Home Uncategorized Developing a Modern Data Architecture Overview White Paper

Uncategorized

Developing a Modern Data Architecture Overview White Paper

For this homework assignment, you are assuming the role of a “Big 4” (KPMG, EY, Deloitte, PwC), where your client, Farmer Consulting, is asking for a white paper discussing the key points, benefits, and components are a modern data architecture. Farmer Consulting is “behind the times” in their infrastructure, and need to make a move towards a modern data architecture. You recently attended a conference and saw a great presentation on this topic, and have a copy of the deck (attached below), that you believe is the basis of your white paper. Your assignment is to use the lecture videos, notes, and presentation below to write a persuasive, informative, and action oriented white paper for your client. The white paper should include an executive summary highlighting the key takeaways that focus points to keep your client excited about reading the paper, and a structured, well flowing paper that will inform their opinion on how to build a modern data architecture.

EVELOPING A
MODERN ENTERPRISE

DATA STRATEGY
Edd Wilder-James, Scott Kurth

March 2017

Data

Science

TODAY’S SCHEDULE
Introduction

Why Have a Data Strategy

Connecting Data with the Business

Understanding Data Gaps

The Data Platform Architecture

Break

Identifying Strategic Workloads

The Chief Data Officer

The Experimental Enterprise

INTRODUCTION

@SVDataScience

To view SVDS speakers and scheduling,
or to receive a copy of our slides, go to:

www.svds.com/StrataCA2017

Silicon Valley Data Science is a boutique
consulting firm focused on transforming
your business through data science and
engineering.

WE DO DATA RIGHT
• We work in cross-functional teams made up of data

scientists, engineers, and solutions architects.

• We combine enterprise know-how with custom
methods derived from Silicon Valley best practices.

• We use an Agile Software Development approach to
make rapid progress against difficult problems that
require flexibility.

• We focus on delivering business value as early as
possible, then iterating toward the larger goal.

OUR SERVICES

DATA
STRATEGY

AGILE
ENGINEERING

AGILE
DATA SCIENCE

ARCHITECTURE

Supports investigative work and
builds a solid layer for production.

Conducts experiments and responds
to the changing environment.

Makes foundational infrastructure
readily accessible.

THE EXPERIMENTAL ENTERPRISE

THE DATA VALUE CHAIN
DRAW VALUE FROM YOUR STRATEGIC DATA

ASSETS

DISCOVER INGEST PROCESS PERSIST INTEGRATE ANALYZE EXPOSE

9 @SVDataScience

WHAT’S ON YOUR MIND?
What is preventing your organization from
realizing its vision?

TODAY’S SCHEDULE
Introduction
Why Have a Data Strategy?
Connecting Data with the Business
Understanding Data Gaps
The Data Platform Architecture
Break
Identifying Strategic Workloads
The Chief Data Officer
The Experimental Enterprise

WHY HAVE A
DATA STRATEGY?

11 @SVDataScience

DATA STRATEGY
is not for the faint of heart*

* Creating an Enterprise Data Strategy by Wayne Eckerson
http://www.enterprisemanagement360.com/white_paper/creating-an-enterprise-data-strategy/

The alternative is to treat
data as a cost of business, to
be minimized.

Data must serve the
strategic imperatives of a
business: the key strategic
aspirations that define the
future vision for an
organization.

IS THERE AN
ALTERNATIVE?

A modern data strategy is a
roadmap to enable data-
driven decision-making and
applications that helps an
enterprise achieve its
strategic imperatives.

An effective data strategy
helps an enterprise make
technology choices,
grounded in business
priorities, to get the most
value from their data.

IS THERE AN
ALTERNATIVE?

CONNECTING TECHNOLOGY AND
BUSINESS VALUE
If you find that:

• you can’t articulate how the cost of your data
systems relates to the benefits to your business, or

• you can’t articulate how your technology philosophy
enables your business aspirations

then your organization would almost certainly benefit
from data strategy.

Poll:
• Is the technology

leadership in your
organization prioritizes
investments to meet the
ambitions of the business?

• Can your organization
clearly articulate the
business impact of the
data and technology
investments it makes?

ARTICULATING THE
BUSINESS IMPACT OF
DATA & TECHNOLOGY

CONNECTING DATA
WITH THE BUSINESS

CLEAN VALIDATE CONTROL PROTECT

CONVENTIONAL DATA STRATEGY

“

WHAT YOU DO TO DATA”

CONVENTIONAL WISDOM:
10 THINGS A DATA STRATEGY SHOULD INCLUDE*
1. What data should be collected?

