�ݺ�ߣ

by
Kimberly NEVALA
best
practices
T H O U G H T P R O V O K I N G B U S I N E S S
business ANALYTICS
a SAS Best Practices white paper
of an ANALYTIC ENTERPRISE
ANATOMY
an
EXAMINATION
of your company’s
ANALYTIC
physique

Anatomy of an Analytic Enterprisebusiness ANALYTICS
2
INTRODUCTION ...................................................................................... 4
It’s Not How Much Data You Have .......................................................... 4
It’s How You Use It.................................................................................. 4
THE ANALYTIC ENTERPRISE ................................................................. 5
THE ANALYTIC MINDSET ....................................................................... 6
We Understand Our Business................................................................. 7
Remember When?.................................................................................. 7
This is How We Do It............................................................................... 8
MAKING THE CASE: Inciting passion & enlisting commitment .......... 10
Informing and Enabling Business Strategy ............................................ 10
Finding Your 1 Percent.......................................................................... 10
DEVELOPING YOUR ANALYTIC MUSCLE ........................................... 12
The Analytic Process............................................................................. 12
The Analytic Community: More Than A Data Scientist........................... 13
The Enabling Technology ...................................................................... 14
table of CONTENTS

Anatomy of an Analytic Enterprise business ANALYTICS
3
MANAGING THE CHANGE : Organizing for success........................... 15
Strategic............................................................................................... 15
Collaborative......................................................................................... 15
Influential............................................................................................... 16
Skilled................................................................................................... 16
Results Oriented ................................................................................... 16
CONCLUSION........................................................................................ 17

4
It’s Not How Much Data You Have
Driven by the big data movement, many conversations about creating a data-driven or-
ganization begin by focusing on the acquisition and storage of data in all its forms. Data
lakes are being created and data galore is flowing in. The issue immediately becomes:
What could and should be done with all that data?
Side note: Storage is cheap. That doesn’t mean managing and maintaining the data
comes free.
In the end, the organization with the most data does not win. It is the organization that
does the most with its data that will ultimately prevail. And herein lies the crux of the is-
sue.
It’s How You Use It
The case for analytics as a competitive differentiator is broadly established. Harken back
to at least 2008 when Competing on Analytics: The New Science of Winning was pub-
lished and subsequently cited by CIO magazine as one of the top 15 most groundbreak-
ing business management books.
Unfortunately, awareness – in theory – does not automatically translate into practice
in reality. In fact, emerging research continues to highlight a growing gap between the
maturity of analytics in an organization and the ability to translate analytic results into
intelligent decisions.
The CIO of a major health care provider stated this point simply: “We’re good at data
crunching. Not so good at making decisions.” A number of reports and studies support
this statement.
A 2013 Beacon Report by Meritalk on big data in government found that only 60 percent
of federal organizations use the data they collect today. And only 40 percent use that
data to make strategic decisions.
Forrester reported similar findings in a 2013 business intelligence and big data survey in
which 54 percent of respondents reported they were successful or very successful mak-
ing informed decisions. Yet only 28 percent reported they were using information to gain
a competitive advantage. Disconcerting, given the preponderance of evidence that a
company’s ability to survive and thrive in the digital economy is proportional to its ability
to use information effectively.
INTRODUCTION
In the end, the
organization
with the most
data does not
win.

5
Becoming an analytic enterprise and embracing data-driven decision making requires
more than analytic tools. Or Hadoop. Or a data scientist.
In a keynote address at the Analytics 2013 conference, Dr. Will Hakes discussed leading
companies who truly compete on analytics. Not just the tech titans such as Amazon and
Google, but companies like Pandora and eHarmony whose business models are predi-
cated upon analytics to find and deliver the answer to some straightforward questions.
On an ongoing basis. And for a profit. He called out “analytic DNA” as the basis for each
company’s go-to-market concept and success.
But can an organization that didn’t spring from analytic DNA develop some? Let’s ex-
plore the anatomy of an analytic enterprise and review some key characteristics and
tactics organizations must adopt in order to become data-driven, analytic competitors.
THE ANALYTIC ENTERPRISE
MIND MUSCLEHEART
figure 1. dimensions of an analytic enterprise

