, worldwide CRM software revenue grew to a total
of US$12 billion in 2011, showing an 82 percent increase from 2006 revenue of $6.6
billion (Gartner 2008; Columbus 2012).
BARRIERS TO THE USE OF KNOWLEDGE DISCOVERY SYSTEMS
Possibly two of the barriers that prevented earlier deployment of knowledge discovery
in the business arena, versus what we have witnessed in the scientific realm, relate
to the prior lack of data in business to support the analysis and the limited comput-
ing power to perform the mathematical calculations required by the DM algorithms.
Clearly, with the advent of more powerful computers at our desktops and the prolif-
eration of relational databases, data warehouses, and data marts, these early barriers
have been overcome. In fact according to the Storage Law (Fayyad and Uthurusamy
2002), the capacity of digital data storage worldwide has doubled every nine months
at twice the rate predicted by Moore’s Law for the growth of computing power. But
by and large, this growing capacity has resulted in phenomena called data tombs
(Fayyad and Uthurusamy 2002) or data stores where data are deposited to “merely
rest in peace.” This means there’s no possibility that these data will be used and the
opportunity to discover new knowledge that could be used to improve services, profits,
or products, will be lost.
In addition, although many of the DM techniques have been around for almost two
decades for scientific applications, only in the past few years have we witnessed the
emergence of solutions that consolidate multiple DM techniques in a single software
offering. Probably one of the most significant barriers to the explosion of the use of
knowledge discovery in organizations relates to the fact that still today implement-
ing a data mining model is still considered an art. Although a number of software
packages exist that bundle data mining tools into one software offering, adequately
implementing the knowledge discovery models requires intimate knowledge of the
algorithmic requirements in addition to familiarity of how to use the software itself and
a deep understanding of the business area and the problem that needs to be solved. In
addition, a successful DM study typically requires a number of actors to partake in the
activity including the project leader, the DM client, the DM analyst, the DM engineer,
and the IT analyst (Jackson 2002). The project leader has the overall responsibility for
the management of the study. The DM client understands the business problem, but
in general doesn’t have the adequate technical skills to carry out the study. The DM
analyst translates the business needs into technical requirements for the DM model.
The DM engineer develops the DM model together with the DM client and analyst.
The IT analyst provides access to hardware, software, and data needed to carry out
the project. In some large projects, a number of DM analysts and engineers may be
involved. Clearly managing the number of actors involved in the study is indeed a
challenging task that must be carefully coordinated by the project leader.
228 CHAPTER 9
Perhaps one of the most interesting dilemmas facing KDD today is its basic defini-
tion of being an “interactive” process versus the notion that for the technology to be
successful it must become “invisible.” KDD can’t be both interactive and invisible at
the same time. Advocates of making KDD invisible argue that DM is primarily con-
cerned with making it easy, convenient, and practical to explore very large databases
without years of training as data analysts (Fayyad and Uthurusamy 2002). In fact, ac-
cording to this view this goal requires that the following challenges be addressed:
1. Scaling analysis to large databases: Current DM techniques require that data
sets be loaded into the computer’s memory to be manipulated. This requirement
offers a significant barrier when very large databases and data warehouses must be
scanned to identify patterns.
2. Scaling to high-dimensional data and models: Typical statistical analysis studies
require humans to formulate a model and then use techniques to validate the model
via understanding how well the data fit the model. But it may be increasingly difficult
for humans to formulate models a priori based on a very large number of variables,
which increasingly add dimension to the problem. Models that seek to understand
customer behavior in retail or Web- based transactions may fall in this category. In
general, current solutions require humans to formulate a lower dimensional abstrac-
tion of the model, which may be easier for humans to understand.
3. Automating the search: DM studies typically require the researcher to enumerate
the hypothesis under study a priori. In the future, it may be ideal for DM algorithms
to be able to perform this work automatically.
4. Finding patterns and models understandable and interesting to users: In the
past, DM projects focused on measures of accuracy (how well the model predicts
the data) and utility (the benefit derived from the pattern, typically money saved).
New benefit measures like understandability of the model and novelty of the results
must also be developed. Also DM techniques should incorporate the generation of
meaningful reports resulting from the study.
Some of these current challenges are being resolved today through the increasing
availability of “verticalized” solutions. For example in CRM software, KDD opera-
tions are streamlined through the use of standardized models which may include
the most widely used data sources. For instance, a standardized model for financial
services would most likely include customer demographics, channel, credit, and card
usage as well as information related to the promotion and actual response. In order
to streamline the KDD process, the metadata type for each table must be predefined
(nominal, ordinal, interval, or continuous) while subsequent KDD operations are
based on this information including the prespecification of algorithms that are ap-
propriate to solve specific business problems (Parsa 2000). For example, based on
the results presented earlier in Table 9.3, a verticalized application to predict which
loan customers are considered a “good” risk will automatically implement the C5.0
algorithm if the input variables are continuous and the outcome is discrete.
An additional limitation in the deployment of DM today is the fact that the suc-
cessful implementation of KDD at any organization may require the integration of
disparate systems, since there are few plug-and-play solutions. All of these require-
KNOWLEDGE DISCOVERY SYSTEMS: SYSTEMS THAT CREATE KNOWLEDGE 229
ments translate into dollars, making many DM solutions sometimes quite expensive.
Making the business case based on realistic estimations of ROI is essential for the
success of the knowledge discovery initiative. Finally, effective application of the
KDD to business applications requires the solution to be seamlessly integrated into
existing environments. This requirement makes the case for vendors, researchers, and
practitioners to adopt standards like the CRISP-DM standard discussed earlier.
A more recent preoccupation for many Chief Information Officers (CIOs) is
the relentless increase in data in most organizations, which has been observed to
be growing by 35 to 50 percent each year. Organizations are said to be processing
more than 60 terabytes of data annually, which represents about one thousand times
more than the amount processed ten years ago (Beath et al. 2012). Interviews with
IT leaders revealed that they are faced with two serious challenges: (1) dealing with
the increased granularity of data and (2) an explosion of unstructured data which is
increasingly challenging to fully exploit. Not surprisingly, the majority of the orga-
nizations appeared to be unsatisfied with their ability to generate significant value
from the data. IT leaders must learn to control data management costs and risks, and
take steps toward promoting innovation in data utilization, including making other
complementary business changes that will enable them to unlock the significant value
inherent in the stored data (Beath et al. 2012).
CASE STUDIES
AN APPLICATION OF RULE INDUCTION TO REAL ESTATE APPRAISAL SYSTEMS
In this section we describe an example of how data mining, specifically rule induction,
can be used to infer new knowledge, specifically the contribution of an individual
piece of data on a data set or the incremental worth of the individual component on
an aggregate set (Gonzalez et al. 1999). For example when performing real estate
appraisals, it is necessary to know the incremental worth that a specific feature may
have on a house—say for example a swimming pool or garage. Property appraisers
face this dilemma, and typically it is the market and not the feature construction cost
that determines the incremental worth of the feature on the house. For instance, the
incremental worth of an additional bedroom when going from 3 to 4 bedrooms may
be different from the incremental worth of going from 4 to 5 bedrooms.
Property appraisals estimate property values via market analysis, which means
comparing a house with other similar houses sold recently in the same area.
Property appraisers typically use databases of recently sold properties to establish
a basis of sales comparables. Market changes are reflected on these databases,
although there may be a lag before the market effects are reflected. Sometimes,
depending on market conditions, it may be difficult to make such comparisons.
For example, during the economic downturn of the housing market it may have
been hard to find comparable sales because homes were just not selling. On the
flipside during the height of the housing market bubble, past sales comparables
didn’t match what the market, based on supply and demand, estimated were ad-
equate property values.
230 CHAPTER 9
The technique that we focus on here is described as calculating the incremental or
relative worth of components of aggregate sets, where an aggregate set represents a
collection, assembly, or grouping of member components that have a common nature
or purpose (Gonzalez et al. 1999). Based on this definition, a house constitutes an
aggregate set of features such as number of bedrooms, number of bathrooms, living
area, and so forth. In this example, the term worth is used to represent the price of
the house, and discovering the incremental worth that a particular feature (e.g., an
extra bedroom) will contribute to the price of a house.
In the research cited in this case study, the authors use the technique called difference-
based induction (DBI)2 to calculate the incremental worth of a feature, given the database
has similar attributes (e.g., bedrooms) that may have dissimilar impact on the price of
the house. The induction algorithm used in this example identifies aggregate sets in the
database that contribute slight differences in their attributes’ values, thereby identifying
which are the most significant attribute/value combination of nearly similar aggregate
sets of individual houses. In this example, value is associated with each attribute (or
feature) in the database, while worth is associated with the value of the aggregate set
(or house). For each house attribute, the value of the attribute is specified (for example
area = 1,500 square feet, or number of bedrooms = 4) as well as how each one contrib-
utes to the house’s worth. In this example, the individual worth of an attribute is the
amount that it contributes to the overall worth of the house.
The procedure to create the decision tree based on the induction techniques pre-
sented earlier is as follows:
1. Data preparation and preprocessing—For example there may be missing
values in the database, there could be data-entering errors, or lack of consis-
tency with other sets. Therefore they will need to be identified and eliminated;
otherwise they could have a distorting effect on the outcome worth of the
house.
2. Tree construction—Houses are progressively assigned to each of the nodes
at each of the tree levels. Branches represent each range of values of the at-
tribute represented in that node, and the branch through which each house is
routed depends on the value that each particular house has for the attribute
represented in the parent node.
3. House pruning—Heuristics are applied to identify and discard those houses
whose worth is not consistent with those of other houses in the same leaf-level
group.
4. Paired leaf analysis—Any difference in worth between houses in two sibling
leaves is directly caused by the difference in their values of the critical attri-
bute. For example, two houses are identical in all features, but one has three
bedrooms and the other four. If the four-bedroom house sells for $75K and
the three-bedroom one for $73K, the extra bedroom is attributed the differ-
ence of $2K.
Figure 9.4 and Table 9.6 represent the partial decision tree resulting from the induc-
tion algorithm applied to a small sample database of 84 sold homes. The incremental
231
F
ig
ur
e
9.
4
P
ar
ti
al
D
ec
is
io
n
T
re
e
R
es
u
lt
s
fo
r
R
ea
l
E
st
at
e
A
p
p
ra
is
al
So
ur
ce
:
G
on
za
le
z
et
a
l.
19
99
.
K
ey
:
a
=
a
ttr
ib
ut
e,
X
=
d
ea
d
en
d
X
232 CHAPTER 9
worth computed with the induction algorithm was validated by opinions of real estate
appraisal experts. For additional details on this case study please refer to Gonzalez
et al. (1999).
AN APPLICATION OF WEB CONTENT MINING TO EXPERTISE LOCATOR SYSTEMS
One application of Web content mining methods is in the construction of expertise
locator knowledge management systems. A KM system that locates experts based on
published documents requires an automatic method for identifying employee names, as
well as a method to associate employee names with skill keywords embedded in those
documents. Although we discuss expertise locator systems in general in Chapter 8,
we include in this section a discussion of the system’s Web text mining component.
An example of an expertise locator KM system is the NASA Expert Seeker Web
Miner,3 which required the development of a name-finding algorithm to identify
names of NASA employees. Traditional IR techniques4 were then used to identify
and match skill keywords with the identified employee names. An IR system typically
uses as input a set of inverted files, which is a sequence of words that reference the
group of documents in which the words appear. These words are chosen according
to a selection algorithm that determines which words in the document are good index
terms. In a traditional IR system, the user enters a query and the system retrieves
all documents that match that keyword entry. Expert Seeker Web Miner is based on
an IR technique that goes one step further. When a user enters a query, the system
initially performs a document search based on user input. However, since the user is
looking for experts in a specific subject area, the system returns the names of those
employees whose names appear in the matching documents (excluding Webmasters
and curators). The employee name results are ranked according to the number of
matching documents in which each individual name appears. The employee informa-
tion is then displayed to the user.
The indexing process was carried out in four stages. First, all the relevant data were
transferred to a local directory for further processing. In this case, the data included
all the Web pages on the NASA-Goddard Space Flight Center domain. The second
stage identifies all instances of employee names by programmatically examining each
Table 9.6
Summary of Induction Results
Attribute Induction Results Expert Estimate Difference
Living Area $15–$31 $15–$25 0–2.4%
Bedrooms $4,311–$5,212 $2,500–$3,500 49–72%
Bathrooms $3,812–$5,718 $1,500–$2,000 154–186%
Garage $3,010–$4,522 $3,000–$3,500 0.3–29%
Pool $7,317–$11,697 $9,000–$12,000 2.5–19%
Fireplace $1,500–$4,180 $1,200–$2,000 25–109%
Year Built 1.2–1.7% 1.0–1.2% 20–42%
Source: Gonzalez et al. 1999.
KNOWLEDGE DISCOVERY SYSTEMS: SYSTEMS THAT CREATE KNOWLEDGE 233
HTML file. The name data are taken from the personnel directory databases (based
on the X.500 standard). All names in the employee database are organized into a
map-like data structure beforehand that is used in the Web content mining process.
This map consists of all employee names referenced by their last name key. In ad-
dition, each full name is stored in every possible form it could appear. For example,
the name John A. Smith is stored as
An individual document is first searched for all last name keys. Subsequently, the
document is again searched using all values of the matching keys. Name data orga-
nized in this way can increase the speed of the text search. Using one long sequence
containing all names in every possible form as search criteria would slow down
processing time.
The third stage involves identifying keywords within the HTML content. This is
done using a combination of word stemming and frequency calculation. First the text
is broken up into individual words through string pattern matching. Any sequence
of alphabetical characters is recognized as a word while punctuation, numbers,
and white-space characters are ignored. The resulting list of words is processed to
determine if a word was included in a stoplist. The resulting list of words was then
processed with a stemming algorithm. This is done to group together words that
may be spelled differently but have the same semantic meaning. A person who types
“astronomical” as a query term would most likely also be interested in documents that
match the term “astronomy.” Once the stemming process is completed, the algorithm
calculates the frequency of each term. Word frequency was used during the keyword
selection process in the determination of good index terms. However, other indexing
algorithms could have been used instead with comparable results.
It is important to note that the degree of relation between an employee name and
a keyword within an individual document is not considered. Rather, expertise is de-
termined based on the assumption that if an employee recurrently appears in many
documents along with a keyword, then that person must have some knowledge of
that term. Theoretically, a large document count for a search query should produce
more accurate results.
The chosen keywords have a twofold purpose. First, they are used to quickly associ-
ate employees with recurring skill terms. These keywords can also be used in future
work for clustering similar documents into topic areas. Finally, knowledge taxonomy
can be constructed from the mined keywords such that an appropriate query relevance
feedback system can be developed that suggests query terms that are related to the
query entered by the user. Details about the role of taxonomies on the development
of expertise locator systems are presented in Chapter 8.
234 CHAPTER 9
Next, we discuss the role that the Web and search engines play in knowledge
discovery.
NOVEL-KNOWLEDGE DISCOVERY ON THE WEB
The Web is a rich source of information for knowledge discovery.5 Organizations and
individuals search the Web for different types of knowledge and to answer different
types of questions. For example, individuals typically use the Web to find specific
answers—also known as focused search—and to uncover patterns and trends about
a topic—also known as scanning. Google (www.google.com) is the most popular
Web-searching tool and is most suited to focused searching and discovering deep
knowledge. Tools such as Kartoo (www.kartoo.com) and Clusty (www.clusty.com),
which provide knowledge maps and clustered categories, are more suitable for scan-
ning and discovering broad knowledge.
In addition to deep and broad knowledge, individuals and organizations sometimes
seek to discover novel knowledge—knowledge that is surprising and unknown to
them yet interesting and relevant. The challenge with discovering novel knowledge
on the Web is that the question posed in the search is usually very vague, therefore
creating appropriate search terms may be virtually impossible. Essentially, it’s like
searching for “what you don’t know you don’t know”! In addition, novel knowledge
may be difficult to locate amongst the vast amount of content returned, one of the
many challenges posed by the information overload that may result from the Web
search. In most instances, the content returned may be highly related to the search
terms provided. Thus, discovering novel knowledge may be more like looking for a
needle in a haystack—except you may not know what the needle looks like. In other
words, because of its novelty, individuals have a difficult time differentiating purely
irrelevant results (“junk”) from surprising and interesting results (“novel knowledge”).
Furthermore, what’s interesting, surprising, and relevant differs from individual to
individual based on what they already know; novel knowledge to one person may be
considered deep knowledge to another person. Thus, novel knowledge is basically
“in the eye of the beholder.”
Novel knowledge can be valuable to organizations for many reasons, for example,
for the discovery of new strategic opportunities, for development of learning capabili-
ties, and to support the creative thinking leading to innovation or solving “wicked”
problems (for more on wicked problems see Chapter 13). Researchers at Queen’s
University in Kingston, Ontario, have been investigating the importance of novel
knowledge to organizations and how they currently go about discovering novel
knowledge. Their findings show that organizations do consider novel knowledge
important, especially in industries where innovation is critical. However, in many of
the organizations investigated, the discovery of novel knowledge was purely seren-
dipitous. Unlike deep and broad knowledge, they found that there are currently no
specialized tools to support the discovery of novel knowledge on the Web, motivating
the development of Athens (Jenkin 2008a; Jenkin et al. 2007; Vats and Skillicorn
2004a; 2004b). The Athens prototype uses text and data mining techniques such
as clustering, singular-value decomposition, and spectral graph partitioning to find
http://www.google.com
http://www.kartoo.com
http://www.clusty.com
KNOWLEDGE DISCOVERY SYSTEMS: SYSTEMS THAT CREATE KNOWLEDGE 235
content on the Web that is indirectly connected yet contextually related to what the
individual knows. In essence, the individual specifies to Athens the specific topic
of interest and Athens finds content on the Web that is literally two steps away from
it, that is, related to the area of knowledge familiar to the individual. Thus, Athens
tries to find what you don’t know you don’t know but would find interesting and
relevant.
An example using a popular knowledge discovery case will help to illustrate how
Athens works. Don Swanson, an information scientist, made an important discovery
about Raynaud’s Syndrome (a condition that results in intermittent restriction of blood
flow to fingers and toes) by exploring the medical literature (Gordon and Lindsay
1996; Swanson 1986, 1990). Swanson did not know what to look for specifically,
so he began by reviewing the Raynaud’s literature and found a connection between
Raynaud’s Syndrome and blood viscosity. He then reviewed the blood-viscosity lit-
erature and found that dietary fish oil lowers blood viscosity. He then hypothesized
that fish oil may be a useful dietary supplement to help decrease the blood viscosity in
humans and therefore alleviate symptoms of Raynaud’s Syndrome (Swanson 1986).
At the time, this was novel knowledge and not explicitly mentioned in any of the
source documents Swanson had previously reviewed. Using this manual approach,
Swanson spent a significant amount of time reviewing different bodies of literature
in order to make this important novel-knowledge discovery.
If Don Swanson had used Athens for this task rather than a manual approach, he
would have provided Athens with terms to describe what he knows or, rather, his focal
topic—Raynaud’s Syndrome. Athens, using an iterative clustering approach, would
discover content that is directly connected to Raynaud’s Syndrome—in this case,
blood viscosity. Next, Athens would repeat this iterative clustering step, using blood
viscosity as the starting point, in order to find content that is indirectly connected to
the original topic of Raynaud’s Syndrome—in this case, fish oil. Thus, the final result
is content that is two steps away from original topic area. It would be up to the user
to hypothesize how Raynaud’s Syndrome and fish oil might be related.
The Athens prototype is an example of a new tool to discover novel knowledge.
However, current research findings show that the innovative aspects of the tool
may present some problems. For most individuals and organizations, the concept
and task of searching for novel knowledge is new. The popularity of Web-search
tools like Google has trained us to view Web searching as a focused search activity.
Thus, the initial reaction when using Athens is to view it as a focused search tool
yielding a specific result, essentially comparing its features and output to those
of Google. Despite these challenges, organizations that have been experimenting
with the Athens prototype see the potential of the tool’s capabilities. For example,
they were impressed with how quickly they were able come up with new and in-
novative ideas for the topics investigated—something that would have taken a lot
of time, effort, and serendipity to do before (Jenkin 2008b). Athens promises to
be a useful tool for organizations interested in discovering novel knowledge on
the Web or within their own internal information repositories. Figures 9.5 and 9.6
show the outcomes of using Athens for a search using the terms digital music and
portable audio player.
236 CHAPTER 9
Figure 9.5 Examples of Search with Terms “Digital Music” and “Portable Audio Player”
Figure 9.6 Novel Clusters Resulting from the Search with Terms “Digital Music” and
“Portable Audio Player”
KNOWLEDGE DISCOVERY SYSTEMS: SYSTEMS THAT CREATE KNOWLEDGE 237
SUMMARY
In this chapter you learned what knowledge discovery systems are, the design con-
siderations for such systems, and specific types of data mining techniques that enable
such systems. Also the chapter discusses the role of DM in customer relationship
management. Three case studies that describe the implementation of knowledge dis-
covery systems are presented, each based on different methodologies and intelligent
technologies. The first system is based on the use of decision trees, or rule induction
as a knowledge-modeling tool, and is described in the context of a real estate appraisal
system. The second system is based on the use of Web-content mining to identify
expertise in an expertise locator system. The third tool presents the use of an innova-
tive tool used to improve the discovery of novel knowledge on the Web. Finally, the
use of socialization in organizational settings is discussed as a mechanism to help
discover new knowledge and catalyze innovation.
REVIEW
1. How do socialization techniques help to discover tacit knowledge?
2. Describe the six steps in the CRISP-DM process.
3. Why is understanding of the business problem essential to knowledge dis-
covery?
4. Describe the three types of Web DM techniques. Which one is used in the
Expert Seeker case study?
5. Describe some of the barriers to the use of knowledge discovery.
APPLICATION EXERCISES
Identify which DM techniques would be selected to solve the following problems.
Explain your answer. Include a description of the input and output variables that
would be relevant in each case. Note that more than one technique may apply for
each of these problems.
1. Predict fraudulent credit card usage based on purchase patterns.
2. Predict instances of fraud related to Medicare claims.
3. Predict which customers are likely to leave their current mobile service
provider.
4. Predict whether a person will renew their insurance policy.
5. Predict who will respond to a direct mail offer.
6. Predict that a generator is likely to fail.
7. Predict which specialized voice services a person is likely to purchase from
their local telecommunications provider.
8. Identify factors resulting in product defects in a manufacturing
environment.
9. Predict the expected revenue from a customer based on a set of customer
characteristics.
238 CHAPTER 9
10. Predict cost of hospitalization for different medical procedures.
11. Create customer segments in a marketing campaign.
12. Segment among university graduates those that are likely to renew their alumni
membership.
NOTES
1. For an extensive review of articles on data mining techniques and applications, see: Bishop 1994;
Smith and Gupta 2000; Widrow et al. 1994; Wong et a1. 1997.
2. DBI builds a classification tree similar to that used in other induction algorithms like ID3 and
C4.5. More details on the tree-building process may be found in Gonzalez et al. 1999.
3. This version of Expert Seeker was developed to support the needs of Goddard Space Flight
Center. Expertise locator systems in general are discussed in detail in Chapter 8.
4. See for example Selection by Discriminant Value in Frakes and Baeza-Yates 1992, an algorithm
for selecting index terms.
5. We acknowledge Tracy Jenkin, Yolande Chan, and David Skillicorn of Queen’s University for
this case study.
REFERENCES
Applegate, L., Piccoli, G., and Dev, C. 2008. Hilton Hotels: Brand differentiation through customer
relationship management. Harvard Business School Case Study #809029, July 23.
Beath, C., Becerra-Fernandez, I., Ross, J., and Short, J. 2012. Finding value in the information explo-
sion. MIT Sloan Management Review, 53(4), 18.
Becerra-Fernandez, I., Zanakis, S., and Walczak, S. 2002. Knowledge discovery techniques for predict-
ing country investment risk. Computers and Industrial Engineering, 43(4), 787–800.
Berry, M., and Linoff, G. 2011. Data mining techniques for marketing, sales, and customer relationship
management. New York: John Wiley & Sons.
Berson, A., Smith, S., and Thearling, K. 2000. Building data mining applications for CRM. New York:
McGraw-Hill.
Bishop, C. 1994. Neural networks and their applications. Review of Scientific Instruments, 65(6),
1803–1832.
Brachman, R., and Anand, T. 1996. The process of knowledge discovering in databases. In Advances
in knowledge discovery and data mining, ed. Usama M. Fayyad et al., 37–57. Menlo Park, CA:
AAAI Press.
Chapman, P., Clinton, J., Kerber, R., Khabaza, T., Reinartz, T., Shearer, C., and Wirth, R. 2000. CRISP-
DM 1.0: Step-by-step data mining guide. SPSS Technical Report.
Chen, H. 2001. Knowledge management systems: A text mining perspective. Tucson, AZ: The Univer-
sity of Arizona.
Columbus, L. 2012. Gartner hype cycle for CRM sales, 2012: Sales turns to the cloud for quick relief.
Forbes.com, July 27. www.forbes.com/sites/louiscolumbus/2012/07/27/gartner-hype-cycle-for-crm-
sales-2012-sales-turns-to-the-cloud-for-quick-relief/ (accessed January 3, 2014).
Coussement, K., and De Bock, K.W. 2013. Customer churn prediction in the online gambling
industry: The beneficial effect of ensemble learning. Journal of Business Research, 66(9),
1629–1636.
Edelstein, H.A. 1999. Introduction to data mining and knowledge discovery (3d ed.). Potomac, MD:
Two Crows Corporation.
———. 2001. Pan for gold in the clickstream. Information Week, March 12.
Fayyad, U., Piatetsky-Shapiro, G., Smyth, P., and Uthurusamy, R. 1996. From data mining to knowl-
edge discovery: An overview. In Advances in knowledge discovery and data mining, ed. Usama M.
Fayyad et al., 1–33. Menlo Park, CA: AAAI Press.
http://www.forbes.com/sites/louiscolumbus/2012/07/27/gartner-hype-cycle-for-crm-sales-2012-sales-turns-to-the-cloud-for-quick-relief/
http://www.forbes.com/sites/louiscolumbus/2012/07/27/gartner-hype-cycle-for-crm-sales-2012-sales-turns-to-the-cloud-for-quick-relief/
KNOWLEDGE DISCOVERY SYSTEMS: SYSTEMS THAT CREATE KNOWLEDGE 239
Fayyad, U., and Uthurusamy, R. 2002. Evolving data mining into solutions for insights. Communica-
tions of the ACM, 45(8), 28–21.
Frakes, W., and Baeza-Yates, R. 1992. Information retrieval: Data structures and algorithms. Upper
Saddle River, NJ: Prentice Hall.
Gartner. 2008. Gartner says worldwide customer relationship management market grew 23 percent in
2007. Press release, July 7. www.gartner.com/it/page.jsp?id=715308 (accessed January 3, 2009).
Gilbert, A. 2002. “Smart” carts on a roll at Safeway. CNET News, October 28. http://news.cnet.com/2100-
1017-963526.html (accessed August 24, 2008).
Gonzalez, A., Daroszweski, S., and Hamilton, H.J. 1999. Determining the incremental worth of mem-
bers of an aggregate set through difference-based induction. International Journal of Intelligent
Systems, 14(3), 275–294.
Gordon, M.D., and Lindsay, R.K. 1996. Toward discovery support systems: A replication, re-exam-
ination, and extension of Swanson’s work on literature-based discovery of a connection between
Raynaud’s and fish oil. Journal of the American Society for Information Science, (47)2, 116–128.
Gray, P., and Watson, H.J. 1998. Decision support in the data warehouse. Upper Saddle River, NJ:
Prentice Hall.
Jackson, J., 2002. Data mining: A conceptual overview. Communications of the Association for Infor-
mation Systems, 8, 267–296.
Jenkin, T.A. 2008a. How IT supports knowledge discovery and learning processes on the Web. In Pro-
ceedings of the 41st Hawaii International Conference on System Sciences (HICSS). Waikoloa, HI.
———. 2008b. Using information technology to support the discovery of novel knowledge in organiza-
tions. PhD Thesis (Management), Queen’s University.
Jenkin, T.A., Chan, Y.E., and Skillicorn, D.B. 2007. Novel-knowledge discovery—challenges and
design theory. In Proceedings of the Annual Conference of the Administrative Sciences Association
of Canada. Ottawa, Canada.
Joseph, D. 2009. Inside Disney’s toy factory. Bloomberg Businessweek, July 1. www.businessweek.
com/innovate/content/jul2009/id2009071_592328.htm (accessed January 3, 2014).
Kovalerchuk, B., Triantaphyllou, E., Ruiz, J., Torvik, V., and Vityaev, E. 2000. The reliability issue of
computer-aided breast cancer diagnosis. Journal of Computers and Biomedical Research, 33(4),
296–313.
Lamont, J. 2002. CRM around the world. KM World, 11(9) (October).
Lawrence, R., Almasi, G., Kotlyar, V., Viveros, M., and Duri, S. 2001. Personalization of supermarket
product recommendations. Data Mining and Knowledge Discovery, 5(1–2), 11–32.
Lee, H., Whang, S., Ahsan, K., Gordon, E., Faragalla, A., Jain, A., Mohsin, A., and Shi, G. 2003. Har-
rah’s Entertainment Inc.: Real-time CRM in a service supply chain. Harvard Business School Case
Study #GS50, October 27.
Le Storti, A.J. 2003. When you’re asked to do the impossible: Principles of business teamwork and
leadership from the U.S. Army’s Elite Rangers. Guilford, CT: The Lyons Press.
Li, S., Lu, W., Wang, C., Yen, D. 2012. Identifying the signs of fraudulent accounts using data mining
techniques. Computers in Human Behavior, 28(3), 1002–1013.
Linden, G., Smith, B., and York, J. 2003. Amazon.com recommendations item-to-item collaborative
filtering. Internet Computing, IEEE, 7(1), 76–80.
Nonaka, I. and Takeuchi, H. 1995. The knowledge creating company. New York: Oxford University
Press.
Parsa, I. 2000. Data mining: Middleware or middleman, panel on KDD process standards (position
statement). In Proceedings from the Sixth ACM SIGKDD International Conference on Knowledge
Discovery and Data Mining, ed. R. Ramakrishnan and S. Stolfo. New York: ACM.
ProFlowers. 2012. ProFlowers honored with top customer satisfaction award from J.D. power and
associates. PR Newswire, December 11. www.prnewswire.com/news-releases/proflowers-honored-
with-top-customer-satisfaction-award-from-jd-power-and-associates-182983321.html (accessed
January 3, 2014).
http://www.gartner.com/it/page.jsp?id=715308
http://news.cnet.com/2100-1017-963526.html
http://news.cnet.com/2100-1017-963526.html
http://www.businessweek.com/innovate/content/jul2009/id2009071_592328.htm
http://www.prnewswire.com/news-releases/proflowers-honored-with-top-customer-satisfaction-award-from-jd-power-and-associates-182983321.html
http://www.prnewswire.com/news-releases/proflowers-honored-with-top-customer-satisfaction-award-from-jd-power-and-associates-182983321.html
http://www.businessweek.com/innovate/content/jul2009/id2009071_592328.htm
240 CHAPTER 9
Salton, G. 1989. Automatic text processing. Reading, MA: Addison-Wesley.
Schwenk, H. 2002. Real-time CRM analytics: The future of BI? KM World, 11(2) (February).
Smith, K.A., and Gupta, J.N.D. 2000. Neural networks in business: Techniques and applications for
the operations researcher. Computers and Operations Research, 27, 1023–1044.
SPSS. 2000. Data mining: Modeling. Chicago, Illinois.
Stevens, L. 2001. IT sharpens data mining’s focus. Internet Week, August 6.
Swanson, D.R. 1986. Fish oil, Raynaud’s Syndrome, and undiscovered public knowledge. Perspectives
in Biology and Medicine, 30(1), 7–18.
———. 1990. Medical literature as a potential source of new knowledge. Bulletin of the Medical
Library Association, 78(1), 29–37.
Vats, N., and Skillicorn, D.B. 2004a. The ATHENS system for novel information discovery. Department
of Computing and Information Science, Queen’s University, Technical Report 2004–489.
———. 2004b. Information discovery within organizations using the Athens system. In Proceedings of
the 2004 Conference of the Centre for Advanced Studies on Collaborative Research, ed. H. Lutfiyya,
J. Singer, and D.A. Stewart, 282–292. Markham, Ontario.
Widrow, B., Rumelhart, D.E., and Lehr, M.A. 1994. Neural networks: Applications in industry, business
and science. Communications of the ACM, 37(3), 93–105.
Witten, I. 2000. Adaptive text mining: inferring structure from sequences. In Proceedings of the 34th
Conference on Information Sciences and Systems. Princeton University, NJ, March 15–17.
Wong, B.K., Bodnovich, T.A., and Selvi, Y. 1997. Neural network applications in business: A review
and analysis of the literature (1988–1995). Decision Support Systems, 19, 301–320.
Zanakis, S., and Becerra-Fernandez, I. 2005. Competitiveness of nations: A knowledge discovery
examination. European Journal of Operations Research, 166(1), 185–211.
Zanasi, A. 2000. Web mining through the Online Analyst. In Proceedings of the first Data Mining
Conference. Cambridge University, Cambridge, UK.
Zdanowicz, J. 2004. Detecting money laundering and terrorist financing via data mining. Communica-
tions of the ACM, 47(5), 53–55.
———. 2009. Trade-based money laundering and terrorist financing. Review of Law and Economics,
5(2), 854–878.
PART III
MANAGEMENT AND THE FUTURE OF
KNOWLEDGE MANAGEMENT
This page intentionally left blank
243
10
Emergent Knowledge
Management Practices
In the last chapter, we discussed knowledge discovery systems. In this chapter we
introduce emergent knowledge management practices. In particular we focus in the
discussion of social networks, how they facilitate knowledge sharing, and how they
benefit from communication technologies. We start with a discussion of emerging
technologies such as social networks, wikis, and blogs, followed by a description of
how they enable collaboration and knowledge sharing. Finally, we present the busi-
ness implications of open source development and virtual worlds.
WEB 2.0
Those of us who were entering the field of computers in the 1980s witnessed a revo-
lution that would forever change the world of computing as we saw it back then.
This revolution consisted of the transformation of the computing platform from
the traditional mainframe to the PC. This change took many by surprise, including
Thomas Watson, then chairman of IBM, who in 1943 uttered one of the most famous
technology quotes: “I think there is a world market for maybe five computers.”
According to Wikipedia, Web 2.0 suggests a new version of the World Wide Web
that refers to cumulative changes in the ways software developers and end users use
the Web. It does not refer to an update to any technical specification. A Web 2.0 site
may allow users to interact and collaborate with each other in a social media dialogue
as creators of user-generated content in a virtual community, in contrast to Web sites
where people are limited to the passive viewing of content. Examples of Web 2.0
include social networking sites, blogs, wikis, video sharing sites, hosted services,
Web applications, mashups, and folksonomies. The term Web 2.0 is credited to Tim
O’Reilly, who used it during the first Web 2.0 conference in 2004 to describe the
Internet as a platform. O’Reilly (2005) first described Web 2.0 via an example that
contrasted a set of popular applications in the new platform to their counterparts in
the Web 1.0 platform. See a comparison of Web 2.0 and Web 1.0 technologies in
Table 10.1 and Figure 10.1.
Web 2.0 was made possible through the development of Ajax (asynchronous
JavaScript and XML), which enabled Web-based applications to work much more
like the desktop ones. Ajax (Garret 2005) refers to several technologies, which com-
bined together provide the possibility to create more interactive applications for the
user. Ajax incorporates:
244 CHAPTER 10
Table 10.1
Web 2.0
Web 1.0 Web 2.0
DoubleClick → Google AdSense
Ofoto → Flickr
Akamai → BitTorrent
mp3.com → Napster
Britannica Online → Wikipedia
personal Web sites → blogging
evite → upcoming.org and EVDB
domain name speculation → search engine optimization
page views → cost per click
screen scraping → Web services
publishing → participation
content management systems → wikis
directories (taxonomy) → tagging (“folksonomy”)
stickiness → syndication
Source: O’Reilly 2005.
Ability to Create Content
Few
Many
Few Many
A
cc
es
s
to
C
on
te
nt
Platform
Channel
Web 2.0
Technologies
Web 1.0
Technologies
Ability to Create Content
Few
Many
Few Many
A
cc
es
s
to
C
on
te
nt
Platform
Channel
Web 2.0
Technologies
Web 1.0
Technologies
Figure 10.1 Web 2.0 Technologies Compared to Web 1.0 Technologies
EMERGENT KNOWLEDGE MANAGEMENT PRACTICES 245
JavaScript binding everything together.
Another essential ingredient of Web 2.0 applications is that this type of computer
application will essentially get better the greater the number of people who use it.
In fact, the most significant aspect of Web 2.0 is its ability to harness collective
intelligence—building databases that represent the shared synergies among its group
of users. For example, Amazon’s ability to garner user reviews of their products im-
proves the experience of someone who is seeking to buy their products. Table 10.2
provides a short definition for types of Web 2.0 applications.
This exciting collaboration infrastructure is not just an entertainment medium
for young adults. Many organizations are beginning to ponder how this new col-
laboration medium can add value to businesses. A 2008 McKinsey & Company
survey (2008) found that businesses confirmed wide support for the use of Web
2.0 applications that many organizations considered “strategic.” More than three-
fourths of the executives who responded confirmed they would maintain or increase
investments in technology that encourages user collaboration such as peer-to-peer
social networking, social networks, and Web services. The report confirmed that
“companies aren’t necessarily relying on the best-known Web 2.0 trends such
as blogs; instead, they place the greatest importance on technologies that enable
automation and networking.” Of the firms surveyed only 32 percent were using
blogging technologies as compared to more popular collective intelligence (48
percent) and peer-to-peer networking (47 percent; McKinsey 2008). Furthermore,
of the firms surveyed 43 percent were not considering the use of blogs, which is
a large percentage as compared to the number of firms that will not be using the
more popular collective intelligence (26 percent) and peer-to-peer networking
technologies (28 percent). These trends remained the same as compared to those
same trends in 2007.
The rise of Web 2.0-like applications within the organization has received the label
of Enterprise 2.0 (McAfee 2006). Traditionally knowledge workers have two main uses
for information technology: as channels, to create and distribute digital information
that is visible by few people via technologies like e-mail and instant messaging; or as
platforms, which enable content to be widely visible via technologies like intranets
and information portals. Enterprise 2.0 technologies focus on search, links, author-
ing (via wikis and blogs, discussed in the next section), tags (to categorize content),
extensions (such as recommendations matched to user preferences), and signals (like
RSS feeds and aggregators). The biggest difference between Enterprise 2.0 technolo-
gies and traditional KM systems is that the latter are typically highly structured and
users cannot influence the structure. Enterprise 2.0 technologies on the other hand
provide users the potential to create intranets built by distributed autonomous peers
and are subject to network effects, which means the more that people participate in
authoring the better the content will get.
246 CHAPTER 10
Web 3.0 will be about the semantic Web, or the meaning of data, which means the
context of content is defined by data. It also refers to a Web capable of reading and
understanding content and context, and in this way the Web can understand content
and thus better satisfy the requests of people and machines. This is how Web 3.0 can
filter the content that is of interest to its users and deliver the right message, at the
right time, to the right person, in the right device. Web 3.0 refers to personalization,
intelligent search, and behavioral advertising among other things. In Web 3.0, search
engines understand who you are, what you have been doing, and where you would
like to go next. From this perspective, Web 3.0 (vs. Web 2.0) will be focused on the
individual (vs. communities), consolidating dynamic content (vs. sharing content),
user behavior or “me-onomy” (vs. tagging or “folksonomy”), user engagement (vs.
Table 10.2
Web 2.0 Applications
Blogs (short for Web logs) are online journals or diaries hosted on a Web site and often distributed
to other sites or readers using RSS (see below).
Collective
intelligence
refers to any system that attempts to tap the expertise of a group rather than an individual to
make decisions. Technologies that contribute to collective intelligence include collaborative
publishing and common databases for sharing knowledge.
Mashups are aggregations of content from different online sources to create a new service. An ex-
ample would be a program that pulls apartment listings from one site and displays them on a
Google map to show where the apartments are located.
Peer-to-peer
networking
(sometimes called P2P) is a technique for efficiently sharing files (music, videos, or text) either
over the Internet or within a closed set of users. Unlike the traditional method of storing a file
on one machine—which can become a bottleneck if many people try to access it at once—
P2P distributes files across many machines, often those of the users themselves. Some
systems retrieve files by gathering and assembling pieces of them from many machines.
Online games include both games played on dedicated game consoles that can be networked and “ massively
multiplayer” games, which involve thousands of people who interact simultaneously through
personal avatars in online worlds that exist independently of any single player’s activity.
Podcasts are audio or video recordings—a multimedia form of a blog or other content. They are often
distributed through an aggregator, such as iTunes.
RSS (Really Simple
Syndication)
allows people to subscribe to online distributions of news, blogs, podcasts, or other
information.
Social networking refers to systems that allow members of a specific site to learn about other members’ skills,
talents, knowledge, or preferences. Commercial examples include Facebook, MySpace, and
LinkedIn. Some companies use these systems internally to help identify experts.
Virtual worlds Virtual worlds, such as Second Life, are highly social, three-dimensional online environ-
ments shaped by users who interact with and receive instant feedback from other users
through the use of avatars.
Web services are software systems that make it easier for different systems to communicate with one
another automatically in order to pass information or conduct transactions. For example, a
retailer and supplier might use Web services to communicate over the Internet and auto-
matically update each other’s inventory systems.