2. How long should data be kept?

3. Where should the data be
stored?

4. How will data privacy and
security be managed?

5. From where can data be
accessed?

6. What data can be displayed?

7. What level of detail should be
retained?

8. Who is responsible for the data
(governance)?

9. How is data integrated?

10. How will data be distributed
(virtualization?)

* 10 Key Elements of your Data Strategy by Mike Schiff
http://www.tdwi.org//Articles/2012/01/17/10-Elements-Data-Strategy.aspx?Page=1

MODERN DATA STRATEGY
“WHAT YOU DO WITH DATA”

TARGET VIP CUSTOMERS ATTRACT NEW CUSTOMERS

AUTOMATE

A NEW ORTHODOXY?
FOUR PRINCIPLES OF A
SUCCESSFUL DATA STRATEGY*

1. How does data generate value?

2. What are our critical data assets?

3. What is our data ecosystem?

4. How do we govern data?

* The 4 Principles of a Successful Data Strategy by Paul Barth
http://www.cioupdate.com/insights/article.php/3936706/The-4-Principles-of-a-
Successful-Data-Strategy.htm

EDW
Governance
Security

NOT ALL DATA IS EQUAL

Conventional data strategy

22 © 2017 SILICON VALLEY DATA SCIENCE LLC. ALL RIGHTS RESERVED. @SVDataScience
EDW
Governance
Security
NOT ALL DATA IS EQUAL

Modern data strategy

WHAT IS A DATA STRATEGY?

Existing data &
technology

Possible data &
technology

Business
strategic

ambitions

Constraints Priorities

Roadmap of
investments

Tools to update
and assess
roadmap

Plan to update
capabilities

Modern Role of Data:
Represents the new role data
and analytics play in the
enterprise.

Outcomes, not Operations: A
strategic notion of maturity
should begin with value
creation before addressing
underlying operational
processes.

Transforming Pragmatically:
Changes are grounded in the
holistic view of the future
state of your enterprise.

A NEW NOTION OF
MATURITY

An organization’s ability to
derive value from its data
defines its maturity.

NEW STAGES, NEW DIMENSIONS

ASSETS

CULTURE

DECISIONS

OUTCOMES

Illustrative

Not just the technology!
• People
• Processes
• Systems

DIMENSIONS OF
DATA MATURITY

CURIOUS WHERE YOU FALL?

ASSETS
CULTURE
DECISIONS
OUTCOMES

IllustrativeMaturity Mini-Assessment
• 20Q survey (5-10 min)

• Identifies your stage and provides
general recommendations

• Creates baseline for future
performance and growth

dmm.svds.com

• Infrastructure is holding
back growth

• Infrastructure is holding
back development

• Analog to digital
transformation

• Changing business models

• Unifying fragmented
offerings

YOU NEED A DATA
STRATEGY WHEN…

BEGIN WITH THE BUSINESS
• First understand what drives your business

• Then make the leap from strategy to tactics

Technologists: This can’t be done without the business
leaders in the room

Business Leaders: This can’t be done without the
technologists in the room

Understand the strategic
imperatives of your
organization:

• Annual report

• Investor updates

• Talk to leadership

STRATEGIC
IMPERATIVES

Break down the strategic
imperatives to make them
tangible, achievable, and
measurable. These become
your business objectives.

Business objectives provide
the guide for many other
analyses in building your data
strategy.

BUSINESS
OBJECTIVES

REAL ESTATE MARKETPLACE: ZILLOW

Business Objectives

• Build and maintain best algorithms for pricing
• Use Hedonic pricing method to incorporate multiple attributes

and ‘nearest neighbors’ to create accurate Zestimate®
• Deploy sophisticated and adaptive models, at scale (over 110

million homes) and at timely interval (3 times / week)

• Use scalable infrastructure (cloud) for rapid analysis

• Build industry’s best real estate data sets
• Increase completeness of data by include public data sets such

as construction listings, foreclosure listings, market context
• Capture unique data with customer reviews and feedback from

real-estate firms
• Manage scale of 110 million properties

and growing

Strategic Imperatives
• Provide products and

services to help
consumers with every
stage of home ownership
– buying, selling, renting,
borrowing, and
remodeling

• Generate more
subscription and ad
revenue

• Drive more unique users
to marketplace

• Become leading real
estate and home-related
information marketplace
on mobile and web

NOTE: Zillow is not an SVDS client.