6
Analytic innovators exhibit common attitudes and traits. Most notably: clarity of vision, a
willingness – indeed, a mandate – to constantly challenge prevailing ideas and wisdom,
and the will to follow where the data leads. Organizations not gifted with analytic DNA
must cultivate a similar mindset. How? By providing education on the art of the possible
and overcoming inherent organizational biases. The first is simple. The latter less so.
Organizations, like individuals, have memories. These memories, and the behaviors they
influence, tend to be long-lived. Rethinking the playing field (or even considering the pos-
sibility) requires acknowledging and breaking down entrenched attitudes and behaviors.
Consider how three types of memory – semantic, procedural and episodic – influence
corporate decision making.
THE ANALYTIC
MINDSET
“WE KNOW
OUR BUSINESS.”
“THIS IS HOW
WE DO THINGS.”
“REMEMBER
WHEN...?”
semantic memory episodic memory
procedural memory
figure 2. barriers to an analytic mindset
• Semantic Memory: Knowledge of ideas, facts and concepts not related to
specific experiences.
• Episodic Memory: Knowledge of specific events and experiences complete
with emotional context.
• Procedural Memory: Implicit or unconscious knowledge of behaviors, habits
or skills: how to.
Author’s note: I am not a neuroscientist. Nor do I play one at SAS. If I have taken liberties
with these concepts, the error is mine. Regardless, you’ll get the gist.

7
We Understand Our Business
Semantic memory supplies our knowledge of ideas, facts and concepts independent of
specific events or experiences. And it is, historically, the foundation on which corporate
seniority is based and decision-making authority is conferred. We’re rewarded, and in
fact promoted, based on our ability to understand the playing field, intuit the next best
move or make the right call.
Simply put, business expertise is grounded – although not solely dependent – upon our
mastery of business semantics. But what happens when the landscape changes and
new ideas enter the fray? Or historical operating parameters or assumptions no longer
hold true?
Malcolm Gladwell, the author of Outliers and The Tipping Point, asserts that many cata-
strophic business failures have been caused by errors of expertise. And while data-
driven decision making cannot eliminate errors of expertise entirely, a willingness to chal-
lenge ingrained ideas is certainly a hedge against them.
This is not to say, of course, that seniority is dead. Data-driven decision makers utilize
their hard won knowledge and business savvy to best effect. Not by predicting the future
and then finding numbers to support it. Rather, they use their hard-won understanding to
home in on the right questions to ask and areas to explore. Ironically, data-driven deci-
sion makers are comfortable with and even court ambiguity. And last but not least, they
have the willingness to learn from and act on the information received.
It should be noted that overturning the inherent bias toward the HiPPO (the highest
paid person’s opinion – Andrew McAfee) is not just an executive or senior management
challenge. Data-driven decision making requires organizations to think differently – as a
collective – and modify the processes and pathways by which questions are raised and
decisions are made. Fostering an analytic mindset and implementing data-driven deci-
sion making is a collaborative effort.
Remember When?
This brings us to the next barrier to embracing a no-holds barred approach to analytic
discovery and the controversial insights that often ensue. Episodic memory, the ghost of
experiences past complete with emotional context: That was great, bad or other!
Data-driven
decision makers
are comfortable
with and even
court ambiguity.