Widgets are programs that allow access from users’ desktops to Web-based content.
Wikis (such as Wikipedia) are systems for collaborative publishing. They allow many authors to
contribute to an online document or discussion.
Source: Mc Kinsey 2007a, 2007b.
EMERGENT KNOWLEDGE MANAGEMENT PRACTICES 247
cost per click), and lifestream—a time-ordered stream of documents that functions
as a diary of your electronic life (vs. blogging).
In the next section, we describe one of the most popular Web 2.0 applications:
social networking.
SOCIAL NETWORKING
User-driven online spaces have redefined the meaning of the word “socializing” in
the United States, primarily due to the emergence of Web 2.0 applications such as
Friendster, Myspace, Facebook, LinkedIn, Orkut, YouTube, Twitter, Skype, Google+,
Pinterest, and FaceTime. In a sense Web 2.0 enables widely available knowledge
management systems that deliver value to communities that have sufficient incentives
to provide content to these popular social sites.
One of the first social networking sites to get significant traction in terms of estab-
lishing a significant user base was Friendster, which was founded in 2002 primarily
as a dating Web site (Wikipedia 2013d). Friendster allowed “friends” to share media
and content online, including messages, comments, photos, and videos. Perhaps one
of the features that enabled Friendster to gain notoriety among all other social net-
working sites was its feature “Who’s Viewed Me,” which allowed users to see who
else had viewed their profile (Mintz 2005). What made social network sites so unique
was their ability to grow in terms of user base through the use of human relationships
without the need for traditional marketing methods (Barnett et al. 2006). In 2011,
Friendster was acquired by a Malaysian company and repositioned its focus as a social
entertainment site for play games and music (Wauters 2011). The company issued
a statement asking users to essentially export all their profile data as the company
moved to wipe out all photos, blogs, comments, and groups created by its users but
keep all accounts alive, along with lists for friends and gaming preferences (see Figure
10.2).
Myspace, which according to its Web site touted itself as “a place for friends,” was
founded in 2003 and quickly became a well-known social networking service that
combined the features of user profiles, blogs, groups, photos, and videos with music
discovery (Figure 10.3). In particular, Myspace established itself as a niche player in
the popular music distribution space by allowing both unsigned artists and major label
musicians to feature music streams and downloads. Myspace was acquired by News
Corporation in July 2005 for $580 million. From 2005 until early 2008, Myspace was
the most visited social networking site in the world and it overtook Google in June
2006 as the most visited Web site in the United States. The power of social network-
ing sites became evident on July 11, 2006, when Myspace became the number one
visited domain by users in the United States (Lagorio, 2006), a phenomenal growth
from 2004 when Myspace.com had represented only 0.1 percent of all Internet visits
to 1.9 percent by 2005 and 4.5 percent of Internet visits by 2006, achieving a 4,300
percent increase in visits over two years and 132 percent increase in visits since the
same time the prior year.
Even after being overtaken by Facebook and surpassing Google in the year 2009,
Myspace generated $800 million in revenue during the 2008 fiscal year. After that,
248 CHAPTER 10
the number of Myspace users started to decline steadily in spite of several redesigns
and modifications of the functionalities. Myspace was ranked 303 by total Web
traffic and 223 in the United States as of June 2013. Myspace employed approxi-
mately 1,600 workers in June 2009. By June 2011, Myspace had reduced its staff
to about 200 employees. In June 2011, Specific Media Group and Danny Trejo
jointly purchased the company for approximately $35 million (see Figure 10.3)
(Wikipedia 2013i).
In 2004, Facebook.com (Wikipedia 2013a; Figure 10.4) was launched by Harvard
student Mark Zuckerberg as a social network that offered school members access to
the profiles of other classmates who were members of the same network. Previously,
paper-based student directories (called facebooks) were circulated among fresh-
man college students as a means to help newcomers establish new friends, and who
knows, perhaps even find a date. Facebook enables users to keep a personal profile
and interact with other people within their same social networks that are organized
by school, workplace, city, and region. Facebook users keep up with their friends
and send them messages, upload photos and links to videos, and simply learn about
people they meet. Probably one of the most captivating features of social network
Figure 10.2 Friendster
EMERGENT KNOWLEDGE MANAGEMENT PRACTICES 249
Figure 10.3 Myspace
Figure 10.4 Facebook
250 CHAPTER 10
sites is the integration of a News Feed functionality, which essentially allow users
to keep up with every little change in their friends’ profile pages, what Zuckerberg
describes as “a stream of everything that’s going on in their lives” (Thompson 2008).
One of the most remarkable characteristics of Facebook is its “stickiness” in the
sense that about two-thirds of its total user base comes back to the site within a day,
a measure of daily retention, which serves as a contrast to most sites, which measure
monthly retention (Barnett et al. 2006). Facebook introduced a messaging platform
on November 15, 2010, that used several methods, such as including a special e-mail
address, text messaging, or messaging via the Facebook Web site or mobile app, to
deliver a message. These multiple methods are contained within single threads in a
unified inbox. Users can adjust from whom they can receive messages with options
like just friends, friends of friends, or from anyone. Facebook also has a dedicated
Facebook Messenger app for mobile devices. Since April 2011, Facebook users have
had the ability to make live voice calls via Facebook Chat, allowing users to chat
with friends from anywhere in the world. Video calling services were launched on
July 6, 2011. Facebook added the ability for users to provide a “Subscribe” button
on their page on September 14, 2011, which allows users to subscribe to the public
post of the user without needing to add them as a friend. In December 2012 the Sub-
scribe button was relabeled as a “Follow” button to avoid the confusion that arose
with its functionalities. This re-labeling will make Facebook more similar to other
social networks with similar functions. As of September 2012, Facebook had over
one billion active users, more than half of whom used Facebook on a mobile device.
Users must be registered before using the site. Registered users can create a personal
profile, add others as friends, and use the messaging service. Additionally, users may
join common-interest user groups, school or college, workplace or any other such
characteristics and categorize their friends into lists such as “People from Work” or
“Close Friends.”
LinkedIn was launched in 2003 as a business-oriented social networking site for
professional networking, allowing users to maintain a detailed list of business contacts
(Wikipedia 2013f). LinkedIn had 33.9 million unique visitors in June 2011, which
surpassed MySpace. It reported more than 255 million users in more than 200 countries
as of June 2013. The site is available in twenty languages (see Figure 10.5).
In 2004 another prominent social network site, which would later capture most
of the Brazilian and Indian market space, made its debut—the Google-owned Or-
kut. This service was designed to help users stay connected with friends and also to
maintain relationships. Orkut is less popular in the United States, but it is one of the
most visited Web sites in India and Brazil (Wikipedia 2013k).
YouTube was launched in 2005 as a videosharing Web site where users could up-
load, share, and view video clips. Essentially, YouTube established the infrastructure
that made it possible for anyone to post a video that millions of users across the world
could view in a matter of minutes (see Figure 10.6).
Twitter is an online social networking service and microblogging service that allows
its users to send and read text-based messages of up to 140 characters (Wikipedia
2013n). These messages are known as “tweets.” Twitter was created in March 2006
by Jack Dorsey, and was launched by July of the same year. The service rapidly
EMERGENT KNOWLEDGE MANAGEMENT PRACTICES 251
gained worldwide popularity, with over 500 million registered users as of 2012,
generating over 340 million tweets daily and handling over 1.6 billion search
queries per day. Since its launch, Twitter has become one of the ten most visited
Web sites on the Internet, and has been described as “the SMS of the Internet.”
Unregistered users can read tweets, while registered users can post tweets through
the Web site interface, SMS, or a range of applications for mobile devices (see
Figure 10.7). Twitter has helped people in different ways, even as a mechanism to
support democracy and freedom of expression. For example, in March 2014, the
Turkish prime minister blocked Twitter after it was used to leak recordings that
implicated him and other government members of electoral fraud (Yeginsu 2014).
This ban was lifted via the country’s highest court, which ruled that the two-week
ban violated freedom of expression. Twitter is also known to have played a role in
the Egyptian revolution of 2011, serving to organize revolutionaries, transmit their
messages to the world, and galvanize international support (Gustin 2011). It is not
that social media caused the revolution, it is that Twitter helped to get the message
out to the broader world, serving to amplify what was happening on the ground.
Perhaps the founders of Twitter never imagined that they would provide inspira-
tion to young activists in repressed countries in the Middle East, and a mechanism
to transmit hope across these countries that would eventually end generations of
violent repression.
Figure 10.5 LinkedIn
252 CHAPTER 10
Figure 10.6 YouTube
Figure 10.7 Twitter
EMERGENT KNOWLEDGE MANAGEMENT PRACTICES 253
Skype is a software application that allows users to make voice and video calls,
as well as instant messaging, over the Internet (Wikipedia 2013m). The application
allows Skype users to place free calls among themselves, while charging a nominal
fee to calls to landline and cellular telephone networks, using a debit-based user
account system. In May 2011, Microsoft announced it would acquire Skype Global
for $8.5 billion in cash, its largest acquisition ever (Sorkin and Lohr 2011). This
purchase was estimated to make Microsoft the leader in Internet communications
(Figure 10.8).
Google+ (pronounced and sometimes written as Google Plus, sometimes abbrevi-
ated as G+) is a multilingual social networking and identity service that is owned
and operated by Google Inc. (Wikipedia 2013e). It was launched on June 28, 2011.
As of May 2013, it had a total of 500 million registered users of whom 235 million
are active on a monthly basis. Unlike other conventional social networks, which are
generally accessed through a single Web site, Google has described Google+ as a
“social layer” consisting of not just a single site, but rather an overarching “layer”
that covers many of its online properties (Figure 10.9).
Figure 10.8 Skype
254 CHAPTER 10
Figure 10.9 Google+
Pinterest is a pinboard-style photo-sharing Web site. It allows users to create and
manage theme-based image collections such as events, interests, hobbies, and more.
Users can browse other pinboards for inspiration, “re-pin” images to their own pin-
boards, or “like” photos (Wikipedia 2013l).
FaceTime is Apple Computer’s video chatting technology, which is preinstalled
as part of the mobile operating system for all Apple devices since the iPhone 4. In
addition, FaceTime can be installed on Macintosh computers running MAC OSX
10.6 or higher (Wikipedia 2013b).
Social network research focuses on the knowledge held by entities (called nodes,
meaning people or information systems) and the relationships between them (ties).
Ties are characterized by type (friendship, advice, professional), strength (intensity
or reciprocity between two nodes, which could be strong or weak), and density (the
ratio of actual ties to the possible number of ties in a network and is defined as high
or low). Nodes could represent individuals as well as organizations, and research-
ers may focus on dyadic analysis (two nodes), ego network (all the relationships
maintained by one node), or whole network analysis (all the nodes and ties among
EMERGENT KNOWLEDGE MANAGEMENT PRACTICES 255
Box 10.1
The Power of Social Networks: Operation Burnt Frost
In December 2007, the U.S. Strategic Command was charged by former President Bush to
mitigate the danger posed by a bus-sized satellite falling out of orbit, commissioning Operation
Burnt Frost. The satellite hurled through space at 17,000 mph and posed a danger to people
around the globe, due to the fact that its fuel tank contained nearly 1,000 pounds of hydrazine
fuel, which is highly toxic. The complex problem required expertise to calculate the satellite’s
trajectory and to understand the likelihood of success. The mission’s success required the
ability to successfully reach out and connect with the appropriate subject-matter experts. A
major factor in the success of the operation was leveraging expertise through individual social
networks. The mission required the successful collaboration of hundreds of subject-matter
experts representing more than two dozen federal agencies that were spread around the United
States. On February 20, 2008, a single Standard Missile 3 (SM3), which was launched from the
Pacific Ocean aboard the USS Lake Erie, shot down a National Reconnaissance Office satellite
that was falling out of orbit. Extensive social networking played a key role in the success of this
mission (Steinhauser and Thon 2008).
them). Knowledge sharing researchers have concluded that strong and weak ties
have different effects on knowledge sharing relationships, and strong ties are bet-
ter for cultivating trust and reliability while weak ties are more appropriate for
searching for different types of knowledge (Alavi and Kane 2008). In addition, the
position of the individual within the network (central or peripheral) is important
to knowledge sharing effectiveness, whereas individuals in central positions act as
knowledge brokers in the network and pose knowledge sharing benefits in terms of
timing, access, and referral of knowledge. LinkedIn uses these concepts to visualize
the strength and span of one’s network.
The recognition that business is built on relationships as much as on information has
spawned a new set of software offerings that range from contact management (both inside
and outside the enterprise) to social network-based enterprise solutions for knowledge
sharing. These new software offerings may, for example, calculate the strength of the
business relationship by measuring the frequency of e-mail message flows outside of the
enterprise. For instance, users could enter the name of an individual or organization and
the software would return a list of people within the firm who have connections with the
target prospect. Social networking solutions have also made an inroad into the world of
talent management, and organizations are increasingly using these products in order to
build relationships with current employees, alumni, and retirees (Lamont 2008). In Box
10.1, we explore an example of how social networks can help solve business problems.
Perhaps social networking has delivered the dreams that KM has forged over the last
decade and the Internet is no longer a collection of static Web sites hijacked by powerful
Web masters. Instead social networking is a mechanism to voice the collective wisdom
of communities that use the Internet to voice their opinions on myriad topics from the
best Web sites, articles, blogs, and music (www.pandora.com), the best restaurants (www.
zagat.com), the best pictures (www.photobucket.com), and those guys you shouldn’t
date (www.dontdatehimgirl.com, now published via Facebook).
In the next section, we discuss two specific types of Web 2.0 technologies used to
generate content: wikis and blogs.
http://www.pandora.com
http://www.zagat.com
http://www.zagat.com
http://www.photobucket.com
http://www.dontdatehimgirl.com
256 CHAPTER 10
COLLABORATIVE CONTENT CREATION VIA WIKIS, BLOGS,
MASHUPS, AND FOLKSONOMIES
Wikipedia is a collaboratively edited, multilingual, free Internet encyclopedia sup-
ported by the nonprofit Wikimedia Foundation. Wikipedia was launched on January
15, 2001, by Jimmy Wales and Larry Sanger. Its 30 million articles, over 4.2 million
in the English Wikipedia alone, are written collaboratively by volunteers around the
world, who are estimated to have an average age of twenty-five. Almost all of its
articles can be edited by anyone who has access to the site, which had about 100,000
active contributors. As of March 2013, there were editions of Wikipedia in 286
languages. It has become the largest and most popular general reference work on
the Internet. A study published in Nature (Giles 2005) found that the differences in
accuracy between Wikipedia and Britannica was not particularly great: the average
science entry in Wikipedia contained around four inaccuracies compared with about
three for Britannica. This particular investigation provides support to the strength of
Collective Intelligence. Articles in Wikipedia are loosely organized according to their
development status and subject matter. A new article often starts as a “stub” which is
a very short page consisting of definitions and few links. On the other extreme, the
most developed articles may be nominated for “featured article” status. One featured
article per day, as selected by editors, appears on the main page of Wikipedia. A
group of Wikipedia editors may form a WikiProject to focus their work on a specific
topic area, using its associated discussion page to coordinate changes across multiple
articles. Contributors to Wikipedia must abide by a set of policies and guidelines,
which fall in one of the following areas (Wikipedia 2014):
Content, which defines the scope of the encyclopedia and the material that is
suitable for it.
Conduct, which describes how editors can successfully collaborate and what
behavior is acceptable.
Deletion, which explains the processes by which pages, revisions, and logs may
be deleted.
Enforcement, which accounts for various means by which standards may be
enforced.
Legal, which includes rules influenced by legal considerations and remedies for
their misuse.
Procedural, which documents various processes by which the English Wikipedia
operates.
Blogs (a contraction between Web and log) are essentially online digital diaries. An
individual makes regular written journal entries that comprise a statement of opinion,
a story, an analysis, description of events, or other material. Blogs typically display
an entry in reverse chronological order, with the most recent entries at the top.
Blogging techniques offer tremendous opportunities for businesses. For example,
firms can take advantage of Web 2.0 technologies such as blogs for agile new product
development.
EMERGENT KNOWLEDGE MANAGEMENT PRACTICES 257
Blogs have played a significant role in chronicling the events of the Iraq war,
including the abuse of prisoners at Abu Ghraib, which was first reported by Chris
Missick, a soldier with the U.S. Army’s 319th Signal Battalion. Missick’s blog, titled
“A Line in the Sand” (Missick 2008), provided an account of the events in Iraq that
summarizes the impact of this particular technology:
Never before has a war been so immediately documented, never before have sentiments
from the front scurried their way to the home front with such ease and precision. Here I sit,
in the desert, staring daily at the electric fence, the deep trenches and the concertina wire
that separates the border of Iraq and Kuwait, and write home and upload my daily reflec-
tions and opinions on the war and my circumstances here, as well as some of the pictures
I have taken along the way. It is amazing, and empowering, and yet the question remains,
should I as a lower enlisted soldier have such power to express my opinion and broadcast to
the world a singular soldier’s point of view? To those outside the uniform who have never
lived the military life, the question may seem absurd, and yet, as an example of what exists
even in the small following of readers I have here, the implications of thought expressed
by soldiers daily could be explosive. (Hockenberry 2005)
Blaming a breach of security and the potential leak of sensitive wartime information,
the U.S. Army ordered in April 2007 that soldiers stop posting to blogs or sending
personal e-mail messages, unless the content was first cleared by a superior (Shacht-
man 2007). A lawsuit ensued under the Freedom of Information Act, which concluded
that the real security breach came from official Army Web sites and not personal
blogs. Another popular blog by an American soldier was Alex Horton’s “Army of
Dude—Reporting on Truth, Justice and the American Way of War” (Horton 2008).
His site won the second place for “best military blog” in the 2007 WebLog Awards,
perhaps due to his crude pictures and unfiltered war accounts, including posts about
a day in which he and other soldiers were being shot at by insurgents with machine
guns, and his open cynicism about the war.
A mashup is a program or Web site that combines two or more online software
products (Wikipedia 2013g). Mashups typically integrate multiple application pro-
gramming interfaces (APIs) and thus combine the data or functionality of the original
services. Mashups can be used to better combine, visualize, or aggregate data. For
example, mashups could combine maps and search locators (like Google Transit); or
user opinions with lookup services for restaurants (Yelp), contractors (Angie’s List),
or music preferences (Pandora).
A folksonomy (a spin on the word taxonomy) is a system of classification created by
regular people (or folk) who collaboratively create and manage tags to classify content
(Wikipedia 2013c). Folksonomies can be broad or narrow. A broad folksonomy allows
multiple users to tag content with a variety of terms from a variety of vocabularies, while
a narrow folksonomy allows few users to tag content with a limited number of terms. The
advantage of folksonomies is to enable the searchability of content by adding a textual
description to an object. Folksonomies are usually used to tag blogs or pictures with tags
that are organized in ways that users can collectively classify and find information.
Online tools, for example, Twitter, are also enabling users to microblog, which
refers to posting frequent tiny (140 characters) updates of what they’re doing. This
258 CHAPTER 10
technology enables users to have what social scientists refer to as online awareness,
which is incessant online contact with friends (and even strangers who may choose
to follow you if you allow them). The continuing rise of online awareness begs the
question: Are these technologies helping increase the number of friends people
have? In 1998, anthropologist Robin Dunbar theorized that humans had an upper
limit on the maximum social connections they had: 150—hence called the Dunbar
number. A closer look at the effect of these tools seems to suggest that perhaps their
biggest contribution is in helping people keep up with their weak ties, for example,
someone they may have met at a conference, high school, or at a party. Sociologists
also have found that weak ties may be important to help people solve problems, for
example finding a new job (Thompson 2008). In the next section, we describe the
open source movement and its role in business today. In Box 10.2, we explore an
example of how a financial institution is using a corporate wiki for employees to ef-
fectively collaborate.
OPEN SOURCE DEVELOPMENT
The Open Source movement was introduced to the general public by Tim O’Reilly,
who organized the first Open Source Summit, as a means to bring together open source
software developers and supporters together from around the world. Open source
software development refers to “Internet-based communities of software developers
who voluntarily collaborate in order to develop software that they or their organiza-
tions need” (von Krogh 2003).
Historically, open source development dates back to the early days of the Inter-
net, when the U.S. Defense Advanced Research Projects Agency (DARPA) helped
launch the first transcontinental high-speed computer network called ARPAnet.
ARPAnet allowed researchers at universities, government agencies, and corporate
Box 10.2
Wikis at Dresdner Kleinwort Wasserstein: Are Bankers and Wikis Ready
for Each Other?
Dresdner Kleinwort Wasserstein (DrKW) investment bank was one of the first firms to launch
a corporate wiki for employees to contribute and edit their corporate Web site without the need
of special permissions or knowledge of HTML or Web-authoring skills. It disappeared as a
brand from the world of investment banking in September 2009. Wiki use at DrKW first started
within the IT department as a means to keep their online training schedules up-to-date (McAfee
and Sjoman 2006). A corporate blog would not provide the needed functionality since it would
still require one person to consolidate all the schedule changes, which was a daunting task.
RSS feeds are an important complement to the wiki, because they can alert all the people
that want to learn about new training. Even though wikis come with the ability for users to edit,
and potentially destroy, the content of the wiki, it would be easy to correct it as the technology
provides the capability to restore the previous version of the wiki page. Following the initial
success of the use of wikis at the IT department, the bank introduced a bankwide wiki that
provided access to most bank employees. Wikis have been successful at DrKW because it had
an initial structure, it provided for a more effective form of collaboration, and it helped reduce
the amount of e-mail.
EMERGENT KNOWLEDGE MANAGEMENT PRACTICES 259
labs to share software code. The Free Software Foundation was established in 1985
as a mechanism to allow developers to preserve the “free” status of their software
in the face of increasing pressure from commercial software firms that sought to
restrict the access to source code. The free software movement became known as
the “open source” software movement, a term that encompassed the right to use it
at no cost, the right to study and modify the source code, and the right to distribute
modified or unmodified versions to others at no cost (von Krogh 2003). The defi-
nition of open source, according to the Open Source Initiative, the steward of the
Open Source Definition, does not denote just free software and encompasses the
following criteria:
1. Free Redistribution—The license shall not require a royalty for any sale.
2. Source Code—The program must include source code, so the programmer
may be able to modify the program.
3. Derived Works—License allows modifications and derived works.
4. Integrity of the Author’s Source Code—License permits the distribution of
software built from modified source code.
5. No Discrimination Against Persons or Groups.
6. No Discrimination Against Fields of Endeavor—License doesn’t restrict
anyone from making use of the program in a specific field.
7. Distribution of License—Rights apply to all to whom the program is redis-
tributed.
8. License Must Not Be Specific to a Product—All parties to whom the program
is distributed to have the same rights.
9. License Must Not Restrict Other Software—No restriction on other software
distributed along the licensed software.
10. License Must Be Technology Neutral—No restrictions based on type of
technology or interface.
But the most interesting fact about open source is that all the projects including
BIND (the Berkeley Internet Name Daemon, which runs the Domain Name System),
Sendmail (which routes Internet e-mail), Apache (the world’s most used Web server),
and Linux (the operating system that became the backbone of the Internet) were
developed by thousands of loosely coordinated individuals (Collis and Montgomery
2004). Other examples of open source projects are Mozilla Firefox (a Web browser
that implements current and anticipated web standards), Google Chromium (founda-
tion on which Google Chrome was developed), and Android (a Linux-based operating
system designed primarily for touch screen devices). The open source movement
is what has allowed the development of protocols that make all of these programs
interoperable. Open source developers embrace cooperation and collaboration as a
source of competitive advantage. In the open source movement value is added through
architecture of participation, which takes advantage of network effects services that
become more valuable the more people use them.
One successful product in the open source space is MySQL. Founded in the early
1990s, by 2004 it was estimated that MySQL held two million out of the twelve million
260 CHAPTER 10
total relational database management systems (DBMS; Wittig and Inkinen 2004). One
of the strategies employed by MySQL to penetrate markets was to use its open source
status as a viral marketing vehicle. MySQL made its products available under a dual-
licensing policy that allowed both open source and commercial users to have access
to the same product either without fees or for purchase for a lower licensing fee than
other commercial DBMS. But commercial users enjoyed special privileges including a
warranty and development support. One of the competitive advantages that companies
like MySQL enjoy is that—unlike first-generation, nonprofit open source products like
Linux in which the code was written by many—second-generation open source com-
panies write and own all the code. Each time MySQL releases a new version, its vast
number of users massively test and debug the product allowing for rapid stabilization
of the product and thus faster development cycles. This means development costs at
MySQL can be significantly lower (even up to 80 percent lower) than in traditional
development environments.
One of the important questions raised by the success of the open source move-
ment is “why should thousands of top-notch programmers contribute freely to the
provision of a public good” (von Krogh 2003). Research into this phenomenon
suggests that open source software development is underpinned by the personal
benefit that developers reap from developing complex software including gaining
expertise, improving their reputation, and the sheer enjoyment of participating in
this communal activity. Collaborating individuals obtain peer evaluation of their
work and gain group-related factors such as indispensability to the team and loy-
alty among peer developers (von Krogh 2003). Other research projects have also
confirmed this finding, namely that a significant predictor of individual knowledge
contribution to electronic networks of practice, such as communities of practice
discussion forums used by individuals to exchange advice and ideas with other
strangers with common interests, is the perception that participation in these net-
works enhances one’s professional reputation. Furthermore, the development of
social capital may play an important role in the underlying knowledge exchange
and the opportunity to master the application of expertise and the understanding
of what is relevant may be what motivates users to contribute (Wasko and Faraj
2005). In summary, the motives for participating and contributing to open source
software include (Shah 2003):
1. Need for product—create, customize, or improve a product or feature.
2. Enjoyment, desire to create and improve—one enjoys it; that is, finds creating
or improving software creative and interesting.
3. Reputation and status within the community—build or maintain reputation
or status within the community.
4. Affiliation—socialize or spend time with like-minded individuals.
5. Identity—reinforce or build a desired self-image.
6. Values and ideology—promote specific ideals, such as the free-software
philosophy.
7. Training, learning, reputation outside the community and career concerns—im-
prove one’s skills, with the belief that it will lead to a better job or promotion.
EMERGENT KNOWLEDGE MANAGEMENT PRACTICES 261
Open source licenses guarantee the rights of future users against appropriation by private
firms, guaranteeing the product will remain a public good. Open source development comes
with its own set of challenges, such as how to monitor the programmers, in particular
when participation is voluntary and not bounded by a formal contract. Also, even though
mechanisms exist to limit “free riding” (those who download code without contributing
to it) typically these measures are not closely watched. On the other hand, open source
communities are known to be “meritocracies,” meaning that technical knowledge and
expertise will determine the impact of a contributor on the software (von Krogh 2003).
Open source development models are redefining business models for software
companies based on the idea that software should be free and resistant to the monopoly
power of large software development companies. Early ventures were not for profit, but
for-profit ventures have now started to embrace the open source business model.
In the next section, we describe a technology that is poised to transform the way
we do business today: virtual worlds.
VIRTUAL WORLDS
Virtual worlds have been touted as a “breakthrough idea” that is destined to transform the
way businesses function. These simulated online environments provide the infrastructure
for innovative operational models as well as alternate realities for working and interacting
with customers. In fact, virtual worlds today are compared to radio and TV broadcasting
from the perspective that when the technologies first emerged they were trivial content
providers but eventually became the dominant medium for advertising (Harvard Business
Review 2008). The metaverse—a compound (known technically as portmanteau) of the
words “meta” (an abstraction of) and “universe”—describes a multiplayer virtual world
in which humans (as avatars) interact with each other and software agents in a three-
dimensional world that is a metaphor of the real world (Wikipedia 2013h).
Virtual worlds such as Second Life (http://secondlife.com/) provide the infrastructure
for users to build three-dimensional immersive virtual worlds in which individuals
represented by avatars socialize, explore, and conduct business. Virtual worlds are prov-
ing to be effective environments for remote users to spontaneously interact, with goals
that span from collaborative learning environments for college students to communi-
ties formed to help autistic children acquire skills they can transfer to the real world.1
Other examples of virtual worlds include League of Legends (www.leagueoflegends.
com), Activeworlds (www.activeworlds.com), Meez (www.meez.com), VirtualWorld
(www.virtualworld.com), Virtual Worlds for Tweens (www.virtualworldfortweens.com),
WeeWorld (www.weeworld.com), and Swinky (www.zwinky.com).
Second Life, a metaverse launched by Linden Lab in 2003, provides its users
(called Residents) the ability to interact via their personalized avatars, and provides
the ability to explore and meet other Residents, socialize and have fun, participate
in individual and group activities, create and trade virtual property and services pur-
chased via its own currency (the Linden dollar), and travel through the world (called
the Grid). Second Life (Figure 10.10) started as sixty-four acres of “virtual land”
space that had grown to over 700 square miles by 2014. Everything in Second Life is
Resident-created, and Residents retain the intellectual property rights in their digital
http://secondlife.com/
http://www.leagueoflegends.com
http://www.leagueoflegends.com
http://www.activeworlds.com
http://www.meez.com
http://www.virtualworld.com
http://www.virtualworldfortweens.com
http://www.weeworld.com
http://www.zwinky.com
262 CHAPTER 10
Figure 10.10 Second Life
creations, which they can buy, sell, and trade with fellow Residents. Residents can
trade their Linden dollars for real-world currency in their LindeX dollar exchange,
while Second Life leases real estate according to their own published land pric-
ing and use fees. Second Life Residents represent an international multiuser
community that spends time in entertainment, education, work, advocacy, and
entrepreneurial activities. In its first ten years, 36 million accounts were created,
US$3.6 billion was spent on virtual assets, and the equivalent total time users
spent was 217,266 years. Each month, more than one million people visit Sec-
ond Life and an average of 400,000 new registrations are created. More than one
million daily transactions were conducted for virtual assets. Users created 2.1
million virtual assets for sale, with the most purchased being women’s hairstyles.
Second Life landmass is nearly 700 square miles (slightly over 1,810 square
kilometers) (Linden Labs 2013). Second Life Residents have a median age of
thiry-five years and are gender neutral by hours of use.2 Figure 10.11 depicts a
creative live music and dance performance by human-led avatars on Second Life
known as Sky Dancers.3
Project Wonderland, launched in 2007 by Sun Microsystems, is a 100 percent Java,
free and open source toolkit for building three-dimensional immersive virtual worlds.
Project Wonderland was developed to overcome some of the collaborative challenges
that Sun employees faced—of whom about 50 percent may be out of the office on any
given day—including remoteness, management issues, difficulty brainstorming, and
lack of social interactions. Project Wonderland was designed with a focus on business
EMERGENT KNOWLEDGE MANAGEMENT PRACTICES 263
and education collaboration and provides a software platform for its users to develop
virtual worlds in which avatars communicate with high-fidelity immersive audio and
share applications like Open Office documents, Web browsers, and games. One of the
advantages that Project Wonderland offers is to allow users to make changes to the
virtual world using XML files instead of needing to modify the source code. The “world
builder” functionality is integrated into the client to let users drag-and-drop external
work and arrange artwork and custom world components. One of the advantages that
Project Wonderland offers is the ease of integration with Web service applications such
as Google Maps and Flickr. On January 27, 2010, Sun Microsystems was acquired
by Oracle who decided to cease funding for Project Wonderland. The development of
Project Wonderland continues as an independent community-supported open-source
project named “Open Wonderland” (Wikipedia 2013j).
One of the challenges that virtual worlds face in becoming pervasive platforms for
business is that computers are not typically designed for three-dimensional inputs, and
the fact that startup costs in developing these environments is high. In fact, people typi-
cally spend a lot more time than they anticipated when designing their avatars because
most people feel a personal sense of identification with their virtual personae. On the
Figure 10.11 Sky Dancers in a Second Life
264 CHAPTER 10
other hand, virtual worlds are beginning to show promise in resolving issues in distrib-
uted collaboration that other technologies like videoconferencing cannot address, such
as supporting multiple simultaneous conversations of remote coworkers. In addition,
virtual worlds can be an important informal communication infrastructure for building
human relationships. And finally, virtual worlds can help its users develop new skills
that can later be transferred to the real world. Is The Matrix potentially a reality?4
THE THREE WORLDS OF INFORMATION TECHNOLOGY:
DOES IT REALLY MATTER?
In this chapter we have presented a group of breakthrough technologies that may be
poised to be transformational; that is, they may disrupt and redefine the way business
is done today. As we look back at the history of another transformational technology,
we learn that at first The Western Union Company rejected Alexander Graham Bell’s
offer to sell the patent for the telephone for a mere $100,000, commenting that the
instrument was nothing but a toy—an offer that would translate to $25 million dollars
only two years later. Similarly, many of these emerging Web 2.0 technologies are fac-
ing increasing resistance from their users as their management struggles with finding
“appropriate” mechanisms to carry out business transactions on these platforms. One
example was Facebook’s release of the marketing initiative Beacon, which allowed
external Web sites to publish a user’s activities to their Facebook profiles in order
to promote their products even if users tried to opt out of this feature. Beacon was
met with resistance from Facebook users and privacy advocate groups who pointed
out that it violated Facebook’s promise that “no personally identifiable information
is shared with an advertiser.” Zuckerberg publicly apologized and promised users a
revamped Beacon where users would opt in, rather than be expected to opt out of
this feature (Perez 2007). On the other hand, companies have already started to use
virtual venues to collaborate and learn. For example, BP created an environment that
allows engineers to roam freely around a future oil pipeline in a virtual representation
of its existing surroundings in order to allow the engineers to effectively identify con-
structability and safety issues (Hemp 2008). Are social networks and virtual worlds,
seen as the entertainment tools of our youngest generation, prepared to be the next
disruptive business technologies in the near future?
It’s easy to infer that IT has a positive influence on KM and organizational learning.
For example, research supports the hypothesis that groupware systems contributed to
improved organizational learning as compared to individuals who did not have access
to those KM systems (Orlikowski 1996). In addition, studies have found that IT-enabled
learning mechanisms facilitate capabilities that have an effect on exploration and exploi-
tation dynamics in organizational learning (Kane and Alavi 2007). So IT can be used to
improve organizational learning, in particular in times of turnover and turbulence.
Some skeptics of the potential transformational role that IT is positioned to play in
the near future point out that the evolution of IT follows the same blueprint of earlier
infrastructural technologies, such as railroads and electric power (Carr 2003). At first,
these technologies provide opportunities for visionary companies that are quick to
integrate them into their business processes and thus gain competitive advantages.
EMERGENT KNOWLEDGE MANAGEMENT PRACTICES 265
But as their cost decreases and their availability become pervasive, the competitive
advantages they offer are similarly eroded, commoditizing their value and becoming
simply the cost of doing business. IT then no longer matters. Have we indeed reached
the point where IT doesn’t matter?
IT spending is facing increasing criticism. A report by Computer Science Corpo-
ration, which surveyed 782 American executives in charge of IT, revealed that 51
percent of large-scale IT efforts finished later than expected and ran over budget, that
only 10 percent of companies believed they were getting high returns from IT invest-
ments, and that 47 percent felt the returns were low or negative or unknown (McAfee
2006). Research has found that executives indeed have a critical role in the selection,
adoption, and exploitation of IT in the organization. Furthermore, executives must
adopt a comprehensive model of what IT does for their organizations and what they
must do to ensure those IT projects are successful (McAfee 2006).
Building a comprehensive model includes understanding not only the uses of IT
innovations but how can they change the manner in which the work is done. IT can
have various impacts on organizations, which fall into three categories:
1. Function IT supports completing specific tasks. This type of IT does not require
modifications to business processes to maximize their utility, although their
impact could increase when this happens. Function IT delivers productivity
and optimization. It is not difficult to implement. The manager’s main role is
to create business processes of maximum utility; spreadsheets and simulation
software, for example.
2. Network IT supports interaction. This type of IT does not require modifica-
tions to business processes, but these may evolve over time in order to gain
additional advantages. Network IT delivers increased collaboration. Adoption
is usually voluntary, examples are wikis and blogs.
3. Enterprise IT specifies business processes. This type must be accompanied
by the specification of new tasks and sequences. Enterprise IT delivers stan-
dardization and monitors work. It is difficult for companies to implement and
employees usually dislike these processes, thus requiring forceful intervention
of the IT leader. Some examples include enterprise resource planning and
customer relationship management.
Successful IT executives successfully complete three tasks: electing IT applica-
tions that deliver needed organizational capabilities, leading their adoption effort,
and shaping their exploitations (McAfee 2006). Successful IT leaders are those that
follow an “inside-out approach [that] puts the spotlight squarely on the business be-
fore evaluating the technology landscape . . . focused on the capabilities that IT can
provide rather than on the technologies themselves” (McAfee 2006).
So we conclude with the analysis of the question: Does IT no longer matter? Even
if software is not rare and is easy to imitate, successful system implementations are
not necessarily easy to replicate. In addition, as disruptive technologies continue to
be developed, new business processes must be discovered for the organizations to
gain new competitive advantages.
266 CHAPTER 10
SUMMARY
In this chapter, you learned what social networks are and how they are redefining the
definition of collaboration both on the Internet and within organizations. We discussed
emerging technologies such as wikis and blogs and how they enable knowledge
management. Then we discussed the importance of open source development and its
emergence as an important new business model for profitable firms across the world.
We followed with a discussion of virtual worlds, including some of the challenges
they face and the value they are already providing to businesses. Finally, we discussed
how IT is poised to act as a transformational technology for businesses that continue
to find ways in which technology can continue to redefine the ways businesses oper-
ate, proving that IT does matter.
REVIEW
1. What is Ajax?
2. Describe the difference between wikis and blogs.
3. Describe the categories for the policies and guidelines for Wikipedia. Who
developed these guidelines?
4. Define open source development.
APPLICATION EXERCISES
1. How could social networking redefine collaboration within organizations?
2. Give an example from the literature of how top organizations are deploying
an enterprise 2.0 architecture.
3. Describe the process to make a new contribution or edit an entry in Wikipedia.
4. Make an entry, either a new contribution or edit an existing article in Wiki-
pedia. Since anyone can modify the contents of Wikipedia, then how can we
be confident that its contents are credible?
NOTES
1. For details, see University of Missouri’s iSocial at http://isocial.missouri.edu/iSocial/ (accessed
July 27, 2014).
2. We acknowledge John Lester of Linden Labs for this information.
3. A video of this performance can be seen at Hayes 2007.
4. The Matrix is a science fiction film released in 1999 in which the reality perceived by hu-
mans is actually a virtual world created by intelligent computers in order to subdue the human
population.
REFERENCES
Alavi, M., and Kane, G. 2008. Social networks and information technology: Evolution and new frontiers.
In Knowledge management: An evolutionary view, ed. I. Becerra-Fernandez and D. Leidner, vol. 12,
Advances in Management Information Systems, 63–85. Armonk, NY: M.E. Sharpe.
Barnett, W., Leslie, M., and Harkey, M. 2006. Facebook. Stanford Graduate School Case E-220, May 3.
http://isocial.missouri.edu/iSocial/
EMERGENT KNOWLEDGE MANAGEMENT PRACTICES 267
Carr, N. 2003. IT doesn’t matter. Harvard Business Review, May, 41–49.
Collis, D., and Montgomery, C. 2004. Sharing to advantage: A new paradigm for business. Strategy
and Innovation, July–August, 3–5.
Garrett, J.J. 2005. Ajax: A new approach to web applications. Adaptive Path, February 18. www.adap-
tivepath.com/ideas/ajax-new-approach-web-applications/ (accessed July 27, 2014).
Giles, H. 2005. Internet encyclopedias go head to head. Nature, December 15, 900–901.
Gustin, S. 2011. Social media sparked, accelerated Egypt’s revolutionary fire. Wired, April 11. http://
www.wired.com/2011/02/egypts-revolutionary-fire/ (accessed July 27, 2014).
Harvard Business Review. 2008. The HBR list: Breakthrough ideas for 2008. Harvard Business Review,
February, 17–45.
Hayes, Gary. 2007. Second Life—ZeroG Sky Dancers—2nd Generation 2007. YouTube, July 3. www.
youtube.com/watch?v=c_gT0YDAkec (accessed July 27, 2014).
Hemp, P. 2008. Getting real about virtual worlds. Harvard Business Review, October.
Hockenberry, J. 2005. The blogs of war. Wired, August 1, 118.
Horton, A. 2008. Army of Dude. http://armyofdude.blogspot.com/ (accessed July 27, 2014).
Kane, G., and Alavi, M. 2007. Information technology and organizational learning: An investigation
of exploration and exploitation process. Organization Science, 18(5), 796–812.
Lagorio, C. 2006. MySpace Pulls Ahead In Page View Race. CBS News, July 12. http://www.cbsnews.
com/news/myspace-pulls-ahead-in-page-view-race/ (accessed July 27, 2014).
Lamont, J. 2008. Contacts and connections: An array of options. KM World, July/August,
20–22.
Linden Lab. 2013. SL10B infographic: Ten years of Second Life. Second Life, June 20. http://commu-
nity.secondlife.com/t5/Featured-News/SL10B-Infographic-Ten-Years-of-Second-Life/ba-p/2053857
(accessed January 3, 2014).
McAfee, A. 2006. Enterprise 2.0: The dawn of emergent collaboration. MIT Sloan Management Review,
47(3), 21–28.
McAfee, A., and Sjoman, A. 2006. Wikis at Dresdner Kleinwort Wasserstein. Harvard Business School
Case 9–606–074, January 29.