HEALTH PROVIDER:
KAISER PERMANENTE

Business Objectives
• Increase data sharing with extended care teams

through secure electronic health record access

• Provide quicker, better diagnoses through evidence-
based medicine techniques

• Provide mobile access to scheduling, pharmacy
interactions, and other related services

• Improve member satisfaction by analyzing web and
mobile user interactions, behavior, and feedback
data

• Share access to knowledge, innovation, and
population data with the public and other health care
leaders

Strategic Imperatives
• Provide seamless,

personalized care
through an integrated
team of care providers

• Enable members to
manage their own care
through easy-to-use
channels

• Transform care and
improve outcomes
through investments in
research and innovation

NOTE: Kaiser Permanente is not an SVDS client.

REAL ESTATE MARKETPLACE: ZILLOW

STRATEGIC IMPERATIVES

• Provide products and services to help consumers
with every stage of home ownership – buying, selling,
renting, borrowing, and remodeling

• Generate more subscription and ad revenue

• Drive more unique users to marketplace

• Become leading real estate and home-related
information marketplace on mobile and web

NOTE: Zillow is not an SVDS client.

REAL ESTATE MARKETPLACE: ZILLOW

BUSINESS OBJECTIVES

1. Build and maintain best algorithms for pricing

• Use Hedonic pricing method to incorporate multiple
attributes and ‘nearest neighbors’ to create accurate
Zestimate®

• Deploy sophisticated and adaptive models, at scale
(over 110 million homes) and at timely interval (3
times / week)

• Use scalable infrastructure (cloud) for rapid analysis
NOTE: Zillow is not an SVDS client.

REAL ESTATE MARKETPLACE: ZILLOW
BUSINESS OBJECTIVES

2. Build industry’s best real estate data sets

• Increase completeness of data by include public
data sets such as construction listings, foreclosure
listings, market context

• Capture unique data with customer reviews and
feedback from real-estate firms

• Manage scale of 110 million properties
and growing

NOTE: Zillow is not an SVDS client.

HEALTH PROVIDER: KAISER PERMANENTE

STRATEGIC IMPERATIVES

• Provide seamless, personalized care through an
integrated team of care providers

• Enable members to manage their own care through
easy-to-use channels

• Transform care and improve outcomes through
investments in research and innovation

NOTE: Kaiser Permanente is not an SVDS client.

HEALTH PROVIDER: KAISER PERMANENTE
BUSINESS OBJECTIVES

• Increase data sharing with extended care teams
through secure electronic health record access

• Provide quicker, better diagnoses through evidence-
based medicine techniques
• Provide mobile access to scheduling, pharmacy
interactions, and other related services
NOTE: Kaiser Permanente is not an SVDS client.

HEALTH PROVIDER: KAISER PERMANENTE
BUSINESS OBJECTIVES
• Improve member satisfaction by analyzing web and
mobile user interactions, behavior, and feedback
data

• Share access to knowledge, innovation, and
population data with the public and other health
care leaders

NOTE: Kaiser Permanente is not an SVDS client.

TODAY’S SCHEDULE
Introduction
Why Have a Data Strategy?
Connecting Data with the Business

Understanding Data Gaps
The Data Platform Architecture

Break
Identifying Strategic Workloads
The Chief Data Officer
The Experimental Enterprise

UNDERSTANDING
DATA GAPS

None of these questions
make sense unless you ask:

For what?

Commonly-asked questions:

• Do I have gaps in my data?

• How good is my data?

• Is my data clean enough?

NO ONE’S DATA IS
PERFECT

FOR WHAT?
• Do I have gaps in my data?

• How good is my data?
• Is my data clean enough?
• Do I have gaps in my data?

…for understanding customer purchase behavior

• How good is my data?

…for predicting quarterly sales

• Is my data clean enough?

…for automating production

• What are you trying to
achieve as a business
[with data]?
These are your business
objectives.

• How do you plan to
achieve it [with data]?
These are your use cases.

UNDERSTAND
YOUR BUSINESS
GOALS

UNDERSTAND YOUR AUDIENCE
Who is going to use this analysis and how?

• CDO? Heads of Business Units? Data Science Directors?
DBAs?

• Project assessment? Operational dashboard?
Continuous improvement plan?

Understanding stakeholders and expectations will
dictate the level of technical analysis required.

UNDERSTAND YOUR AUDIENCE
What are the dimensions of requirements that matter
to your audience?

• For a technical application, it might be depth, breadth,
latency, frequency.

• For an executive perspective, it might be higher-order
requirements like ease of integration or coverage.

What are the questions your audience needs
answered? Select the dimensions that provide visibility
into those questions.

• Start with an effective
catalog of your data.

• Organize the data to be
effective. Think about how
data is produced AND how
it gets used in your
organization.

• By data source?

• By entity?

• By organization?

• By data owner?