8
Yes, organizations herald and reward spectacular success. But in the long run it’s the
failures and even the near misses where the lessons lie. In the realm of analytics this
includes the specter of historical business intelligence and analytic investments or data
warehouse projects that ended with a fizzle. Or worse, never ended – but never created
value either.
Ultimately, analytics is about exploration. And whether drilling for oil, gold or the next
business insight, it can take a lot of holes to find the proverbial nugget. As a result,
adopting an analytic mindset often requires organizations to redefine success.
Enter the concept of failing to succeed. In the analytic enterprise, disproving a hypoth-
esis or getting a null result doesn’t translate to a failing mark. Analytic organizations
recognize that results are neither good nor bad. They are just results. And every result
(or the perceived lack thereof) is a clue to what to investigate or not investigate, do or
not do next.
Of course, this assumes that each analytic experiment arrives at a result. Any result. Yes,
analytic enterprises embrace the fact that failure is a requisite part of success. But they
also ensure analytic practices are disciplined enough to redirect attention and refocus
efforts before too many resources have been consumed in vain. This practice is often
referred to as “failing fast.”
Perhaps John W. Holt Jr., co-author of Celebrate Your Mistakes, said it best: “If you’re
not making mistakes, you’re not taking risks, and that means you’re not going anywhere.
The key is to make mistakes faster than the competition so you have more chances to
learn and win.”
This is How We Do It
When in doubt, organizations, like individuals, operate by rote, falling back into routines
they may not even be conscious of. This is procedural memory – the unconscious knowl-
edge of past behaviors and habits – at work. For example, an individual with amnesia
may not be able to recall his or her name, but can sign on the dotted line without think-
ing. Or play the piano with no memory of learning.
When it comes to our business practices the analogy applies. In lieu of a conscious and
concerted effort to challenge the status quo, things proceed as they ever were. Estab-
lished business practices and processes are held inviolate – regardless of whether the
original business justification or drivers (if they are even known) still apply. In the IT world
this equates to the legacy system that can’t be phased out because no one can remem-
ber what it does or how it does it.
Adopting an
analytic mindset
often requires
organizations
to redefine
success.

9
Creating a data-driven culture requires companies to cultivate a mindset that both ac-
knowledges and is open to the need for change. A deliberate plan of attack – complete
with incentives – is also necessary to make it stick.
While few companies have the luxury of starting with a clean slate, rethinking ingrained
beliefs and practices doesn’t necessarily require razing existing operating models to
the ground. It does, however, require companies to constantly challenge themselves: Is
this the best it can be? Is there another way? What do I not know today that I could use
tomorrow?
In many cases, there is much to be learned from simply taking a peek outside traditional
industry or market boundaries. As an example, government agencies are taking a page
from the public sector and retailers in particular. These agencies have begun using a
broad array of data sources to segment their customers (previously known as citizens
and the populations they serve) and tailor their interactions accordingly. Utilizing this
approach, revenue agencies in Europe have reported up to a 10 percent increase in rev-
enue collected. This is a significant chunk of change that can – let us hope – be used to
further the public good.
There is much
to be learned
from simply
peeking outside
traditional
industry
boundaries.

10
Analytic enterprises manage analytics as a core business practice no different from sales
or marketing. And like established business functions, analytics must enable and con-
tribute to the company’s core goals and objectives. It’s a point worth repeating.
Interesting models are, well, interesting. But if the business doesn’t see the value or isn’t
willing to put action to insight, the exercise is moot.
Informing and Enabling Business Strategy
Analytic innovators gain commitment (and fans) by speaking to the business in its own
terms about the things that matter to it. Or, as one executive said: “[They] follow the
money.”
To that end, the analytic enterprise is hyper focused on analytic projects and capabilities
that will:
• Inform corporate strategy.
• Enable or drive operational execution and decision making.
• Stimulate innovation.
Developing an analytic agenda that addresses these needs without veering far afield
isn’t as hard as it might seem.
Most organizations have well-defined processes for corporate strategy development
and operational planning. Aligning analytic programs and research projects with identi-
fied business objectives and goals is the single best way to achieve this goal.
Finding Your 1 Percent
Of course, for most organizations the issue isn’t necessarily finding a problem to solve.
It’s picking just one. Or two.
Analytic competitors are masters of focus, concentrating their attention on questions
that tie directly to what makes them tick. That being said, it isn’t necessary to solve all
your company’s woes in a single pass. When thinking about where to start and what to
do next, the following analogy is useful.
MAKING THE CASE
inciting passion & enlisting commitment
Analytics must
enable and
contribute to
the company’s
core goals and
objectives.