McKinsey Co. 2007a. How businesses are using Web 2.0: A McKinsey Global Survey. The McKinsey
Quarterly, March. http://hosting-qa.mckinsey.com/insights/marketing_sales/how_businesses_are_us-
ing_web_20_a_mckinsey_global_survey (accessed March 4, 2013).
———. 2007b. How companies are marketing online: A McKinsey Global Survey. The McKinsey
Quarterly, September. http://hosting-qa.mckinsey.com/insights/marketing_sales/how_compa-
nies_are_marketing_online_a_mckinsey_global_survey (accessed March 4, 2013).
———. 2008. Building the web 2.0 enterprise: McKinsey Global Survey Results. The McKinsey
Quarterly, July.
Mintz, J. 2005. Friendster’s “eww” moment. The Wall Street Journal, December 8. http://online.wsj.
com/news/articles/SB113400562217716935 (accessed July 27, 2014).
Missick, C. 2008. http://www.missick.com/warletters.htm (accessed July 27, 2014).
O’Reilly, T. 2005. What is Web 2.0? Design patterns and business models for the next generation of
software. Communications & Strategies, January 1, 17.
Orlikowski, W. 1996. Improving organizational transformation over time: A situated change perspec-
tive. Information Systems Research, 7(1), 63–92.
Perez, J. 2007. Facebook’s Beacon more intrusive than previously thought. PC World, November 30.
www.pcworld.com/article/140182/article.html (accessed July 27, 2014).
Shachtman, N. 2007. Army squeezes soldier blogs, maybe to death. Wired, May 2. http://archive.
wired.com/politics/onlinerights/news/2007/05/army_bloggers?currentPage=all (accessed July 27,
2014).
Shah, S. 2003. Understanding the nature of participation and coordination in open and gated
source software. In Proceedings of the HBS-MIT Sloan Free/Open Source Software Confer-
ence 2003.
http://www.adap-tivepath.com/ideas/ajax-new-approach-web-applications/
http://www.adap-tivepath.com/ideas/ajax-new-approach-web-applications/
http://www.wired.com/2011/02/egypts-revolutionary-fire/
http://www.wired.com/2011/02/egypts-revolutionary-fire/
http://armyofdude.blogspot.com/
http://www.cbsnews.com/news/myspace-pulls-ahead-in-page-view-race/
http://commu-nity.secondlife.com/t5/Featured-News/SL10B-Infographic-Ten-Years-of-Second-Life/ba-p/2053857
http://commu-nity.secondlife.com/t5/Featured-News/SL10B-Infographic-Ten-Years-of-Second-Life/ba-p/2053857
http://hosting-qa.mckinsey.com/insights/marketing_sales/how_businesses_are_us-ing_web_20_a_mckinsey_global_survey
http://hosting-qa.mckinsey.com/insights/marketing_sales/how_businesses_are_us-ing_web_20_a_mckinsey_global_survey
http://hosting-qa.mckinsey.com/insights/marketing_sales/how_compa-nies_are_marketing_online_a_mckinsey_global_survey
http://hosting-qa.mckinsey.com/insights/marketing_sales/how_compa-nies_are_marketing_online_a_mckinsey_global_survey
http://online.wsj.com/news/articles/SB113400562217716935
http://www.missick.com/warletters.htm
http://www.pcworld.com/article/140182/article.html
http://archive.wired.com/politics/onlinerights/news/2007/05/army_bloggers?currentPage=all
http://www.cbsnews.com/news/myspace-pulls-ahead-in-page-view-race/
http://archive.wired.com/politics/onlinerights/news/2007/05/army_bloggers?currentPage=all
http://online.wsj.com/news/articles/SB113400562217716935
268 CHAPTER 10
Sorkin, A.R., and Lohr, S. 2011. Microsoft to buy Skype for $8.5 billion. New York Times, May 10, 2011.
type=blogs&_r=0 (accessed July 27, 2017).
Steinhauser, L., and Thon, S. 2008. Operation Burnt Frost: The power of social networks. NASA’s
Academy of Sharing Knowledge (ASK) Magazine, Summer, 17–19.
Thompson, C. 2008. Brave new world of digital intimacy. New York Times, September 7. www.nytimes.
com/2008/09/07/magazine/07awareness-t.html?pagewanted=all (accessed July 27, 2014).
von Krogh, G. 2003. Open-source software development. MIT Sloan Management Review, 44(3),
14–18.
Wasko, M., and Faraj, S. 2005. Why should I share? Examining social capital and knowledge contribu-
tion in electronic networks of practice. MIS Quarterly, 29(1), 35–57.
Wauters, R. 2011. Social network pioneer Friendster to erase all user photos, blogs and more on May
31. TechCrunch, April 26. http://techcrunch.com/2011/04/26/social-network-pioneer-friendster-to-
erase-all-user-photos-blogs-and-more-on-may-31/ (accessed July 27, 2014).
Wikipedia. 2013a. Facebook. http://en.wikipedia.org/wiki/Facebook (accessed July 11, 2013).
———. 2013b. FaceTime. https://en.wikipedia.org/wiki/FaceTime (accessed July 11, 2013).
———. 2013c. Folksonomies. http://en.wikipedia.org/wiki/Folksonomy (accessed July 11, 2013).
———. 2013d. Friendster. http://en.wikipedia.org/wiki/Friendster (accessed July 11, 2013).
———. 2013e. Google+. https://en.wikipedia.org/wiki/Google%2B (accessed July 11, 2013).
———. 2013f. LinkedIn. http://en.wikipedia.org/wiki/LinkedIn (accessed July 11, 2013).
———. 2013g. Mashup. https://en.wikipedia.org/wiki/Mashup_(web_application_hybrid) (accessed
July 11, 2013).
———. 2013h. Metaverse. http://en.wikipedia.org/wiki/Metaverse (accessed March 4, 2013).
———. 2013i. MySpace. http://en.wikipedia.org/wiki/Myspace (accessed July 11, 2013).
———. 2013j. Open Wonderland. http://en.wikipedia.org/wiki/Open_Wonderland (accessed July 11,
2013).
———. 2013k. Orkut. http://en.wikipedia.org/wiki/Orkut (accessed July 11, 2013).
———. 2013l. Pinterest. https://en.wikipedia.org/wiki/Pinterest (accessed July 11, 2013).
———. 2013m. Skype. https://en.wikipedia.org/wiki/Skype (accessed July 11, 2013).
———. 2013n. Twitter. http://en.wikipedia.org/wiki/Twitter (accessed July 11, 2013).
———. 2014. Wikipedia: List of policies. http://en.wikipedia.org/wiki/Wikipedia:List_of_policies
(accessed January 2, 2014).
Wittig, C., and Inkinen, S. 2004. MySQL open source database in 2004. Stanford Graduate School of
Business Case, SM-124, June.
Yeginsu, C. 2014. Turkey lifts Twitter ban after court calls it illegal. New York Times, April 3. www.
newsdiffs.org/article-history/www.nytimes.com/2014/04/04/world/middleeast/turkey-lifts-ban-on-
twitter.html (accessed July 27, 2014).
Social Network Pioneer Friendster To Erase All User Photos, Blogs And More On May 31
Social Network Pioneer Friendster To Erase All User Photos, Blogs And More On May 31
http://en.wikipedia.org/wiki/Facebook
https://en.wikipedia.org/wiki/FaceTime
http://en.wikipedia.org/wiki/Folksonomy
http://en.wikipedia.org/wiki/Friendster
https://en.wikipedia.org/wiki/Google%2B
http://en.wikipedia.org/wiki/LinkedIn
https://en.wikipedia.org/wiki/Mashup_
http://en.wikipedia.org/wiki/Metaverse
http://en.wikipedia.org/wiki/Myspace
http://en.wikipedia.org/wiki/Open_Wonderland
http://en.wikipedia.org/wiki/Orkut
https://en.wikipedia.org/wiki/Pinterest
https://en.wikipedia.org/wiki/Skype
http://en.wikipedia.org/wiki/Twitter
http://en.wikipedia.org/wiki/Wikipedia:List_of_policies
http://www.newsdiffs.org/article-history/www.nytimes.com/2014/04/04/world/middleeast/turkey-lifts-ban-on-twitter.html
http://www.newsdiffs.org/article-history/www.nytimes.com/2014/04/04/world/middleeast/turkey-lifts-ban-on-twitter.html
http://www.newsdiffs.org/article-history/www.nytimes.com/2014/04/04/world/middleeast/turkey-lifts-ban-on-twitter.html
269
11
Factors Influencing Knowledge
Management
In the last chapter, we discussed some of the emerging trends in knowledge
management. Earlier, in Chapter 5, we examined the impacts KM can have on
companies and other private or public organizations. These impacts result either
directly from KM solutions or indirectly through knowledge created by KM so-
lutions. KM solutions include KM processes and systems, which were discussed
in Chapter 4. In this chapter, we argue that various KM solutions may have dif-
ferent impacts on performance depending on the circumstances, and we examine
the key factors affecting the suitability of KM solutions. This perspective, which
is called contingency perspective, is discussed next and the overall approach in
this chapter is outlined. The subsequent sections examine the effects of several
important factors.
A CONTINGENCY VIEW OF KNOWLEDGE MANAGEMENT
According to Becerra-Fernandez and Sabherwal (2001), previous literature promoted
a universalistic view of knowledge management, which would imply that there is a
single best approach of managing knowledge that should be adopted by all organiza-
tions in all circumstances. This seems to be implicit in the literature on knowledge
management; for example, knowledge sharing is recommended as being useful to all
organizations, although we believe that the use of direction may sometimes represent
an equally effective and more efficient alternative. In contrast to this universalistic
view is the contingency view of KM, which has previously been used, for example,
in the literature on organization design. This view suggests that no one approach is
best under all circumstances. Whereas a universalistic view focuses on identifying
a single path to successful performance, a contingency theory considers the path to
success to include multiple alternative paths with success being achieved only when
the appropriate path is selected. For instance an organization design with few rules
or procedures is considered appropriate for small organizations, whereas one with
extensive rules and procedures is recommended for large organizations.
Similarly, Sabherwal and Sabherwal (2007) argued that not all knowledge
management efforts would necessarily end in better results in terms of firm per-
formance. Instead, they argued that a KM effort that is aligned to the firm’s busi-
ness strategy would lead to an improvement in the firm’s business performance
whereas a KM effort that is not aligned with the firm’s business strategy would
270 CHAPTER 11
not. Based on an empirical study of the stock market responses to the announce-
ment of KM efforts, they found this expectation to be empirically supported.
More specifically, the stock market reacted better to the announcement of a KM
effort as the alignment between the announced KM effort and the firm’s business
strategy increased.
We recommend the use of an approach based on contingency theory for identify-
ing KM processes and solutions. When asked what kind of a KM solution should
an organization use, we often find ourselves responding, “it depends,” rather than
unequivocally recommending a specific solution. We need to understand the specific
circumstances within the organization and the ones surrounding it in order to identify
the KM solution that would be most beneficial for those circumstances. This indicates
that each KM solution is contingent upon the presence of certain circumstances—
hence, the name. A contingency perspective for KM is supported by prior empirical
research (Becerra-Fernandez and Sabherwal 2001; Sabherwal and Sabherwal 2005).
For example based on a detailed study of Nortel Networks Corporation, Massey et
al. (2002, p. 284) conclude: “Thus, a key finding of our study is that successful KM
initiatives like Nortel’s cannot be disentangled from broader organizational factors
and changes.”
Figure 11.1 summarizes the way in which the relationship between the contingency
factors and KM solutions is examined in this chapter. As discussed in Chapters 3
and 4, KM solutions include KM systems and KM processes. Much of this chapter
focuses on knowledge management processes, with the choice of appropriate KM
process depending on contingency factors, as shown by arrow 1 in Figure 11.1. Once
the appropriate KM processes are recognized, the KM systems needed to support
them can be identified as well. Thus, the contingency factors indirectly affect KM
systems and the mechanisms and technologies enabling KM systems, as shown us-
ing arrows 2 and 3. Moreover, the KM infrastructure supports KM mechanisms and
technologies (arrow 4), which in turn affect KM systems (arrow 5) and KM systems
support KM processes (arrow 6). Thus, the KM infrastructure indirectly affects KM
processes (arrow 7).
Several contingency factors influence the choice of KM processes. They include
characteristics of the tasks being performed, the knowledge being managed, the
organization, and the organization’s environment. Figure 11.2 summarizes these cat-
egories of contingency factors affecting KM processes. In the forthcoming sections,
we will examine the effects of task characteristics and knowledge characteristics,
respectively. And, we will also describe the effects of organizational and environ-
mental characteristics.
In general, the contingency factors and KM infrastructure affect the suitability of
KM processes in two ways: (a) by increasing or reducing the need to manage knowl-
edge in a particular way; and (b) by increasing or reducing the organization’s ability
to manage knowledge in a particular way. For example, larger organizations have
a greater need to invest in knowledge sharing processes, whereas an organization
culture characterized by trust increases the organization’s ability to use knowledge
sharing processes. Consequently, the benefits from a KM process would depend on
the contingency factors (Sabherwal and Sabherwal 2005).
271
K
M
F
ou
nd
at
io
ns
K
M
S
ol
ut
io
ns
K
M
I
n
fr
a
st
ru
ct
u
re
rg
a
ni
za
tio
n
C
ul
tu
re
rg
a
ni
za
tio
n
S
tr
uc
tu
re
T
fr
as
tr
uc
tu
re
om
m
on
K
no
w
le
d
ge
hy
si
ca
l E
nv
iro
nm
en
t
K
M
S
ys
te
m
s
no
w
le
dg
e
D
is
co
ve
ry
S
ys
te
m
s
no
w
le
dg
e
C
ap
tu
re
S
ys
te
m
s
no
w
le
dg
e
S
h
ar
in
g
S
ys
te
m
s
no
w
le
dg
e
A
p
pl
ic
at
io
n
S
ys
te
m
s
K
M
M
ec
h
an
is
m
s
K
M
T
ec
h
n
o
lo
g
ie
s
K
M
P
ro
ce
ss
e
s
no
w
le
dg
e
D
is
co
ve
ry
no
w
le
dg
e
C
ap
tu
re
no
w
le
dg
e
S
h
ar
in
g
no
w
le
dg
e
A
p
pl
ic
at
io
n
C
on
tin
ge
nc
y
F
ac
to
rs 1
2
3
4
5
6
7
K
M
F
ou
nd
at
io
ns
K
M
S
ol
ut
io
ns
K
M
I
n
fr
a
st
ru
ct
u
re
rg
a
ni
za
tio
n
C
ul
tu
re
rg
a
ni
za
tio
n
S
tr
uc
tu
re
T
fr
as
tr
uc
tu
re
om
m
on
K
no
w
le
d
ge
hy
si
ca
l E
nv
iro
nm
en
t
K
M
S
ys
te
m
s
no
w
le
dg
e
D
is
co
ve
ry
S
ys
te
m
s
no
w
le
dg
e
C
ap
tu
re
S
ys
te
m
s
no
w
le
dg
e
S
h
ar
in
g
S
ys
te
m
s
no
w
le
dg
e
A
p
pl
ic
at
io
n
S
ys
te
m
s
K
M
M
ec
h
an
is
m
s
K
M
T
ec
h
n
o
lo
g
ie
s
K
M
P
ro
ce
ss
e
s
no
w
le
dg
e
D
is
co
ve
ry
no
w
le
dg
e
C
ap
tu
re
no
w
le
dg
e
S
h
ar
in
g
no
w
le
dg
e
A
p
pl
ic
at
io
n
C
on
tin
ge
nc
y
F
ac
to
rs 1
2
3
4
5
6
7
F
ig
ur
e
11
.1
C
o
n
ti
n
g
en
cy
F
ac
to
rs
a
n
d
K
n
o
w
le
d
g
e
M
an
ag
em
en
t
272 CHAPTER 11
THE EFFECTS OF TASK CHARACTERISTICS
The underlying argument here is that the KM processes that are appropriate for an
organizational subunit (e.g., a department, a geographic location, etc.) depend on the
nature of its tasks (Becerra-Fernandez and Sabherwal 2001; Haas and Hansen 2005).
This involves viewing each subunit at the aggregate level based on the predominant
nature of its tasks. This approach has considerable support in prior literature. For
example, Van de Ven and Delbecq (1974) offered a contingency view of the rela-
tionship between subunit tasks and organization structure. They suggested that the
structure appropriate for a subunit depends on task difficulty, or on the problems in
analyzing the work and stating performance procedures and task variability, or on
the variety of problems encountered in the tasks. Lawrence and Lorsch (1967) also
focused on a task characteristic—task uncertainty—at the subunit level, and found
subunits that perform certain, predictable tasks to be more effective when they were
formally structured. Thus, a number of task characteristics have been studied at the
level of organizational subunits. Here, two task characteristics—task uncertainty and
task interdependence—are considered as influencing the appropriate KM processes
(Spender 1996).
Consistent with Lawrence and Lorsch (1967), greater task uncertainty is argued to
reduce the organization’s ability to develop routines, and hence knowledge application
Figure 11.2 Categories of Contingency Factors
Environmental Characteristics
Organizational Characteristics
Task Characteristics
Knowledge Characteristics
Knowledge Management
Environmental Characteristics
Organizational Characteristics
Task Characteristics
Knowledge Characteristics
Knowledge Management
FACTORS INFLUENCING KNOWLEDGE MANAGEMENT 273
would depend on direction. Moreover when task uncertainty is high, externalization
and internalization would be more costly due to changing problems and tasks. Under
such circumstances knowledge is more likely to remain tacit, inhibiting the ability
to use combination or exchange. Therefore under high task uncertainty, direction or
socialization would be recommended. For example, individuals responsible for product
design when customer tastes are expected to change frequently would benefit most
from socializing with, and receiving directions from, each other.
On the other hand, when the tasks are low in uncertainty routines can be devel-
oped for the knowledge supporting them. Moreover, the benefits from externalizing
or internalizing knowledge related to any specific task would accumulate through
the greater occurrence of that task. Finally, exchange and combination would be
useful due to the externalization of potentially tacit knowledge. Therefore under low
task uncertainty, routines, exchange, combination, internalization, or externalization
would be recommended. These conclusions are summarized in the bottom part of
Figure 11.3.
For example, for individuals performing tasks related to credit and accounts receiv-
ables, considerable benefits would be obtained from the use of routines (e.g., those
for credit-checking procedures), exchange (e.g., sharing of standards and policies),
combination (e.g., integration of explicit knowledge that different credit analysts may
have generated from their experiences), and from externalization and internalization
(e.g., to facilitate training and learning of existing policies by new credit analysts).
The second important task characteristic is task interdependence, which indicates
the extent to which the subunit’s achievement of its goals depends on the efforts of
other subunits (Jarvenpaa and Staples 2001). Performing tasks that are independent
of others primarily requires the knowledge directly available to the individuals within
the subunit. These tasks rely mainly on distinctive units of knowledge, such as “func-
tional knowledge embodied in a specific group of engineers, elemental technologies,
information processing devices, databases, and patents” (Kusonaki et al. 1998).
They often require deep knowledge in a particular area. With internalization, such
as when individuals acquire knowledge by observing or by talking to others, as well
as with externalization, such as when they try to model their knowledge into analo-
gies, metaphors, or problem-solving systems, the learning processes are personal and
individualized. Through externalization, the individual makes the knowledge more
agreeable and understandable to others in the group while through internalization the
individual absorbs knowledge held by others in the group (Maturana and Varela 1987).
Internalization and externalization are thus fundamental to KM in an independent task
domain. Performance of interdependent tasks relies mainly on dynamic interaction
in which individual units of knowledge are combined and transformed through com-
munication and coordination across different functional groups. This creates greater
causal ambiguity, since knowledge is being integrated across multiple groups that
may not have a high level of shared understanding. Socialization and combination
processes, both of which help synthesize prior knowledge to create new knowledge,
are therefore appropriate for interdependent tasks (Grant 1996).
The left portion of Figure 11.3 shows that internalization and externalization should
be preferred for independent tasks whereas exchange, combination, and socialization
274 CHAPTER 11
should be preferred for interdependent tasks. Moreover, directions and routines can be
used for independent as well as interdependent tasks; their suitability depends more
on task uncertainty, as already discussed.
Combining the arguments regarding the effects of task uncertainty and task inter-
dependence, we obtain the four-cell matrix in Figure 11.3. As shown in the matrix,
direction is recommended for uncertain, independent tasks; direction and socializa-
tion are recommended for uncertain, interdependent tasks; exchange, combination,
and routines are recommended for certain, interdependent tasks; and internalization,
externalization, and routines are recommended for certain, independent tasks.
THE EFFECTS OF KNOWLEDGE CHARACTERISTICS
Three knowledge characteristics—explicit versus tacit, procedural versus declarative,
and general versus specific—were examined in Chapter 2. The first two of these knowl-
edge characteristics directly affect the suitability of KM processes. The underlying
contingency argument is that certain KM processes may have greater impact on the
value that one type of knowledge contributes to the organization, while some other
KM processes might affect the value of another type of knowledge (Spender 1996).
Figure 11.4 shows the KM processes that were presented earlier in Chapter 3 while
also depicting the effects of the two knowledge classifications. The difference between
Figure 11.3 Effects of Task Characteristics on KM Processes
Direction
Exchange
Combination
Routines
Internalization
Externalization
Routines
Direction
Socialization
Task Uncertainty
T
as
k
In
te
rd
ep
en
de
nc
e
High
Low High
Low
Routines
Internalization
Externalization
Exchange
Combination
Direction
Socialization
Exchange
Combination
Socialization
Direction
Routines
Internalization
Externalization
Direction
Routines
Direction
Exchange
Combination
Routines
Internalization
Externalization
Routines
Direction
Socialization
Task Uncertainty
T
as
k
In
te
rd
ep
en
de
nc
e
High
Low High
Low
Routines
Internalization
Externalization
Exchange
Combination
Direction
Socialization
Exchange
Combination
Socialization
Direction
Routines
Internalization
Externalization
Direction
Routines
275
F
ig
ur
e
11
.4
E
ff
ec
ts
o
f
K
n
o
w
le
d
g
e
C
h
ar
ac
te
ri
st
ic
s
o
n
K
M
P
ro
ce
ss
es
C
ap
tu
re
ac
it
:
E
xt
er
na
liz
at
io
n
xp
li
ci
t:
I
nt
er
na
liz
at
io
n
S
h
ar
in
g
ac
it
:
S
oc
ia
liz
at
io
n
xp
li
ci
t:
E
xc
ha
ng
e
A
p
p
li
ca
ti
o
n
ac
it
xp
li
ci
t:
D
ir
ec
tio
n
ac
it
xp
li
ci
t:
R
ou
tin
es
D
is
co
v
er
y
xp
li
ci
t:
C
om
bi
n
at
io
n
ac
it
:
S
oc
ia
liz
at
io
nP
ro
ce
du
ra
l o
r D
ec
la
ra
tiv
e
P
ro
ce
du
ra
l
C
ap
tu
re
ac
it
:
E
xt
er
na
liz
at
io
n
xp
li
ci
t:
I
nt
er
na
liz
at
io
n
S
h
ar
in
g
ac
it
:
S
oc
ia
liz
at
io
n
xp
li
ci
t:
E
xc
ha
ng
e
A
p
p
li
ca
ti
o
n
ac
it
xp
li
ci
t:
D
ir
ec
tio
n
ac
it
xp
li
ci
t:
R
ou
tin
es
D
is
co
v
er
y
xp
li
ci
t:
C
om
bi
n
at
io
n
ac
it
:
S
oc
ia
liz
at
io
nP
ro
ce
du
ra
l o
r D
ec
la
ra
tiv
e
P
ro
ce
du
ra
l
276 CHAPTER 11
KM processes appropriate for explicit and tacit knowledge is based directly on the
main difference between these knowledge types.
For knowledge discovery, combination would be appropriate for integrating multiple
streams of explicit knowledge, for example with knowledge discovery systems, where
socialization would be suitable for integrating multiple streams of tacit knowledge. For
knowledge capture, externalization would be appropriate for tacit knowledge as it helps
convert tacit knowledge into explicit knowledge, for example in knowledge capture
systems; whereas internalization would be appropriate for explicit knowledge, as it helps
convert explicit knowledge into tacit knowledge, for example in learning. For knowledge
sharing, exchange helps transfer explicit knowledge whereas socialization is needed
for tacit knowledge. These intuitively obvious recommendations are also based on the
logic that a KM process would contribute much to the value of knowledge if it is both
effective and efficient for managing that knowledge (Gupta and Govindarajan 2000).
Some KM processes might not contribute to the value of a given type of knowledge
either because they are not effective in managing it (e.g., combination and exchange
would not be effective for managing tacit knowledge), or because they are too expensive
or too slow—that is, an alternative process would be able to integrate it more quickly or
at a lower cost (e.g., socialization would be too expensive and slow for sharing explicit
knowledge, especially in comparison to exchange).
No difference between the suitability of direction and routines is expected between
tacit and explicit knowledge. In other words, either direction or routines could be used
to apply either tacit or explicit knowledge. This is the case because no knowledge is
being transferred in either direction or routines; only recommendations based on the
expert’s knowledge (whether tacit or explicit) are being transferred. Both direction
and routine processes are appropriate to be used mainly for procedural knowledge, or
“know-how,” which focuses on the processes or means that should be used to perform
the required tasks—for instance how to perform the steps in performing a specific
process, such as installing a piece of software. This is shown in the right portion of
Figure 11.4. Procedural knowledge differs from declarative knowledge, substantive
knowledge, or “know what,” which focuses on beliefs about relationships among
variables, as we discussed in Chapter 2. As shown in the left part of Figure 11.4, all
the KM processes supporting knowledge discovery, capture, and sharing can be used
for both declarative and procedural knowledge.
Thus, either direction or routines could be used to apply procedural knowledge,
whether tacit or explicit. KM processes used to discover, capture, or share knowledge
are the same for both procedural and declarative kinds of knowledge. However, these
processes differ between tacit and explicit knowledge, as discussed and shown in
Figure 11.4 within the boxes for knowledge discovery, capture, and sharing.
THE EFFECTS OF ORGANIZATIONAL AND ENVIRONMENTAL CHARACTERISTICS
Two organizational characteristics—size and strategy—and one environmental
characteristic—uncertainty—affect the suitability of various knowledge management
processes. Table 11.1 summarizes the effects of environmental and organizational
characteristics.
FACTORS INFLUENCING KNOWLEDGE MANAGEMENT 277
Organization size affects KM processes by influencing the choice between the two
processes supporting knowledge application (direction, routines) and the two processes
supporting knowledge sharing (socialization, exchange). For knowledge application,
large and more bureaucratic organizations would benefit more from routines because
of their greater use of standards and their potential for reuse of these routines. Small
organizations, on the other hand, are usually not very bureaucratic and have less po-
tential for reusing processes and procedures coded as routines. They would therefore
benefit more from direction, which does not rely on standardization and rules. The
circumstances needed for direction—for example, the knowledge user’s trust in the
individual providing direction (Conner and Prahalad 1996)—are also more likely to
exist in smaller organizations. Large organizations are often globally distributed and
therefore knowledge sharing across greater distances would be needed in large orga-
nizations; whereas knowledge is more likely to be shared across shorter distances in
smaller co-located organizations. Therefore, knowledge sharing through exchange is
recommended for large distributed organizations while socialization is recommended
for small co-located organizations (Boh 2007). Socialization for knowledge discovery
is also recommended for small organizations, although combination could be used
in either small or large organizations. Finally, small and large organizations do not
differ in terms of the suitability of the alternative knowledge capture processes (ex-
ternalization, internalization).
For example, a small financial consulting firm with 25 employees would have only
a few experts in any area—for instance, customer relations practices. Consequently,
others in the organization are likely to trust these experts and depend on their direc-
tion. Moreover, the small number of employees would have frequent opportunities
Table 11.1
Effect of Environmental and Organizational Characteristics on KM Processes
Characteristic Level/Type Recommended KM Processes
Organization Size Small
Large
Business Strategy Low cost
Differentiation
Environmental Uncertainty Low
High
278 CHAPTER 11
to interact with each other thereby enabling greater use of socialization for knowl-
edge discovery as well as knowledge sharing. On the other hand, a large consulting
firm with over 5,000 employees would find it infeasible or overly expensive to rely
on socialization, especially across large distances. Instead, in such an organization,
knowledge sharing would rely more on exchange of knowledge explicated in reports,
lessons learned documents, and so on. In addition, large organizations may find it
more likely to reuse knowledge that has been explicated previously, for example in
lessons learned systems or best practices databases as described in Chapter 8. Fur-
thermore, this firm would find it beneficial to develop and use routines for applying
knowledge. Routines would be more economical due to their greater frequency of
use in such larger firms and also needed by more individuals within the organization
who may be seeking help.
The effect of business strategy may be examined using Porter’s (1980, 1985)
popular typology of low cost and differentiation strategies.1 Organizations pursuing
a low-cost strategy should focus on applying existing knowledge rather than creat-
ing new knowledge, whereas organizations following a differentiation strategy are
more likely to innovate (Langerak et al. 1999), seek new opportunities (Miles and
Snow 1978), and frequently develop new products (Hambrick 1983). They would
therefore benefit more from knowledge discovery and capture processes (combination
and socialization). Organizations pursuing either low-cost or differentiation strategy
would benefit from knowledge capture and sharing processes, as these processes can
be used to capture or share knowledge on ways of reducing costs as well as innovat-
ing with products or services.
For example, a supermarket chain competing through a low-cost strategy would
seek to reuse prior knowledge about ordering, inventory management, supplier rela-
tions, pricing, and so on. This company would therefore use organizational routines
(in case the company is large) or direction (if the company is small) to support the
application of prior knowledge. In contrast, an exclusive fashion boutique trying to
differentiate itself from its competitors would seek new knowledge about attracting
competitors’ customers, and retaining its own customers, developing innovative
products, and so on. This boutique would significantly benefit from socialization and
combination processes for creating new knowledge about these aspects, using prior
tacit and explicit knowledge, respectively.
The environmental uncertainty encountered by the organization also affects
knowledge management (Hsu and Wang 2008), and the suitability of various KM
processes (Sabherwal and Sabherwal 2005). Environmental uncertainty, which re-
fers to the business context in which the firm operates, should not be confused with
task uncertainty which refers to not having a priori knowledge of details involved in
the steps required by a task. When the organization faces low levels of uncertainty,
knowledge sharing and knowledge capture processes would be recommended because
the captured and shared knowledge would be relevant for longer periods of time. On
the other hand, under higher uncertainty, knowledge application and discovery would
be recommended. Knowledge application contributes in an uncertain environment
by enabling individuals to address problems based on existing solutions indicated
by those possessing the knowledge, instead of more time-consuming processes like
FACTORS INFLUENCING KNOWLEDGE MANAGEMENT 279
sharing knowledge (Alavi and Leidner 2001; Conner and Prahalad 1996). Knowledge
discovery processes contribute by enhancing the organization’s ability to develop
new innovative solutions to emergent problems that may not have been faced before
(Davenport and Prusak 1998).
For example, the environment would be rather certain and predictable for an auto-
mobile-manufacturing firm that has a relatively stable product line and competes with
a small number of competitors especially when each firm has its own, clear market
niche. For such an organization, knowledge about product design, manufacturing,
marketing, sales, and so forth would be generally stable, benefiting from the sharing
of prior knowledge through socialization or exchange and the capture of knowledge
through internalization and externalization. Knowledge sharing, as well as internaliza-
tion and externalization, would have long-term benefits, as the knowledge remains
inherently stable. On the other hand, an international mobile phone manufacturer
having a dynamic product line and evolving customer base would face a highly
uncertain environment. This organization would seek to create new knowledge and
quickly apply existing knowledge by investing in combination and socialization
for knowledge discovery and routines and direction for knowledge application. For
example, Canon Inc.’s success has been attributed to both creating new innovations,
for example in the photography industry, and quickly applying these innovations to
other relevant products like fax and copy machines.
The effects of contingency factors on the selection of knowledge management
solutions at one leading consumer-goods company, Groupe Danone, is described in
Box 11.1.
IDENTIFICATION OF APPROPRIATE KNOWLEDGE MANAGEMENT SOLUTIONS
Based on the above discussion, we recommend a methodology for identifying ap-
propriate KM solutions. The methodology includes the following seven steps:
1. Assess the contingency factors.
2. Identify the KM processes based on each contingency factor.
3. Prioritize the needed KM processes.
4. Identify the existing KM processes.
5. Identify the additional needed KM processes.
6. Assess the KM infrastructure and identify the sequential ordering of KM
processes.
7. Develop additional needed KM systems, mechanisms, and technologies.
These seven steps are now discussed.
STEP 1. ASSESS THE CONTINGENCY FACTORS
This step requires assessing the organization’s environment in terms of the contin-
gency factors—characterizing the tasks, the knowledge, the environment, and the
organization—and how they contribute to uncertainty.
280 CHAPTER 11
Box 11.1
Networking Attitude Fosters Knowledge Sharing and Creation at Danone
Groupe Danone is a leading consumer-goods company, with headquarters in Paris. Known
as Dannon in the United States, it is among the world’s leading producers of dairy products,
bottled water, cereals, and baby foods. Danone is a fast-moving and entrepreneurial company,
emphasizing differentiation rather than cost reduction. Although it is spread across 120 coun-
tries, it is smaller than its competitors and operates in a decentralized fashion with consider-
able emphasis on responsiveness to the needs of local markets. Most of the knowledge used
at Danone is held in tacit form by its employees. Danone’s business strategy (differentiation),
environmental uncertainty (which is high because of its operating in numerous countries with
changing products), and focus on tacit knowledge make it important for the company to share
knowledge across countries and across divisions. However, the use of exchange through IT did
not seem promising due to the tacit nature of their knowledge. Moreover, Danone’s employees
did not make much use of portals and information technologies in general.
Senior executives at Danone, including Frank Mougin, Danone’s executive vice president of hu-
man resources, recognized that using IT to share knowledge would not work as well for the com-
pany. Therefore, they have concentrated their KM efforts on using socialization for sharing existing
knowledge as well as creating new knowledge using processes akin to their firm’s characteristics.
As a result, they launched the Networking Attitude Conference in the fall of 2002. Networking
Attitude was presented to the company’s general managers as a mechanism for people in units
distant from each other to share their knowledge. It focused on the following initiatives:
for best practices obtained from “givers,” using a “check” and a facilitator tracking the num-
ber of checks acquired by each giver and using it as a way of evaluating the relevance of
each best practice.
potential “givers,” and with no observer to facilitate more spontaneous networking.
writing suggestions and problems on the front and back of their T-shirts, respectively.
met regularly (every six months or so).
Although there was no formal tracking of its impact, the Networking Attitude Conference ap-
pears to have worked well for Danone, with people participating actively by sharing important
knowledge and providing each other recommendations on directions. People seemed to like it,
and marketplaces had especially become popular in Mexico and Hungary. From 2004 to 2007,
Danone employees shared about 640 best practices with each other and made useful knowledge
available to about 5,000 (out of a total of about 9,000) Danone managers around the world.
The success of Networking Attitude led to a demand for introducing additional networking oppor-
tunities. Danone was considering three ways of extending Networking Attitude: deeper (i.e., involving
all the employees rather than only managers), wider (i.e., extending the use of networks to custom-
ers, consumers, suppliers, and partners), and richer (i.e., using it more explicitly for innovation by in-
viting employees to network with each other with the goal of identifying new products or processes).
Source: Compiled from Edmondson et al. 2008; Groupe Danone; Mougin and Benanti 2005.
The variety of tasks for which KM is needed should be characterized in terms of
task interdependence and task uncertainty. Furthermore, the kind of knowledge those
tasks require should be classified as general or specific, declarative or procedural,
and tacit or explicit. Environmental uncertainty may arise from changes in the firm’s
competition, government regulations and policies, economic conditions, and so on.
FACTORS INFLUENCING KNOWLEDGE MANAGEMENT 281
Additionally, the organization’s business strategy—low-cost or differentiation—
should be identified. Lastly, the organization should be classified as small or large
relative to its competitors. In some instances, it may be labeled as midsized, in which
case the KM processes would be based on considerations of both small and large
organizations.
In using these contingencies, it is important to use the appropriate unit of analysis,
which could be either the entire organization or a subunit depending on the specific
context for which the KM solution would be developed. When deciding on KM
processes that are intended to improve KM within a subunit, such as the account-
ing department of the organization, the contingency factors should be evaluated for
that subunit. On the other hand, when deciding on KM processes that are intended
to improve KM for the entire company, the contingency factors should be evaluated
for the entire company.
STEP 2. IDENTIFY THE KM PROCESSES BASED ON EACH CONTINGENCY FACTOR
Next, the appropriate KM processes based on each contingency factor should be
identified. In doing this, Table 11.2, which summarizes the effects of various contin-
gency factors, should be useful. This table shows the seven contingency factors and
the effects they have on the KM processes. It is important to note, however, that this
table only provides some of the most important factors that need to be considered in
making this choice. There are several other factors, such as the information intensity
of the organization’s industry, that would also affect the appropriateness of KM pro-
cesses, but they have been excluded to simplify the presentation.
STEP 3. PRIORITIZE THE NEEDED KM PROCESSES
Once the KM processes appropriate for each contingency factor have been identified,
they need to be considered together in order to identify the needed KM processes.
In doing so, it is useful to assign a value of 1.0 to situations where a KM process is
appropriate for a contingency variable and 0.0 where it is not appropriate. Moreover,
where a KM process is appropriate for all possible states of a contingency variable,
a value of 0.5 could be assigned. As a result, a prioritization of the importance of
various KM processes can be developed, and a Cumulative Priority Score can be
computed. For example, if KM process A has a composite score of 6.0 based on the
seven contingency factors whereas another one (B) has a composite score of 3.0,
greater attention is needed toward KM process A rather than B. This computation is
shown in greater detail using an illustrative example in the next section.
STEP 4. IDENTIFY THE EXISTING KM PROCESSES
Next, the KM processes that are currently being used should be identified. In doing
so, a short survey of some of the employees assessing the extent to which each KM
process is being used may be helpful. Possible approaches for such assessments are
discussed in detail in the next chapter.
282
Ta
bl
e
11
.2
A
p
p
ro
p
ri
at
e
C
ir
cu
m
st
an
ce
s
fo
r
V
ar
io
u
s
K
M
P
ro
ce
ss
es
K
M
P
ro
ce
ss
es
C
on
tin
ge
nc
y
Fa
ct
or
s
C
om
bi
na
tio
n
S
oc
ia
liz
at
io
n
fo
r
K
no
w
le
dg
e
D
is
co
ve
ry
S
oc
ia
liz
at
io
n
fo
r
K
no
w
le
dg
e
S
ha
rin
g
E
xc
ha
ng
e
E
xt
er
na
l-
iz
at
io
n
In
te
rn
al
iz
at
io
n
D
ire
ct
io
n
R
ou
tin
es
Ta
sk
u
nc
er
ta
in
ty
Lo
w
H
ig
h
H
ig
h
Lo
w
Lo
w
Lo
w
H
ig
h
Lo
w
Ta
sk
in
te
rd
ep
en
de
nc
e
H
ig
h
H
ig
h
H
ig
h
H
ig
h
Lo
w
Lo
w
H
ig
h/
lo
w
H
ig
h/
lo
w
E
xp
lic
it
(E
)
or
ta
ci
t (
T
)
kn
ow
ed
ge
E
T
T
E
T
E
T
/E
T
/E
P
ro
ce
du
ra
l (
P
)
or
d
ec
la
ra
tiv
e
(D
)
kn
ow
le
dg
e
P
/D
P
/D
P
/D
P
/D
P
/D
P
/D
P
P
O
rg
an
iz
at
io
na
l s
iz
e
S
m
al
l/l
ar
ge
S
m
al
l
S
m
al
l
La
rg
e
S
m
al
l/l
ar
ge
S
m
al
l/l
ar
ge
S
m
al
l
La
rg
e
B
us
in
es
s
st
ra
te
gy
(
Lo
w
c
os
t—
LC
; D
iff
er
en
tia
tio
n—
D
)
D
D
LC
/D
LC
/D
LC
/D
LC
/D
LC
LC
E
nv
iro
nm
en
ta
l u
nc
er
ta
in
ty
H
ig
h
H
ig
h
Lo
w
Lo
w
Lo
w
Lo
w
H
ig
h
H
ig
h
FACTORS INFLUENCING KNOWLEDGE MANAGEMENT 283
STEP 5. IDENTIFY THE ADDITIONAL NEEDED KM PROCESSES
Based on the needed KM processes (identified in step 3) and the existing KM pro-
cesses (identified in step 4), the additional needed KM processes can be identified.
This comparison might also find some of the existing KM processes to not be very
useful. In other words, if a KM process is identified as needed (step 3) but it is not
currently being used (step 4), it should be added; whereas if a KM process is not
identified as needed (step 3) but it is currently being used (step 4), it could potentially
be dropped—at least based on knowledge management considerations.
STEP 6. ASSESS THE KM INFRASTRUCTURE AND IDENTIFY THE SEQUENTIAL
ORDERING OF KM PROCESSES
The KM infrastructure indirectly affects the KM processes as we discussed earlier.
Specifically, organization culture, organization structure, and the physical environ-
ment can facilitate or inhibit knowledge sharing and creation. Additionally, informa-
tion technologies can support all KM processes, and organizing knowledge can help
enhance the efficiency of knowledge sharing (e.g., through common language and
vocabulary) and application processes (e.g., by enhancing recognition of individual
knowledge domains). These aspects of the KM infrastructure should be considered
with respect to the additional KM processes needed (as identified in step 5) to identify
the KM processes for which supporting infrastructure, mechanisms, and technologies
currently exist. This step is especially important when deciding the sequence in which
KM processes that are nearly equal in importance (step 3) should be developed.