UNDERSTAND
YOUR DATA

LINK IT ALL TOGETHER

Business
Objectives

Use Cases

Requirements

Data

VISUALIZE YOUR GAPS

SO… WHAT IS A ”GAP”?
Two schools of thought:

• Purists: If a requirement isn’t met, it’s a gap.

• Pragmatists: If you can still get the job done,
it isn’t a gap.

Both views can be valuable ways of looking at your
analysis.

THE DATA PLATFORM
ARCHITECTURE

51 @SVDataScience

WHY BIG DATA?
1. New Capabilities

2. Economic Scalability

Edmunds.com wanted to reduce time-
to-market by speeding creation of
attribute data for new car models.

We developed a new capability to
automatically extract vehicle features
from specification guides and
categorize the features into
appropriate vehicle classes.

DATA
PLATFORMS
FOR NEW
CAPABILITIES

Existing revenue streams:
• Ads
• Price quotes (leads)

Shopping is the focus:
• Need real-time

inventory
• Accurately described

VINs

DATA PLATFORMS
FOR ECONOMIC
SCALABILITY
at NetApp

NOTE: NetApp is not an SVDS client.
http://blogs.wsj.com/cio/2012/06/12/netapp-cio-uses-big-data-to-assess-product-performance

UP VS. OUT — SAAS EDITION

$,
€

, ¥

, £

Users

Revenue

Cost to serve

Scale-out cost

Profit

Loss

UP VS. OUT — ENTERPRISE EDITION
$,

€
, ¥

, £

Data Resource Usage

Scale-up cost

Scale-out cost

UC2

UC4

UC5

57 @SVDataScience

BIG DATA
… it’s really about agility

• Linear scale-out cost

• Opex vs. capex

• Ease of purchase

BUYING AGILITY

59 @SVDataScience

Scale-out systems move us from managing scarcity to
promoting utility.

• Architectural factors
• Schema on read
• Rapid deployment
• Mirror production setup
• Executes faster

• Programmer factors
• Fun to program
• Concision
• Easier to test
• Faster to write

DEVELOPMENT
AGILITY

WHAT IS DOCKER?
• Container technology: bundles every part of an

application
• Provides isolation for each application without the

overhead of running a virtual machine
• Ships only the parts that are needed—leaves out the

operating system

WHY SHOULD BUSINESS CARE?

• Better use of server resource than virtual machines
• A fast and reliable way of deploying applications

• It’s the ideal packaging mechanism for scale-out
distributed systems

• Easy for developers to work in an environment
identical to production
• Sharing containers leads to innovation

WHAT IS APACHE KAFKA?

• Scale-out fault-tolerant messaging system
• Comes from LinkedIn
• Supported by Confluent

USE CASES

• Stream

processing

• Log aggregation
• Creating decoupled evented

architecture

WHY SHOULD BUSINESS CARE?

• Scalability in a critical area of distributed applications
• Online reliability, compared to alternatives
• Will be a core building block of distributed data

architecture

WHAT IS APACHE SPARK?

• In-memory distributed computing platform
• Comes from Berkeley AMPlab
• In production with early adopters, now integral to

every commercial Hadoop distribution
• Doesn’t need Hadoop, but runs easily on top

USE CASES

• Managing a major retailer’s inventory across a
diverse network of entities in near real time

• Managing and processing event streams for online
gaming

• Supporting data science initiatives across massive
data sets at a media analytics company

WHY SHOULD BUSINESS CARE?

• Enables use cases Hadoop didn’t provide, all in one
platform
• streaming, interactive analytics, machine learning,

graphs

• Fast
• Iteration time down, more productive

• Use existing cluster investment
• Sits on HDFS, can run under YARN

(or use Amazon S3, or Cassandra)

WHY SHOULD BUSINESS CARE?

• SparkSQL
• Use SQL skills and tools, e.g. Tableau
• Dataframes integrate external data sources into one

context: RDBMS, Hive, JSON…

• Developer-friendly
• Concise and fluid to program
• Language integration: Scala, R, Python, Java

WHAT ARE NOTEBOOKS?
• Interactive documents that contain a program and

its output
• Long history: Mathematica

• Particularly successful with data science
• Projects to watch

• Jupyter — https://jupyter.org/
• Apache Zeppelin —

https://zeppelin.incubator.apache.org/

WHY SHOULD BUSINESS CARE?
• Easy collaboration and sharing of data science

• Think “Docker for analysis”

• Easy access to data and compute resource
• A building block for more self-service analytical

capabilities

Commercial version of Notebooks + Spark is the
Databricks Cloud

@SVDataScience

ENTERPRISE DATA
ARCHITECTURE

Towards a production

DATA
PLATFORM

Data Management
Security, Operations, Data Quality, Meta Data Management and Data Lineage