11
GE coined the term “industrial Internet” to encapsulate the combination of intelligent
machines with advanced analytics and enlightened subject matter experts (SME) to op-
timize how people work. GE’s hypothesis is that relatively small changes – single-digit
improvements in operational effectiveness and efficiency – can drive HUGE economic
and social benefit.
The savings GE projects across industries are impressive. But the real point is to ask
yourself and your business partners: What’s our 1 percent? It really can be that simple.
And effective.
A multinational insurance company focused on optimizing its processes to highlight
claims that appear low-risk but result in costly litigation later. How? They combine and
visualize:
• Customer profiles and demographics.
• Behavioral data from call logs, Web visits and social media up to and at the
time a claim is submitted.
• Relationships between individuals and companies suspected of fraudulent
practices via social network analysis.
The payoff? A 1 percent reduction in the billions of dollars spent annually to settle work-
ers’ compensation claims.
Is this innovation in the groundbreaking, state-of-the-art, never-before-seen sense of the
word? It’s safe to say no. But it is a new way of doing business. And tens of millions in
retained revenue isn’t chump change. More importantly, the project was a powerful proof
point for the power of analytics and data-driven decision making. Bigger, cutting-edge
projects followed suit.
What’s your 1
percent?

12
An appropriate mindset and commitment to the analytics cause is for naught if your or-
ganization doesn’t have the ability to execute on the vision. To succeed you must devel-
op and nurture the skills, processes and tools that enable data-driven decision making.
The Analytic Process
Data-driven decision making is about much more than just development of an analytic
model. And yet, as shown in the figure , discussions regarding the analytic life cycle have
typically focused on the process by which the analytic model will be created and man-
aged (orange), almost to the exclusion of the overarching business process and decision
points (blue) that surround it.
DEVELOPING YOUR
ANALYTIC MUSCLE
IDENTIFY /
FORMULATE
PROBLEM
DATA
PREPARATION
DATA
EXPLORATION
TRANSFORM
& SELECT
BUILD
MODEL
VALIDATE
MODEL
DEPLOY
MODEL
EVALUATE /
MONITOR
RESULTS
ANALYTIC BUSINESS PROCESS
ANALYTIC MODEL LIFECYCLE
business hypothesis
or problem space
probable business
value & utility
• go/no go
• “go to market”
strategy
• update policies & SOP
• business process
& application integration
• monitor business
outcomes
• refine business
processes
figure 3. analytic business process

13
But as hard as finding the nugget is, moving from insight to action is harder. Turning in-
sight into action requires a holistic approach to both creation of analytic models and the
care and feeding of the insights they provide.
Analytic leaders address data-driven decision making as an integrated business pro-
cess.
The Analytic Community: More Than A Data Scientist
There is a lot of talk about the shortage of trained analytic workers. In 2013 Gartner pre-
dicted the creation of 5.5 million data scientist-related jobs in the next five years – and
predicted only 30 percent would be filled. The numbers are staggering. But the fact re-
mains that a data scientist in a vacuum does not an intelligent organization make.
As shown above, data-driven decision making requires input and participation from us-
ers across the spectrum, from executives to business analysts. Change management is
also critical, as enterprises new to analytics must:
• Educate executives and decision makers on the benefits and enlist their sup-
port.
• Enculturate analytic skills and awareness from the C-suite to the feet on the
street.
• Proactively engage business owners to identify areas for investigation and to
apply found insight.
• Redefine existing decision-making practices and protocols within both op-
erational and strategic business processes.
Data scientists are a unique breed and possess a necessary and rare skill set. But in the
analytic enterprise, executives, business decision makers and knowledge workers also
need to step up to the plate.
A data scientist
in a vacuum
does not an
intelligent
organization
make.