STEP 7. DEVELOP ADDITIONAL NEEDED KM SYSTEMS, MECHANISMS,
AND TECHNOLOGIES
Steps 1 through 6 have helped identify the KM processes and the order in which they
should be developed. Now the organization needs to undertake steps to initiate the
creation of KM systems, mechanisms, and technologies that would support those
KM processes. This might require creation of teams, acquisition of technologies,
development of systems, and so on. In the long run, these systems, mechanisms, and
technologies would also contribute to the KM infrastructure.
ILLUSTRATIVE EXAMPLE
As an illustration, which is kept somewhat simple to prevent this discussion from be-
coming overly complex, let us consider the fictional Doubtfire Computer Corporation,
a manufacturer of low-end personal computers for home users. A small player in this
industry, Doubtfire has recently undergone some difficult times due to new competition
for its product line. Competitors make frequent changes in technology in an attempt to
gain the upper hand in the marketplace with more state-of-the-art products. Having belat-
edly recognized this, Doubtfire recently hired a new president and a new sales manager
to turn the situation around. The new president called a meeting of the staff to discuss
284 CHAPTER 11
possible strategies for the financial turnaround of the company. The main thrust of this
presentation was that the staff needed to better manage knowledge so as to creatively
identify areas where new technology could improve the company’s products and opera-
tions. Based on inputs from the senior management, the president hired a knowledge
management consulting firm, KM-Consult Inc., to help improve its KM strategy.
A team of consultants from KM-Consult Inc. conducted an in-depth study of
Doubtfire, using interviews with several employees and examination of company
documents. Based on their investigation, they concluded that Doubtfire is a small
organization that has pursued a low-cost business strategy to operate in an uncertain
environment as is typical of high-tech firms. Knowledge management is needed for
its tasks, which are highly interdependent and also highly uncertain due to changing
components in the computer industry. Doubtfire relies mainly on the tacit, procedural
knowledge possessed by its employees rather than seeking the explication of that
knowledge or management of declarative knowledge. Then, based on Table 11.2, the
consulting team arrived at the following conclusions.
First, based on Doubtfire’s small organization size, socialization (for knowledge
sharing or knowledge discovery) and direction processes would be appropriate. In
addition, combination, internalization, and externalization could be used regardless
of organization size. However, exchange and routines would be inappropriate due to
Doubtfire being a small organization.
Moreover, considering Doubtfire’s low-cost business strategy, direction and routines
would be appropriate. In addition, socialization (for knowledge sharing), exchange,
internalization, and externalization could be used regardless of strategy. However,
combination and socialization (for knowledge discovery) would be inappropriate
because they are not suitable for firms pursuing a low-cost strategy.
The consulting team also concluded that, based on the uncertain environment in
which Doubtfire operates—which is characteristic of firms in the high-tech sector—
direction, combination, and socialization (for knowledge discovery) would be ap-
propriate. However, the remaining processes would be inappropriate as they are more
suitable for certain, predictable environments.
The high task interdependence in Doubtfire suggests that socialization (for
knowledge sharing or knowledge discovery), combination, and exchange would
be appropriate. In addition, direction and routines could be used regardless of task
interdependence. However, externalization and internalization would not be as use-
ful. The high task uncertainty suggests that socialization (for knowledge sharing or
knowledge discovery) and direction would be appropriate. However, the remaining
processes would be less suitable.
The procedural nature of knowledge indicates that direction and routines would
be useful for managing this knowledge. The tacit nature of knowledge suggests that
socialization (for knowledge sharing or knowledge discovery) and externalization
would be appropriate. In addition, direction and routines could be used regardless of
tacit or explicit nature of knowledge.
Table 11.3 shows the results of this analysis by KM-Consult Inc. The cells in the
columns for each contingency factor show the suitability of the KM process in that
row for that contingency variable. More specifically, “Yes” indicates that KM process
285
Ta
bl
e
11
.3
P
ri
o
ri
ti
zi
n
g
K
M
P
ro
ce
ss
es
f
o
r
D
o
u
b
tfi
re
C
o
m
p
u
te
r
C
o
rp
o
ra
ti
o
n
K
M
P
ro
ce
ss
es
C
on
tin
ge
nc
y
Fa
ct
or
s
C
om
bi
na
tio
n
S
oc
ia
liz
at
io
n
fo
r
K
no
w
le
dg
e
D
is
co
ve
ry
S
oc
ia
liz
at
io
n
fo
r
K
no
w
le
dg
e
S
ha
rin
g
E
xc
ha
ng
e
E
xt
er
na
l-
iz
at
io
n
In
te
rn
al
iz
at
io
n
D
ire
ct
io
n
R
ou
tin
es
Ta
sk
u
nc
er
ta
in
ty
=
H
ig
h
N
o
Ye
s
Ye
s
N
o
N
o
N
o
Ye
s
N
o
Ta
sk
in
te
rd
ep
en
de
nc
e
=
H
ig
h
Ye
s
Ye
s
Ye
s
Ye
s
N
o
N
o
O
K
O
K
Ta
ci
t k
no
w
le
dg
e
N
o
Ye
s
Ye
s
N
o
Ye
s
N
o
O
K
O
K
P
ro
ce
du
ra
l k
no
w
le
dg
e
O
K
O
K
O
K
O
K
O
K
O
K
Ye
s
Ye
s
O
rg
an
iz
at
io
na
l s
iz
e
=
S
m
al
l
O
K
Ye
s
Ye
s
N
o
O
K
O
K
Ye
s
N
o
B
us
in
es
s
st
ra
te
gy
=
L
ow
c
os
t
N
o
N
o
O
K
O
K
O
K
O
K
Ye
s
Ye
s
E
nv
iro
nm
en
ta
l u
nc
er
ta
in
ty
=
H
ig
h
Ye
s
Ye
s
N
o
N
o
N
o
N
o
Ye
s
Ye
s
N
um
be
r
of
“
Ye
s”
2
5
4
1
1
0
5
3
N
um
be
r
of
“
O
K
”
2
1
2
2
3
3
2
2
N
um
be
r
of
“
N
o”
3
1
1
4
3
4
0
2
C
um
ul
at
iv
e
pr
io
rit
y
sc
or
e
(“
Ye
s”
=
1
;
“O
K
” =
0
.5
; “
N
o”
=
0
)
3.
0
5.
5
5.
0
2.
0
2.
5
1.
5
6.
0
4.
0
286 CHAPTER 11
in that row is appropriate for the contingency variable in that column, which converts
to a score of 1.0; “No” indicates that KM process in that row is inappropriate for
the contingency variable in that column, which converts to a score of 0.0; and “OK”
indicates that KM process in that row can be used for all possible values of the con-
tingency variable in that column, which converts to a score of 0.5.
The last four columns of the table show the computation of the Cumulative Priority
Score for each KM process, based on the number of “Yes,” “OK,” and “No” responses
for the suitability of that KM process for the seven contingency variables. Based on
this analysis, direction has the highest Cumulative Priority Score (6.0), followed by
socialization for knowledge discovery (5.5) and then socialization for knowledge shar-
ing (5.0). Routines are at an intermediate level of priority with a Cumulative Priority
Score of 4.0 whereas combination, externalization, exchange, and internalization
have low Cumulative Priority Scores (3.0 or less).
Thus, the consideration of the contingency variables led KM-Consult Inc., to
conclude that Doubtfire should focus its KM efforts primarily on direction and social-
ization (for both knowledge discovery and knowledge sharing), with attention being
given to combination and routines if the resources so allow. However, recognizing
the financial difficulties Doubtfire was facing, KM-Consult Inc., recommended that
Doubtfire should focus its efforts on direction and socialization. Moreover, KM-
Consult Inc. had found that the current KM initiative at Doubtfire was making little
use of both socialization and direction. Therefore, KM-Consult Inc. recommended that
Doubtfire should try to enhance the use of direction and socialization for knowledge
management. Their report also identified the specific technologies and systems for
Doubtfire to pursue. It recommended the establishment and use of communities of
practice to support socialization and an expertise locator system to support direction. It
also recommended that Doubtfire should enhance socialization through more frequent
meetings, rituals, brainstorming retreats, and more. The consultants argued that this
socialization would also enhance mutual trust among Doubtfire’s employees, thereby
increasing their willingness to provide and accept direction. Moreover, KM-Consult
Inc., found Doubtfire to be currently making considerable use of internalization,
and spending considerable resources on employee training programs. In the light of
the low cumulative score for internalization, KM-Consult Inc., advised Doubtfire to
consider reducing the budget allocated towards employee training.
SUMMARY
Following our discussion of knowledge management impacts in Chapter 5, we have
described how an organization can seek to enhance these impacts by targeting its
KM solutions according to the circumstances in which KM is being used. In doing
so we have examined the variety of KM processes, systems, mechanisms, as well as
technologies discussed in Chapters 3 and 4, while focusing mainly on the KM pro-
cesses. Table 11.2 summarizes the conclusions regarding the suitability of the KM
processes under various circumstances. A methodology for effectively targeting the
KM solutions has also been described and illustrated using a detailed example. The
next chapter examines how we can evaluate the contributions of KM solutions.
FACTORS INFLUENCING KNOWLEDGE MANAGEMENT 287
REVIEW
1. What is the contingency view of knowledge management? How does it differ
from the universalistic view of knowledge management?
2. What do you understand by the terms task uncertainty and task interde-
pendence?
3. What are the knowledge characteristics that affect the appropriateness of
knowledge management processes? Explain why.
4. How does organizational size affect knowledge management processes?
5. In what way do organizational strategy and environmental uncertainty affect
knowledge management processes?
6. What steps would one take in identifying appropriate knowledge management
solutions? Briefly describe them.
7. Explain how a large organization operating in a highly uncertain environment
should pursue a low-cost business strategy using knowledge management.
State the assumptions made to arrive at your answer.
8. In the seven steps of identifying appropriate KM solutions, Cumulative Priority
Score was computed. Describe the function of the score and its application.
APPLICATION EXERCISES
1. Visit local area companies to study their knowledge management practices.
Determine how they decided on the type of KM solution they use.
2. Consider reasons why an organization would choose the universalistic view
of KM over the contingency view.
3. Visit an organization with a high level of task uncertainty in their business.
Explore the extent to which KM is helping or could help them.
4. Similarly visit an organization with high levels of task interdependence among
the subunits. Explore the ways in which they have implemented KM to the
benefit of the organization.
5. Visit any three organizations and identify their major areas of organizational
knowledge and the prominently used KM processes. Next, classify the charac-
teristics of their organizational knowledge into: explicit or tacit; procedural or
declarative; and general or specific. Based on the data you collect, determine
how appropriate their KM processes are.
6. Collect information from the Internet, Business Week, Fortune, and others on
either Toyota Motor Corporation or Apple Inc. about the nature of these orga-
nizations. Based on this information and the contingency approach presented
in this chapter, identify how knowledge should be managed at this company.
7. You are a KM consultant for BP (www.bp.com). BP is one of the world’s
largest petroleum and petrochemicals groups. Its main activities are explora-
tion and production of crude oil and natural gas; refining, marketing, supply,
and transportation of oil and gas; and selling fuels and related products. Due
to current worldwide financial problems, environmental uncertainty is said
to be relatively high these days.
http://www.bp.com
288 CHAPTER 11
a. Gather information on BP and decide its task uncertainty and task inter-
dependence whether high or low. Provide the reasons for your decision.
b. What types of knowledge does BP use most? Suggest the appropriate KM
process for each of these types of knowledge.
c. Assess (i) the organization size of BP (Small or Large); (ii) Business
strategy (Low cost or Differentiation); and (iii) Environmental uncertainty
(High or Low).
d. Next, compute the Cumulate Priority Score of each KM processes dis-
cussed in this chapter. Based on this analysis, what is your recommendation
to BP of appropriate KM solutions?
NOTE
1. Another popular classification of business strategy focuses on classifying firms into Defenders,
Analyzers, and Prospectors (Miles and Snow 1978). Defenders, Analyzers, and Prospectors have been
found to differ according to the kind of KM efforts that would be most suitable in terms of their effects
on the firm’s stock market performance (Sabherwal and Sabherwal 2007).
REFERENCES
Alavi, M., and Leidner, D. 2001. Knowledge management and knowledge management systems:
Conceptual foundations and research issues. MIS Quarterly, 25(1), 107–136.
Becerra-Fernandez, I., and Sabherwal, R. 2001. Organizational knowledge management processes:
A contingency perspective. Journal of Management Information Systems, 18(1) (Summer),
23–55.
Boh, W.F. 2007. Mechanisms for sharing knowledge in project-based organizations. Information and
Organization, 17, 27–58.
Conner, K.R., and Prahalad, C.K. 1996. A resource-based theory of the firm: Knowledge versus op-
portunism. Organization Science, 7(5), 477–501.
Davenport, T.H., and Prusak, L. 1998. Working knowledge: How organizations manage what they know.
Boston: Harvard Business School Press.
Edmondson, A., Moingeon, B., Dessain, V., and Jensen, D. 2008. Global knowledge management at
Danone. Harvard Business Publishing, 9–608–107, April 16.
Grant, R.M. 1996. Toward a knowledge-based theory of the firm. Strategic Management Journal, 17
(Winter), 109–22.
Gupta, A.K., and Govindarajan, V. 2000. Knowledge management’s social dimension: Lessons from
Nucor steel. Sloan Management Review (Fall), 71–80.
Haas, M.R., and Hansen, M.T. 2005. When using knowledge can hurt performance: The value of
organizational capabilities in a management consulting company. Strategic Management Journal,
26, 1–24.
Hambrick, D.C. 1983. Some tests of the effectiveness and functional attributes of Miles and Snow’s
strategic types. Academy of Management Journal, 26(1), 5–26.
Hsu, I-C., and Wang, Y-S. 2008. A model of intraorganizational knowledge sharing: Development and
initial test. Journal of Global Information Management, 16(3), 45–73.
Jarvenpaa, S.L., and Staples, D.S. 2001. Exploring perceptions of organizational ownership of informa-
tion and expertise. Journal of Management Information Systems, 18(1) (Summer), 151–184.
Kusonaki, K., Nonaka, I., and Nagata, A. 1998. Organizational capabilities in product development
of Japanese firms: A conceptual framework and empirical findings. Organization Science, 9(6),
699–718.
FACTORS INFLUENCING KNOWLEDGE MANAGEMENT 289
Langerak, F., Nijssen, E., Frambach, R., and Gupta, A. 1999. Exploratory results on the importance
of R&D knowledge domains in businesses with different strategies. R&D Management, 29(3),
209–217.
Lawrence, P.R., and Lorsch, J.W. 1967. Organization and environment: Managing differentiation and
integration. Cambridge: Harvard University Press.
Massey, A.P., Montoya-Weiss, M.M., and O’Driscoll, T.M. 2002. Knowledge management in pursuit
of performance: Insights from Nortel Networks. MIS Quarterly, 26(3) (September), 269–289.
Maturana, H., and Varela, F. 1987. The tree of knowledge. Boston: New Science Library.
Miles, R.E., and Snow, C.C. 1978. Organizational strategy, structure, and process. New York: McGraw-
Hill.
Mougin, F., and Benenati, B. 2005. Story-telling at Danone: A Latin approach to knowledge manage-
ment. Report by Sylvie Cherier, trans. Rachel Marlin. Report, École de Paris du management.
Nahapiet, J., and Ghoshal, S. 1998. Social capital, intellectual capital, and organizational advantage.
Academy of Management Review, 23(2), 242–266.
Nonaka, I. 1994. A dynamic theory of organizational knowledge creation. Organization Science, 5
(1), 14–37.
Pisano, G.P. 1994. Knowledge, integration, and the locus of learning: An empirical analysis of process
development. Strategic Management Journal (Winter), 85–100.
Porter, M.E. 1980. Competitive strategy. New York: Free Press.
———. 1985. Competitive advantage. New York: Free Press.
Sabherwal, R., and Sabherwal, S. 2005. Knowledge management using information technology: De-
terminants of short-term impact on firm value. Decision Sciences, 36(4), 531–567.
———. 2007. How do knowledge management announcements affect firm value? A study of firms
pursuing different business strategies. IEEE Transactions on Engineering Management, 54(3),
409–422.
Spender, J.C. 1996. Making knowledge the basis of a dynamic theory of the firm. Strategic Manage-
ment Journal, 17 (Winter), 45–63.
Van de Ven, A., and Delbecq, A. 1974. The effectiveness of nominal, delphi, and interacting group
decision-making processes. Academy of Management Journal, 17, 314–318.
290
12
Leadership and Assessment of
Knowledge Management
In Chapter 3 we discussed knowledge management foundations including infrastructure,
mechanisms, and technologies, and how organizations manage it. In Chapter 4, we ex-
amined how organizations manage KM solutions, including KM systems and processes.
In Chapter 11, we examined how organizations should consider the contingency factors
in selecting KM solutions. We also examined the management of specific KM systems
in Chapters 6, 7, 8, and 9. To complement these chapters and better understand the
overall management of KM in an organization, this chapter examines the leadership of
KM and the ways in which the value of KM can be assessed in an organization.
This chapter begins with a discussion of the overall leadership of KM in an organiza-
tion. Next, it examines the reasons why a KM assessment is needed. It subsequently de-
scribes the alternative approaches to assessing KM in the organization, first for evaluating
various aspects related to KM and then for the overall evaluation of KM effectiveness.
LEADERSHIP OF KNOWLEDGE MANAGEMENT
The Chief Executive Officer (CEO) and the executive board have a direct impact on how
the organization views KM. In order for KM to be practiced across the organization, lead-
ers at the top must endorse and stress the importance of KM programs (DeTienne et al.
2004). The CEO must be involved in the knowledge sharing efforts so that others in the
organization can follow (Kluge et al. 2001). Also, “if KM doesn’t permeate all levels of
an organization, beginning at the top, it is unlikely that KM programs will ever catch on or
be effective” (DeTienne et al. 2004, p. 34). In summary, the role of the CEO is critical to
the success of KM in the organization: first to articulate a “grand theory” of the organiza-
tion’s vision for KM; second to incorporate this vision into the organization’s objectives;
and third to identify which KM initiatives support that strategy (Takeuchi 2001).
The CEO designates the leadership of the knowledge management function to another
senior executive who could be the Chief Knowledge Officer (CKO), Chief Learning Of-
ficer, and in some cases the Chief Information Officer. Whereas the Chief Knowledge Of-
ficer is usually expected to balance social and technical aspects of KM, the Chief Learning
Officer and the Chief Information Officer are generally charged with KM in organizations
where the emphasis is on the social aspects and technical aspects respectively.
Some CEOs might consider adding that responsibility of leading KM to the role of
the Chief Information Officer (CIO). However, this may not be an appropriate decision:
“While some CIOs might have the capabilities for the model CKO—entrepreneur, con-
LEADERSHIP AND ASSESSMENT OF KNOWLEDGE MANAGEMENT 291
sultant, environmentalist, and technologist—most will score high on the technologist and
consultant dimensions but be less accomplished on the entrepreneur and environmentalist
dimensions. And CIOs are oriented toward directing a substantial function, rather than
toward nurturing and leading a transitory team. Most CIOs have demanding enough
agendas without adding the ambiguities of the CKO role” (Earl and Scott 1999, p. 38).
The Chief Learning Officer (CLO) is the “business leader of corporate learning”
(Bersin 2007). At organizations like CIGNA Corporation (Conz 2008), HP (Kiger
2007), PriceWaterhouseCoopers LLP (Cencigh-Albulario 2008), Accenture Ltd (Meis-
ter and Davenport 2005), and many others the CLO is the business executive who leads
the organization’s learning and development strategy, processes, and systems. Thus,
the CLO usually focuses on human resource development and employees’ learning
and training. For example, PriceWaterhouseCoopers hired a Chief Learning Officer
in 2007 to work with its human resource development team and to lead the further
study and training of its personnel (Cencigh-Albulario 2008). CLOs usually focus on
people and on social aspects of KM, although the CLO’s role increasingly involves
utilizing ITs to improve KM, often in collaboration with the CIO.
Organizations that recognize the importance of knowledge management as a criti-
cal function that goes beyond either information management or human resource
development appoint a CKO and charge that individual with the management of the
organization’s intellectual assets and knowledge management processes, systems, and
technologies (Kaplan 2007). Appointing a CKO is “one way of galvanizing, directing,
and coordinating a knowledge management program” (Earl and Scott 1999, p. 37).
A study of twenty CKOs in North America and Europe (Earl and Scott 1999) found
that many of the CKOs were appointed by CEOs more through intuition and instinct
than through analysis or logic, based on the understanding about the increasing im-
portance of knowledge in value creation and the recognition that companies are not
good about managing it. Therefore, CKOs were named with the purpose of correct-
ing a perceived corporate deficiency: lack of formal management of knowledge in
operations, failure to leverage knowledge in business development, inability to learn
from past failures and successes, and not creating value from existing knowledge
assets. This study also found that many CKOs did not have a formal job description
and that their position was perceived as somewhat transient (three to five years), cul-
minating in the expectation that KM would be embedded in all organizational work
processes. Many of these individuals were tasked first with articulating a customized
KM program. This study characterized a model CKO as being both a technologist and
an environmentalist. CKOs are technologists because they invest in IT, and they are
environmentalists because they also create social environments that stimulate conver-
sations and knowledge sharing. In addition the model CKOs are also an entrepreneur,
because they are visionary and starting a new activity; and at the same time they are
consultants, because they match new ideas with managers’ business needs.
Two other studies, one based on a survey of 41 organizations in the United States
and Canada (McKeen and Staples 2003) and the other based on announcements of
23 newly created CKO positions during the period 1995–2003 (Awazu and Desouza
2004), found additional insights regarding the backgrounds, roles, and challenges for
CKOs. They found that CKOs usually possess postgraduate education in business
292 CHAPTER 12
or an allied discipline and include many former academics, mainly professors in the
areas of information and knowledge management (Awazu and Desouza 2004). An
analysis of the background of CKOs revealed that most had a nice blend of techni-
cal and management skills. Many CKOs spent their formative years in areas such
as KM, management consulting, corporate planning, change management, customer
research, marketing, human resource planning, and IT. Organizations were equally
likely to promote from within for the CKO position or make an external hire for the
job. In either case, the average CKO had about ten years of experience in the industry
in which the organization operates (Awazu and Desouza 2004).
CKOs’ budgets and staff are modest, because KM initiatives are typically cor-
porately funded and they may have divisional knowledge managers appointed on
a dotted-line basis (Earl and Scott 1999; McKeen and Staples 2003). But the most
important resource for CKOs is CEO support and sponsorship. The critical success
factors for CKOs to achieve their goal of managing knowledge in organizations are
(Awazu and Desouza 2004; Earl and Scott 1999):
–
edge flows.
The CKO continues to be an important position in contemporary organizations. It
is sometimes combined with other important positions. For example, at Colliers Inter-
national, the president of U.S. Brokerage Services also serves as the Chief Knowledge
Officer and was instrumental in setting up Colliers University, which is the company’s
business development and training division (Business Wire 2009). Similarly, global
management consulting firm Booz & Company announced the position of Chief
Marketing and Knowledge Officer and the Symbio Group in China, which provides
outsourced software development for companies such as IBM, Mercedes-Benz (a
division of Daimler AG), and MasterCard, announced the appointment of their Chief
Knowledge Officer (Knowledge Management Review 2008).
Box 12.1 describes the experience of the Chief Knowledge Officer (CKO) and
Chief Operating Officer (COO) at Atlantis Systems International.
Another important component in the management of an organization’s KM is the man-
agement of knowledge communities of practice within the organization. Two aspects are
important in this realm—the management of the community at the broad level and leader-
ship from within the community (Fallah 2011). Management is important in terms of the
sponsorship of the community to enable the resources, technologies, infrastructure, and
incentives needed for the launch and sustained use of the community, whereas leadership
within the community, which is often distributed, is crucial for the exchange of knowledge
within the community in a cooperative and respectful fashion (Denning 2009).
LEADERSHIP AND ASSESSMENT OF KNOWLEDGE MANAGEMENT 293
Box 12.1
Management of KM at Atlantis Systems International
In an interview in September 2006, Blake Melnick, the chief knowledge officer (CKO) and
chief operating officer (COO) of Atlantis Systems, discussed his role in the company’s
knowledge management efforts and the way in which Atlantis deployed its strategic knowl-
edge advantage to recover following the September 11 terrorist attacks. Melnick stated that
he has “always been an active knowledge builder,” but his formal KM work started in 1995
during graduate studies at the University of Toronto, where he helped to found the Institute
for Knowledge Innovation and Technology. He served as the head of external relations and
workplace research for this institute for five years before joining Atlantis.
After September 11, 2001, the company faced difficult financial times and changed owner-
ship in 2004. The new CEO, Andrew Day, a strong supporter of KM, hired Melnick as CKO
and COO to develop and deploy KM initiatives intended to achieve internal change manage-
ment. Melnick considered measuring and demonstrating the ROI associated with KM to be a
major challenge. Another challenge related to externalizing KM and using it as a discrimina-
tor in Atlantis’ current and targeted markets.
Melnick described an integrated systems approach to KM, Knowledge Exchange (KX),
which has been developed at Atlantis. The KX system integrates facilities for content man-
agement, collaborative discourse, performance management, mentoring, and employee
and customer satisfaction and is the key to KM at Atlantis. The KX system is supported
by several analytic tools that track usage, idea development, knowledge clusters, and so
forth.
Melnick champions several established KM methods at Atlantis:
the company’s direction.
best practices and capture ideas for improvement as they perform their tasks and
activities.
knowledge.
contribute to the company’s knowledge base.
Melnick has learned three lessons about KM that might be valuable to others:
1. KM is not all about technology. Instead, it involves both information management (tech-
nology) and knowledge-building (people).
2. We cannot really “manage knowledge,” but we can manage the process that helps
convert information into knowledge.
3. For KM in any organization, it is essential to address the employees’ primary concern,
such as, “What’s in it for me.”
Melnick believes that since he joined Atlantis, KM at the company has progressed in terms
of a greater appreciation for the “human dynamic” as an essential ingredient to successful
KM implementation. Atlantis became more successful after making the various improve-
ments in KM. The revenues grew by over 200 percent during the three years prior to 2007.
During the same period the number of employees increased from 102 to 210, the retention
rates remained stable at 3 percent, and the company successfully leveraged its knowledge
of the aerospace sector to enter the nuclear energy sector.
Source: Compiled from Knowledge Management Review 2006; Melnick 2006, 2007.
294 CHAPTER 12
IMPORTANCE OF KNOWLEDGE MANAGEMENT ASSESSMENT
In any aspect of organizational or individual task performance, it is imperative to track
whether the efforts are enabling the organization or the individual to achieve the under-
lying objectives. Without such assessment, it would be impossible to determine either
the contribution of those efforts or whether and where improvements are needed. More
specifically, a knowledge management assessment is aimed at evaluating the need for
KM solutions—the knowledge these solutions can help discover, capture, share, or
apply—and the impact they will have on individual or organizational performance. A KM
assessment can help establish the baseline for implementing those KM solutions, includ-
ing the existing infrastructure and technologies that can help support those efforts.
Overall, the assessment of knowledge management is a critical aspect of a KM
implementation; what is not measured can’t be managed well. A survey by Ernst &
Young (1997) indicated that measuring the value and contribution of knowledge as-
sets ranks as the second most important challenge faced by companies, with changing
people’s behavior being the most important. However, only 4 percent of the firms
surveyed by Ernst & Young claimed to be good or excellent at “measuring the value
of knowledge assets and/or impact of knowledge management.” Several reasons attest
to the need for conducting a KM assessment, as described below.
1. A KM assessment helps identify the contributions being currently made by
KM. It helps answer the question: Is KM improving the individual’s or the
organization’s ability to perform various tasks and thereby enhancing effi-
ciency, effectiveness, and/or innovativeness?
2. A KM assessment enhances the understanding of the quality of the efforts being
put into KM as well as the intellectual capital produced through these efforts.
It helps answer the questions: Are the KM solutions being employed adequate
for the needs of the individual or the organization? Do these efforts produce
the intellectual capital required to perform individual or organizational tasks?
3. A KM assessment helps understand whether the costs of the KM efforts are justi-
fied by the benefits they produce. It helps answer the question: Do the direct and
indirect benefits from KM together exceed or equal the various costs incurred?
This is an important benefit for the overall KM solutions as well as the solutions
pursued in a specific KM project. Thus, the overall KM solutions as well as spe-
cific KM projects can be cost-justified through careful KM assessment.
4. A KM assessment helps recognize the gaps that need to be addressed in the
KM efforts by individuals or the organization. It helps answer the question:
What kind of potentially valuable KM solutions do the individual and the
organization currently lack? What potentially important knowledge is not
adequately supported by the KM efforts?
5. Finally, a KM assessment can also help in making a business case to senior
executives in an organization for additional investments in KM efforts. Based
on the benefits currently provided by the organization’s KM solutions (Point 1
above) and the gaps in the organization’s KM efforts (Point 4 above), a business
case can be built for the development of solutions that address these gaps.
LEADERSHIP AND ASSESSMENT OF KNOWLEDGE MANAGEMENT 295
Thus, knowledge management assessments are important because of several rea-
sons, as described above. We next examine the different types of KM assessments
and then examine the alternative KM assessment approaches in some detail.
TYPES OF KNOWLEDGE MANAGEMENT ASSESSMENT
KM assessments can be classified in a number of different ways. Three possible ways
of considering alternative KM assessments are described here. They are related to
the following aspects: (1) When is KM assessed? (2) How is KM assessed? (3) What
aspects of KM are assessed?
THE TIMING OF KM ASSESSMENT
A KM assessment can be performed on different occasions. Three possibilities are
especially noteworthy. First, a KM assessment may be performed periodically for an
entire organization or a subunit. The objective of such an assessment is to evaluate
the overall quality of KM solutions, intellectual capital, and their impacts. This could
help identify any areas that need improvement in KM. Such an assessment can be
performed, for example by surveying employees and inquiring about their degree of
agreement with statements such as in Box 12.2.
KM assessments may also be conducted at the start of a KM project to build a
business case for it. The purpose of such an assessment is to identify the gap in cur-
rent KM at the organization and delineate the potential benefits of the proposed KM
project. For example, for a firm focusing on new products and increasing market
Box 12.2
An Illustrative Tool for Assessing KM
Please indicate your level of agreement with each of the following statements by selecting a
number from 1 (Strongly Disagree) to 5 (Strongly Agree).
1. I am satisfied with the availability of knowledge for my tasks.
2. It is easy for me to locate information I need to perform my job.
3. I always know where to look for information.
4. The available knowledge improves my effectiveness in performing my tasks.
5. My supervisor encourages knowledge sharing within my subunit.
6. The members of my group consistently share their knowledge.
7. I am satisfied with the management of knowledge in my subunit.
8. The available knowledge improves my subunit’s effectiveness.
9. The organization directly rewards employees for sharing their knowledge.
10. The organization publicly recognizes employees who share their knowledge.
Responses to the above statements may be averaged across a number of employees from
various subunits of the organization. Averages for each subunit and the entire organization
would then show where the individual subunits, as well as the overall organization, perform in
terms of the overall quality of KM. Each assessment would be in terms of a number ranging
from 1 (poor KM) to 5 (excellent KM).
296 CHAPTER 12
share, so that R&D represents a major cost center, the business plan might include
the following statement describing the value of the proposed KM project:
The target for the KM project will be to cut cycle time on specific new projects by 20 percent.
In addition, . . . The project will identify cost savings and time savings for scientists in the
unit of 25 percent. (Wilson 2002, p. 17)
The above example illustrates the outcome of a KM assessment conducted at the
start of the project. It indicates there are currently problems in KM within the R&D
function, which is a critical component of the organization, and that addressing these
problems through the proposed project would be highly beneficial.
A KM assessment may also be done following the conclusion of a KM project. Such
assessment aims to determine the impacts of the KM project and may focus on the
entire organization or a specific subunit. It may be necessary to establish historical KM
performance in order to evaluate the effects produced by the KM project. Following are
some of the aspects that can be evaluated during such a post-project assessment:
project.
project.
etc.), either for KM function itself or for the entire organization.
to the organization.
and knowledge management.
For example, a KM project at one large consulting firm caused a major transfor-
mation of the organization. This transformation was significant in both breadth and
depth of impact across the organization. The KM project required line managers to
re-engineer their business processes to draw heavily from the organization’s centralized
knowledge by accessing earlier client presentations, work plans, system specifica-
tions, and other important documents. Consequently, the consulting firm’s “win rate”
in client proposals increased as well (Davenport and Prusak 1998, p. 152). Box 12.3
describes a KM assessment that relies on measuring the effectiveness of one specific
KM mechanism—communities of practice.
THE NATURE OF KM ASSESSMENT
KM assessments are also differentiated on the basis of the way in which KM assess-
ment is done. There are two distinct and important methods to perform KM assess-
ments: qualitative and quantitative.
LEADERSHIP AND ASSESSMENT OF KNOWLEDGE MANAGEMENT 297
Qualitative KM assessments aim to develop a basic understanding of whether the
KM efforts are producing positive results. Qualitative assessments focus on signs, text,
language, and so on instead of focusing on numbers, as is the case with quantitative as-
sessments. Qualitative assessments focus involve such simple tasks as walking around
the halls and buildings of the organization and informally chatting with the employees
about how things are going for them. They also include more formal interviews based
on semistructured or structured interview guides, individually conducted with a care-
fully selected set of employees. Regardless of the formality of these conversations,
they are inherently qualitative, surfacing anecdotes about how well the KM efforts
seem to be working as well as examples of situations where the KM efforts did not
produce the desired results. Such anecdotes of successes (or problems) may concern
the quality of decisions, innovations, and technology transfer at the organizational
level. In addition they may point out issues related to career development, visibility,
confidence, and staying up-to-date technologically at the individual level. Further-
more, such qualitative assessments can be performed at certain periodic intervals,
such as at the start of a project or at the conclusion of a project as discussed before.
Consequently, they may focus on the organization’s overall strategy for KM or on
more specific aspects, such as the development of a KM system like a community of
practice or an expertise locator system.
Quantitative KM assessments, on the other hand, produce specific numerical scores
indicating how well an organization, an organizational subunit, or an individual is per-
forming with respect to KM. Such quantitative assessments may be based on a survey,
such as the one described in Box 12.2. Alternatively, such quantitative KM assessments
may be in financial terms, such as the ROI or the cost savings from a KM project. Fi-
nally, quantitative measures also include such ratios or percentages as employee reten-
tion rate (i.e., the percentage of employees most essential to the organization retained
during the preceding year) or training expenditures as a proportion of payroll (i.e., total
expenditures on training as a percent of the organization’s annual payroll).
Box 12.3
Assessment of KM Through Communities of Practice
1. What was the overall value of this community to you and your team?
2. When your community discussed “topic A,” what specific knowledge, information, or data
did you use?
3. What was the value of this knowledge, information, or data for you as an individual? Can
you express the value in numeric terms such as time saved?
4. Can you estimate the value of this knowledge, information, or data to your business unit in
cost savings, reduced cycle time, improved quality of decision-making, or lower risk?
5. What percentage of this value was obtained directly from the community? What is the likeli-
hood you would have learned it without the community?
6. How confident are you of the above estimate?
7. Who else in your team used this knowledge, information, or data?
Source: Compiled from Wilson 2002.
298 CHAPTER 12
Quantitative measures are more difficult to develop during an organization’s
early experiences with knowledge management. During initial stages, qualitative
assessments should be preferred, with greater use of quantitative measures as the
organization gains experience with knowledge management. This is depicted in
Figure 12.1. However, it is important to note that even when an organization is
very experienced with KM, it can obtain considerable benefits from using qualita-
tive assessment especially in uncertain environments that reduce the benefits from
quantitative measurement.
DIFFERENCES IN THE ASPECTS OF KM ASSESSED
The third way of viewing knowledge management assessments, which is used to
structure the rest of this chapter, focuses on the aspect being assessed. As discussed in
Chapter 5, KM can directly or indirectly impact organizational performance at several
levels: people, processes, products, and the overall organizational performance. These
impacts either come about directly from the KM solutions or from the knowledge
produced and shared through the KM solutions. Therefore, the KM assessment can
focus on: (a) the KM solutions, as discussed in the next section—“Assessment of
KM Solutions”; (b) the knowledge produced or shared through KM solutions, as
discussed in the section entitled “Assessment of Knowledge”; and (c) the impacts of
KM solutions or knowledge on performance (including individuals or employees,
processes, products, and the overall organizational performance), as discussed in the
section entitled “Assessment of Impacts.”
Figure 12.1 Qualitative and Quantitative Assessments of KM
L
ev
el
o
f
U
se
o
f
Q
u
al
it
at
iv
e
an
d
Q
u
an
ti
ta
ti
ve
M
ea
su
re
s
Lo
w
Low High
H
ig
h
Level of Experience with Knowledge Management
Quantitative Measures
Qualitative Measures
LEADERSHIP AND ASSESSMENT OF KNOWLEDGE MANAGEMENT 299
Table 12.1
Illustrative Measures of Key Aspects of KM Solutions
Dimension Illustrative Measures
Knowledge
Discovery organizational subunits
number of employees
Capture
Sharing
lessons learned
Application
are listed in the corporate directory
ASSESSMENT OF KNOWLEDGE MANAGEMENT SOLUTIONS
Assessment of knowledge management solutions involves evaluating the extent to
which knowledge discovery, capture, sharing, and application processes—discussed
in Chapter 3—are utilized and how well they are supported by KM technologies and
systems. Table 12.1 provides some illustrative measures of the four aspects of KM
solutions—discovery, capture, sharing, and application. Although most of the measures
given in this table are easy to quantify, some (e.g., extent of use of learning by doing)
involve perceptions to some extent. Moreover, further research is needed to establish
these measures, but some of them are based on prior empirical research.
Collison and Parcell (2001) describe another way of viewing KM solutions, especially
for knowledge sharing in organizations that focus on organizational subunits and the
key activities they perform. Such organizational activities include increased morale
and motivation; plan, schedule, and work execution; management of spare parts and
stores; and more. Once these activities are identified for the organization, interviews
with managers from each subunit are used to evaluate each subunit’s target performance
as well as actual performance for each activity. This process helps identify, for each
combination of subunit and activity, the gap between actual and target performance.
For each activity, actual as well as target performance for various subunits can then be
placed along a matrix as shown in Figure 12.2. Subunits that show a high level of actual
300 CHAPTER 12
performance and a high level of target performance, such as SU-1 in Figure 12.2, are
the ones that both consider that activity as important and perform it well. These subunits
should be emulated by subunits, such as SU-2, that consider that activity as important but
perform it poorly as shown by a high level of target performance combined with a low
level of actual performance. Therefore, the organization would benefit from knowledge
sharing between these two kinds of subunits (SU-1 and SU-2), which both consider that
activity as important (high level of target performance) but differ in actual performance.
Subunits (such as SU-3) that consider the activity as less important (low level of target
performance) may also benefit from knowledge sharing with subunits that consider that
activity as important in case the focus of their operations changes.
Some specific tools for assessing KM solutions have also been proposed. One example
is “Metrics that Matter”1 from Knowledge Advisors, a Chicago-based company (PR
Newswire 2001b), which provides a comprehensive solution to help training organizations
measure their learning investments. This approach using metrics has three components—
learner-based, manager-based, and analyst-based. Each component helps measure learning
across five levels of evaluation: (1) did they like it? (2) did they learn? (3) did they use it?
(4) what were the results? and (5) what is the return on investment?
ASSESSMENT OF KNOWLEDGE
Assessment of knowledge requires: (a) the identification of the various areas of
knowledge that are relevant to the organization or a specific subunit, followed by (b)
an evaluation of the extent to which knowledge in each of these areas is available.
Figure 12.2 Identifying Knowledge Sharing Opportunities
Target Performance
A
ct
ua
l P
er
fo
rm
an
ce
1
1
2 3 4 5
2
3
4
5 SU-1
SU-2SU-3
Target Performance
A
ct
ua
l P
er
fo
rm
an
ce
1
1
2 3 4 5
2
3
4
5 SU-1
SU-2SU-3
SU-1
SU-2SU-3
LEADERSHIP AND ASSESSMENT OF KNOWLEDGE MANAGEMENT 301
The first of these steps—identification of the relevant areas of knowledge—may be
performed using interviews with managers and other employees of that organization
or subunit. In this step, it may be useful to first identify the critical success factors
for the organization or the subunit. Critical success factors have been defined as “the
limited number of areas in which results, if they are satisfactory, will insure success-
ful performance for the organization” (Rockart 1979, p. 85). Organizations should
therefore give special attention to them, trying to perform exceedingly well in the
few areas they represent rather than seeking to perform a larger number of tasks only
reasonably well. Asking the senior executives to identify six to eight critical success
factors, and then asking them to identify the knowledge needed to succeed with respect
to each critical success factor can thus obtain the most important knowledge areas.
Once the relevant knowledge areas have been identified, the extent and quality of
available knowledge in each area needs to be assessed. This is often a very tricky
issue, for such knowledge can reside in individuals’ minds, corporate databases and
documents, organizational processes, and so on. To some extent, such measurement
of available knowledge may be conducted through surveys or interviews of organi-
zational employees, asking them to evaluate items such as the ones in Box 12.4.