Analytics

Lo
w

L
at

en
cy

A
cc

es
s

Data

Ingest

Data
Repository

Persistence

Offline
Processing

Real-Time
Processing

Batch
Processing

Data
Services

External
Systems

Data Acquisition

Internal External

CHOICES: TOOLS

Graph Document Key-Value Columnar
Social networks

Ontologies
Knowledge, Property

Logging
Document archive

Web content

Shopping Cart
Session Data

Sensors
Network devices

Internet of Things

Technical Use Cases

CHOICES: DATABASES

Graph Document Key-Value Columnar
Social networks
Ontologies
Knowledge, Property
Logging
Document archive
Web content
Shopping Cart
Session Data
Sensors
Network devices
Internet of Things

CHOICES: DATABASES
SPECIALIZED

Graph Document Key-Value Columnar
Social networks
Ontologies
Knowledge, Property
Logging
Document archive
Web content
Shopping Cart
Session Data
Sensors
Network devices
Internet of Things

CHOICES: DATABASES
GENERAL PURPOSE

CHOICES: VELOCITY
SVDS R&D TRAINS
Batch:

• Using FFT transformed
frequency data, identify the
train based around
fundamental frequencies of
train whistle.

• Construct the decision tree
for train classifier based on
minimum and maximum
fundamental frequencies

Real-Time:

• Apply FFT to audio signal

• Extract min and max
fundamental frequencies

• Classify the train into local
or express

• Send data to the Event
Detector to alert the APP

• Store results in HBase

80 @SVDataScience

[Amazon] do services because
they’ve come to understand that
it’s the Right Thing. There are
without question pros and cons to
the SOA approach, and some of
the cons are pretty long. But
overall it’s the right thing because
SOA-driven design enables
Platforms. …
You wouldn’t really think that an
online bookstore needs to be an
extensible, programmable
platform. Would you?

+Steve Yegge

CHOICES: SERVICES

https://plus.google.com/112678702228711889851/posts/eVeouesvaV
X

CHOICES: DATA RESILIENCY

Hard Failure: If the data
source is broken, so is
the app.

Stovepipe: One-to-one
relationship from data
source to product.

Multi-sourced:
Redundancy of
overlapping data
sources makes your
products more
resilient.

Graceful
Degradation: If a data
source breaks, there is
a backup and your app
continues to function.

Production data
services abstract the
probabilistic
integration of
overlapping data
sources. We call this
model a Data Mesh.

CHOICES: EXTERNAL
SYSTEMS
Applications, visualization, business
intelligence

üIncremental revenue

üTime to market

üEconomically viable
implementation

üCost avoidance

üBrand benefit

üEcosystem friendliness

DEFINING
SUCCESS

@SVDataScience

BREAK

IDENTIFYING
STRATEGIC

WORKLOAD

HOW SVDS DOES DATA STRATEGY
• We work with your stakeholders to analyze and articulate a data

strategy.

• The data strategy provides an actionable roadmap that generates
immediate value and serves as the foundation for future
capability investments.

• We work to understand your current business and technology
landscapes in order to unlock untapped business opportunities.

• Our collaborative approach ensures that your business, product,
and technology teams become effective advocates within your
organization.

BUSINESS MODEL
TRANSFORMATION

PRODUCT RESEARCH &
RECOMMENDATION COMPANY

A product research and recommendation
company is transforming their core business
from content and information services to a
referrer of high-value transactions to partners.

SVDS devised a data strategy that enables new
analytical capabilities core to their retail
ambitions, addressing critical accuracy and
timeliness issues with unstructured data.

Based on this data strategy, they are building a
solution for near real-time product inventory
that increases their value to partners in a
complex, multi-tier market.

PERSONALIZED
USER EXPERIENCE

MEDIA & ENTERTAINMENT COMPANY

A media and entertainment company seeks to
deliver personalized content directly to users on
digital entertainment devices.

SVDS developed a data strategy and architecture
that enables real-time data ingestion, deeper
customer insight, and highly-personalized
content recommendations.

The data strategy and architecture design now
serve as the foundation for iterative, new
product development and guide technology
investments and acquisitions.