Another important aspect of KM assessment is the value each area of knowledge
contributes to the organization. Assessment of value of knowledge is one way of at-
tributing a tangible measure of benefits resulting from knowledge, which is often
intangible (Sullivan 2000). In general, value has two monetary measures—cost and
price. Price represents the amount a purchaser is willing to pay in exchange for the
utility derived from that knowledge, whereas cost is the amount of money required
to produce that knowledge. Both cost and price are direct, quantitative measures of
value, but there are also other nonmonetary or indirect measures of value, such as the
improvement in the quality of decisions enabled by this knowledge. Some of these
Box 12.4
Assessment of Available Knowledge
On a 5-point scale (ranging from 1 = Strongly disagree to 5 = Strongly agree), I can easily
access knowledge in this area. Statements 1–4 are coded such that a high score indicates
excellent availability of this knowledge, whereas statements 5 and 6 are reverse-coded so that
a low score on these items indicates excellent availability of this knowledge. Therefore, ratings
on items 5 and 6 should be subtracted from 6 and the results can then be averaged with the rat-
ings on items 1 to 4. The resulting average would range from 1 to 5, with 5 indicating excellent
availability of this knowledge.
1. I can easily access knowledge in this area.
2. Everyone in the organization (or the subunit) recognizes the experts in this area of
knowledge.
3. Available knowledge in this area is of a high quality.
4. Available knowledge in this area helps improve the organization’s (or subunit’s) performance.
5. I often have to perform my tasks without being able to access knowledge in this area.
6. The performance of this organization (or subunit) is often adversely affected due to the
lack of knowledge in this area.
302 CHAPTER 12
benefits of knowledge are discussed in the next section. The Intangible Assets Monitor
approach focuses on intangible measures of knowledge, This approach, and its use
to evaluate the value of intellectual capital, is discussed later in the section entitled
“Overall Approaches for KM Assessment.”
ASSESSMENT OF IMPACTS
As we discussed in Chapter 11, KM solutions and the knowledge they help to create,
capture, share, and apply can impact individuals, products, processes, and the overall
performance of organizations. A KM assessment, therefore, involves not only the
evaluation of KM solutions and knowledge but also an evaluation of their impacts.
This section describes how these impacts may be assessed.
ASSESSMENT OF IMPACTS ON EMPLOYEES
KM can impact an organization’s employees by facilitating their learning from each
other, from prior experiences of former employees, and from external sources. KM
can also enable employees to become more flexible by enhancing their awareness of
new ideas, which prepares them to respond to changes and also by making them more
likely to accept change. These impacts, in turn, can cause the employees to feel more
satisfied with their jobs due to the knowledge acquisition and skill enhancement and
their enhanced market value. Thus, KM can enhance learning, adaptability, and job
satisfaction of employees. Some illustrative measures of impacts on each of these
three dimensions are given in Table 12.2.
ASSESSMENT OF IMPACTS ON PROCESSES
KM can improve organizational processes—for example marketing, manufacturing,
accounting, engineering, public relations, and so forth. These improvements can occur
along three major dimensions: effectiveness, efficiency, and degree of innovation of
the processes as discussed in Chapter 4. For example, at HP, a KM system for computer
resellers enhanced efficiency by considerably reducing the number of calls for human
support and enabling the number of people needed to provide this support (Davenport
and Prusak 1998). Table 12.3 lists some illustrative measures of the impacts that KM
and organizational knowledge can have along each of these dimensions.
ASSESSMENT OF IMPACTS ON PRODUCTS
KM can also impact the organization’s products by helping to produce either value-
added products or inherently knowledge-based products. Value-added products are
new or improved products that provide a significant additional value as compared to
earlier products. Inherently knowledge-based products refer for example to products
from the consulting and software development industries. These impacts were dis-
cussed in Chapter 4. Table 12.4 provides some examples of possible measures of the
impacts that knowledge management can have on these two dimensions.
LEADERSHIP AND ASSESSMENT OF KNOWLEDGE MANAGEMENT 303
Table 12.2
Illustrative Measures of Impacts on People
Dimension Illustrative Measures
Employee learning
employee
–
ers within the organization
Employee adaptability
the area in which they currently work) for more than one year
proportion of the total number of countries in which the organization
conducts business
Employee job satisfaction
organization and their jobs
experience filled in the previous year
Table 12.3
Illustrative Measures of Impacts on Organizational Processes
Dimension Illustrative Measures
Efficiency
Effectiveness
Innovativeness
ASSESSMENT OF IMPACTS ON ORGANIZATIONAL PERFORMANCE
KM can impact overall organizational performance either directly or indirectly.
Direct impacts concern revenues and/or costs, and can be explicitly linked to the or-
ganization’s vision or strategy. Consequently direct impact can be observed in terms
of increased sales, decreased costs, and higher profitability or return on investment.
For example, Texas Instruments Inc. generated revenues by licensing patents and
intellectual property (Davenport and Prusak 1998). However, it is harder to attribute
304 CHAPTER 12
revenue increases to KM than cost savings (Davenport et al. 2001). Indirect impacts
on organizational performance come about through activities that are not linked to
the organization’s vision, strategy, or revenues and cannot be associated with trans-
actions. As discussed in Chapter 5, indirect impacts include economies of scale and
scope, and sustainable competitive advantage. Table 12.5 provides some examples
of possible measures of these direct and indirect impacts that knowledge management
can have on overall organizational performance.
The value of a KM investment should be evaluated based on how it affects dis-
counted cash flow. Improved problem-solving, enhanced creativity, better relation-
ships with customers, and employee’s more meaningful work can all eventually be
linked to real cash flows. Therefore, organizations can enhance their cash flow in the
following ways:
organization.
regulation.
(Clare 2002; Wilson 2002)
It is important to keep the above drivers in mind during the implementation of
knowledge management projects. In other words, if KM initiatives are observed to
help increase the company’s cash flow, executives will listen and therefore find a
viable way to fund them.
CONCLUSIONS ABOUT KNOWLEDGE MANAGEMENT ASSESSMENT
We have examined and provided illustrative measures for KM assessments. We also
discussed the direct and indirect impacts that KM assessments can have on the overall
Table 12.4
Illustrative Measures of Impacts on Organizational Products
Dimension Illustrative Measures
Value-added Products
–
ucts offered by the organizations
Knowledge-based Products
organization places on the Internet
LEADERSHIP AND ASSESSMENT OF KNOWLEDGE MANAGEMENT 305
organizational performance. In this section, we examine and provide a broader dis-
cussion of KM assessment including a discussion of who performs KM assessment,
some overall approaches for KM assessment, the approach for the implementation
of a KM assessment, and some caveats regarding KM assessments.
WHO PERFORMS KM ASSESSMENT?
In order to perform a KM assessment, it is helpful to form a team that includes internal
and external members. The internal members provide the necessary context and help
retain within the organization the knowledge acquired from the assessment, whereas
the external members can help identify KM-related assumptions and opportunities that
may be missed by internal members. Overall, a KM assessment should incorporate:
(a) peer review of internal performance; (b) external appraisal (by customers, suppli-
ers, etc.) of the organization and its outputs; (c) business evaluation of effectiveness,
efficiency, and innovativeness; and (d) evaluation of the knowledge assets created
(Quinn et al. 1996).
The following example illustrates how these perspectives could be included and
effectively integrated. Following each project a major investment banking firm asks
all team members, the team leader, and its customer group to rank all project partici-
pants in terms of their exhibited knowledge, specific contributions to the project, and
support for the team. Customers also rate their overall satisfaction with the firm as
Table 12.5
Illustrative Measures of Impacts on Organizational Performance
Illustrative Measures
Direct Impacts
previous year
–
ous year
Indirect Impacts –
tion) change in total cost per unit sold as compared to the previous year
–
tion) change in the number of different products a salesperson can sell as
compared to the previous year
–
tion) of the difference between the price of the organization’s product and
the mean price of competing products
products produced in the organization’s manufacturing plants and the av-
erage number of different products produced in the manufacturing plants
of its main competitors
organization and its key competitors
have been buying the organization’s products/services
–
tracts in the previous year
306 CHAPTER 12
well as with the specific project. Annual surveys, ranking the firm against competitors
on 28 key dimensions, complement these evaluations. The firm also measures costs
and profits for each project and allocates them among participating groups based on
a simple, pre-established formula. Annually, for each division the firm computes the
net differential between its market value (if sold) and its fixed asset base. This net
intellectual value of the division is tracked over time as an aggregate measure of how
well the division’s management is building its intellectual assets.
OVERALL APPROACHES FOR KM ASSESSMENT
In the preceding sections we have discussed a number of measures that can be used for
KM assessment. Overall KM assessment approaches usually combine several of these
measures, as illustrated above with the investment banking firm’s example. One such
approach involves the use of benchmarking, or comparing KM at an organization or
subunit with other organizations or subunits. Adopted as a systematic technique for
evaluating a company’s performance in reaching its strategic goals, benchmarking
is based on the recognition that best practices are often the same within a company
or even within an industry. Benchmark targets could therefore include other units
within the same company, competing firms, the entire industry, or in some cases, suc-
cessful companies in other industries. For example, a leading manufacturer identifies
outstanding operating units, formally studies them, and then replicates their practices
throughout the rest of the company. This approach produced sales that exceed goals
by five percent (PR Newswire 2001a). Box 12.5 provides information on a cross-
industry survey that may be used as a benchmark in the arena of KM.
Another overall approach for KM assessment utilizes the Balanced Scorecard,
which was originally developed by Kaplan and Norton (1996) to provide a more “bal-
anced view” of internal performance rather than for KM assessment. The Balanced
Scorecard provides a way of maintaining a balance between short-term and long-term
objectives, financial and nonfinancial measures, lagging and leading indicators, and
external and internal perspectives. It examines the goals, metrics, targets, and initia-
tives for the following four different perspectives (Tiwana 2002):
1. The Customer Perspective: How should our customers perceive us?
2. The Financial Perspective: What is the face that we want to present to our
shareholders?
3. The Internal Business Perspective: Are our internal operations efficient and
effective and performing at their best?
4. The Learning and Growth Perspective: How can we sustain our competitive
advantage over time?
In employing the Balanced Scorecard for KM assessment, the above four perspec-
tives are used in a series of four steps performed over time. The first step involves
translating the KM vision (i.e., Why are we managing knowledge, and what is our
vision for KM?). In the second step, this vision is communicated within the organiza-
tion with rewards linked to knowledge use and contribution. The third step involves
LEADERSHIP AND ASSESSMENT OF KNOWLEDGE MANAGEMENT 307
business planning, including the establishment of goals and the alignment of metrics
and rewards to them. The fourth step—learning and feedback on whether KM is
working and whether it can be improved—then feeds back to the first step to begin
the cycle again. The above four complementary criteria from the Balanced Scorecard
are used during each of these steps.
Like the Balanced Scorecard, the Intangible Assets Monitor Framework (Sveiby
2000) also recognizes the importance of examining intangible knowledge assets rather
than focusing only on financial or monetary assets. The Intangible Assets Monitor
considers a firm’s market value to depend on tangible net book value and intangible
assets which include external structure (including relationships with customers and
suppliers, brand names, trademarks, and image or reputation); internal structure
(including the patents, concepts, models, and systems); and the competence of the
organization’s individual employees (including skills, education, experience, values,
and social skills). Based on these factors, WM-Data, a Swedish computer software and
consulting company,2 designed a set of nonmonetary indicators that top management
uses to supervise their operations on a weekly, monthly, and annual basis. The Intan-
gible Assets Monitor Framework (Sveiby 1998) evaluates growth, renewal, efficiency,
and stability for tangible assets (financial value), external structure (customer value),
internal structure (organizational value), and individuals’ competence (individual
value). Following are some of the questions that may be used to evaluate growth:
1. Is the existing customer base growing in value?
2. Are the support staff and administrative management improving their com-
petence?
3. Are our tools and processes growing in value?
The Skandia Navigator method is another approach to KM assessment that gives
considerable attention to intangible assets (Edvinsson and Malone 1997). A Swedish
Box 12.5
The Most Admired Knowledge Enterprises Survey
The Annual Most Admired Knowledge Enterprises (MAKE) survey by Teleos, an independent
KM research company and leaders of the KNOW network, is based on a ranking of firms by a
panel of CKOs and leading KM practitioners along eight criteria (Teleos 2009):
1. Ability to create and sustain an enterprise knowledge-driven culture.
2. Ability to develop knowledge workers through senior management leadership.
3. Ability to develop and deliver knowledge-based products/services/solutions.
4. Ability to manage and maximize the value of enterprise intellectual capital.
5. Ability to create and sustain an enterprise-wide, collaborative knowledge sharing
environment.
6. Ability to create and sustain a learning organization.
7. Ability to manage customer knowledge to create value and enterprise intellectual capital.
8. Ability to transform enterprise knowledge into shareholder value (or societal value for non-
profits and the public sector).
308 CHAPTER 12
company, Skandia Insurance Company Ltd., developed this method in 1993 under
the leadership of Leif Edvinsson, although it preferred using the term intellectual
capital rather than knowledge. The Skandia Navigator included a number of ratios
in which it looks at the past, present, and future. In the Skandia Navigator approach,
the past is examined with an emphasis on financial aspects, the present is examined
by focusing on customers, people, and processes, and the future is examined in terms
of renewal and development.3
KM assessment can also benefit from the real options approach, which views KM
initiatives as a portfolio of investments (Tiwana 2002). This approach focuses on the
value-to-cost ratio—that is, the ratio of the net value to the total cost for each invest-
ment and the volatility faced by each investment. Using this approach, KM projects
can be placed on an option space as shown in Figure 12.3. A clockwise move from
region 1 to region 3 in the option space implies a shift from projects that are low-risk
and attractive to projects that are fairly attractive. Continuing further to region 6, the
projects reduce further in attractiveness. Thus, projects A and B in the figure are at-
tractive, and projects C and D are not, with project A being the most attractive and
project D being the least attractive. Such real options analysis combines strategic and
financial approaches to evaluating investments. In positioning projects on the option
space, it can benefit from the techniques discussed earlier in the chapter, especially
for identifying the value-to-cost ratio. To conclude this section on overall KM assess-
ment, Box 12.6 provides a summary of KM assessment at Siemens AG.
Figure 12.3 KM Projects Mapped on the Option Space
Value-to-cost Ratio
Less than 1.0 More than 1.01.0
6. Definite failure, should be avoided 1. Low risk, should be pursued
2. Semi-attractive
3. Almost attractive
5. Semi-unattractive
4. Almost unattractive
Project A
Project B
Project C
Project D
Value-to-cost Ratio
Less than 1.0 More than 1.01.0
6. Definite failure, should be avoided 1. Low risk, should be pursued
2. Semi-attractive
3. Almost attractive
5. Semi-unattractive
4. Almost unattractive
Project A
Project B
Project C
Project D
LEADERSHIP AND ASSESSMENT OF KNOWLEDGE MANAGEMENT 309
Finally, another overall approach for KM assessment is to evaluate the organization
in terms of the maturity of its KM. For example, Yokell (2010) defined five stages of
maturity: (1) initiation, where awareness of KM is growing within the organization;
(2) development, where the organization introduces localized and repeatable prac-
tices; (3) standardization, where the organization moves toward common processes
and approaches; (4) optimization, where the organization measures the consequences
of KM and accordingly adapts its KM approach; and (5) innovation, where the
organization continually improves its KM practices. Using archival data from the
American Productivity and Quality Center (APQC), Yokell found firms that were
higher in KM maturity to significantly outperform those with lower KM maturity in
terms of return on sales as well as return on assets. The experience of another large
Box 12.6
KM Assessment at Siemens
Siemens, a large firm in the electronics industry, has benefited considerably over the years from
one of its knowledge sharing systems. Called “ShareNet,” this system serves as the founda-
tion for several communities of practice. To estimate ROI, Siemens computes the costs of a
community of practice including labor, meetings, facilities, and the effort spent by KM experts. It
also considers costs of the incentive program. Siemens then decides how much effort has been
saved through the sharing of solutions in the community.
Siemens also considers subcommunities and their generation of solutions in terms of com-
munity projects. If a group needs a solution and embarks on a knowledge-creation effort, it
can determine the savings in time-to-market, competitive positioning, and so forth. To further
determine the value of KM, Siemens has developed a master plan of KM metrics that contains
measures for each of four dimensions of its holistic KM system:
Knowledge community: the organization, community, and people dimensions
Knowledge marketplace: the IT involved in knowledge management
Key KM processes: sharing and creation
Knowledge environment: all of the above
Siemens has realized that it can assess the success of its communities and marketplaces
with measures such as how much knowledge comes in or out of the community and the quality
of feedback. Contracts that had been gained with the support of other divisions, or savings ob-
tained through knowledge shared using the knowledge communities, were also included in the
benefits. A contribution key, determined through a survey form that the ShareNet Managers fill
out, indicated the proportion that ShareNet had contributed to the success of each initiative.
Siemens believes communities are the heart of their KM systems, and it has spent a great
deal of time on communities-of-practice assessments—questionnaires for community members
that provide ideas on how to improve each of the communities. Siemens has tried to check the
health of KM processes to determine the performance of the sharing process. Ideally, the mea-
sures evaluate whether a person has managed the process correctly and set the right limits on
it. This provides Siemens a good way to look at the marketplace and also to examine how much
sharing and creation is taking place.
To monitor the entire KM systems, Siemens performs a KM maturity assessment that
defines whether KM is still ad hoc and chaotic or has progressed to an optimized state. To
do this, Siemens measures its four dimensions and 16 enablers, each of which has a set of
questions.
Source: Compiled from Davenport and Probst 2002; MacCormack et al. 2002; Voelpel et al. 2005.
310 CHAPTER 12
company—Shell—in computing the return on investment for its expenditure in KM
communities of practice is described in Box 12.7.
FURTHER RECOMMENDATIONS FOR KM ASSESSMENT
So far in this chapter we have described a number of KM metrics and assessment
approaches. In developing these measures and approaches, the following eight sug-
gestions should be carefully considered (Tiwana 2002; Wilson 2002).
1. Remember why you are doing KM: When proposing a KM project, it is critical
to define its measures of success based on things the organization cares about,
such as: reducing waste, lowering costs, enhancing the customer experience,
and so forth.
2. Establish a baseline: It is important to identify and develop a baseline measure
when you begin efforts, rather than scrambling after the effort is completed
to try to determine measures of success. Establishing a baseline is essential
to prove successful results down the line.
3. Consider qualitative methods: KM is a qualitative concept and qualitative
methods of measurement, such as analyzing the value of social networks,
telling success stories, and others, should not be ignored.
Box 12.7
Evaluating Returns on Knowledge Management at Shell
Oil exploration often involves extrapolating from sketchy data and comparing new exploration
sites to sites that are already known. This allows geoscientists to decide if enough reserves
exist on a site to make developing it worthwhile. For example, one site contained layers of oil-
bearing sand that were less than an inch think. A Shell exploration team needed to decide if thin
sand beds could extend over a large enough area for the oil in them to be efficiently pumped
out. This would normally require drilling and testing a number of exploratory wells. The team
asked one of Shell’s communities of practice, including geoscientists from several disciplines,
for help. By comparing this site to others, the community helped in the team’s analysis of where
to drill more accurately, resulting in fewer exploratory wells.
Community members estimated that the discussions of such comparisons enabled
them to drill and test three fewer wells a year, saving US$20M in drilling and an additional
US$20M in testing costs for each well, for an annual savings of US$120M. It is possible that
they might have reached the same conclusions on where to drill, but the leader estimated
that the community could claim 25 percent of the savings and was 80 percent sure of this
estimate. So the community may have saved 25 percent of 80 percent of US$120M, or
US$24M annually. Since it cost between US$300K and US$400K annually to run the com-
munity, this represented an annual return of 40 times the investment. This was not the only
benefit, but it was sufficient to address the senior executives’ need to know whether the
community was worth the investment. Overall, Shell International Exploration and Produc-
tion estimated that its use of KM resulted in more than $200 million in reduced costs and
new income in 2000 (King 2001).
Source: Compiled from King 2001; Wilson 2002.
LEADERSHIP AND ASSESSMENT OF KNOWLEDGE MANAGEMENT 311
4. Keep it simple: An organization does not need hundreds of measures. A handful
of the relevant, robust, and easily assessable ones are better in demonstrating
to yourself and your organization that KM is indeed adding value.
5. Avoid KM metrics that are hard to control: KM assessment should use met-
rics that are specific and within the control of the organization’s employees.
Broad and grand statements such as “enable the firm to become a one of the
top learning organizations in the World by 2010” are visionary but impossible
to control or measure.
6. Measure at the appropriate level: Measure at the project or application level
in the beginning. Organizations that have implemented KM initiatives for a
long time can then try to measure the total organizational value of KM or
their program.
7. Link rewards to KM assessment results: KM assessment should not be an
end in itself. Instead, it’s the results of KM assessments that should be used
to provide rewards and incentives thereby motivating improved KM results
in the future.
8. Be conservative in your claims: When calculating a figure like ROI for KM
projects, it is better to err on the higher side when estimating costs and to err
on the lower side when estimating value in order to make the results more
believable to management.
SUMMARY
This chapter complements our earlier discussion of management of KM foundations
and KM solutions in Chapters 3 and 4, respectively; the effects of how contingency
factors on KM solutions; and the management of specific KM systems in Chapters
6, 7, 8, and 9. To better understand the overall management of knowledge manage-
ment in an organization, we have examined in this chapter the leadership of KM
and the ways in which the value of KM can be assessed in an organization. We have
discussed the assessment of KM systems and the impacts that assessments can have.
We have also summarized some overall KM assessment approaches and how a KM
assessment can be implemented.
REVIEW
1. Distinguish between the roles of Chief Knowledge Officer and Chief Learning
Officer.
2. Why is it important to perform KM assessment? Identify and discuss any
three reasons.
3. Describe the different types of KM assessment in terms of (a) the timing of
KM assessment; and (b) the aspects assessed.
4. What are the differences between quantitative and qualitative assessments of
KM assessment? How does their use depend upon the organization’s experi-
ence with KM?
5. Briefly describe some financial measures that can be used for KM assessment.
312 CHAPTER 12
6. Briefly describe some nonfinancial measures that can be used for KM
assessment.
7. Briefly discuss how the different impacts of KM on employees can be
assessed.
8. How can the impacts of KM on efficiency, effectiveness, and innovation be
evaluated?
9. What is KM maturity? Identify the various stages from one KM maturity
model.
10. How do the measures of the direct impacts of KM differ from the measures
of its indirect impacts?
APPLICATION EXERCISES
1. Visit a local area firm to study its KM assessment process. Determine how
they decided on the type of KM solution they use.
2. How would you conduct KM assessment at the firm you visited? Describe
the suggested approach in some detail, making sure to connect this approach
to the approaches described in this chapter.
3. Study how knowledge is managed at either your family physician’s office or
your dentist’s office through 15-minute conversations with a few individuals
who work at that office. Then recommend an approach for assessing KM at
this office. Discuss the suggested approach with some senior employees (e.g.,
the family physician or the dentist) at this office, and seek their feedback
concerning your suggestions.
4. Visit any three organizations of varying sizes and different industries. Identify
who leads the KM function at each organization, and examine how these or-
ganizations perform their KM assessments. Compare the three organizations
in terms of whether they use a Chief Knowledge Officer, a Chief Learning
Officer, a Chief Information Officer, or an individual in some other position
to lead the KM function. Discuss why the organizations might differ or be
similar with respect to the leaders of their KM function.
5. For each of the organizations you visited in Question 4 above, examine how
consistent the organization’s KM assessment approach is with the recom-
mendations in this chapter. Which organization seems most consistent with
the recommended approach? Of the three organizations, is this organization
the one that has the most experience with KM?
NOTES
1. More details about this approach may be found at www.knowledgeadvisors.com/metrics-that-
matter/.
2. Since October 10, 2006, WM-Data has become a subsidiary to LogicaCMG (www.logica.com/).
3. Edvinsson left Skandia in 1999. Although intellectual capital remained an important focus of
Skandia’s corporate philosophy until early 2000, Skandia has undergone considerable changes as a
company, with a merger with Storebrand in 1999, followed by its acquisition by Old Mutual in 2006.
See Value Based Management.net 2013 and Skandia 2014.
www.knowledgeadvisors.com/metrics-thatmatter/
www.knowledgeadvisors.com/metrics-thatmatter/
www.logica.com/
LEADERSHIP AND ASSESSMENT OF KNOWLEDGE MANAGEMENT 313
REFERENCES
Awazu, Y., and Desouza, K. 2004. The knowledge chiefs: CKOs, CLOs and CPOs. European Manage-
ment Journal, 22(3), 339–344.
Bersin, J., 2007. The new Chief Learning Officer: 2008 and beyond. Bersin and Associates, July 21.
http://joshbersin.com/2007/07/21/the-new-chief-learning-officer-2008-and-beyond/ (accessed
February 17, 2009).
Business Wire. 2009. Colliers International appoints top executives to its USA management team.
Business Wire, February 2.
Cencigh-Albulario, L. 2008. Now everyone is joining the rush to learn. Weekend Australian, November
22, 11.
Clare, M. 2002. Solving the knowledge-value equation: Part one. Knowledge Management Review
(May/June).
Collison, C., and Parcell, G. 2001. Learning to fly. Milford, CT: Capstone Publishing.
Conz, N. 2008. CIGNA has launched a series of public-facing e-learning modules that leverage Web
technologies to teach consumers about the basics of the health insurance system. Insurance and
Technology, December 1, 13.
Davenport, T.H., Harris, J.G., De Long, D.W., and Jacobson, A.L. 2001. Data to knowledge to results:
Building an analytic capability. California Management Review, 43(2), 117–138.
Davenport, T.H., and Probst, G. (Eds.). 2002. Knowledge management case book: Siemens best practises
(2d ed.). New York: John Wiley & Sons.
Davenport, T.H., and Prusak, L. 1998. Working knowledge: How organizations manage what they know.
Boston: Harvard Business School Press.
Denning, S. 2009. Communities for knowledge management. www.stevedenning.com/Knowledge-
Management/communities-of-practice.aspx (accessed January 31, 2014).
DeTienne, K., Dyer, G., Hoopes, C., and Harris, S. 2004. Toward a model of effective knowledge man-
agement and directions for future research: Culture, leadership and CKOs. Journal of Leadership
and Organizational Studies, 10(4), 26–43.
Earl, M., and Scott, I. 1999. What is a chief knowledge officer? Sloan Management Review (Winter),
29–38.
Edvinsson, L., and Malone, M.S. 1997. Intellectual capital: The proven way to establish your company’s
real value by measuring its hidden values. London: Piatkus.
Ernst & Young. 1997. Executive perspectives on knowledge in the organization. Cambridge, MA: The
Ernst & Young Center.
Fallah, N. 2011. Distributed form of leadership in communities of practice (CoPs). International Journal
of Emerging Science, 1(3) (September), 357–370. http://ijes.info/1/3/42541310 .
Kaplan, B. 2007. Creating long-term value as chief knowledge officer. Knowledge Management Review,
September/October, 30–33.
Kaplan, R.S., and Norton, D.P. 1996. The balanced scorecard. Boston: Harvard Business School
Press.
Kiger, P.J. 2007. Precision-targeted development at HP. Workforce Management, June 25, 38.
Kluge, J., Shein, W., and Licht, T. 2001. Knowledge unplugged: The McKinsey and Company Global
Survey on knowledge management. New York: Palgrave Macmillan.
Knowledge Management Review. 2006. Using knowledge as a competitive differentiator at Atlantis
Systems. Knowledge Management Review, 9(4), 4.
———. 2008. Recent appointments demonstrate KM’s continued influence. Knowledge Management
Review, 11(5) (November/December), 4.
MacCormack, A., Volpel, S., and Herman, K. 2002. Siemens ShareNet: Building a knowledge network.
Boston: Harvard Business School Publishing, Case 9–603–036, November 5.
McKeen, J.D., and Staples, D.S. 2003. Knowledge managers: Who they are and what they do. In Hand-
book on knowledge management, Vol. 1, ed. C.W. Holsapple, 21–41. Berlin: Springer.
http://joshbersin.com/2007/07/21/the-new-chief-learning-officer-2008-and-beyond/
http://ijes.info/1/3/42541310
www.stevedenning.com/Knowledge-Management/communities-of-practice.aspx
www.stevedenning.com/Knowledge-Management/communities-of-practice.aspx
314 CHAPTER 12
Meister, D., and Davenport, T. 2005. Knowledge management at Accenture. Ivey Management Services,
University of Western Ontario.
Melnick, B., 2006. Using knowledge as a competitive differentiator at Atlantis Systems. Knowledge
Management Review, 9(5), 8–13.
———. 2007. Case study: Atlantis System International—Using KM principles to drive productivity
and performance, prevent critical knowledge loss and encourage innovation. International Atomic
Energy Agency. http://inis.iaea.org/search/search.aspx?orig_q=RN:38067992 (accessed August 3,
2014).
PR Newswire. 2001a. Knowledge management: Best practice sharing overcomes barriers. PR News-
wire, November 1.
———. 2001b. Microsoft teams with knowledge advisors to measure training investments. PR News-
wire, November 27.
Quinn, J.B., Anderson, P., and Finkelstein, S. 1996. Leveraging intellect. Academy of Management
Executive, 10(3), 7–27.
Rockart, J.H. 1979. Chief executives define their own data needs. Harvard Business Review, 57(2),
81–92.
Skandia. 2014. About Skandia: Our history speaks for itself. www2.skandia.co.uk/Adviser/About-
Skandia/Our-History/.
Sullivan, P.H. 2000. Value-driven intellectual capital. New York: John Wiley & Sons.
Sveiby, K.-E. 1998. The intangible assets monitor framework. In Measuring Intangibles and Intellctual
Capital—An Emerging First Standard. Sveiby Knowledge Associates, August 5. www.sveiby.com/
articles/EmergingStandard.html#TheIntang.
———. 2000. Measuring intangibles and intellectual capital. In Knowledge management: classic and
contemporary works, ed. D. Morey, M. Maybury, and B. Thuraisingham, 337–353. Cambridge,
MA: The MIT Press.
Takeuchi, H. 2001. Toward a universal management concept of knowledge. In Managing industrial
knowledge, ed. I. Nonaka and D. Teece, 314–329. London: Sage Publications.
Teleos. 2009. Most admired knowledge enterprises. The KNOW Network. www.knowledgebusiness.
com/knowledgebusiness/Screens/MakeSurvey.aspx?siteId=1&menuItemId=43 (accessed February
18, 2009).
Tiwana, A. 2002. The knowledge management toolkit: Orchestrating IT, strategy, and knowledge plat-
forms. Upper Saddle River, NJ: Prentice Hall.
Value based management.net. 2013. Measuring intangible assets: Summary of Skandia Navigator.
www.valuebasedmanagement.net/methods_skandianavigator.html.
Voelpel, S.C., Dous, M., and Davenport, T.H. 2005. Five steps to creating a global knowledge-sharing
system: Siemens’ ShareNet. Academy of Management Executive, 19(2) (May), 9–23.
Wilson, J. (Ed.). 2002. Knowledge management review: The practitioner’s guide to knowledge manage-
ment. Chicago: Melcrum Publishing.
Yokell, M.R. 2010. A quantitative correlational study of the relationship between knowledge manage-
ment maturity and firm performance. Unpublished doctoral dissertation, University of Phoenix.
ProQuest, UMI Dissertations Publishing. http://search.proquest.com/docview/878892492.
http://inis.iaea.org/search/search.aspx?orig_q=RN:38067992
http://search.proquest.com/docview/878892492
www2.skandia.co.uk/Adviser/About-Skandia/Our-History/
www2.skandia.co.uk/Adviser/About-Skandia/Our-History/
www.sveiby.com/articles/EmergingStandard.html#TheIntang
www.sveiby.com/articles/EmergingStandard.html#TheIntang
www.knowledgebusiness.com/knowledgebusiness/Screens/MakeSurvey.aspx?siteId=1&menuItemId=43
www.knowledgebusiness.com/knowledgebusiness/Screens/MakeSurvey.aspx?siteId=1&menuItemId=43
www.valuebasedmanagement.net/methods_skandianavigator.html
315
13
The Future of Knowledge
Management
As we have seen throughout this book, knowledge management goals are for the mem-
bers of an organization to discover, capture, share, and apply knowledge. However,
the nature of KM is undergoing a dramatic change partly as a result of the emergence
of Web 2.0 technologies, as discussed in Chapter 10. In this concluding chapter, we
identify and discuss five critical issues for the future of KM. These five issues are
discussed in the next five sections followed by some concluding remarks.
USING KNOWLEDGE MANAGEMENT AS A DECISION-MAKING PARADIGM
TO ADDRESS WICKED PROBLEMS
The development of management information systems, decision support systems, and
KM systems has been influenced by the works of five influential philosophers, namely
Leibniz, Locke, Kant, Hegel, and Singer (Churchman 1971). Based on Churchman’s
definition of inquiring organizations,1 a new paradigm for decision-making in today’s
complex organizational contexts has been developed (Courtney 2001).
In the conventional decision-making process, the emphasis is first on recognizing the
problem, then on defining it in terms of a model. Alternative solutions are then analyzed,
and the best solution is selected and implemented. Thus, KM systems have successfully
supported solving semistructured problems, those characterized by a limited number of
factors and a certain future. Recent developments in KM have helped extend the reach of
those involved in the solution. But the jury is still out on how well KM systems support
problems that are characterized as wicked (Rittel and Webber 1973). Wicked problems
are unique and difficult to formulate. Their solutions are good or bad (rather than true
or false) and generate waves of consequence over time. Solutions to wicked problems
are accomplished in one-shot occurrences, and so there is no opportunity to learn from
prior mistakes and solutions cannot be undone. Moreover, solutions to wicked problems
are not a numerable set of solutions, and many may have no solutions.
For example, a project plan for an enterprise resource planning (ERP) system
implementation is a wicked problem. ERP systems’ implementations are one-shot
occurrences, in the sense that organizations will typically only implement them once.
Therefore, there’s no opportunity to learn over time how to successfully implement
these systems. Usually organizations only find out if their implementation was “good”
or “bad” on the deployment or “go-live” date, and at this point “bad” implementations
result in disastrous economic consequences for the organization.
316 CHAPTER 13
The fact is that as globalization expands, the number of stakeholders affected by
the organization increases, each one affected by different customs, laws, behaviors,
and environmental concerns. Globalization also leads to wicked planning problems
for organizations, and methods to help make decisions in such situations are greatly
needed. The new paradigm for KM support, suggested by Courtney (2001), defines the
decision-making process as starting with the recognition that the problem exists, but
then rather than proceeding immediately into analysis the process consists of develop-
ing multiple perspectives. These multiple perspectives consider the following:
1. Technical perspective: Consists of analyzing the alternatives and implementing
the chosen alternative. This perspective is the only one relevant to existing
decision support and KM systems.
2. Personal and individual perspective: Complex problems involve a multiplicity
of actors. Each sees the problem differently and generates a different perspective
based on individual experiences, intuition, personality, and attitudes about risk.
3. Organizational and social perspective: Complex problems involve various
organizations. Organizations also each view the problem in a different fashion,
and thus generate a different perspective. Organizations may also consist of
diverse members with different interests.
4. Ethics and aesthetics perspective: Complex problems involve business ethics and
aesthetic issues that are so high that they require the involvement of key stake-
holders since there are no simple solutions. Perhaps the utilitarian emphasis of the
Industrial Age neglected the spirituality of the “rational man” and contributed to
the demise of ethics and aesthetics in decision-making today (Courtney 2001).
This new paradigm bases decisions on the use of these multiple perspectives. The
prior view of decision-making environments minimizes the importance of relation-
ships, collaboration, and trust in the organization. Personal relationships define orga-
nizational boundaries to a large extent. The future calls for the development of KM
systems that support the human aspects of decisions: the personal, organizational,
ethical, and aesthetic perspective. Thus, KM systems should help decision makers
make more humane decisions and enable them to deal with wicked problems.
Further work is needed on how KM can be used to address problems that are not only
wicked but also critical to humanity. Two of the biggest threats currently being encountered
are the staggering global financial crisis and the escalating rise in terrorism. Could the kind
of Web 2.0 technologies that have been successfully used in such diverse tasks as enabling
common people to design T-shirts (online), creating an entire encyclopedia of knowledge
and building grassroots support in the U.S. presidential elections, play a role in generating
ideas and taking actions needed to address such wicked and crucial problems?
PROMOTING KNOWLEDGE SHARING WHILE PROTECTING
INTELLECTUAL PROPERTY
Knowledge sharing could also bring forth certain risks, namely that the knowledge
falls into the wrong hands either maliciously or accidentally. The same communica-
THE FUTURE OF KNOWLEDGE MANAGEMENT 317
tion technologies that support the sharing of knowledge within an organization also
enable the knowledge to leak outside the organization to its competing firms. Given
the value of the knowledge, and the reliance that an organization places on this
knowledge, losing this knowledge could have severe negative consequences for the
organization. It is therefore critical for organizations to manage knowledge such that
knowledge sharing is enhanced but knowledge leakage is controlled. This is not an
easy balance to achieve. Below we discuss some of the ways in which knowledge
leakage can be controlled.
Intellectual property (IP) can be defined as any result of a human intellectual process
that has inherent value to the individual or organization that sponsored the process.
It includes inventions, designs, processes, organizational structures, strategic plans,
marketing plans, computer programs, algorithms, literary works, music scores, and
works of art, among many other things. KM enables the effective use of IP, but it
could also lead to loss of IP, which can damage the organization just as much as los-
ing real capital property. In fact, in many cases IP is an organization’s most valuable
asset. One of the KM initiatives actively pursued by The Dow Chemical Company
was harvesting little-used patents and intellectual assets (Davenport and Prusak 1998).
Box 13.1 describes Dow Chemical’s approach to KM.
As discussed in prior chapters, organizations often capture knowledge from docu-
ments stored in Web-based repositories. The more codifiable this knowledge is and the
more it is documented and distributed, the greater the risk of losing this knowledge.
IP losses can happen in many ways including the following:
1. Employee turnover. The employee may leave the organization to be hired
by a competitor. The employee may deliberately or accidentally share her
knowledge with her new employer.
2. Physical theft of sensitive proprietary documents, either by outsiders or by
insiders.
3. Inadvertent disclosure to third parties without a nondisclosure agreement.
4. Reverse engineering or close examination of company’s products.
Box 13.1
Dow Chemical’s KM Initiative Captures Big Returns
In October 3, 1994, Dow Chemical’s KM initiative under the direction of then director of intellec-
tual asset management Gordon Petrash achieved worldwide notoriety on the cover of Fortune
Magazine (Stewart 1994). Petrash recognized that the intellectual capital represented in the
company’s 29,000 unused patents presented an underutilized opportunity that could bring back
huge returns to the organization.
Unused patents can represent a sizable investment to an organization, since the expense
associated with keeping the patents current could be quite high. Petrash’s group first decided
to develop a concerted effort to evaluate these patents. Patents were assessed to determine
if they could be used, sold, or abandoned. During this process, Petrash was able to save the
organization close to $US1 million in licensing fees that were being expended on patents that
would return no value to the organization.
318 CHAPTER 13
5. The Web repository security is breached, and unauthorized access to the
proprietary documents takes place.
6. Unauthorized parties intercept electronic mail, fax, telephone conversation,
or other communications for the purpose of illicitly acquiring knowledge.
7. Attempts by insiders or outsiders to corrupt documents or databases with false
data, information, or knowledge. This could be done directly via hacking into
a database and effecting unauthorized modifications or indirectly via a virus.
This is a variation of the electronic breach of data problem in item 5, but it
is somewhat different in that the actions can destroy the system in question.
There are significant criminal implications with this act.
Note that the first four types of IP loss are not related to technology, while the last
three are. Also, some of the intellectual capital losses are related to legal practices
used to acquire sensitive competitive intelligence (items 1, 3, and 4), while the law
prosecutes others (items 2, 5, 6, and 7). Clearly the losses related to technology are
easier to prove and therefore easier to prosecute. Companies can take a number of
steps to protect their organization against IP losses as follows.
1. NONDISCLOSURE AGREEMENTS
A nondisclosure agreement is a contract between an organization that owns the IP
and outside individuals to whom the organization’s sensitive and proprietary informa-
tion is disclosed on the condition that they maintain it as confidential. Divulging this
knowledge to a third party constitutes a breach of confidentiality, and the offending
party can be sued for damages. Employees of the organization owning the IP are by
definition expected to maintain confidentiality not only while they are actively em-
ployed by that organization, but also after they terminate their association for whatever
reason. Nondisclosure agreements can serve to protect against loss of knowledge via
employee turnover as well as via covered disclosure to outsiders.
2. PATENTS
Patents are the oldest and most traditional means of protecting inventions. They
grew out of the need to encourage exceptionally bright people to invent products and
processes that benefit humankind. Patents do this by giving exclusive rights (a mo-
nopoly) to any product containing the patented works to the inventor, with all rights
therewith. This means that an inventor, for a fixed period of either 16 or 20 years from
patent issuance, can control the duplication of the patented works or process. Patent
law can be quite complex in what can and cannot be patented, at what time, and for
how long. However, as long as a patent is not overturned, it provides the most secure
of protections. Unfortunately, this protection is only exercised through court action
taken by the patent holder against the individual or organization allegedly infringing
the patent. In some cases, small inventors holding valid patents cannot successfully
sue large corporate entities with large legal staffs. Patents are excellent vehicles for
protecting knowledge about technical innovations and products. They can protect
THE FUTURE OF KNOWLEDGE MANAGEMENT 319
against reverse engineering of a product as well as unauthorized acquisition of any
design or other documents that detail the nature of the invention. In fact, loss of such
documents is considered immaterial since the design of the patented invention is
already part of the public record by virtue of its patented nature.
3. COPYRIGHTS
While patents protect the ideas behind the invention (the so-called claims), copyrights
protect the expression of the work. They have been traditionally used to protect liter-
ary works, works of art, architecture, and music. However, they can also be used to
protect computer programs albeit weakly. The advantage is that while patents require
a rather rigorous process to be granted, a copyright can be done by merely stating on
a copy of the body of the work that it is copyrighted. Registration of the copyrighted
work with the government in the United States is not required, although it’s highly
advisable. Other countries require registration. This is done with the symbol ©. Copy-
rights typically last for the life of the creator plus up to 50 more years, depending on
the country of filing. A copyright holder maintains the rights to publish, broadcast,
reproduce, or copy the work. She has the exclusive right to translate it into another
language, either wholly or in part. Copyrights can protect stolen or illicitly obtained
IP only if it is valuable in its expression. For example, computer programs may fall
into that category.
4. TRADE SECRETS
An organization may choose not to patent an invention but instead keep it as a trade
secret. This invention may not fulfill all the criteria for patentability. Alternatively, the
organization may want to avoid the legal process required to protect IP. Stealing trade
secrets is illegal and punishable by law if the damaged organization takes legal action.
However, said organization must make a strong effort to maintain confidentiality in
order to maintain its legal rights. Organizations may accomplish this by instituting
reasonable safeguards of its IP. Lacking that, a court may decide that it was not a very
important trade secret to begin with.
We have discussed some legal avenues of IP protection. However, once the or-
ganization resorts to legal remedies, the damage has already been done and it most
likely can only aspire to damage recovery. An effective KM initiative must include
institutionalizing policies and safeguards that will prevent the loss of IP in the first
place. Installing firewalls in computer systems, access controls, and protecting all the
sensitive information through encryption can go a long ways toward this. Furthermore,
organizations should clearly educate their employees of their responsibility for con-
fidentiality and the consequences they could suffer if they violate this confidentiality
whether accidentally or purposely.
As mentioned earlier, these avenues for IP protection should be used with some
caution. Although they do help in preventing knowledge leakage, they may also inhibit
the ability of the organization’s own employees to seek knowledge from individuals
outside the organization who may be able to provide them with helpful advice.
320 CHAPTER 13
INVOLVING INTERNAL AND EXTERNAL KNOWLEDGE CREATORS
Two interrelated and emerging aspects of knowledge management focus on involv-
ing collaborations of a large numbers of individuals: (a) from across various levels
within the organization; and (b) from outside the organization to share and create
knowledge.
THE VALUE OF GRASSROOTS CONTRIBUTIONS
As discussed in Chapter 12, the democratization of knowledge refers to providing
every employee within an organization with the ability to make grassroots contribu-
tions that stand the chance to influence the company’s direction, much like Blake
Melnick has championed at Atlantis Systems International. This ability is becoming
increasingly valuable because people at the lowest levels of the organization are the
ones who most commonly interact with customers, and are often the ones involved
to the least extent in important decisions. Web 2.0 technologies enable widespread
participation in decision-making, or at least the consideration of ideas from across the
organization, by drastically reducing the time that it takes for such ideas to be com-
municated and aggregated. This could be done through blogs or wikis, for example,
as they allow individuals across the organization to contribute to content and for
individuals across the organization to access that content. Box 13.2 illustrates how
such widespread contribution is being obtained at Northwestern Mutual Financial
Network.
Moreover, the success of Wikipedia, open-source software development, and in-
novative companies, such as Threadless, led to the recognition of the potential for
value creation through contributions from beyond traditional organizational bound-
aries (Cook 2008). Indeed, this has led to the emergence of the phenomenon called
crowdsourcing or community-based design, which refers to the outsourcing of
tasks to an , large group of people or community in the form of an open
call (Howe, 2006; Kaufman 2008; Wikipedia 2014a). Box 13.3 illustrates the use of
such crowdsourcing, which has also been labeled as collective intelligence (Bonbeau
2009). Box 13.4 describes the use of crowdsourcing for the purpose of raising funds,
also known as crowdfunding.
WHERE DO WE SOURCE KNOWLEDGE? LOOK AT THE CLOUD!
Recent industry trends reveal that organizations are increasingly adopting cloud
computing to manage their IT resources. Cloud computing refers to the ability to
access software and/or hardware resources from remote locations through the use of
browsers or thin-client application interfaces. Some analysts view Cloud computing
as the use of virtual servers over the Internet, while others use a broader definition:
any computing resource that can be consumed as a service outside the organization’s
firewall (Knorr and Gruman 2008). John McCarthy defined the paradigm for cloud
computing: “computation may someday be organized as a public utility” (Garfinkel
2011), which refers to the elastic provision for modern computing and compares it to
THE FUTURE OF KNOWLEDGE MANAGEMENT 321
Box 13.2
Spurring Grassroots Collaboration at Northwestern Mutual
Until 2005, one of the problems facing Northwestern Mutual was that the company’s formal
hierarchical structure and communication channels adversely affected information flow
across departments. The employees were using e-mail and structured reporting systems
to transfer information up the chain of command, while hoping that the senior executives
would take appropriate actions and disseminate the results back down and across to other
departments as well. This frequently led to one department not knowing about related ongo-
ing projects in other departments, thereby inhibiting coordinated efforts or learning across
departments.
Charged with the task of addressing such communication problems and promoting open
communication, the Assistant Director of Corporate Communication found a solution in the
form of corporate blogging, which would put information out in the open so that anyone could
find it. In October 2005, she appeared in front of Northwestern Mutual’s public affairs commit-
tee (a cross-functional steering committee created to track consumer and government trends
that might affect the business) to discuss blogging opportunities and threats. As she described
it, the problem with e-mail is one of reach: “You may be aware of only some subset of people
that may have an interest in what you’re working on.” Sharing information through a blog helps
in keeping everyone informed. “You’re not determining and limiting who your potential audience
may be,” she remarked.
The committee found these ideas intriguing, and the Assistant Director of Corporate Com-
munication returned in December to suggest alternative courses of action including various
internal blogs and an external public-facing corporate blog. Although the communications
department had been concentrating on external blogging, the CEO enthusiastically supported
the notion of internal blogs. For Northwestern Mutual, blogging fit well into an overall corporate
communications strategy that, according to the CEO, was intended to “open the windows” and
foster a more honest and open dialogue within the company including between management
and employees as well as among employees.
However, since it belonged to a highly regulated industry, Northwestern Mutual needed
to be ready to produce a comprehensive record of all communications whenever needed.
In January 2006, Northwestern Mutual selected the Customer Conversation System (from
Awareness), which combines Web-based blogging and content management with enterprise
security, workflow and regulatory compliance tools, and extensive versioning capabilities. A
blogging solution (Mutualblog) was up and running four months later and was rolled out to
5,000 users across the organization in June 2006. About 100 people were actively blogging
by September 2006. Although Northwestern was still in the early phase of its experience with
blogging, it seemed to be working well in jump-starting collaboration and sparking a larger
change in the corporate culture. “This is the first time we’ve had a grassroots application that
allowed employees to share what they’re working on directly,” the Assistant Director of Corpo-
rate Communication remarked.
Source: Compiled from Spanbauer 2006; Young et al. 2007.
the electricity industry with the illusion of an infinite supply. The underlying concept of
Cloud computing dates to the 1950s, when large mainframe computers were accessed
via dumb terminals to share the CPU time. The concept of Cloud computing gained
impetus when Amazon launched Web services in 2006, in an effort to increase the
utilization of their data centers by providing Cloud computing to external customers.
Today other vendors offer on-demand access to storage and virtual servers, such as
IBM’s SmartCloud network.
322 CHAPTER 13
The National Institute of Standards and Technology defines Cloud computing as:
a model for enabling ubiquitous, convenient, on-demand network access to a shared pool
of configurable computing resources (e.g., networks, servers, storage, applications, and
services) that can be rapidly provisioned and released with minimal management effort or
service provider interaction. This cloud model is composed of five essential characteristics,
[and] three service models . . .
Essential Characteristics:
On-demand self-service. A consumer can unilaterally provision computing capabilities,
such as server time and network storage, as needed automatically without requiring human
interaction with each service provider.
Broad network access. Capabilities are available over the network and accessed through
standard mechanisms that promote use by heterogeneous thin or thick client platforms (e.g.,
mobile phones, tablets, laptops, and workstations).
Resource pooling. The provider’s computing resources are pooled to serve multiple
consumers using a multi-tenant model, with different physical and virtual resources dynami-
cally assigned and reassigned according to consumer demand. There is a sense of location
independence in that the customer generally has no control or knowledge over the exact
location of the provided resources but may be able to specify location at a higher level
Box 13.3
Open Innovation at InnoCentive
InnoCentive was founded in 2001, when pharmaceutical company Eli Lilly and Company funded
its launch as a way to connect with people outside the company who could help in develop-
ing drugs and speeding them to market. InnoCentive is open to other firms eager to access the
network’s community of ad hoc experts. It connects public sector and nonprofit organizations,
companies, and academic institutions all looking for breakthrough innovations, with a global
network of over 160,000 creative thinkers (including engineers, scientists, inventors, and business
people with expertise in life sciences, engineering, chemistry, math, computer science, and entre-
preneurship) across the world. These creative thinkers, called “Solvers,” join a community called
“InnoCentive Solver” to address some of the world’s toughest challenges posted by “Seeker”
organizations, who offer registered Solvers considerable financial awards (anywhere from $5,000
to $100,000 per solution) for the best solutions. InnoCentive manages the entire process, while
keeping the identities of Seekers and Solvers completely confidential and secure.
Companies like Boeing, DuPont, and Procter & Gamble also pay InnoCentive a fee to par-
ticipate. They post their most tricky scientific problems on InnoCentive’s Web site, and anyone
in InnoCentive’s Solver community can try solving them. The Solvers are quite diverse. Many of
them are hobbyists, such as a University of Dallas undergraduate student who identified which
chemical to use in an art restoration project, or a patent lawyer from Cary, North Carolina, who
developed a creative way of mixing large batches of chemical compounds.
Dr. Karim R. Lakhani examined 166 problems posted by 26 research labs over four years on
the InnoCentive site and found that an average of 240 people examined each problem, an aver-
age of 10 people offered answers, and about 30 percent of the problems were solved (Wessell
2007). During its seven-year history, InnoCentive has paid out more than $3.5 million in awards
to over 300 winning Solvers (InnoCentive 2009).
Source: Compiled from the InnoCentive website at www.innocentive.com; Howe 2006; Wessell
2007.
http://www.innocentive.com
THE FUTURE OF KNOWLEDGE MANAGEMENT 323
of abstraction (e.g., country, state, or datacenter). Examples of resources include storage,
processing, memory, and network bandwidth.
Rapid elasticity. Capabilities can be elastically provisioned and released, in some cases
automatically, to scale rapidly outward and inward commensurate with demand. To the
consumer, the capabilities available for provisioning often appear to be unlimited and can
be appropriated in any quantity at any time.
Measured service. Cloud systems automatically control and optimize resource use by
leveraging a metering capability at some level of abstraction appropriate to the type of ser-
vice (e.g., storage, processing, bandwidth, and active user accounts). Resource usage can
be monitored, controlled, and reported, providing transparency for both the provider and
consumer of the utilized service. (Mell and Grance 2011)
The three service models refer to:
1. Software as a Service (SaaS)–which provides customers the ability to use
the provider’s software running on their own cloud infrastructure, accessed
via a web browser, thus eliminating the need for the customer to manage the
underlying hardware and software infrastructure.
Box 13.4
A New Way to Raise Capital: Crowdfunding
Crowdfunding refers to a collective effort to pool many small amounts of money together, via the
Internet, to invest in a project or venture started by others. There are many crowdfunding platforms
including Kickstarter, Indiegogo, RocketHub, and RockThePost (Prive 2012). The idea is similar
across all these platforms: project creators launch a profile and a short video that describes the proj-
ect, the rewards per donation, and a convincing sales pitch that can raise money to fund their ideas.
Crowdfunding Web sites allow anyone to post their projects online and ask the online com-
munity to help fund it. Different from the traditional methods of raising capital, project sponsors
don’t necessarily expect equity in the project in return; and instead of traditional investors,
crowdfunding campaigns are funded by the general public. The creator of the project can offer
different incentives depending on the amount of money being donated, or sponsors can simply
donate their money with the satisfaction of having contributed to the success of the project.
Projects are not limited to a certain category or scope and can vary anywhere from art, comics,
dance, design, fashion, film and video, food, games, music, photography, publishing, technology,
and theater among others. Also crowdfunding is used to raise funds for philanthropic and disaster
relief efforts such as Kiva, which provides support for Microcredit crowdfunding. After providing a
detailed explanation of the project and pitching the idea (in the form of a video), the creator can
then set a goal for the necessary funds as well as an end date for the fundraising campaign. Most
campaigns last anywhere from 30 to 60 days. The campaign’s progress is public and updated
in real time so potential sponsors can see how much money is needed to successfully fund the
project. If the goal is reached, the creator gets all of the moneys raised minus a small fee (about 5
percent) from the crowdfunding Web site. In the event that the project does not meet its goal, the
creator does not receive any funds and the money is returned to the sponsors.
As an example, Pebble Technology, the maker of a digital watch compatible with Android and
iPhone smartphones, had set out to raise $100,000 on Kickstarter.com. It goes without saying
that the more interesting and feasible the project, the more money it will receive. In their case, the
projects exceeded the goal and their campaign ended up raising $10.3 million (Svensson 2012).
Source: Compiled from Prive 2012; Svensson 2012.
324 CHAPTER 13
2. Platform as a Service (Paas)–which provides customers the ability to deploy
their own applications onto the Cloud hardware and software infrastructure
3. Infrastructure as a Service (IaaS)–which provides customers the ability to
control the software systems, like for example firewalls, in the underlying
Cloud infrastructure (Mell and Grance 2011).
The main benefits of using Cloud computing relate to optimizing the utilization
of hardware and software infrastructure while minimizing the upfront investment.
This allows customers to invest in resources as needed; for example, as the number
of transactions increases, the required infrastructure to appropriately support those
transactions can be augmented. Furthermore, sharing the cost of infrastructure among
a number of customers can bring economies of scale to all the companies sharing the
Cloud. Finally, customers can take advantage of the established IT infrastructure in
the Cloud and focus their efforts on the business and innovation.
Cloud computing supports the growth of big data in organizations, which in turn
creates new opportunities for smart analytics and improved customer service, but
this can only happen if the IT and business units work together. Big data refers to
collecting large data sets in organizations, a phenomenon of increasing concern to IT
leaders. Box 13.5 describes why business leaders must take the lead in making better
use of big data to find value in information.
ADDRESSING BARRIERS TO KNOWLEDGE SHARING AND CREATION
Although it is important to involve internal and external knowledge creators and Web
2.0 technologies provide some interesting ways of doing so, it may not be easy to
accomplish this. Three main problems constrain individuals’ contributions to KM:
(a) privacy concerns; (b) concerns related to “knowledge as power”; and (c) senior
executives’ reluctance to adapt. Each of these barriers is discussed next.
PRIVACY CONCERNS
Perceived threats to privacy may inhibit individuals from contributing knowledge
both within the organization as well as across organizational boundaries. An indi-
vidual may be less forthcoming with honest opinions if she believes that the rec-
ommendations she is providing for a decision would be compiled and potentially
viewed in the future in the light of all other comments she makes over time, either
within the organization or over the Internet, all the decisions she makes, and her
demographic and other personal information. This is especially true when the in-
dividual perceives that the comments might not be liked by a powerful individual
within the organization. This applies to comments sent via e-mail as well as posted
on blogs (DePree and Jude 2006).
People expect a certain level of privacy even beyond traditional organizational
boundaries, such as in social networking sites, and are likely to react strongly to
threats to privacy (McCreary 2008, Pottie 2004). Box 13.6 on page 326 provides one
illustrative example.
THE FUTURE OF KNOWLEDGE MANAGEMENT 325
Box 13.5
Data Explosion: Finding Value in “Big Data”
Modern enterprises are inundated with information, and in most of the organizations the volume of data has been
increasing at a rate of 35 percent to 50 percent each year. Companies today process more than 60 terabytes of
data annually, which is 1,000 times more than a decade ago. Most of the data collected today by the companies
is unstructured (as opposed to in structured databases) and captured in word processing documents, spread-
sheets, images, and videos, which often makes it difficult to retrieve or interpret. Despite the buzz around “big
data,” most organizations are focused on challenges of storing, protecting, and accessing massive amounts of
data, which are primarily the responsibilities of the IT unit. In general, most organizations have underexploited
the opportunities that could be made possible from exploiting such data. Only a few CIOs and other IT executives
reported that their organizations are generating significant business benefits from the stored data.
Studies show that on average, data in organizations grows at a rate that nearly doubles the volume of
data stored every two years. The highest growth rates were primarily observed at research universities and
hospitals. Contrary to the expectation of increased decision-making opportunities based on the volume of
data, most organizations are still in the aspiring stage of generating business value from the data. For a bet-
ter understanding of the challenge in producing business value from the information in the organizations, it is
helpful to understand various segments of the data sources.
1. Growth of Structured Data
As companies invest heavily in enterprise resource planning systems, customer relationship management
systems, radio-frequency identification tags, and other technologies, the amount of data collected per trans-
action is multiplying. For example, companies collect data on inventory levels, transportation movements,
and financial transactions enabling them to communicate more accurately with customers, accelerate supply
chains, manage risks, and identify business opportunities. Increased structured data can be a double-edged
sword: increased granularity creates opportunities for analytics that could lead to improved business processes
and customer service, but duplicated and conflicting data can undermine service delivery resulting in conflicts
over whose data are more accurate. A large percentage of stored data serves no useful purpose in practice be-
cause management has not specified how it will be used or who will make what decisions on the provided data.
2. Explosion of Unstructured Data
Documents, images, videos, and e-mail make up a significant percentage of the data stored in most organi-
zations. This growth is the result of many factors such as increased regulatory requirements. Some organiza-
tions are finding that unstructured data collected through social media enabled greater sharing of informa-
tion and knowledge. However, realizing business value from this unstructured data typically requires indexing
and reorganizing of some information.
At the tactical level, IT units can take the lead in ensuring safe, reliable, cost-effective data storage and
access, but that will not necessarily lead to business success. In order to achieve the maximum value from
data, senior management must commit to three practices:
so forth constitute most of the data. This information can provide different opportunities for different
segments of the business. By scoping the “sacred data,” management can clarify how the business will
set the parameters for the organization’s enterprise architecture on which IT can build.
data, management needs to define the workflows that create, retrieve, change, and reuse documents,
messages, images, and other unstructured data. In particular manual or automated processes need to
be defined for adding the “metadata”—tags that categorize unstructured data.
process. Improving business processes and service can lead to richer data allowing innovation and
efficiencies.
Source: Compiled from Beath et al. 2012.
326 CHAPTER 13
CONCERNS RELATED TO “KNOWLEDGE AS POWER”
KM mechanisms and technologies that help capture and store employee’s knowledge
reduce knowledge loss when expert employees leave the organization. However, this may
lead to employees, who are not nearing retirement and are concerned about job security at
the organization, being concerned about sharing their knowledge with others in the organi-
zation. The perception that “knowledge is power” could lead to the belief that by sharing
one’s privately held knowledge, one might become dispensable or lose some influence
within the organization. Uniqueness is widely considered to be an important determinant
of power within the organization (Hickson et al. 1971), and by sharing private knowledge
individuals risk losing their unique contribution to the organization (Gray 2001).
Box 13.6
Privacy Concerns at Facebook
Facebook is a social networking Web site that provides free access to people who can join
networks organized by interest, region, city, place of work, or academic institution, and connect
and interact with others. Users can also add friends and communicate with them and update
personal profiles to inform friends about themselves. Facebook’s mission is “to give people the
power to share and make the world more open and connected” (Facebook 2014). According to
Alexa, Facebook’s ranking among all Web sites increased from 60th to 7th in worldwide traffic,
from September 2006 to September 2007, and is currently second (Alexa 2014). Operated and
privately owned by Facebook, Inc., it had over 1.16 billion monthly and 727 million daily average
active users worldwide, 80 percent from outside the United States and Canada, as of Septem-
ber 2013 (Facebook 2014).
In 2007, Facebook encountered problems with the default settings that provide advertisers
the product preferences that its users share with their friends. Users reacted negatively to this
“privacy intrusion,” and Facebook management had to change the default settings from “yes” to
“no” (i.e., Facebook would not be able share these product preferences; McCreary 2008).
Following this incident, Facebook decided to change its policy on user-generated content,
such that it would have a “perpetual” license to use any material uploaded by users for advertis-
ing or other venues, even if the user had subsequently deleted the content or even cancelled
the account. However, Facebook rescinded the decision upon learning that the Electronic Pri-
vacy Information Center (EPIC), an advocacy group based in Washington, DC, was planning to
file a formal complaint with the Federal Trade Commission over the changed license agreement.
The Executive Director of EPIC remarked: “What we sensed was taking place was that Face-
book was asserting a greater legal authority over the user-generated content. It represented a
fundamental shift in terms of how the company saw its ability to exercise control over what its
users were posting, and that really concerned us.” Credited for this protest against Facebook
was a grassroots effort by Julius Harper Jr., a 25-year-old who formed the “People Against the
New Terms of Service” Facebook group, with over 80,000 members. This group started as a
simple protest, and then submitted its major concerns to the service’s legal team. The most
significant concerns included why the revised terms of service appeared to give Facebook the
right to use members’ photos if the company had no intention of utilizing them, and what would
occur if Facebook were to be acquired by another corporation in the future and the new owner
wanted to utilize the user-generated content in ways the current Facebook leaders were not
considering.
Source: Compiled from Alexa 2014; Facebook 2014; McCreary 2008; Raphael 2009; Wikipedia
2014.
THE FUTURE OF KNOWLEDGE MANAGEMENT 327
When job security is low due to adverse economic conditions and frequent layoffs,
workers have an even stronger motivation to attach greater utility to, and consequently
withhold, their private knowledge (Davenport and Prusak 1998). By contrast, when the
level of trust or care among employees in an organization is high, they may perceive a lower
psychological cost of knowledge sharing due to greater concern about how they contribute
to solving organizational problems and are useful to others (Constant et al. 1996).
A similar perception may inhibit knowledge sharing across departments. Therefore,
managers need to be sensitive to issues related to power and control (Gray 2001).
Incentives might help to some extent and only in some situations. Despite the prog-
ress in IT and in the field of KM, motivating employees to share private knowledge
remains a critical issue.
SENIOR EXECUTIVES’ RELUCTANCE TO ADAPT
As discussed above, individuals possessing the knowledge might be reluctant to share
it due to concerns about privacy and perceived loss of power. In order to convince
employees to share their knowledge and contribute to the creation of new organiza-
tional knowledge, senior executives need to play an important role. Senior executives
should make considerable changes in organizational forums as well as in their own
attitudes. Some of the important changes are:
1. The creation of a flatter organizational structure, because “most companies
have hierarchical structures, and differences in status among people impede
the exchange of ideas.” (Amabile and Khaire 2008, p. 102)
2. The incorporation of diverse and multiple, and often starkly conflicting, per-
spectives to facilitate knowledge creation (Yoo 2008).
3. The willingness to allow redundancy and slack resources needed for the
incorporation of these diverse and multiple perspectives.
4. The willingness to let go of their power, much more so than the traditional notion
of “empowerment” might imply, as is needed for a true knowledge democracy.
Unfortunately, a number of senior executives are unwilling to recognize the im-
portance of the above changes. A knowledge democracy in an organization cannot be
achieved unless senior executives truly believe that people across the organization,
including at the very lowest levels where interactions with customers most often oc-
cur, might have truly valuable ideas that would influence the organizational strategy,
or product roadmap. For example, Scott Cook, the cofounder of Intuit Inc., wonders
whether management is “a net positive or a net negative for creativity”: “If there is a
bottleneck in organizational creativity, might it be at the top of the bottle?” (Amabile
and Khaire 2008, p. 102).
CONCLUDING REMARKS
The benefits from knowledge management are considerable, and progress in infor-
mation technology as well as the experience gained within the field of knowledge
328 CHAPTER 13
management implies that there are also some valuable ways of managing knowledge
so as to increase efficiency, effectiveness, and innovation for both organizations and
individuals. However, KM is not easy, and encounters numerous challenges related
to the adoption of technologies, motivation of individuals within and outside the or-
ganization, and the integration of people and technologies within the KM processes.
In this book, we have tried to provide the reader with a comprehensive overview
of the foundations of KM, the opportunities and challenges, as well as some of the
important emerging and future directions.
In conclusion, the future of KM is one where people and advanced technology will
continue to work together, enabling knowledge integration across diverse domains and
with potentially high payoffs. However, the new opportunities and greater benefits
will require careful management of people and technologies, synthesis of multiple
perspectives, and effectively dealing with a variety of tradeoffs. Even though interest-
ing challenges lie ahead for knowledge managers, the future of KM is clearly exciting
because of the opportunities it promises for generations to come.
REVIEW
1. Identify the one issue you consider most important for the future of knowledge
management. Why?
2. How do you see organizations changing in the future, especially in terms of
knowledge management but also in terms of their structure, as a result of Web
2.0 technologies?
3. Based on both Chapters 3 and 13, what role do you see top management
playing in knowledge management in organizations in the future?
4. How might privacy concerns affect knowledge management in the future,
both within organizations as well as in social networks?
5. Describe Cloud computing and its benefits.
6. Briefly identify any three ways in which you see employees changing their
behaviors related to knowledge management in the future.
APPLICATION EXERCISES
1. Select any three topics on Wikipedia (http://en.wikipedia.org/) and track the
identified pages for a week (visiting each page twice a day) to see how the
knowledge changes on these pages. What lessons do you learn from this
experience?
2. Select any one organization that utilizes crowdsourcing (such as Cambrian
House [http://www.cambrianhouse.com/] or CrowdSpirit [http://www.crowd-
spirit.com/]), and find our more information about whose knowledge is be-
ing managed at this organization through crowdsourcing. Then summarize
the lessons you learn from this organization about the relationship between
crowdsourcing and knowledge management.
3. Visit a local area firm to study how knowledge is being managed there at
present, examine whether Web 2.0 technologies are being used, and talk to
http://en.wikipedia.org/
http://www.cambrianhouse.com/
http://www.crowd-spirit.com/
http://www.crowd-spirit.com/
THE FUTURE OF KNOWLEDGE MANAGEMENT 329
some of its people regarding how knowledge management might change in
the future.
4. Visit a small local area firm and a large local organization, and talk to two
individuals at each organization to examine whether concerns related to
privacy and “knowledge is power” affect knowledge management at this
organization.
5. Identify one wicked problem that seems really important to you. Investigate
ways in which knowledge management, including knowledge management
through crowdsourcing or collective intelligence, might help address this
wicked problem.
NOTE
1. They have also been called learning organizations.
REFERENCES
Alexa. 2014. Alexa: The Web Information Company. www.alexa.com/ (accessed January 7, 2014).
Amabile, T.M., and Khaire, M. 2008. Creativity and the role of the leader. Harvard Business Review,
October, 100–109.
Bonbeau, E. 2009. Decisions 2.0: The power of collective intelligence. Sloan Management Review,
50(2) (Winter), 45–52.
Churchman, C. 1971. The design of inquiring systems: Basic concepts of systems and organization.
New York: Basic Books.
Constant, D., Sproull, L., and Kiesler, S. 1996. The kindness of strangers: The usefulness of electronic
weak ties for technical advice. Organization Science, 7(2), 119–135.
Cook, S. 2008. The contribution revolution: Letting volunteers build your business. Harvard Business
Review (October), 123–131.
Courtney, J. 2001. Decision making and knowledge management in inquiring organizations: Toward a
new decision-making paradigm for DSS. Decision Support Systems, 31, 17–38.
Davenport, T., and Prusak, L. 1998. Working knowledge: How organizations manage what they know.
Boston: Harvard Business School Press.
DePree Jr., C.M., and Jude, R.K. 2006. Who’s reading your office e-mail? Is that legal? Strategic
Finance, 87(10).
Facebook. 2014. Company Info. http://newsroom.fb.com/company-info/ (accessed July 27, 2014).
Garfinkel, S. 2011. The cloud imperative. MIT Technology Review, December. www.technologyreview.
com/news/425623/the-cloud-imperative/ (accessed July 27, 2014).
Gray, P. 2001. The impact of knowledge repositories on power and control in the workplace. Informa-
tion Technology and People, 14(4), 368–384.
Hickson, D.J., Hinings, C.R., Lee, C.A., Schneck, R.E., and Pennings, J.M. 1971. A strategic contingen-
cies theory of intra-organizational power. Administrative Science Quarterly, 16, 216–229.
Howe, J. 2006. The rise of crowdsourcing. Wired (June). www.wired.com/wired/archive/14.06/crowds.
html (accessed July 27, 2014).
InnoCentive. 2009. Frequently asked questions. www.innocentive.com/faq/Solver (accessed July 27,
2014).
Kaufman, W. 2008. Crowd sourcing turns business on its head. NPR, August 20. www.npr.org/templates/
story/story.php?storyId=93495217 (accessed July 27, 2014).
Knorr, E., and Gruman, G. 2008. What cloud computing really means. Infoworld, April, 7. www.infoworld.
com/d/cloud-computing/what-cloud-computing-really-means-031(accessed July 27, 2014).
http://www.alexa.com/
http://newsroom.fb.com/company-info/
http://www.technologyreview.com/news/425623/the-cloud-imperative/
http://www.wired.com/wired/archive/14.06/crowds.html
http://www.innocentive.com/faq/Solver
http://www.npr.org/templates/story/story.php?storyId=93495217
http://www.infoworld.com/d/cloud-computing/what-cloud-computing-really-means-031
http://www.technologyreview.com/news/425623/the-cloud-imperative/
http://www.npr.org/templates/story/story.php?storyId=93495217
http://www.infoworld.com/d/cloud-computing/what-cloud-computing-really-means-031
330 CHAPTER 13
McCreary, L. 2008. What was privacy? Harvard Business Review, October, 123–131.
Mell, P., and Grance, T. 2011. The NIST Definition of Cloud Computing. Gaithersburg, MD: National
Institute of Standards and Technology. Special Publication 800–145, September. http://csrc.nist.gov/
publications/nistpubs/800-145/SP800-145 (accessed January 3, 2014).
Pottie, G.J. 2004. Privacy in the global e-village. Communication of the ACM, 47(2) (February),
21–23.
Prive, T. 2012. What is crowdfunding and how does it benefit the economy?, Forbes, November.
www.forbes.com/sites/tanyaprive/2012/11/27/what-is-crowdfunding-and-how-does-it-benefit-the-
economy/ (accessed July 27, 2014).
Raphael, J.R. 2009. Facebook’s privacy flap: What really went down, and what’s next. PC World,
February 18. www.pcworld.com/article/159743/facebook.html (accessed July 27, 2014).
Rittel, H., and Webber, M. 1973. Dilemmas in a general theory of planning. Policy Sciences, 4,
155–169.
Spanbauer, S. 2006. Modern knowledge management applications. CIO.com, December 1. www.cio.
com/article/27087/Modern_Knowledge_Management_Applications (accessed July 27, 2014).
Stewart, T. 1994. Intellectual capital: Your company’s most valuable asset. Fortune, 10, 68–73.
Svensson, P. 2012. Kickstarter projects generate millions of dollars. Associated Press, August 17.
http://bigstory.ap.org/article/kickstarter-projects-generate-millions-dollars-0 (accessed January 4,
2014).
Wessel, D. 2007. Prizes for solutions to problems play valuable role in innovation. Wall Street Jour-
nal, January 25. http://online.wsj.com/news/articles/SB116968486074286927 (accessed July 27,
2014).
Wikipedia. 2014a. Crowdsourcing. http://en.wikipedia.org/wiki/Crowdsourcing.
———. 2014b. Facebook. http://en.wikipedia.org/wiki/Facebook.
Yoo, Y. 2008. Mobilizing knowledge in a Yu-Gi-Oh! world. In Knowledge management: An evolution-
ary view, vol. 12, Advances in Management Information Systems, ed. I. Becerra-Fernandez and D.
Leidner, 127–144. Armonk, NY: M.E. Sharpe.
Young, G.O., Brown, M., Holmes, B.J., and Lawson, A. 2007. Forrester Research: How Northwestern
Mutual deployed enterprise web 2.0. CIO Zone.com. www.ciozone.com/index.php/Editorial-
Research/Forrester-Research-How-Northwestern-Mutual-Deployed-Enterprise-Web-2.0.html
( accessed February 20, 2009).
http://csrc.nist.gov/publications/nistpubs/800-145/SP800-145
http://www.forbes.com/sites/tanyaprive/2012/11/27/what-is-crowdfunding-and-how-does-it-benefit-the-economy/
http://www.forbes.com/sites/tanyaprive/2012/11/27/what-is-crowdfunding-and-how-does-it-benefit-the-economy/
http://www.pcworld.com/article/159743/facebook.html
http://www.cio.com/article/27087/Modern_Knowledge_Management_Applications
http://bigstory.ap.org/article/kickstarter-projects-generate-millions-dollars-0
http://online.wsj.com/news/articles/SB116968486074286927
http://en.wikipedia.org/wiki/Crowdsourcing
http://en.wikipedia.org/wiki/Facebook
http://www.ciozone.com/index.php/Editorial-Research/Forrester-Research-How-Northwestern-Mutual-Deployed-Enterprise-Web-2.0.html
http://www.ciozone.com/index.php/Editorial-Research/Forrester-Research-How-Northwestern-Mutual-Deployed-Enterprise-Web-2.0.html
http://csrc.nist.gov/publications/nistpubs/800-145/SP800-145
http://www.cio.com/article/27087/Modern_Knowledge_Management_Applications
331
Glossary
Access Control. Refers to mechanisms and policies that restrict access to computer
resources.
Active Web Documents. Web documents received by a client from the WWW server,
which contain programs that execute on the client’s computer, continually changing
the display.
Adaptation. The process of modifying a historical solution to solve the current prob-
lem when the current problem is not identical to the historical problem associated
with that solution.
Algorithm. A step-by-step problem-solving procedure for solving a problem in a
finite number of steps.
Application Linking. Refers to when the enterprise shares business processes and
data between two or more IT applications.
Artificial Intelligence (AI). The branch of computer science concerned with
making computers behave like humans. John McCarthy coined the term in 1956
while at the Massachusetts Institute of Technology. It refers to the science that
provides computers with the ability to solve problems not easily solved through
algorithmic models.
Artificial Neural Networks (ANN). A type of artificial intelligence that attempts
to simulate human intelligence by recreating the connective physiology of the
human brain. It is a mechanism (either implemented in computer hardware or
simulated using computer software) that can learn to map information between two
vector spaces: input and output. Neural nets can be used to solve a large variety of
problems, provided that it is possible to formulate the problem in terms of vector
space-mapping.
Associational Expertise. Knowledge or heuristic ability acquired mostly through
human experience and elicited through the knowledge engineering process.
Authorizer’s Assistant. Assists the credit authorization staff to determine the credit
level for credit card customers. The system takes information from a number of data-
bases and approves or disapproves a telephone request from a merchant to authorize
a large purchase from a cardholder.
332 GLOSSARY
Auxiliary Memory. The transfer of pages of data between a computer’s main memory
and a secondary medium of memory.
Avatar. Image of person in virtual reality; a movable three-dimensional image that
can be used to represent somebody in cyberspace, for example, an Internet user.
Back-propagation Algorithm. An algorithm for efficiently calculating the error
gradient of a neural network, which can then be used as the basis of learning.
Backward Reasoning. Reasoning from conclusions, or goals, to the inputs.
Benchmarking. A test used to measure performance.
Best Practices. An assessment recommending the most appropriate way of handling
a certain type of task based on an observation of the way that several organizations
handle that task.
Bi-directional Reasoning. Uses forward reasoning to propagate belief from the
inputs and generate conclusions and backward reasoning to confirm the conclusions
generated dynamically.
Blocks World. Early AI system. It used a robot arm to set blocks on a table. Dem-
onstrated the feasibility of automated task planning.
Blogs. A contraction between Web and log, it refers to a form of online digital diary.
In essence it is a Web site where an individual makes regular written journal entries
that comprise a statement of opinion, a story, an analysis, description of events, or
other material.
Brainstorming Retreats. Company-organized social retreats designed to foster in-
novative ideas.
Business Intelligence (BI). Using information technologies to provide decision-
makers with valuable information and knowledge by utilizing a variety of sources of
data and structured and unstructured information, via the discovery of the relationships
that may exist between these sources of data and information.
Business Process. A network of activities performed by resources that transform
inputs into outputs.
Campaign Management Software. Software used to manage and monitor a com-
pany’s communications with its customers.
Case. A documented historical occurrence used for comparison to current problems,
which is the basis of case-based reasoning (CBR).
Case-based Reasoning (CBR). An intelligent reasoning process that seeks to solve
new problems based on solutions of similar past problems, much like a lawyer seeks
a new outcome on a trial based on precedents established by past legal cases.
Case Library. A database of historical cases containing the universe of knowledge
in a CBR system.
GLOSSARY 333
Choosing Parameters from a Menu. In this method, the database system presents
a list of parameters from which you can choose. This is perhaps the easiest way to
pose a query because the menus guide you, but it is also the least flexible.
Churn. The turnover of users, for example on an online service, especially after the
expiration of a free-trial period.
Classification. The grouping together of data sets according to some predefined
similarity shared by all members of that set.
Clickstream Analysis. A virtual trail that a user leaves behind while surfing the In-
ternet. A clickstream is a record of a user’s activity on the Internet, including every
Web site and every page of every Web site that the user visits, how long the user was
on a page or site, in what order the pages were visited, any newsgroups that the user
participates in, and even the e-mail addresses of mail that the user sends and receives.
Both ISPs and individual Web sites are capable of tracking a user’s clickstream.
Client. The client part of the client-server architecture. Typically, a client is an ap-
plication that runs on a personal computer or workstation and relies on a server to
perform some operations. For example, an e-mail client is an application that enables
you to send and receive e-mail.
Close-ended Questions. Questions asked of an expert that require a short answer or
a number for answers. Used for gathering specific problem-solving knowledge.
Cluster Analysis. The grouping together of data into groups or clusters according
to a natural but parameter shared by all members of the set or based on
empirical computations of proximities across the members of the data set.
Codifiability. Reflects the extent to which knowledge can be articulated or codi-
fied, even if the resulting codified knowledge might be difficult to impart to another
individual.
Collaborative-based Filtering. Recommends items (similar to the one being studied)
that people have liked in the past.
Collaborative Computing. Software systems designed to help people involved in a
project or common task collaborate, typically through the Web, to achieve their goals.
Includes e-mails, calendars, text and chat, wikis, etc.
Collective Intelligence (or crowdsourcing or community-based design). Refers to
the outsourcing of tasks to an , large group of people or community in the
form of an open call.
Combination. Involves the synthesis of multiple bodies of explicit knowledge (and/
or data and/or information) to create new, more complex sets of explicit knowledge.
It is a process of systemizing concepts into a knowledge system. This may take place
during activities such as sorting, adding, combining, and categorizing knowledge.
Common Knowledge. Refers to the organization’s cumulative experiences in com-
prehending a category of knowledge and activities and the organizing principles that
334 GLOSSARY
support communication and coordination. It provides unity to the organization and
includes a common language and vocabulary, recognition of individual knowledge
domains, common cognitive schema, shared norms, and elements of specialized
knowledge common across individuals sharing knowledge.
Common Virtual System. The pinnacle of EAI; all aspects of enterprise computing
are tied together so that they appear as a unified application.
Community-based Design (or crowdsourcing or collective intelligence). Refers to
the outsourcing of tasks to an , large group of people or community in the
form of an open call.
Community of Practice (CoP). An organic and self-organized group of individuals
who are dispersed geographically or organizationally but communicate regularly to
discuss issues of mutual interest.
Competitive Advantage. A situation where the focal organization outperforms its
competitors, usually due to superior ability in some area.
Complex Knowledge. Draws upon multiple distinct areas of expertise. It in-
volves mastering several different pieces of knowledge organized in the form of
a system.
Computer-aided Design (CAD) System. A combination of hardware and software
that enables engineers and architects to design everything from furniture to airplanes.
In addition to the software CAD systems require a high-quality graphics monitor,
a mouse, light pen, or digitizing tablet for drawing, and a special printer or plotter
for printing design specifications. CAD systems allow an engineer to view a design
from any angle with the push of a button and to zoom in or out for closeups and
long-distance views. In addition, the computer keeps track of design dependencies
so that when the engineer changes one value, all other values that depend on it are
automatically changed accordingly.
Concept Learning System. An algorithm that classifies a set of example data by
building an inductive tree and distributing the examples throughout the tree.
Concept Maps. These aim to represent knowledge through concepts and are enclosed in
circles or boxes of some types, which are related via connecting lines or propositions.
Constraint-based Reasoning. An artificial intelligence technique that uses essentially
“what cannot be done” to guide the process of finding a solution.
Constrained Processing Tasks. Artificial tasks given to an expert for the purpose of
observing him and learning from his performance. The expert is typically constrained
in terms of time available for the solution.
Context. The background or environmental circumstances that surround a particular
event or situation.