ACTION PLAN
& ROADMAP

OUR METHOD FOR DATA STRATEGY

IDENTIFY

STRATEGIC
IMPERATIVES

DEFINE

BUSINESS
OBJECTIVES

DEFINE DATA
REQUIREMENTS

IDENTIFY GAPS
IN CURRENT
SYSTEMS &
TECHNOLOGY

MAP BUSINESS
OBJECTIVES TO
USE CASES

RATIONALIZE USE CASES
INTO WORKLOADS

USE CASE

IDENTIFY YOUR STRATEGIC WORKLOADS

USE CASE

WORKLOAD

A
WORKLOAD

WORKLOAD

B
WORKLOAD

C
USE CASE

3
WORKLOAD

B
WORKLOAD
D

AN EXAMPLE

DATA STRATEGY FOR THE DOGS

NOTE: PetSmart is not an SVDS client. This is a fictional example based on public information.
http://risnews.edgl.com/retail-news/PetSmart-Leverages-Analytics-for-Personalized-Experience91783

AN EXAMPLE
DATA STRATEGY FOR THE DOGS

We’ve been investing in new
capabilities to help us capture
and use customer and pet data,
and this year, we will deliver on
new methods to use this data to
drive growth.

— David Lenhardt
PetSmart CEO

“

@SVDataScience © 2017 SILICON VALLEY DATA SCIENCE LLC. ALL RIGHTS RESERVED.
STRATEGIC IMPERATIVES
STRATEGIC
IMPERATIVES
BUSINESS
OBJECTIVES
USE CASES
WORKLOAD

Our strategy:
“To be the preferred
provider for the
lifetime needs of
pets.”

Connect with pet parents in
a personalized way

Attract and retain our most
valuable customers

Provide innovative products
& services at fair prices

Drive consistent execution
in our stores

@SVDataScience © 2017 SILICON VALLEY DATA SCIENCE LLC. ALL RIGHTS RESERVED.
AN EXAMPLE

Connect with pet parents
in a personalized way

Deliver personalized
recommendations and offers

Recommendation
Engine

Recommend new pet
products based on past

purchases at point of sale

Recommend upcoming
store/community events

based on customer
preferences

STRATEGIC
IMPERATIVES
BUSINESS
OBJECTIVES
USE CASES
WORKLOAD

@SVDataScience © 2017 SILICON VALLEY DATA SCIENCE LLC. ALL RIGHTS RESERVED.
BUSINESS OBJECTIVES

Illustrative

Connect with pet parents in
a personalized way

Learn from consumer
interactions

Optimize consumer journeys
based on insights

Deliver personalized
content to customers

1
2
3

. . .

STRATEGIC
IMPERATIVES
BUSINESS
OBJECTIVES
USE CASES
WORKLOAD

@SVDataScience © 2017 SILICON VALLEY DATA SCIENCE LLC. ALL RIGHTS RESERVED.
USE CASES
Deliver personalized
content to customers

1. Identify customers 2. Profile behaviors

4. Anticipate behaviors

. . .

3. Understand context

5. Optimize personalization

Illustrative
STRATEGIC
IMPERATIVES
BUSINESS
OBJECTIVES
USE CASES
WORKLOAD

WORKLOADS
Data Value Chain Example Workloads

Acquire • Capture mobile app transactions• Accessing streaming web activity data

Ingest • Flexible data ingestion• Ingest unstructured data

Process • Data validation• Omnichannel data integration

Persist • Heterogeneous data storage• Scalable data storage

Analyze • Probabilistic data integration

• Predictive modeling

Expose • Service based data access• Interactive visualization

STRATEGIC
IMPERATIVES
BUSINESS
OBJECTIVES
USE CASES
WORKLOAD

TECHNICAL WORKLOADS

Acquire

Ingest

Process

Persist

Analyze

Expose

1. Identify customers Technical Workload

Customer data (Acquire,
Ingest, Persist)

• Acquire multiple data sources & formats
• Flexible data ingestion
• Flexible & scalable data storage and