Content-based Filtering. Recommendations based on what a person has liked in
the past.
GLOSSARY 335
Context-based Reasoning (CxBR). A human behavior representation paradigm
specifically designed to effectively represent human tactical behavior.
Context-Specific Knowledge. Refers to the knowledge of particular circumstances
of time and place in which work is to be performed. It pertains to the organization
and the organizational subunit within which tasks are performed.
Contingency View of KM. Suggests that no “one approach” to managing knowledge
is best under all circumstances.
Cooker. Assists in the maintenance of soup-making equipment. It uses a personal
computer as the delivery platform.
Critic. Rules used for adapting the solution of a similar historical case to the current
problem.
Cross-Industry Standard Process for Data Mining (CRISP-DM). An industry-
neutral and tool-neutral standard process for data mining. Starting from the embry-
onic knowledge discovery processes used in industry today and responding directly
to user requirements, this project defined and validated a data mining process that is
applicable in diverse industry sectors.
Crowdsourcing (or community-based design or collective intelligence). Refers to
the outsourcing of tasks to an , large group of people or community in the
form of an open call.
Customer Relationship Management (CRM). CRM entails all aspects of interac-
tion a company has with its customer, whether it be sales or services related. Com-
puterization has changed the way companies are approaching their CRM strategies
because it has also changed consumer-buying behavior. With each new advance in
technology, especially the proliferation of self-service channels like the Web, more
of the relationship is being managed electronically. Organizations are therefore look-
ing for ways to personalize online experiences (a process also referred to as mass
customization) through tools such as help desk software, e-mail organizers, and Web
development applications.
Customer Touchpoints. Refers to the steps in a business process, like a purchase
transaction, or a software where it interfaces directly with the customer.
Daemons. Functions attached to a frame that assist in obtaining values for slots or
to maintain consistency in the frame system.
Data. Comprise facts, observations, or perceptions (which may or may not be correct).
By itself, data represent raw numbers or assertions and may therefore be devoid of
context, meaning, or intent.
Database (DB). A large collection of data organized for rapid search and retrieval. Pro-
grams that manage data and can be used to store, retrieve, and sort information. You can
think of a database as an electronic filing system. A relational database is a database that
stores data in the form of related tables. Relational databases are powerful because they
336 GLOSSARY
require few assumptions about how data are related or how they will be extracted from
the database. As a result, the same database can be viewed in many different ways.
Database Linking. Databases that share information and duplicate information as
needed.
Data Mining (DM). A class of database applications that look for hidden patterns in
a group of data that can be used to predict future behavior. For example, data mining
software can help retail companies find customers with common interests. The term
is commonly misused to describe software that presents data in new ways. True data
mining software doesn’t just change the presentation but actually discovers previously
unknown relationships among the data.
Data Mining Group (DMG). An independent, vendor led consortium that develops
data mining standards, such as Predictive Model Markup Language (PMLL) supported
by approximately 20 vendors.
Data Warehouse. A collection of data designed to support management decision-
making. Data warehouses contain a wide variety of data that present a coherent picture
of business conditions at a single point in time. The development of a data warehouse
includes development of systems to extract data from operating systems plus instal-
lation of a warehouse database system that provides managers flexible access to the
data. The term data warehousing generally refers to combining many different data-
bases across an entire enterprise. A data mart is defined as a database, or collection of
databases, designed to help managers make strategic decisions about their business.
Whereas a data warehouse combines databases across an entire enterprise, data marts
are usually smaller and focus on a particular subject or department. Some data marts,
called dependent data marts, are subsets of larger data warehouses.
Declarative Knowledge. Focuses on beliefs about relationships among variables.
Characterized as “know what,” it can be stated in the form of propositions, expected
correlations, or formulas relating concepts represented as variables.
Deep Expertise. More theoretical knowledge acquired through formal training and
hands-on problem-solving.
Deployment. Implementing the live model within an organization to aid the decision-
making process.
Development Environment. A program used to develop the knowledge for the
knowledge-based system and provides the inference mechanism used to exercise the
knowledge to solve a problem or answer a question posed by the end user.
Diagrammatic Reasoning. An artificial intelligence technique that aims to understand
concepts and ideas using diagrams that represent knowledge.
Direction. The process through which the individual possessing the knowledge
guides the action of another individual without transferring to him the knowledge
underlying the direction.
GLOSSARY 337
Disjunctions. Otherwise identical cases with different solutions.
Document Management Systems. Computer systems that provide a Web-based
repository accessible from multiple points. The systems also provide a collaborative
environment for several clients to work on electronic documents simultaneously.
Domain Expert. An individual who is both experienced and knowledgeable about
a particular application domain.
Domain Knowledge. Relevant knowledge about a problem domain; knowledge
embedded in the operators of the solution space.
Downsizing. To reduce a unit’s size, such as that of an organization in terms of its
number of employees.
Dynamic. Refers to actions that take place at the moment they are needed rather
than in advance.
Dynamic Web Documents. Web pages created by the Web server in response to the
specific request by the client.
Economy of Scale. A firm’s output is said to exhibit economy of scale when as the
amount of its output is increased, average costs (i.e., total costs divided by the output)
decline.
Economy of Scope. A firm’s output exhibits economy of scope when the total cost
of that same firm producing two (or more) different products is less than sum of the
costs that would be incurred if each product was produced separately by a different
company.
Effectiveness. Performing the most suitable processes and making the best possible
decisions.
Efficiency. Performing the processes quickly and in a low-cost fashion.
Eliza. Early AI implementation. It used a natural language interface to act as an ar-
tificial psychoanalyst, carrying on a dialogue with a patient.
EMYCIN. Knowledge-based system shell. Developed by removing the domain-
specific knowledge from MYCIN.
Encryption. The translation of data into a secret code. Encryption is the most effec-
tive way to achieve data security. To read an encrypted file, you must have access to
a secret key or password that enables you to decrypt it. Unencrypted data is called
plain text; encrypted data is referred to as cipher text.
End User. The person for whom the product was designed by the person who pro-
grams, services, or installs the product.
Enterprise Application Integration (EAI) Technology. EAI is the unrestricted
sharing of data and business processes throughout the networked applications or
data sources in an organization. Early software programs in areas such as inventory
338 GLOSSARY
control, human resources, sales automation, and database management were designed
to run independently with no interaction between the systems. They were custom
built in the technology of the day for a specific need being addressed and were often
proprietary systems. As enterprises grow and recognize the need for their informa-
tion and applications to have the ability to be transferred across and shared among
systems, companies are investing in EAI in order to streamline processes and keep
all the elements of the enterprise interconnected.
Enterprise Resource Planning (ERP) System. A business management system that
integrates all facets of the business including planning, manufacturing, sales, and
marketing. As the ERP methodology has become more popular, software applications
have emerged to help business managers implement ERP in business activities such as
inventory control, order tracking, customer service, finance, and human resources.
Enterprise System. Literally, a business organization. In the computer industry,
the term is often used to describe any large organization that utilizes computers. An
Intranet, for example, is a good example of an enterprise computing system.
Evaluation. Evaluation of a case in the case library for similarity with the current
problem.
Exchange. Used for communicating or transferring explicit knowledge among indi-
viduals, groups, and organizations.
Expected Value. The value a variable is expected to have, based on the probability
distribution of its various observed values. Computed by summing the product of
each possible value for a variable and its probability of occurrence.
Experience Management. Encompasses the processes governing creation, storage,
reuse, maintenance, dissemination, and evaluation of experience relevant to a par-
ticular situation or problem-solving context.
Expertise. Refers to knowledge of higher quality—that is, specific knowledge at its
best. One who possesses expertise is able to perform a task much better than those
who do not.
Expertise Locator System (ELS). A system to catalog knowledge competencies,
including information not typically captured by human resources systems, in a way
that could later be queried across the organization.
Explicit Knowledge. Refers to knowledge that has been expressed into words and num-
bers. Such knowledge can be shared formally and systematically in the form of data, speci-
fications, manuals, drawings, audio and videotapes, programs, patents, and so forth.
Exploratory Analysis of Data (with OLAP). Refers to the use of online analytical
processing (OLAP) for creating complex queries across a multidimensional data
model for the purpose of business reporting.
Externalization. Involves converting tacit knowledge into explicit forms such as words,
concepts, visuals, or figurative language (e.g., metaphors, analogies, and narratives).
GLOSSARY 339
Facets. Subdivisions of a frame slot that contain various types of information related
to the slot.
Facilitator. Leader or chairperson of a brainstorming session.
Fact Base. A data structure that holds all assertions made either by the system or
provided as inputs. These assertions serve as facts for matching premises in an infer-
ence chain.
Fault Diagnosis. Software used to try to determine the causes of a malfunction, also
known as a fault, in particular in a piece of equipment.
Firewall. A system designed to prevent unauthorized access to or from a private
network. Firewalls can be implemented in both hardware and software or a com-
bination of both. Firewalls are frequently used to prevent unauthorized Internet
users from accessing private networks connected to the Internet, especially In-
tranets. All messages entering or leaving the Intranet pass through the firewall,
which examines each message and blocks those that do not meet the specified
security criteria.
Flat Case Libraries. Case library organization where all cases lie at the same hier-
archical level in the case library.
Folksonomy. Refers to a system of classification through the collaborative creation
and translation of tags to annotate and categorize content. For example, multiple us-
ers can tag content with a variety of terms that constitute the metadata for indexing
that content. The folksonomy enables the searchability of content through the textual
description.
Forward Reasoning. Reasoning from inputs to conclusions.
Frames. Structured framework for storing and retrieving knowledge best organized
as attribute value pairs. Composed of slots.
Game-playing. Programming computers to play games such as chess and checkers.
GenAID. Remotely monitors and diagnoses the status of large electrical generators
in real time. It issues a diagnosis with a confidence factor whenever the machine is
operating outside its normal operating conditions. It is presently in commercial op-
eration at various sites throughout the United States.
Generalization. The opposite of specialization. When a frame is more general than
its child frame. This is the normal situation.
General Knowledge. Is possessed by a large number of individuals and can be trans-
ferred easily across individuals.
General Knowledge-gathering Interview Sessions. Interview sessions designed to
elicit general domain knowledge from the expert.
General Problem Solver (GPS). Early AI implementation, which demonstrated
ability to solve problems by searching for an answer in a solution space.
340 GLOSSARY
Goal State. Final desired state of a problem in the solution space.
Graphical User Interface (GUI). A program interface that takes advantage of the
computer’s graphics capabilities to make the program easier to use. Well-designed
graphical user interfaces can free the user from learning complex command languages.
On the other hand, many users find that they work more effectively with a command-
driven interface, especially if they already know the command language.
Groupware. A class of software that helps groups of colleagues (workgroups) at-
tached to a network organize their activities, in particular collaborate on tasks and
achieve common goals.
GUIDON. An instructional program for teaching students therapy for patients with
bacterial infections. GUIDON is a descendant of MYCIN and was developed as a
research tool at Stanford University.
Hacking. The pejorative sense of the term hacker is becoming more prominent largely
because the popular press has co-opted the term to refer to individuals who gain unau-
thorized access to computer systems for the purpose of stealing and corrupting data.
Help Desk Technologies. Software used to aid the operations of a help desk, which
is used to troubleshoot problems with computers or other similar products. The help
desk technologies help to track requests for service as well as to diagnose potential
reasons for the problem.
Heterogeneous Networks. Networks consisting of computers with different proces-
sors and/or different operating systems.
Heuristic Functions. Used in solution space searches to compute the desirability
of moving on to each of the possible next states based on some general knowledge.
These states are ranked in order of decreasing desirability.
Heuristics. Common-sense knowledge drawn from experience to solve problems.
Represents rules-of-thumb and other such shortcuts to the solution that are only learned
through experience. This is in contrast to algorithmic programming, which is based
on a deterministic sequence of steps procedures. Heuristic programs do not always
reach the very best result but usually produce a good result.
Heuristic Search. A search that uses heuristic functions as a guide to determine where
in the problem space to search next.
Human Computer Interface (HCI). The interface between a human and a computer; for
example, a command line interface, a graphical user interface, virtual reality interfaces.
Hyperlink. In a computer document, refers to a word, phrase, or picture on which a
reader may click or hover over, to move automatically to another part of the docu-
ment, another document, or web page.
Hypertext Markup Language (html). A standard representation for text and graphics
that allows the browser to interpret the intentions of the Web page designer. Hypertext
is text with Hyperlinks.
GLOSSARY 341
Hypertext Transfer Protocol (http). A transfer protocol used for exchanging hy-
pertext.
Indexing. The act of classifying and providing an index in order to make items easier
to retrieve.
Inference Chain. A sequence of rules in a rule-based system where the assertions of
an upstream rule serve as the facts to match the premises of downstream rules.
Inferential DM. Models that explain the relationships that exist in data. They may
indicate the driving factors for stock market movement or show failure factors in
printed circuit-board production.
Information. A subset of data, only including those data that possess context, rel-
evance, and purpose. Information typically involves the manipulation of raw data to
obtain a more meaningful indication of trends or patterns in the data.
Information Retrieval. The science of searching for documents, information, or
metadata within documents in the Web, as well as for data within databases.
Information Technology (IT). The broad subject concerned with all aspects of man-
aging and processing information, especially within a large organization or company.
Because computers are central to information management, computer departments
within companies and universities are often called IT departments. Some companies
refer to this department as IS (Information Systems) or MIS (Management Informa-
tion Systems).
Inheritance. The ability in frames and objects to conserve representational effort
by having “children frame” contain all attributes and values possessed by its “parent
frame.”
Initial State. Starting problem definition in a solution space.
Innovation. Performing the processes in a creative and novel fashion that improves
effectiveness and efficiency—or at least marketability.
Innovators. Those who brainstorm the solutions to the customer’s problem.
Intellectual Capital. Knowledge that can be exploited for some moneymaking or
other useful purpose. The term combines the idea of the intellect or brainpower with
the economic concept of capital—that is, that intellect like capital can be used in
service of the saving of entitled benefits so that they can be invested in producing
more goods and services.
Intelligent Program. A concept where the end user sees the knowledge-based system
as a black box that provides intelligent problem-solving capability without the ability
to see its components. It is composed of a knowledge base, an inference engine, and
a development environment.
Internalization. The conversion of explicit knowledge into tacit knowledge. It rep-
resents the traditional notion of “learning.”
342 GLOSSARY
Internet. A computer network protocol able to interconnect heterogeneous net-
works.
Inter-page Structures. Evaluate the arrangement of the various HTML or XML tags
that connect one page to another.
Interviews. The time of interaction with an expert for the purposes of eliciting his/
her knowledge.
Intra-page Structures. Evaluate the arrangement of the various HTML or XML
tags within a page.
Iterative. This is a programming term. It refers to a process that can be described by
a fixed number of variables and a set of rules, which describe what happens to those
variables to achieve the next step of the process. If the process is interrupted, it can
be continued if the state of all the variables is known. Contrast this to a recursive
process. Iteration is a single step.
JavaScript. A dynamic computer language typically used as part of web browsers,
which allow client-side scripts for Web pages to interact with the user and communi-
cate asynchronously, among other functionality frequently required for gaming and
mobile applications.
Just-in-time (JIT) Manufacturing. An ideal method of manufacturing with minimal
waste, short cycle times, and fast communication that can respond rapidly to chang-
ing circumstances.
Kickoff Interview. The first interview in the process of knowledge elicitation.
Knowledge. Intrinsically different from information. Knowledge in an area is
defined as justified beliefs about relationships among concepts relevant to that
particular area.
Knowledge Application. The process through which knowledge is utilized within
the organization to make decisions and perform tasks, thereby contributing to orga-
nizational performance.
Knowledge-based Systems. A computerized system that uses domain knowledge
to arrive at a solution to a problem within that domain. This solution is essentially
the same as one decided upon by a person knowledgeable about the domain when
confronted with the same problem.
Knowledge Capture. The process of eliciting knowledge (either explicit or tacit)
that resides within people, artifacts, or organizational entities and representing
it in an electronic form such as a knowledge-based system, for later reuse or
retrieval.
Knowledge Creation. An activity that catalyzes the innovation of knowledge.
Knowledge Discovery. The development of new tacit or explicit knowledge from
data and information or from the synthesis of prior knowledge.
GLOSSARY 343
Knowledge Discovery in Databases (KDD). The process of data selection, data
cleaning, transfer to a DM technique, applying the DM technique, validating the
results of the DM technique, and finally interpreting them for the user.
Knowledge Elicitation. The process of obtaining tacit knowledge from an expert for
the purposes of making that knowledge explicit.
Knowledge Engineering. The process of developing a knowledge-based system.
Knowledge Management (KM). Can be defined as performing the activities involved
in discovering, capturing, sharing, and applying knowledge in terms of resources,
documents, and people skills, so as to enhance, in a cost-effective fashion, the impact
of knowledge on the unit’s goal achievement.
Knowledge Management Foundations. The broad organizational aspects that sup-
port knowledge management in the long-term. They include knowledge management
infrastructure, knowledge management mechanisms, and knowledge management
technologies.
Knowledge Management Infrastructure. The long-term foundation on which knowl-
edge management resides. It includes five main components: organization culture,
organization structure, information technology infrastructure, common knowledge,
and physical environment.
Knowledge Management Mechanisms. Organizational or structural means used to
promote knowledge management. They may (or may not) utilize technology, but they
do involve some kind of organizational arrangement or social or structural means of
facilitating KM.
Knowledge Management Processes. The broad processes that help in discovering,
capturing, sharing, and applying knowledge.
Knowledge Management Solutions. The ways in which specific aspects of knowl-
edge management (discovery, capture, sharing, and application of knowledge) can be
accomplished. Knowledge management solutions include knowledge management
processes and knowledge management systems.
Knowledge Management Systems (KMS). Integrate technologies and mechanisms
to support KM processes.
Knowledge Repository. A system to share knowledge. Also known as a document
management system and content management systems and serve primarily to share
unstructured information.
Knowledge Sharing. The process through which explicit or tacit knowledge is com-
municated to other individuals.
Lateral Thinking. Using an entirely different approach to solve a problem. In any
self-organizing system there is a need to escape from a local optimum in order to
move towards a more global optimum. The techniques of lateral thinking, such as
provocation, are designed to help that change.
344 GLOSSARY
Learning. The process of improving one’s performance by experiencing an activity
or observing someone else experience that activity. Learning has been one goal of
artificially intelligent systems.
Lessons Learned Systems (LLS). The goal of LLS is to capture and provide lessons
that can benefit employees who encounter situations that closely resemble a previous
experience in a similar situation.
Limited Information Tasks. Artificial tasks given to an expert to perform with lim-
ited information. Used to place the expert in a situation where he will be challenged.
Serve to observe the expert in problem-solving without danger of failure.
Linear Regression. A statistical procedure for predicting the value of a dependent
variable from an independent variable when the relationship between the variables
can be described with a linear model.
Machine Learning. Machine learning refers to the ability of computers to automati-
cally acquire new knowledge—learning from, for example, past cases or experience,
from the computer’s own experiences, or from exploration.
Many-on-Many Interviews. Interview sessions in which several knowledge engineers
interact with several experts.
Mashup. A program or Web site that combines two or more online software products,
typically integrating multiple application programming interfaces, and thus combining
the data or functionality of the original services.
Market Basket Analysis. An algorithm that examines a long list of transactions in
order to determine which items are most frequently purchased together.
Metadata. Data about data. Metadata describes how, when, and by whom a particular
set of data was collected and how the data are formatted. Metadata are essential for
understanding information stored in data warehouses.
Metaphor. One thing conceived as representing another—a symbol.
Method of Least Squares. A statistical method of finding the best-fitting straight line
or other theoretically derived curve for a group of experimental data points.
Methodology. A body of practices, procedures, and rules used by those who work in
a discipline or engage in an inquiry; a set of working methods.
Metrics. A set of measures used to evaluate performance of a task or an organiza-
tional unit.
Mission-Critical Objectives. A set of organizational goals that include both the
purpose of the organization and its scope of operations critical to operationalize and
accomplish the set of organizational goals. They must be measurable, specific, ap-
propriate, realistic, and timely.
Model-based Reasoning (MBR). An intelligent reasoning technique that uses a
model of an engineered system to simulate its normal behavior. The simulated opera-
GLOSSARY 345
tion is compared to the behavior of a real system and noted discrepancies can lead
to a diagnosis.
Motor Skills Expertise. Physical rather than cognitive knowledge. This type of
knowledge is difficult for knowledge-based systems to emulate. Examples include
riding a bicycle and hitting a baseball.
MYCIN. Early knowledge-based system developed in the early 1970s. Developed
to diagnose and specify treatments for blood disorders through a Q&A session with
a physician. It is the most significant and renowned research system, for it pioneered
the separation of the knowledge from the way it is used.
Narrative. A narrated account; a story.
Narrowcast. To send data to a specific list of recipients. Cable television is an ex-
ample of narrowcasting since the cable TV signals are sent only to homes that have
subscribed to the cable service. In contrast, network TV uses a broadcast model in
which the signals are transmitted everywhere, and anyone with an antenna can re-
ceive them.
Natural Language. Programming computers to understand natural human languages.
Natural Language Processing (NLP). A branch of artificial intelligence that deals
with analyzing, understanding, and generating the languages that humans use natu-
rally in order to interface with computers in both written and spoken contexts using
natural human languages instead of computer languages.
Normalization. In data processing, a process applied to all data in a set that
produce a specific statistical property. For example, monthly expenditures can
be divided by total expenditures to produce a normalized value that represents
a percentage.
Object Oriented Programming (OOP). Refers to a special type of programming
that combines data structures with functions to create reusable objects.
Observables. A physical property, such as weight or temperature, that can be observed
or measured directly as distinguished from a quantity, such as work or entropy, that
must be derived from observed quantities.
Observational Elicitation. The process of observing an expert perform a task in
order to learn his/her process for performing the task. Can be quiet observation or
interactive.
One-line Diagram. Drawing depicting the connectivity of an engineered system.
One-on-Many Interviews. Interview sessions in which one knowledge engineer
interacts with several experts.
One-on-One Interview. An interview in which one knowledge engineer and one
expert participate.
346 GLOSSARY
Online Analytical Processing (OLAP). A category of software tools that provides analysis
of data stored in a database. OLAP tools enable users to analyze different dimensions of
multidimensional data. For example, it provides time series and trend analysis views. The
chief component of OLAP is the OLAP server, which sits between a client and a database
management system (DBMS). The OLAP server understands how data are organized in
the database and has special functions for analyzing the data. There are OLAP servers
available for nearly all the major database systems. In essence OLAP enables to sum-
marize, aggregate, or selectively extract data from different points of view.
Ontology. Description of the concepts and relationships that can exist for an agent
or a community of agents.
Open-ended Questions. Questions asked of the expert, which require a narrative
and/or long explanation. Used for gathering general domain knowledge.
Open Source. A broad general type of software license that makes source code avail-
able to the general public with relaxed or nonexistent copyright restrictions.
Output-Input-Middle Method. Method used to organize knowledge elicited during
expert interviews.
Predictive Data Mining. Models may or may not explain relationships. Primarily,
they make predictions of output conditions given a set of input conditions.
Probability. A number expressing the likelihood that a specific event will occur,
expressed as the ratio of the number of actual occurrences to the number of possible
occurrences.
Procedural Knowledge. Focuses on beliefs relating sequences of steps or actions to
desired (or undesired) outcomes. It may be viewed as “know-how.”
Proposition. A statement that affirms or denies something.
Prospector. A system that assists geologists in identifying geological formations that
may contain mineral deposits. The program elicited, preserved, and reused geologic
formation knowledge to assist in mineral exploration.
Prototype. An original type, form, or instance serving as a basis or standard for later
stages.
Proxy Server. A server that sits between a client application, such as a Web browser,
and a real server. It intercepts all requests to the real server to see if it can fulfill the
requests itself. If not, it forwards the request to the real server.
Qualiative KM Assessments. Evaluation of knowledge management using percep-
tions and interpretations rather than based on numerical scores.
Quantitative KM Assessments. Produce specific numerical scores indicating how
well an organization, an organizational subunit, or an individual is performing with
respect to KM. They may be based on a survey, in financial terms, such as the ROI or
cost savings, or may include such ratios or percentages as employee-retention rate.
GLOSSARY 347
Query. A request for information from a database.
Query by Example (QBE). In this method, the system presents a blank record and
lets you specify the fields and values that define the query.
Query Language. Many database systems require you to make requests for informa-
tion in the form of a stylized query that must be written in a special query language.
This is the most complex method because it forces you to learn a specialized language,
but it is also the most powerful.
Radio Frequency Identification (RFID). The combination of radio broadcast with
radar technology and consists of two parts. First is an RFID tag, which is an inte-
grated circuit that modulates and demodulates a radio frequency signal and processes
and stores information. The second part is an antenna that receives and transmits the
signal.
Random Search. A search that has no specific pattern, purpose, or objective.
Recontextualized Knowledge. Existing knowledge is re-created using alternative
and innovative knowledge technologies and mechanisms.
Repertory Grids. Table associating attributes of several subjects with respect to two
diametrically opposed extremes. Used to organize elicited knowledge. Its use can be
easily automated.
Retrieval. Obtaining a case from the library that matches the description of the cur-
rent problem.
Reverse Engineering. The process of recreating a design by analyzing a final product.
Reverse engineering is common in both hardware and software.
Rocketry. The science and technology of rocket design, construction, and flight.
Role Reversal. Elicitation technique where the expert and the knowledge engineer
exchange roles, and the expert interviews the knowledge engineer. Can serve to verify
already elicited knowledge.
Routines. The utilization of knowledge embedded in procedures, rules, and norms
that guide future behavior.
Search-Retrieve-Propose. The process upon which case-based reasoning is founded.
A case is sought, compared to the current problem, retrieved if it is similar, and its
solution proposed as the solution to the current problem.
Security. Refers to techniques for ensuring that data stored in a computer cannot be
read by unauthorized users or compromised. Most security measures involve data
encryption and passwords. Data encryption is the translation of data into a form that
is unintelligible without a deciphering mechanism. A password is a secret word or
phrase that gives a user access to a particular program or system.
Semantic Networks. A directed graph in which concepts are represented as nodes
and relations between concepts are represented as links.
348 GLOSSARY
Server. A computer or device on a network that manages network resources. For
example, a file server is a computer and storage device dedicated to storing files. Any
user on the network can store files on the server. A print server is a computer that man-
ages one or more printers, and a network server is a computer that manages network
traffic. A database server is a computer system that processes database queries.
Shell. Development environment designed to exercise domain knowledge expressed
as rules and to arrive at solutions or answers to questions.
Shifting character or context. Fictional anecdotes where the characters may be
shifted to study the new perspective of the story.
Simple Knowledge. Knowledge related to one basic area.
Slot. The attribute of a frame to which a value or set of values is assigned. Consists
of facets.
Socialization. The integration of multiple streams of tacit knowledge for the creation
of new knowledge through social means. Socialization is the process of sharing ex-
periences and thereby creating tacit knowledge, such as shared mental models and
technical skills.
Social Networks. A group of individuals with ties or interpersonal relationships that
may be based on information technology.
Solution Space. Contains the actions, states, or beliefs that represent the status of the
problem. The solution is a sequence of steps through these actions, beliefs, or states
starting from the initial state to the goal state.
Specialization. Could be used when a child frame is more specific than its parent
frame; the basis for inheritance, or in an organizational context where individuals
focus on, and gain expertise in, a particular area.
Specific Knowledge. Knowledge that is particular to a situation such as a context
area or a specific context. Is usually possessed by a limited number of individuals
and is expensive to transfer.
Specific Problem-solving Interview Sessions. Interview sessions where the objec-
tive is to gather specific problem-solving knowledge.
Stakeholder. One who has a share or an interest as in an enterprise.
Static Web Documents. Web documents designed a priori, consisting of nonactive
HTML.
Stemming Algorithm. Used to remove the suffix of a word.
Stoplists. Used to eliminate words that are not good concept descriptions. A group
of words that are not considered to have any indexing value. These include common
words such as “and,” “the,” and “there.”
Storage Law. Data storage capacity doubles every 9 months. This law has been in
operation for over 10 years now. Storage Law is related to Moore’s Law, which states
that the number of transistors on CPUs doubles about every 18 months.
GLOSSARY 349
Storytelling. The act or practice of telling a story.
Strategic Knowledge. Knowledge regarding the long-term positioning of the orga-
nization in terms of its corporate vision and strategies for achieving that vision.
Structured Knowledge. Knowledge that is best represented through attribute value
pairs; knowledge not conditional in nature.
Support Knowledge. Relevant knowledge, usually related to organizational infra-
structure and facilitates day-to-day operations.
Surrogates. A substitute that replaces someone or something equivalent.
Symbol Manipulation. Refers to using symbols for solving problems; basis of
symbolic AI.
Synergy. The interaction of two or more agents or forces so that their combined effect
is greater than the sum of their individual effects.
Systematic Blind Search. Follows a systematic, exhaustive method to find target.
Does not use any knowledge. Can be very time-consuming.
Tacit Knowledge. Includes insights, intuitions, and hunches. It is difficult to express
and formalize and therefore difficult to share.
Tactical Knowledge. Knowledge used to decide course of action to achieve a specific
goal in a dynamically changing environment; pertains to the short-term positioning
of the organization relative to its markets, competitors, and suppliers.
Talking Head. A talking head or avatar is an image selected to represent oneself.
Talking heads could be a photograph, a cartoon character, or an animated image driven
by the user’s voice including lip-synchronization.
Task Interdependence. The extent to which the subunit’s achievement of its goals
depends on the efforts of other subunits.
Task Uncertainty. The extent to which the organizational subunit encounters difficulty
in predicting the nature of its tasks. High uncertainty implies changing problems and
tasks, which reduces the unit’s ability to develop routines.
Teachability. Reflects the extent to which the knowledge can be taught to other in-
dividuals through training, apprenticeship, and so on.
Technically Specific Knowledge. Deep knowledge about a specific discipline or
content area. It includes knowledge about the tools and techniques that may be used
to address problems in that area.
Ten-fold Cross Validation. Cross-validation is a method for assessing how the re-
sults of a statistical analysis will generalize to a data set. In particular for predictive
models, part of the data set is used to train the model, and the remaining is used to
test the model. For tenfold cross validation the data set of n observations is divided,
with random selection of examples into ten partitions (folds) of equal sizes, each
350 GLOSSARY
of size n/10, of which 9 partitions are used to train the model and one is used to test
the model.
Term Frequency Inverse Document Frequency (TFIDF). Highlights terms that are fre-
quently used in one document but infrequently used across the collection of documents.
Text Mining. Automatically reading large documents of text and deriving knowledge
from the process.
Touchpoints. Refers to the steps in a business process, like a purchase transaction,
or software when it interfaces with the customer directly.
Uniform Resource Locator (URL). A format for specifying Internet addresses; the
global address of documents and other resources on the World Wide Web.
Universalistic View of KM. Implies that there is a single best approach of managing
knowledge that should be adopted by all organizations in all circumstances.
User Interface. The screen and dialogue format seen by the user when working with
a particular computer program.
Value-added Products. New or improved products that provide a significant ad-
ditional value as compared to earlier products. Value-added products benefit from
knowledge management due to increased knowledge or enhanced organizational
process innovation.
Variable. A symbol or name that represents a concept that can take one or more dif-
ferent values.
Virus. A program or piece of code that is loaded onto your computer without your
knowledge and runs against your wishes. Viruses can also replicate themselves. All
computer viruses are manmade. A simple virus that can copy itself over and over
again is relatively easy to produce. Even such a simple virus is dangerous because it
will quickly use all available memory and bring the system to a halt. An even more
dangerous type of virus is one capable of transmitting itself across networks and
bypassing security systems.
Weak-theory Domains. Domains where robust theoretical explanations do not exist,
or if they exist, they contain uncertainty.
Web 2.0. A term describing changing trends in the use of World Wide Web technology
and Web design that aims to enhance creativity, information sharing, collaboration,
and functionality of the Web.
Web 3.0. Refers to the semantic Web and personalization, or the use of autonomous
agents to perform some tasks for the user.
Web Content Mining. Discovers what a Web page is about and how to uncover new
knowledge from it.
Web Crawling. Refers to the use of computer programs that visit Web sites continu-
ously and regularly, acquiring information for use in search engines.
GLOSSARY 351
Web Mining. Web crawling with online text mining.
Web Structure Mining. Examines how Web documents are structured; attempts to
discover the model underlying the link structures of the Web.
Web Usage Mining. Identification of patterns in user navigation through Web surfing.
Wicked Problem. Describes a problem that is one-of-a-kind and is difficult or impossible
to solve, or one that may have contradicting requirements. For example, ERP implemen-
tations offer have both financial and accuracy requirements that contradict each other:
accuracy requires more time to implement, but low cost requires expedited completion.
Wiki. A page or collection of Web pages designed to enable anyone who accesses it
to contribute or modify content using a simplified markup language.
Workflow Management System (WfMS). A system that provides procedural au-
tomation of a business process by managing the sequence of work activities and by
managing the required resources (people, data, and applications) associated with the
various activity steps. Computer programs that provide a method of capturing the
steps, which lead to the completion of a project within a fixed time frame.
World Wide Web (WWW). A format that enables large-scale storage of documents
to be easily accessed by a user via a browser.
XCON. One of the earliest commercially successful systems, XCON assists in the
configuration of newly ordered VAX computer systems. Developed by Digital Equip-
ment Corporation (DEC) in conjunction with Carnegie-Mellon University XCON
elicited, preserved, and reused the knowledge of human configurators of computer
systems in order to automate and duplicate their functions.
352 GLOSSARY
SOURCES
The American Heritage Dictionary of the English Language, 4th ed. 2000. Boston: Houghton Mifflin.
Cambridge Dictionaries Online. Cambridge University Press. http://dictionary.cambridge.org/.
CRoss Industry Standard Process for Data Mining (CRISP-DM). http://www.crisp-dm.org/.
Dictionary.com. http://dictionary.reference.com/.
Flake, Gary William. 2002. Glossary. In The computational beauty of nature: Computer explorations of frac-
tals, chaos, complex systems, and adaptation. Cambridge, MA: The MIT Press. http://mitpress.mit.edu/
sites/default/files/titles/content/cbnhtml/glossary-intro.html.
Geek.com. www.geek.com.
Knowledgepoint Australia. www.knowledgepoint.com.au.
Nonaka, Ikujiro, and Takeuchi, Hirotaka. 1995. The knowledge creating company: How Japanese companies
create the dynamics of innovation. New York: Oxford University Press.
OneLook® Dictionary Search. www.onelook.com.
The Oxford American desk dictionary and thesaurus, 2d ed. 2001. New York: Berkley Books.
Stottler Heck Associates, Inc. 2002. Glossary. http://www.stottlerhenke.com/ai_general/glossary.htm.
Webopedia. www.webopedia.com.
Whatis.com. http://whatis.techtarget.com/.
Wikipedia—The Free Encyclopedia. http://en.wikipedia.org/wiki/Main_Page.
YourDictionary.com—The last word in words. www.yourdictionary.com.
353
About the Authors
Dr. Irma Becerra-Fernandez is the Vice President for Engagement and Professor of
Management Information Systems at Florida International University College of Business
Administration. She is the former Vice-Provost for Academic Affairs and the Founder and
Program Co-Chair of the 2010 and 2011 Americas Venture Capital Conference. She is
also a Fellow and former Director of the Pino Global Entrepreneurship Center. She was
the MIT Sloan Visiting Scholar with the Center for Information Systems Research in the
spring of 2009. Also, she was the 2007 Kauffman Entrepreneurship Professor.
Her research focuses on knowledge management (KM), KM systems, business
intelligence, enterprise systems, disaster management, and IT entrepreneurship. She
has studied and advised organizations, in particular NASA, about KM practices.
She founded the FIU Knowledge Management Lab ten years ago, and has obtained
funding as principal investigator for approximately $2 million from the National Sci-
ence Foundation, NASA (Headquarters, Kennedy, Ames, and Goddard Space Flight
Center), and the Air Force Research Lab to develop organizational KM strategies
and innovative KM systems.
She has published extensively in leading journals including the Journal of MIS,
Decision Sciences, Communications of the ACM, European Journal of Operational
Research, IEEE Transactions on Engineering Management, ACM Transactions on
Internet Technology, Knowledge Based Systems, International Journal of Expert
Systems Research & Applications, and others. She is an author of the books Business
Intelligence: Practices, Technologies, & Management (Wiley, 2010), and Knowledge
Management: Systems and Processes (M.E. Sharpe, 2015, 2010), and Knowledge
Management: Challenges, Solutions, and Technologies (Prentice Hall, 2004) and co-
editor of Knowledge Management: An Evolutionary View (M.E. Sharpe, 2008). She
has delivered many invited presentations and keynote speeches at many NASA Cen-
ters, the NAVY Research Lab, universities around the world, and many international
conferences with both an academic and a practitioner focus. She was the prior faculty
director for the Masters in MIS and the MIS PhD Program, and serves on the editorial
board of the International Journal of Knowledge Management, International Journal
of Knowledge and Learning and the International Journal of Mobile Learning and
Organisation, among others. She also served as the Americas Region representative
to the Association for Information Systems Executive Council.
Dr. Becerra-Fernandez was the recipient of the 2013 Educator of the Year by
Great Minds in STEM, the 2012 In the Company of Women Award for Educa-
354 ABOUT THE AUTHORS
tion and Research by the Mayor of Miami-Dade County, the 2011 Engineer of the
Year Award by the Association of Cuban Engineers, the 2004 Outstanding Fac-
ulty Torch Award, presented by the FIU Alumni Association, the 2006 FIU Faculty
Teaching Award and the 2001 FIU Faculty Research Award. Finally, Dr. Becerra-
Fernandez was the first female to receive a PhD from FIU’s Engineering Program.
She earned her PhD in 1994 in electrical engineering and her Master’s and Bachelor’s
degrees, also in electrical engineering, from the University of Miami.
Dr. Rajiv Sabherwal is the Department Chair of Information Systems, the Edwin &
Karlee Bradberry Chair in Information Systems, and the Executive Director of the
Information Technology Institute in the Walton College of Business at the University
of Arkansas. He previously taught at University of Missouri–St. Louis (where he was
appointed University of Missouri System Curators Professor in 2003), Florida State
University, and Florida International University. He was a Southeastern Conference
Academic Leadership Development Program Fellow for 2013–2014. He received the
National Impact Award at the Gateway to Innovation Conference in 2010, and was
selected for, and participated in, the University of Missouri Leadership Development
Program during 2003–2004.
Dr. Sabherwal is the editor-in-chief (since 2010) of IEEE Transactions on Engineering
Management, and serves as a member of the IEEE Technology Management Council’s
Board of Governors. He was the conference co-chair for the International Conference
on Information Systems (ICIS) (2010), and chair of the ICIS Executive Committee
(2011–2012). He serves or has previously served as senior editor for MIS Quarterly, senior
editor for a special issue of Information Systems Research, department editor for IEEE
Transactions on Engineering Management, and associate editor or editorial review board
member for several journals, including Management Science, MIS Quarterly, Informa-
tion Systems Research, IEEE Transactions on Engineering Management, and Journal
of Management Information Systems. He has also served as the program chair, doctoral
consortium chair, and chair for panels for several conferences, and as director of the PhD
program in business (2000–2011), chair and member of the University Student Grievance
Committee, member of the University Faculty Senate, and chair of Campus Review Team
for the Five-Year Review of Department of Mathematics and Computer Science.
Dr. Sabherwal has authored 47 refereed journal publications, chaired seven doctoral
dissertations, presented at conferences and universities worldwide, and created sev-
eral new courses at various levels. He has coauthored three textbooks on knowledge
management and business intelligence. His research interests include strategic align-
ment, information systems planning, knowledge management, business intelligence,
and social aspects of systems development. He received his Bachelor’s degree in
engineering (electronics) from Regional Engineering College, Bhopal, India, a post-
graduate diploma degree in management from the Indian Institute of Management,
Calcutta, and a PhD in business administration from the University of Pittsburgh. He
is a member of AIS, Academy of Management, IEEE, and INFORMS.
Dr. Sabherwal was the 2009 Fulbright-Queen’s School of Business Research
Chair at Queen’s University in Ontario, Canada. He is a Fellow of the Association of
Information Systems (inducted in 2008).
355
Index
Absorptive capacity, 49
Activeworlds, 261
Adaptability of employee, 76–77, 302, 303t
Advanced Quality and Reliability Information
System (AQUARIS), 109–13
Ajax, 243, 245
Alert system, 163, 166t
Algorithmic characteristics, 95
All Partners Access Network (U.S. Department of
Defense), 193 (box), 194
Amazon, 159, 206 (box), 245
Analogy-based reasoning, 98
Android, 259
Apriori knowledge, 219, 278
ARPAnet, 258–59
Artificial Intelligence (AI)
knowledge application systems, 94–96
knowledge management technologies, 51
Artificial neural networks (ANN), 205, 206
Assessment of knowledge management
approaches to, 306–10
Balanced Scoreboard, 306–7
benchmarking, 306
Intangible Assets Monitor Framework, 307
knowledge management maturity, 309–10
Most Admired Knowledge Enterprises
(MAKE) survey, 307 (box)
organizational examples, 309 (box), 310 (box)
real options approach, 308
Skandia Navigator, 307–8, 312n3
assessment team, 305–6
classification of, 295–98
assessment focus, 298
assessment method, 296–98
assessment timing, 295–96
Artificial neural networks
classification of (continued)
assessment tool, 295 (box)
communities of practice (CoP), 297 (box)
qualitative KM assessment, 297, 298f
quantitative KM assessment, 297–98
importance of, 294–95
knowledge assessment, 300–302
knowledge management impacts, 302–4, 305t
competitive advantage, 304, 305t
direct effects, 304, 305t
employee impacts, 302, 303t
indirect effects, 304, 305t
organizational impacts, 302, 303t
organizational performance, 303–4, 305t
organizational products, 302, 304t
knowledge management processes, 299–300
metrics, 300
recommendations for, 310–11
study guide
application exercises, 312
research summary, 311
review questions, 311–12
Associational expertise, 29
Association technique, 219, 221t
Associative networks, 136
Athens, 234–36
Autonomous Land Vehicle in a Neural Network
(ALVINN), 147
Avatars, 261
Back-propagation algorithm, 207
Balanced Scoreboard, 306–7
Bank of Montreal, 207–8
Behavioral cloning, 147
356 INDEX
Bell, Alexander Graham, 264
Benchmarking, 306
Best practices
knowledge management impacts, 83
knowledge sharing systems, 164, 166t
Big Blue, 95
Big data, 324, 325 (box)
BIND (Berkeley Internet Name Daemon), 259
Blogs, 256–57
BlueReach (IBM), 184–87
Boeing Corporation, 322 (box)
BP Amoco Chemical Company, 52 (box)
Brainstorming, 199–203, 204 (box)
Brainstorming camps, 65
Brainstorming retreats, 40
Bridgestone Corporation, 79
British Petroleum (BP), 80 (box), 81, 264
British Telecom, 84
Buckman, Bob, 77
Buckman Laboratories International, Inc., 77, 83
Business intelligence (BI), 40, 41t
Business processes, 164
CABER (Lockheed Martin Corporation), 102
Campaign management software, 226
Case-based design aid (CBDA), 101
Case-based reasoning (CBR), 97–99
Case-Method Cycle, 99–101
Center for Advancing Microbial Risk Assessment
(CAMRA), 155, 157 (box)
Central Agricultural University, India, 103
Chevron Corporation, 84
Chief Information Officer (CIO), 290–93
Chief Knowledge Officer (CKO), 290, 291–92,
293 (box)
Chief Learning Officer (CLO), 290, 291
Churn reduction, 206
Cisco Systems, Inc., 67 (box)
CiteSeer, 180
CLAIM project (GE Healthcare), 104, 114,
116–18
Classification research, 96
CLAVIER (Lockheed Corporation), 102
Clickstream analysis, 224
Clients, 160
Cloud computing, 320–24, 325 (box)
Clustering technique, 219, 221t
Clusty, 234
CmapTools, 136, 138–39, 140f, 151n1
Cognitive imitation, 148
Collaborations
emergent knowledge management practices,
256–58
knowledge management future, 320–24, 325
(box)
knowledge sharing systems, 158–59
Collaborative Agile Knowledge Engine (CAKE),
103
Collaborative computing, 158–59
Collective Intelligence, 245, 256
Colliers International, 292
Combination process, 59, 64–65, 199
Common knowledge, 48–49, 51t
absorptive capacity, 49
shared domain knowledge, 49
Communities of practice (CoP)
assessment of knowledge management, 297
(box)
employee learning impact, 75–76
knowledge management infrastructure, 45–47
as knowledge reservoir, 33
knowledge sharing systems, 189–94
Community-based design, 320
Compaq Computer Corporation, 101–2
Competitive advantage
assessment of knowledge management, 304, 305t
knowledge management impacts, 86–87
lessons learned system (LLS), 164, 165 (box),
166
theoretical knowledge, 30 (box)
Complex knowledge, 30
Computer-aided design (CAD), 133–34
Concept maps, 134–39
CONNEX (Hewlett-Packard), 172t, 173 (box)
Constraint-based reasoning, 98, 99t
Context-and-technology-specific knowledge, 27
Context-based Intelligent Tactical Knowledge
Acquisition (CITKA), 143–45
Context-based Reasoning (CxBR), 139–45
Contextually specific knowledge, 27, 28t, 31
Contingency view, 269–70
Copyrights, 319
Cross-Industry Standard Process for Data Mining
(CRISP-DM), 210, 214, 215f, 216f
INDEX 357
Cross-links, 136
Crowdfunding, 323 (box)
Crowdsourcing, 320, 321 (box)
Customer Relationship Management (CRM), 208,
225–27
Customer touchpoints, 226
Daimler-Chrysler, 45, 210
Dannon, 54, 280 (box)
Darty, 104, 113–14, 115f
Data, 17–22
defined, 17
illustrative models, 20f, 21f
knowledge-information relationship, 19–22
Database, 47
Data Mining Group (DMG), 214
Data mining techniques
knowledge discovery, 59
knowledge discovery systems, 203–229
knowledge sharing systems, 174
Data preparation, 212–13, 230
Data warehouse, 47, 211
Decision-making paradigm, 315–16
Decision trees, 218t, 219, 220t, 230, 231f,
232
Declarative knowledge, 24–25, 28t
DeepGreen Financial, 67 (box)
Derr, Kenneth T., 84
Deutsche Bank, 84
Development Forum (World Bank), 191–92
Diagrammatic reasoning, 98, 99t
Direct effects, 84, 304, 305t
Direction process, 63–64, 66
Disjunctions, 120
Disney Corporation, 201 (box)
Document management system, 155–57
Dolan, Tom, 76 (box)
Domain expert, 96
Domain knowledge, 96
Domain Name System, 259
Dorsey, Jack, 250
Dow Chemical Company, 317 (box)
Downsizing, 7
Dresdner Kleinwort Wasserstein, 258 (box)
Dunbar, Robin, 258
DuPont, 322 (box)
Dynamic organizational capabilities, 82
eBags, 208
Economy of scale, 85–86
Economy of scope, 85–86
Edge platform (Verdande Technology), 103
Effectiveness of organizations, 78, 79, 80 (box),
302, 303t
Efficiency of organizations, 78, 80–81, 302,
303t
Electronic Privacy Information Center (EPIC),
326 (box)
Eli Lilly and Company, 322 (box)
Emergent knowledge management practices
avatars, 261
collaborative content creation
blogs, 256–57
Collective Intelligence, 245, 256
folksonomy, 246–47, 257
mashup, 257
microblogs, 257–58
online awareness, 257–58
Wikipedia, 256
Wikis, 258 (box)
groupware, 264
information technology (IT), 264–65
JavaScript, 245
open source software development, 258–61
social networking, 247–55
application of, 255 (box)
Facebook, 248, 249f, 250, 264
FaceTime, 254
Friendster, 247, 248f
GooglePlus, 253, 254f
LinkedIn, 250, 251f
Myspace, 247–48, 249f
Orkut, 250
Pinterest, 254
Skype, 253
Twitter, 250–51, 252f, 257–58
YouTube, 250, 252f
study guide
application exercises, 266
research summary, 266
review questions, 266
virtual worlds, 261–64
Web 1.0 technologies, 244f
Web 2.0 technologies, 243–47
Web 3.0 technologies, 246–47
358 INDEX
Employees
adaptability of, 76–77, 302, 303t
assessment of knowledge management, 302, 303t
job satisfaction, 77–78, 302, 303t
knowledge management impacts, 75–78
learning impact, 75–76, 302, 303t
organizational downsizing, 7
turnover, 7–8
Encryption, 319
End-user identification, 119
Enterprise application integration (EAI), 225
Enterprise resource planning (ERP), 48, 315
Ernst & Young, 155, 156 (box)
Eureka (Xerox Corporation), 168 (box)
Exchange process, 62, 66
Exemplar-based reasoning, 98
Experience management, 9
Expertise
associational expertise, 29
defined, 27–28
motor skills expertise, 29
tacit knowledge, 26
theoretical expertise, 29, 30 (box)
Expertise locator system (ELS)
knowledge discovery systems, 232–33
knowledge management infrastructure, 48
knowledge sharing systems, 154, 163, 170–87
Expert Seeker (NASA), 172t, 180, 182–84, 185f,
232–34
Explicit knowledge, 25–26, 28t, 30
Externalization process
employee learning impact, 75
knowledge capture, 61, 65–66
Facebook, 248, 249f, 250, 264, 326 (box)
FaceTime, 254
Facilitator, 199, 201
Fault diagnosis systems, 102–3
File transfer protocol (FTP), 177–78
Firestone Corporation, 79
Firewalls, 319
Flat case library, 101
Folksonomy, 246–47, 257
Ford Motor Company, 79, 82–83
FormTool (General Electric), 102–3
Frames, 96
Freedom of Information Act, 257
Free Software Foundation, 259
Friendster, 247, 248f
GenAID (Westinghouse Electric Corporation), 96,
104–5
General Electric
CLAIM project, 104, 114, 116–18
FormTool, 102–3
Generalized Rule Induction (GRI), 219
General knowledge, 26, 28t
General Problem Solver (GPS), 95
Giant Eagle, Inc., 44
Google, 234
Google Alerts, 163
Google Chrome, 259
Google Chromium, 259
Google-Docs, 156
GooglePlus, 253, 254f
Graham, Ted, 31 (box)
Grants.gov, 163
Graphical user interface (GUI), 117
Group Decision Support Software (GDSS), 204
(box)
Groupe Danone, 54, 280 (box)
Groupware, 158–59, 264
Halliburton Company, 44
Harrah’s Entertainment, 226
Help desk technologies, 101–2
Heuristics
knowledge application systems, 107
knowledge discovery systems, 230
Hewlett-Packard, 172t, 173 (box)
Hill and Knowlton, 31 (box), 45 (box)
Hilton Hotels, 226
Honda Motor Company, 65
Horton, Alex, 257
Human capital, 5, 34–35
Human computer interface (HCI), 223
Hyperlinks, 160
Hypertext Markup Language (HTML), 160
IBM Corporation
BlueReach (IBM), 184–87
communities of practice (CoP), 191
organizational storytelling, 129–30
ICL, 83, 88n2
INDEX 359
IDC, 74, 88n1
IDEATECTS Problem Solving Model, 203 (box),
205f
Incident report database, 163, 166t
Indirect effects, 84–87, 304, 305t
Information, 17–22
defined, 17–18
illustrative models, 20f, 21f
knowledge-data relationship, 19–22
Information retrieval (IR), 222
Information technology (IT)
emergent knowledge management practices,
264–65
knowledge management infrastructure, 47–48, 51t
in knowledge management systems, 8
Infrastructure as a Service (IaaS), 324
InnoCentive, 322 (box)
Innovation of organization, 78, 81–82, 302, 303t
Innovators, 199, 201
Inquiring organizations, 315, 329n1
Instance-based reasoning, 98
Instituto Boliviano de Comercio Exterior (IBCE),
53 (box)
Intangible Assets Monitor Framework, 307
Intellectual capital
defined, 5
human capital, 5, 34–35
as knowledge reservoir, 17, 34–35
organizational capital, 5, 34–35
social capital, 5, 34–35
Intellectual property rights, 316–19
Internalization process
employee learning impact, 75
knowledge capture, 61, 65–66
Internet
knowledge management infrastructure, 47
website resources, 234, 261
Inter-page structure, 224
Intra-page structure, 224
Intraspect Software, Inc., 31 (box)
Iterative process, 219
JavaScript, 245
jCOLIBRI platform, 103
Job satisfaction, 77–78, 302, 303t
J.P. Morgan Chase & Company, 81
Just-in-time manufacturing, 109
Kaidara, Inc., 114
Kartoo, 234
Kasparov, Boris, 95
Katzenbach Partners, 52 (box), 54
Knowledge
assessment of, 300–302
classification of, 24–31
combinations, 27, 28t
complex knowledge, 30
context-and-technology-specific knowledge,
27
contextually specific knowledge, 27, 28t, 31
declarative knowledge, 24–25, 28t
expertise, 26, 27–30
explicit knowledge, 25–26, 28t, 30
general knowledge, 26, 28t
knowledge specificity, 30–31
procedural knowledge, 24–25, 28t
simple knowledge, 30
specific knowledge, 26–27, 28t, 30–31
strategic knowledge, 30
summary, 28t
support knowledge, 30
tacit knowledge, 25–26, 28t, 30
tactical knowledge, 30
technically specific knowledge, 26–27, 28t,
31
data, 17–22
defined, 17–22
expertise, 26, 27–30
associational expertise, 29
motor skills expertise, 29
theoretical expertise, 29, 30 (box)
information, 17–22
knowledge reservoirs, 32–35
artifacts, 32–33
communities of practice (CoP), 33
groups of people, 32
illustrative model, 33f
individual people, 32
intellectual capital, 17, 34–35
interorganizational networks, 34
knowledge repository, 33
organizational practices, 32
organizational technologies, 32
organizational units, 33
organizations, 33–34
360 INDEX
Knowledge (continued)
objective perspective
as capability, 24
illustrative model, 23f
as information access, 23, 24
as objects, 23, 24
organizational examples, 26–27, 30 (box), 31 (box)
probability relationship, 19
study guide
application exercises, 36–37
research summary, 35
review questions, 35–36
subjective perspective
illustrative model, 23f
as practice, 22–23, 24
as state of mind, 22, 24
Knowledge accidents, 50
Knowledge and Skills Management System
(National Security Agency), 172t, 173 (box)
Knowledge application, 40, 62–64, 69
Knowledge application systems
Artificial Intelligence (AI)
algorithmic characteristics, 95
characteristics of, 94–96
classification research, 96
General Problem Solver (GPS), 95
machine learning research, 96
natural language, 95
solution space, 95
case-based reasoning (CBR), 97–99
analogy-based reasoning, 98
constraint-based reasoning, 98, 99t
diagrammatic reasoning, 98, 99t
exemplar-based reasoning, 98
instance-based reasoning, 98
model-based reasoning, 98, 99t
organizational example, 94 (box)
case studies, 104–21
CLAIM project (GE Healthcare), 104, 114,
116–18
Darty, 104, 113–14, 115f
GenAID (Westinghouse Electric Corporation),
96, 104–5
Out-of-Family Disposition process (NASA),
99, 118–21
SBIR/STTR Online System (SOS) Advisor, 96,
104, 106–9
Knowledge application systems (continued)
Total Recall System (National Semiconductor
Corporation), 99, 109–13
classification of
fault diagnosis systems, 102–3
help desk technologies, 101–2
organizational examples, 101–4
defined, 66
developmental process
case-based design aid (CBDA), 101
Case-Method Cycle, 99–101
flat case library, 101
disjunctions, 120
end-user identification, 119
graphical user interface (GUI), 117
heuristics, 107
just-in-time manufacturing, 109
limitations of, 122
organizational examples, 67 (box)
rule-based systems
characteristics of, 96
domain expert, 96
domain knowledge, 96
frames, 96
production rules, 96
summary, 99t
study guide
application exercises, 123
research summary, 122–23
review questions, 123
technologies for, 94–99
Knowledge assessment. See Assessment of
knowledge management
Knowledge-based products, 82, 83, 302, 304t
Knowledge capture, 40, 60–61, 62 (box), 69
Knowledge capture systems
computer-aided design (CAD), 133–34
concept maps, 134–39
associative networks, 136
CmapTools, 136, 138–39, 140f, 151n1
cross-links, 136
knowledge representation, 134, 135f,
136
learning psychology theory, 136
propositions, 134, 135f, 136
semantic networks, 134, 135f, 136
system example, 137f
INDEX 361
Knowledge capture systems (continued)
Context-based Reasoning (CxBR), 139–45
knowledge representation, 139–43
mission context, 142–43
system example, 143–45
defined, 11, 65–66, 127–29
knowledge-elicitation techniques, 132–33
narratives, 132–33
organizational example, 128 (box)
prototypes, 128
research trends, 146–50
observational learning, 146–48
radio frequency identification (RFID),
148–50, 151n3
stakeholders, 133
storytelling, 127, 128 (box), 129–33
study guide
application exercises, 150–51
research summary, 150
review questions, 150
system design, 133–34
tacit knowledge, 129–32
utilization barriers, 145–46
Knowledge creation, 40
Knowledge discovery, 40, 59–60, 69
Knowledge discovery in databases (KDD), 204
See also Data mining techniques
Knowledge discovery systems
case studies, 229–36
Athens, 234–36
expertise locator system (ELS), 232–33
Expert Seeker (NASA), 232–34
real estate appraisal systems, 229–32
combination process, 199
Customer Relationship Management (CRM),
208, 225–27
applications of, 225–27
campaign management software, 226
customer touchpoints, 226
enterprise application integration (EAI), 225
data mining guidelines, 214, 216–21
apriori knowledge, 219
association technique, 219, 221t
clustering technique, 219, 221t
decision trees, 218t, 219, 220t, 230, 231f, 232
Generalized Rule Induction (GRI), 219
heart disease diagnostic, 219–20, 221t
Knowledge discovery systems
data mining guidelines (continued)
iterative process, 219
Memory-based Reasoning (MBR), 218t, 219
non-statistical techniques, 216, 217, 218t, 219
statistical techniques, 216–17
data mining techniques, 203–229
applications, 205–6
artificial neural networks (ANN), 205, 206
back-propagation algorithm, 207
churn reduction, 206
Customer Relationship Management (CRM),
208, 225–27
data preparation, 230
guidelines for, 214, 216–21
heuristics, 230
knowledge discovery in databases (KDD), 204
market basket analysis, 205, 206 (box)
organizational examples, 206 (box), 209
(box), 210 (box)
paired leaf analysis, 230
predictive DM techniques, 205–6
query, 232
tree construction, 230
Web content mining, 208, 224, 232–34
defined, 11, 64–65, 198–99
recontextualized knowledge, 199
socialization process for, 199–203, 204 (box)
brainstorming, 199–203, 204 (box)
facilitator, 199, 201
innovators, 199, 201
lateral thinking, 201–2
organizational examples, 200 (box), 201
(box), 202 (box), 203 (box), 204 (box)
Storage Law, 227
study guide
application exercises, 237–38
research summary, 237
review questions, 237
system design, 209–14
business understanding requirement, 210–11
Cross-Industry Standard Process for Data
Mining (CRISP-DM), 210, 214, 215f,
216f
data analysis, 212
data collection, 211
data description, 211–12
362 INDEX
Knowledge discovery
system design (continued)
Data Mining Group (DMG), 214
data preparation steps, 212–13
data understanding requirement, 211–12
data verification, 212
data warehouse, 211
model deployment, 214
model validation, 213–14
rule induction algorithms, 213–14
ten-fold cross validation, 213–14
tacit knowledge, 199–203, 204 (box)
utilization barriers, 227–29
Web mining applications
clickstream analysis, 224
inter-page structure, 224
intra-page structure, 224
Online Analytical Processing (OLAP), 224
proxy server, 224
variables, 224
Web content mining, 224
Web Structure Mining, 223–24
Web usage mining, 224
Web mining techniques, 221–24
applications, 223–24
human computer interface (HCI), 223
information retrieval (IR), 222
linguistic analysis, 222, 223
natural language processing (NLP), 222, 223
statistical/co-occurrence analysis, 223
statistical/neural networks clustering, 223
stemming algorithm, 223, 233
stoplist, 223
term frequency inverse document frequency
(TFIDF), 222
text mining, 222
Web crawling, 222
Knowledge-elicitation techniques, 132–33
Knowledge engineering, 65–66
Knowledge management
defined, 4–5, 39–41
downsizing, 7
driving forces of, 5–8
domain complexity, 6
employee turnover, 7–8
market volatility, 6
responsiveness speed, 6
Knowledge management (continued)
experience management, 9
implementation challenges, 9–10
information technology (IT), 8
intellectual capital, 5
knowledge capture systems, 11
knowledge discovery systems, 11
knowledge management systems, 8–9
knowledge sharing systems, 11
methodologies, 7
mission-critical objectives, 7
organizational examples, 3, 4–5, 9 (box)
research overview, 10–12
study guide
application exercises, 12–13
research summary, 12
review questions, 12
Knowledge management foundations
brainstorming retreats, 40
business intelligence (BI), 40, 41t
illustrative model, 42f
knowledge creation, 40
knowledge management defined, 39–41
knowledge management infrastructure, 43–50,
51t, 53–54
common knowledge, 48–49, 51t
communities of practice (CoP), 45–47
components of, 43–50
database, 47
data warehouse, 47
defined, 41, 42–43
enterprise resource planning (ERP), 48
expertise locator system (ELS), 48
information technology (IT), 47–48, 51t
Internet, 47
management of, 53–54
organizational culture, 43–45, 51t
organizational examples, 45 (box),
46 (box)
organizational structure, 45–47, 51t
physical environment, 49–50, 51t
summary, 51t
knowledge management mechanisms, 50, 52
(box), 53–54
defined, 41, 50
management of, 53–54
organizational examples, 52 (box)
INDEX 363
Knowledge management foundations (continued)
knowledge management technologies, 50–54
Artificial Intelligence (AI), 51
defined, 41, 51–52
management of, 53–54
organizational example, 53 (box)
Web 2.0 technologies, 51
management of, 53–54
overview, 41–43
study guide
application exercises, 55
research summary, 54
review questions, 54–55
Knowledge management future
collaborations, 320–24, 325 (box)
big data, 324, 325 (box)
cloud computing, 320–24, 325 (box)
community–based design, 320
crowdfunding, 323 (box)
crowdsourcing, 320, 321 (box)
grassroots contributions, 320
organizational examples, 321 (box), 322
(box), 323 (box), 325 (box)
service models, 323–34
decision–making paradigm, 315–16
enterprise resource planning (ERP), 315
inquiring organizations, 315, 329n1
intellectual property rights, 316–19
copyrights, 319
encryption, 319
firewalls, 319
nondisclosure agreements, 318
organizational example, 317 (box)
patents, 318–19
trade secrets, 319
knowledge sharing barriers, 324, 326–27
knowledge–as–power, 326–27
leadership reluctance, 327
privacy concerns, 324, 326 (box)
study guide
application exercises, 328–29
research summary, 327–38
review questions, 328
wicked problems, 315
Knowledge management impacts
assessment of knowledge management, 302–4,
305t
Knowledge management impacts (continued)
employees, 75–78
adaptability, 76–77
communities of practice (CoP), 75–76
externalization process, 75
illustrative model, 78f
internalization process, 75
job satisfaction, 77–78
learning tradition, 75–76
organizational example, 76 (box)
socialization process, 75–76
illustrative models, 74f, 75f
organizational performance, 84–87
competitive advantage, 86–87
direct effects, 84
economy of scale, 85–86
economy of scope, 85–86
illustrative model, 86f
indirect effects, 84–87
organizational example, 85 (box)
organizational processes, 78–82
dynamic capabilities, 82
effectiveness, 78, 79, 80 (box)
efficiency, 78, 80–81
illustrative model, 82f
innovation, 78, 81–82
organizational example, 80 (box)
organizational products, 82–83
best practices, 83
illustrative model, 83f
knowledge–based products, 82, 83
value–added products, 82–83
study guide
application exercises, 88
research summary, 87
review questions, 87
Knowledge management influences
contingency view, 269–70
environmental characteristics, 277t, 278–79
knowledge management processes, 277t,
278–79
summary, 277t
illustrative models, 271f, 272f
knowledge characteristics
illustrative model, 275f
knowledge management processes, 274–76
knowledge management foundations, 270, 271f
364 INDEX
Knowledge management influences (continued)
illustrative model, 271f
knowledge management processes, 270, 271f
identification methodology, 279–83
illustrative example, 283–86
illustrative model, 271f
knowledge characteristics, 274–76
organizational characteristics, 276–78
organizational example, 280 (box)
task characteristics, 272–74
organizational characteristics, 277–78
business strategy, 277t, 278
knowledge management processes, 276–78
organizational size, 277–78
study guide
application exercises, 287–88
research summary, 286
review questions, 287
task characteristics
illustrative model, 274f
knowledge management processes, 272–74
task interdependence, 273–74
task uncertainty, 272–73, 274f
Knowledge management infrastructure, 43–50,
51t, 53–54
Knowledge Management Magazine, 74
Knowledge management mechanisms, 50, 52
(box), 53–54
Knowledge management processes
assessment of knowledge management, 299–300
knowledge management solutions, 58–64, 68t
See also Knowledge management influences
Knowledge management solutions
business strategy alignment, 70
illustrative model, 42f
knowledge application
defined, 62–63
direction process, 63–64, 66
knowledge management processes, 40, 62–64,
69
knowledge substitution process, 63–64
management of, 69
routines, 64, 66
knowledge capture
defined, 60–61
externalization process, 61, 65–66
internalization process, 61, 65–66
Knowledge management mechanisms
knowledge capture (continued)
knowledge management process, 40, 60–61,
62 (box), 69
learning tradition, 61, 65–66
management of, 69
knowledge discovery
combination process, 59, 64–65
data mining techniques, 59
defined, 59
knowledge management process, 40, 59–60,
69
management of, 69
organizational example, 60 (box)
socialization process, 59–60, 64–65
knowledge management processes, 58–64, 68t
components of, 58t, 58–64
defined, 41
illustrative model, 59f
knowledge application, 40, 62–64, 69
knowledge capture, 40, 60–61, 62 (box), 69
knowledge discovery, 40, 59–60, 69
knowledge sharing, 40, 42 (box), 61–62, 63
(box), 69
management of, 69–70
summary, 68t
knowledge management systems, 64–67, 68t
defined, 42
knowledge application systems, 66, 67 (box)
knowledge capture systems, 65–66
knowledge discovery systems, 64–65
knowledge engineering, 65–66
knowledge sharing systems, 66
management of, 69–70
organizational examples, 67 (box)
summary, 68t
knowledge sharing
exchange process, 62, 66
knowledge management processes, 40, 42
(box), 61–62, 63 (box), 69
management of, 69
organizational examples, 42 (box), 63 (box)
overview, 41–43
study guide
application exercises, 72
research summary, 70, 71f
review questions, 70–71
INDEX 365
Knowledge management systems, 8–9, 64–67, 68t
Knowledge management technologies, 50–54
Knowledge markets, 155
Knowledge repository
defined, 33
knowledge sharing systems, 154, 155–56
Knowledge sharing, 40, 42 (box), 61–62, 63 (box),
69
Knowledge sharing systems
case studies, 176–87
BlueReach (IBM), 184–87
Expert Seeker (NASA), 172t, 180, 182–84,
185f
Postdoc system (NASA), 176–77
Searchable Answer Generating Environment
(SAGE) Expert Finder, 172t, 177–80, 181f
classification of, 163–66
alert system, 163, 166t
best practices database, 164, 166t
business processes, 164
expertise locator system (ELS), 154, 163,
170–87
incident report database, 163, 166t
lessons learned system (LLS), 154, 164, 165
(box), 166–70
summary, 166t
Clients, 160
collaborative computing, 158–59
communities of practice (CoP), 189–94
organizational examples, 189, 190 (box),
191–92, 193 (box)
Thematic Groups (TGs), 190 (box)
computer–mediated knowledge, 159–60
defined, 11, 66, 155–59
document management system, 155–57
expertise locator system (ELS), 154, 163, 170–
87
characteristics of, 170–71, 172t
data mining techniques, 174
ontology of, 171, 173–74
organizational example, 173 (box), 175t
summary, 172t
taxonomy of, 171, 173–74, 175f
groupware, 158–59
Hyperlinks, 160
Hypertext Markup Language (HTML), 160
hypertext transfer protocol, 160
Knowledge sharing (continued)
knowledge markets, 155
knowledge repository, 154, 155–56
lessons learned system (LLS), 154, 164, 165
(box), 166–70
competitive advantage, 164, 165 (box), 166
defined, 166
essential tasks, 167–70
illustrative model, 167f
organizational examples, 165 (box), 168 (box)
limitations of, 187–89
not–invented–here syndrome, 161
organizational examples, 155, 156 (box), 157
(box)
organizational memory, 154
study guide
application exercises, 194–95
research summary, 194
review questions, 194
synergies, 182
system design, 160–61
tacit knowledge, 189–94
Uniform Resource Locator (URL), 160
utilization barriers, 161–62
workflow management system (WfMS), 157–58
World Wide Web (WWW), 154–55, 160
Knowledge specificity, 30–31
Knowledge substitution process, 63–64
Lakhani, Karim R., 322 (box)
Lateral thinking, 201–2
Leadership of knowledge management, 290–93
Chief Information Officer (CIO), 290–93
Chief Knowledge Officer (CKO), 290, 291–92,
293 (box)
Chief Learning Officer (CLO), 290, 291
Chief Operating Officer (COO), 292, 293 (box)
organizational example, 293 (box)
League of Legends, 261
Learning psychology theory, 136
Learning tradition
employee impacts, 75–76, 302, 303t
knowledge capture, 61, 65–66
LeaseCo, 86
Lessons learned system (LLS), 154, 164, 165
(box), 166–70
Linguistic analysis, 222, 223
366 INDEX
LinkedIn, 250, 251f
Lockheed Corporation, 102
Lockheed Martin Corporation, 102
Machine learning research, 96
Market basket analysis, 205, 206 (box)
Mashup, 257
MasterCard, 292
Matsushita Corporation, 83
McCarthy, John, 95
Meez, 261
Melnick, Blake, 293 (box)
Memory–based Reasoning (MBR), 218t, 219
Mercedes–Benz, 292
Methodologies, 7
Metrics, 300
Microblogs, 257–58
Microsoft Corporation, 4–5, 172t, 173 (box)
Minsky, Marvin, 95
Missick, Chris, 257
Mission context, 142–43
Mission–critical objectives, 7
Model–based reasoning, 98, 99t
Montgomery Watson Harza (MWH), 46 (box), 54
Morrisons Supermarkets, 208, 209 (box)
Most Admired Knowledge Enterprises (MAKE)
survey, 307 (box)
Motor skills expertise, 29
Mougin, Frank, 280 (box)
Mozilla Firefox, 259
Myspace, 247–48, 249f
MySQL, 259–60
Narratives, 132–33
National Aeronautics and Space Administration
(NASA)
common knowledge, 49
Expert Seeker (NASA), 172t, 180, 182–84, 185f,
232–34
Goddard Space Flight Center, 174, 175f
knowledge management, 3
Out–of–Family Disposition process, 99, 118–21
Postdoc system, 176–77
Remote Agent, 173, 195n3
technically specific knowledge, 26–27
National Security Agency (NSA), 172t,
173 (box)
National Semiconductor Corporation, 99, 109–13
Natural language, 95
Natural language processing (NLP), 222, 223
NCR Systems Engineering Copenhagen, 210
NEC Corporation, 94 (box)
Newell, Allan, 95
News Corporation, 247
New Zealand Mountain Safety Council, 163
Nondisclosure agreements, 318
Northwestern Mutual, 321 (box)
Not–invented–here syndrome, 161
Objective view of knowledge, 23–24
Observational learning, 146–48
OBSERVER system, 147
OBSERVO–SOAR system, 147
OHRA Verzegeringen en Bank Groep B.V., 210
Online Analytical Processing (OLAP), 224
Online awareness, 257–58
Ontology, 171, 173–74
Open source software development, 258–61
Operation Burnt Frost (2007), 255 (box)
O’Reilly, Tim, 243, 258
Organizational capital, 5, 34–35
Organizational culture, 43–45, 51t
Organizational memory, 154
Organizational performance impacts, 84–87,
303–4, 305t
Organizational process impacts, 78–82
Organizational products, 82–83, 302, 304t
Organizational structure, 45–47, 51t
Orkut, 250
Orr, Julian, 60 (box)
Out–of–Family Disposition process (NASA), 99,
118–21
Paired leaf analysis, 230
Panasonic Corporation, 83
Patents, 318–19
Phonak, Inc., 52 (box)
Physical organizational environment, 49–50, 51t
Pinterest, 254
Platform as a Service (PaaS), 324
Platform independence, 154
POPS ELS, 184, 185f
Postdoc system (NASA), 176–77
Predictive DM techniques, 205–6
INDEX 367
Predictive Model Markup Language (PMML),
214
Privacy concerns, 324, 326 (box)
Probability relationship, 19
Procedural knowledge, 24–25, 28t
Process Diagnosis System (PDS), 104–5
Procter & Gamble, 322 (box)
Product impacts, 82–83
Production rules, 96
Product Quality Analysis (PQA), 109–10
Proflowers, 208
Project Wonderland, 262
Propositions, 134, 135f, 136
Prototypes, 128
Proxy server, 224
PTC Servigistics, 114
Purvis, Kent, 76 (box)
Qualitative KM assessment, 297, 298f
Quantitative KM assessment, 297–98
Query, 232
QuickSource (Compaq Computer Corporation),
102
Radio frequency identification (RFID), 148–50,
151n3
active tags, 149, 151n4
knowledge capture systems, 146–50148–50,
151n3
passive tags, 149, 151n3
real–time location system (RTLS), 149
semi–passive tags, 149
Raynaud’s Syndrome, 235
Real–time location system (RTLS), 149
Redecard, 225
Remote Agent (NASA), 173, 195n3
RightAnswers Unified Knowledge Platform,
85 (box)
Routines, 64, 66
RS Information Systems, Inc. (RSIS), 165
Rule–based systems, 96
Rule induction algorithms, 213–14
Sam’s Club, 149
Sanger, Larry, 256
SBIR/STTR Online System (SOS) Advisor,
96, 104, 106–9
Searchable Answer Generating Environment
(SAGE) Expert Finder, 172t, 177–80, 181f
Second Life, 261–62, 263f
Semantic networks, 134, 135f, 136
Shared domain knowledge, 49
Sharepoint, 156
Shell Oil Company, 44, 310 (box)
Simon, Herbert, 95
Simple knowledge, 30
Skandia Navigator, 307–8, 312n3
Skills Planning and Development (Microsoft),
172t, 173 (box)
Skype, 253
Small Business Innovation Research (SBIR), 96,
104, 106–9
Small Business Technology Transfer Research
(STTR), 96, 104, 106–9
SMART system (Compaq Computer Corporation),
101–2
Social capital, 5, 34–35
Socialization process
employee learning impact, 75–76
knowledge discovery, 59–60, 64–65
knowledge discovery systems, 199–203, 204
(box)
Social networking, 247–55
Software as a Service (SaaS), 323
Software quality control advisor (SQUAD), 94
(box)
Solution space, 95
Specific knowledge, 26–27, 28t, 30–31
Specific Media Group, 248
SPSS/Integral Solutions Ltd., 210
Stakeholders, 133
Steelcase, Inc., 83
Stemming algorithm, 223, 233
Stoplist, 223
Storage Law, 227
Storytelling, 127, 128 (box), 129–33
Strategic knowledge, 30
Subjective view of knowledge, 22–23
Sun Microsystems, 83
Support knowledge, 30
Swanson, Don, 235
Swinky, 261
Symbio Group, 292
Synergies, 182
368 INDEX
Tacit knowledge
classification of, 25–26, 28t, 30
expertise, 26
knowledge capture systems, 129–32
knowledge discovery systems, 199–203, 204
(box)
knowledge sharing systems, 189–94
Tactical knowledge, 30
Taxonomy, 171, 173–74, 175f
Technically specific knowledge, 26–27, 28t, 31
Ten–fold cross validation, 213–14
Term frequency inverse document frequency
(TFIDF), 222
Text mining, 222
Theoretical expertise, 29, 30 (box)
3M Company, 128 (box), 129
Total Recall System (National Semiconductor
Corporation), 99, 109–13
Toyota Motor Corporation, 80
Trade secrets, 319
Tree construction, 230
Trejo, Danny, 248
Twitter, 250–51, 252f, 257–58
Uniform Resource Locator (URL), 160
U.S. Chemical Safety Board, 163
U.S. Defense Advanced Research Projects Agency
(DARPA), 258–59
U.S. Department of Defense, 193 (box), 194
Value–added products, 82–83, 302, 304t
Variables, 224
Verdande Technology, 103
Veteran’s Health Administration (VHA), 63 (box),
81–82
Viant, 52 (box), 62 (box)
VIDUR (Central Agricultural University, India),
103
VirtualWorld, 261
Virtual worlds, 261–64
Virtual Worlds for Tweens, 261
Wales, Jimmy, 256
Walmart, 149
Watson, Thomas, 243
Web content mining, 208, 224, 232–34
Web crawling, 222
Web mining techniques, 221–24
Web Structure Mining, 223–24
Web technologies
1.0, 244f
2.0, 51, 243–47
3.0, 246–47
Web usage mining, 224
WeeWorld, 261
Western Union Company, 264
Westinghouse, George, 200 (box)
Westinghouse Electric Corporation
GenAID, 96, 104–5
knowledge discovery systems, 200 (box), 202
(box)
Wicked problems, 315
Wikipedia, 256
Wikis, 258 (box)
Wolfensohn, James, 190 (box)
Workflow management system (WfMS), 157–58
World Bank, 51–52, 189, 190 (box), 191–92
Worldwide Quality Network, 109
World Wide Web (WWW), 154–55, 160
Xerox Corporation, 45, 60 (box), 76 (box), 168
(box)
YouTube, 250, 252f
Zuckerberg, Mark, 248
Cover
Half Title
Title Page
Copyright Page
Table of Contents
Preface and Acknowledgments
1. Introducing Knowledge Management
What Is Knowledge Management?
Forces Driving Knowledge Management
Knowledge Management Systems
Issues in Knowledge Management
Text Overview
Summary
Review
Application Exercises
Note
References
Part I. Principles of Knowledge Management
2. The Nature of Knowledge
What Is Knowledge?
Alternative Views of Knowledge
Different Types of Knowledge
Locations of Knowledge
Summary
Review
Application Exercises
References
3. Knowledge Management Foundations: Infrastructure, Mechanisms, and Technologies
Knowledge Management
Knowledge Management Solutions and Foundations
Knowledge Management Infrastructure
Knowledge Management Mechanisms
Knowledge Management Technologies
Management of Knowledge Management Foundations (Infrastructure, Mechanisms, and Technologies)
Summary
Review
Application Exercises
Note
References
4. Knowledge Management Solutions: Processes and Systems
Knowledge Management Processes
Knowledge Management Systems
Managing Knowledge Management Solutions
Alignment Between Knowledge Management and Business Strategy
Summary
Review
Application Exercises
References
5. Organizational Impacts of Knowledge Management
Impact on People
Impact on Processes
Impact on Products
Impact on Organizational Performance
Summary
Review
Application Exercises
Notes
References
Part II. Knowledge Management Technologies and Systems
6. Knowledge Application Systems: Systems that Utilize Knowledge
Technologies for Applying Knowledge
Developing Knowledge Application Systems
Types of Knowledge Application Systems
Case Studies
Limitations of Knowledge Application Systems
Summary
Review
Application Exercises
Notes
References
7. Knowledge Capture Systems: Systems that Preserve and Formalize Knowledge
What Are Knowledge Capture Systems?
Knowledge Management Mechanisms for Capturing Tacit Knowledge: Using Organizational Stories
Techniques for Organizing and Using Stories in the Organization
Designing the Knowledge Capture System
Concept Maps
Context-based Reasoning
Knowledge Capture Systems Based on Context-based Reasoning
Barriers to the Use of Knowledge Capture Systems
Research Trends
Summary
Review
Application Exercises
Notes
References
8. Knowledge Sharing Systems: Systems that Organize and Distribute Knowledge
What Are Knowledge Sharing Systems?
The Computer as a Medium for Sharing Knowledge
Designing the Knowledge Sharing System
Barriers to the Use of Knowledge Sharing Systems
Specific Types of Knowledge Sharing Systems
Lessons Learned Systems
Expertise Locator Knowledge Sharing Systems
The Role of Ontologies and Knowledge Taxonomies in the Development of Expertise Locator Systems
Case Studies
Shortcomings of Knowledge Sharing Systems
Knowledge Management Systems that Share Tacit Knowledge
Summary
Review
Application Exercises
Notes
References
9. Knowledge Discovery Systems: Systems that Create Knowledge
Mechanisms to Discover Knowledge: Using Socialization to Create New Tacit Knowledge
Technologies to Discover Knowledge: Using Data Mining to Create New Explicit Knowledge
Designing the Knowledge Discovery System
Guidelines for Employing Data Mining Techniques
Discovering Knowledge on the Web
Data Mining and Customer Relationship Management
Barriers to the Use of Knowledge Discovery Systems
Case Studies
Summary
Review
Application Exercises
Notes
References
Part III. Management and the Future of Knowledge Management
10. Emergent Knowledge Management Practices
Web 2.0
Social Networking
Collaborative Content Creation via Wikis, Blogs, Mashups, and Folksonomies
Open Source Development
Virtual Worlds
The Three Worlds of Information Technology: Does It Really Matter?
Summary
Review
Application Exercises
Notes
References
11. Factors Influencing Knowledge Management
A Contingency View of Knowledge Management
The Effects of Task Characteristics
The Effects of Knowledge Characteristics
The Effects of Organizational and Environmental Characteristics
Identification of Appropriate Knowledge Management Solutions
Illustrative Example
Summary
Review
Application Exercises
Note
References
12. Leadership and Assessment of Knowledge Management
Leadership of Knowledge Management
Importance of Knowledge Management Assessment
Types of Knowledge Management Assessment
Assessment of Knowledge Management Solutions
Assessment of Knowledge
Assessment of Impacts
Conclusions About Knowledge Management Assessment
Summary
Review
Application Exercises
Notes
References
13. The Future of Knowledge Management
Using Knowledge Management as a Decision-Making Paradigm to Address Wicked Problems
Promoting Knowledge Sharing While Protecting Intellectual Property
Involving Internal and External Knowledge Creators
Addressing Barriers to Knowledge Sharing and Creation
Concluding Remarks
Review
Application Exercises
Note
References
Glossary
About The Authors
Index