processing

Identity resolution • Probabilistic data integration

Data cleansing • Data validation

Householding • Probabilistic data integration

Relationship context • Detailed views of entities

Life-time Value • Feature engineering

Illustrative

2. Profile behaviors Technical Workload

360 degree view of customer • Detailed views of entities

Views of historical transactions • Time series analysis

Determination of ‘favorites’ • Predicting customer behavior

Map to archetype

• Stream processing

Evaluate previously unseen
transactions and classify

• Stream processing

Update archetypes • Feature extraction
• Analyze customer behavior

TECHNICAL WORKLOADS
Acquire
Ingest
Process
Persist
Analyze
Expose
Illustrative

3. Understand context Technical Workload

Characterize temporal customer
behavior

• Feature engineering
• Analyze customer behavior

Determine goal of next
interaction

• Predictive modeling

Categorize content needs • Predictive modeling

TECHNICAL WORKLOADS
Acquire
Ingest
Process
Persist
Analyze
Expose
Illustrative

4. Anticipate behaviors Technical Workload

Score product offers with
likelihood to respond

• Integrate internal systems
• Service based data access

Score content options with
likelihood to respond

• Integrate internal systems
• Service based data access

Identify next best action • Third party structured data
integration

• Business rules execution

TECHNICAL WORKLOADS
Acquire
Ingest
Process
Persist
Analyze
Expose
Illustrative

5. Optimize personalization Technical Workload

Apply business rules, constraints
to personalization options

• Business rule execution

Select optimal personalization to
achieve goal

• Optimization execution

TECHNICAL WORKLOADS
Acquire
Ingest
Process
Persist
Analyze
Expose
Illustrative

PRIORITIES DIMENSIONS OVERCOME YOUR
ASSUMPTIONS

FOCUS ON THE VALUE

DEVELOPMENT HORIZONS
Illustrative

TECHNICAL WORKLOAD PRIORITIZATION

TECHNICAL WORKLOAD
STRATEGIC
VALUE

TECHNICAL
FEASIBILITY

ACCESSIBILITY
OF REQUIRED
SKILLS

ARCHITECTURAL
FIT

PROD ROLL-
OUT EFFORT

Real time recommendations 10

Omnichannel data integration 10

Predictive modeling 9

Unstructured text analytics 8

Behavioral analytics 7

Data quality monitoring 6

Pattern recognition 5

Heterogeneous data storage 3

Data ingestion 3

Illustrative

DEFINE YOUR ROADMAP

Plan Prove Pilot Production

We define a project plan to build a specific capability.

For each capability, we describe a project to build
technical workloads that implement use cases that
address high-priority business objectives.

Silicon Valley Data Science employs an agile development
processes as we work with our clients from planning and
proof-of-concepts to pilot implementations and finally
full scale production systems.

PROJECT ACTION PLAN

Plan Prove Pilot Production Agile Build
Process

Illustrative

PATH FORWARD

Horizon I Horizon II Horizon III Horizon IV

2-3
months

5-6
months

3-4
months

0 months

Illustrative

DEVISING A PROJECT PLAN:
INPUTS & APPROACH

Technical Workload
AssessmentData Gaps

Project Roadmaps

Workload Rationalization Development Horizons

RINSE REPEATLATHER

MAKE SURE IT’S FLEXIBLE
• Technology moves incredibly fast, and competitive

landscapes are highly dynamic.

• Your data strategy should be a living document,
revisited often and revised as conditions change.

MAKE SURE IT’S ACTIONABLE
• If it isn’t clear how you’re going to execute your

strategy, then you don’t have the right one.

• Must work within the realm of the possible and
practical.

FROM IDEA TO PRODUCTION
We identify the business goals, distill those into use cases, and then
work in short, iterative cycles to achieve tangible gains.

Plan Prototype Pilot Production

What can we do with data?

MODERNIZING DATA TECHNOLOGY
HEALTH MANAGEMENT COMPANY

Aging data infrastructure and brittle application
integration was inhibiting growth and business
insight for a health management company.
Their data strategy focused on creating a
concrete roadmap for migrating to a new data
platform so that technology and infrastructure
are no longer a barrier to growth and
transparency.
Based on this data strategy, they are building a
new data platform in stages that allows them to
add new products and services to capture more
market opportunity.

Case Study: Data Strategy
Major Pharmaceutical Company

Defined Data Strategy that will help enable business growth and enable
expansion into new markets

Challenge
• Ongoing need to improve

discovery and better predict new
targets for drug development

• Difficulty to integrate new data
sources into identification &
discovery processes

• Inability to connect business
strategy & aims with specific,
tangible projects

Solution
• SVDS devised a data strategy with a

concrete roadmap for migrating to
a new data platform

• Recommended data technology &
architecture which supports highest
value projects

• Outlined cultural, technological,
organizational, and collaboration
challenges & objectives

Results
• Identified specific opportunity areas

to increase GTM efficiency
• Prescribed Common Data and

Analytics Platform for Commercial
and R&D operations

• Recommended projects for
Predictive Modeling & Data
Exploration

DATA STRATEGY CHECKLIST
¨ Identify your business objectives
¨ Go from objectives to tactics
¨ Include all stakeholders in the conversation
¨ Look at how technology can support strategic

workloads
¨ Exploit patterns and reuse
¨ Prioritize the possibilities to figure out where to start
¨ Define your roadmap with an end-point in mind
¨ Lather, rinse, repeat

THE CHIEF DATA
OFFICER

118

DO YOU NEED
EXECUTIVE HELP?

To download a free PDF, go to:
www.svds.com/CDOreport

EMERGENCE OF THE CDO
• Started with heavily regulated industries such as

government and finance

• Now becoming common in “disruptable” industries
such as retail and telecommunications

RESPONSIBILITIES OF THE CDO
Centralization:

• Data from internal silos

• Data from external APIs and real-time streams

• The organization’s priorities

RESPONSIBILITIES OF THE CDO
Evangelization:

• Technical chops, business savvy, and the
diplomacy skills to translate between the two

RESPONSIBILITIES OF THE CDO
Facilitation:

• Coordinate stakeholders across the organization

• Free up resources and lower barriers

• Offer tools and training to help others succeed

CHALLENGES FOR THE CDO
Building technical bridges:

• Working with data in different silos, formats, etc.

Mining for business value:

• “If you don’t have good business questions it
doesn’t matter what kind of technology you
have.” — Joy Bonaguro

UNDERSTANDING THE CDO
“While technology is inevitably involved when working
with data, the defining goal of the CDO is not
technological, but business-oriented. The ideal CDO
exists to drive business value.”

— Julie Steele

DECIDING TO HIRE A CDO
Know why you want one:

• Are you part of a regulated industry?
• Do you need to move from being product-centric

to customer-centric?
• Could you add products or services?
• Could your current processes and outcomes be

optimized even further?
• Are there insights in one part of your company

that could benefit others?

DECIDING TO HIRE A CDO
Look for the right skill set:

• Technical chops

• Business savvy

• Diplomacy and political skills

• Executive-level experience

THE AVAILABILITY GAP
“The spike in demand for Chief Digital Officers has been
felt globally. In Europe, the number of search requests
for this role has risen by almost a third in the last 24
months. The United States has seen the same growth in
half that time.”

— Russell Reynolds Associates

PREPPING FOR SUCCESS
Companies that are eager and prepared for real change
will be the most appealing to qualified CDO candidates.

THE EXPERIMENTAL
ENTERPRISE

“…let’s seek to understand how
the new generation of
technology companies are
doing what they do, what the
broader consequences are for
businesses and the economy.”

– Marc Andreesen

DIGITAL NERVOUS SYSTEM

133 @SVDataScience

Data is your business.

Disruptive
Change

Cloud
Computing

Customer
Content

Internet of
Things

User
Experience

SAAS & Apps

Business
Intelligence Consumer IT

Regulation

Employees
Partners

Contractors
Suppliers

FROM: Innosight Executive Briefing Winter 2012

SILICON VALLEY’S DATA MACHINE

UP VS. OUT
$,

€
, ¥
, £
Data Resource Usage
Scale-up cost
Scale-out cost

UC1

UC2

UC3

UC4
UC5

139 @SVDataScience

The legacy of big data is business agility.

• Make it cheap

• Failure as a feature

• Ask good questions

• Make it quick

• Both learning and
adaptation

• Enable the feedback loop

• Don’t break things

• Make operations a
platform for innovation

• APIs, platforms, simulation

BUILD FOR
EXPERIMENTS

THE EXPERIMENTAL ENTERPRISE

Supports investigative work and
builds a solid layer for production.
Conducts experiments and responds
to the changing environment.
Makes foundational infrastructure
readily accessible.

LEAD A DATA REVOLUTION
• You can only win with situational awareness

• New architectures offer new opportunities

• Creation of data-driven value requires new approach

• Create an Experimental Enterprise

• Business must lead, and understand the potential of
the technology

To view SVDS speakers and scheduling,
or to receive a copy of our slides, go to:
www.svds.com/StrataCA2017

THANK YOU

Ask how we can help
info@svds.com

Edd Wilder-James (@edd)

Scott Kurth (@ScottWKurth)

March 2017

Turn in your highest-quality paper
Get a qualified writer to help you with

“ Developing a Modern Data Architecture Overview White Paper ”

Get high-quality paper

NEW! AI matching with writer

Hire a Writer

Client Reviews

4.9

Sitejabber

4.6

Trustpilot

4.8

Our Guarantees

100% Confidentiality

Information about customers is confidential and never disclosed to third parties.

Original Writing

We complete all papers from scratch. You can get a plagiarism report.

Timely Delivery

No missed deadlines – 97% of assignments are completed in time.

Money Back

If you're confident that a writer didn't follow your order details, ask for a refund.

New to Your Trusted Assignment Help Service? Sign up & Save

Calculate the price of your order

Type of paper needed:

Pages:

You will get a personal manager and a discount.

Academic level:

We'll send you the first draft for approval by at

Total price:

$0.00

Power up Your Academic Success with the
Team of Professionals. We’ve Got Your Back.

Power up Your Study Success with Experts We’ve Got Your Back.

Order Now Order Now