14
The Enabling Technology
The analytic enterprise recognizes that data-driven decision making requires input and
participation from multiple user communities. It also recognizes that the enabling tech-
nology is not one solution, but an ecosystem in which different solutions enable different
functions and stages of the analytic life cycle. The good news here is that the technolo-
gies to support the full life cycle now exist. And this is really the big deal about big data.
So technology, while a key ingredient, is no longer the key barrier to data-driven decision
making.
DELIVERY & CONSUMPTION
ANALYTIC SOLUTIONS
PROCESSING
ADVANCED ANALYTICS
DATA MANAGEMENT & ACCESS
CAPTURE & STORAGE
reporting OLAP visualization knowledge
discovery
model
management
business
applications
Hadoop NoSQL columnarSQL relational
databases
files
(XML, docs)
text voice video natural
language
sentiment network geospatial
in-memory in-database complex
event
event
streaming
machine
learning
Map-
Reduce
DM
(DQ,
Metadata)
MDM/
RDM
data
access
data
integration
data
virutalization/
federation
data
security &
privacy
visual
analytics
data
mining
forecasting optimization machine
learning
figure 4. the analytic ecosystem

15
Organizations are recognizing the need to manage analytics as a program and make the
journey toward data-driving decision making like any other business practice. Thus the
advent of chief analytic officers and data scientists as well as the re-emergence of ana-
lytic centers of excellence. Regardless of the approach, successful analytic programs
exhibit common characteristics. They are:
Strategic
They develop an analytic agenda based on a broad, cross-functional view of the organi-
zation. Why? In order to develop a focused analytic agenda based on corporate strategy
and objectives that will meet the needs of multiple functions. In this way, the center of
excellence ensures it is driving a transformative agenda and does not become a help
desk or order taker.
The implication here is that analytics needs to be governed. Creating awareness, engen-
dering organizational buy-in and staying the course requires a steady – and authoritative
– hand on the wheel.
Collaborative
Effective analytic groups engage the customer in the process. If your model is interest-
ing, but the business doesn’t see the value, the entire exercise is moot.
Acknowledging this point, a national financial services institution has established an ana-
lytics lab where the business consumers work side by side with analytic experts and
data management teams from knowledge discovery through to model validation and
business process re-engineering. The organization has also borrowed from agile devel-
opment philosophies to create staged decision gates at each point in the life cycle to
encourage failing fast.
Yes, conflicts will inevitably arise between the long-in-the-tooth executive or business
user who knows how things run (inside out) and the data scientist who takes a different
view (outside in). But rather than discouraging this discourse, we need to encourage it.
Because it’s in this margin that innovation occurs.
Leading organizations also partner with IT or data management teams to deliver both
the required data environments and the suite of analytic tools to exploit them. That be-
ing said, IT cannot own analytics. IT can partner with the business to enable data-driven
decision making. IT can and should serve as a thought leader educating the business on
the art of the possible. In fact, this is critical. But the business must have both the willing-
MANAGING THE CHANGE
organizing for success
Manage the
journey toward
data-driven
decision making
like any other
business
practice.

16
ness to explore and the will to act on generated insights, which translates into the need
for the business to own and operate the analytic engine like any other business process,
be it sales, marketing or supply chain management.
Influential
Analytic innovators develop deep relationships and forge productive partnerships with
different constituencies, including the executives.
They also recognize and embrace their roles as change agents and the importance of
consistent communication, expectation management, education and enlistment. And
they ensure that skill sets to play this role are present and accounted for.
Last but not least, they see their mission as one of enablement, not world domination.
This means facilitating the growth of embedded skill sets and developing a critical mass
of analytic talent, both inside and outside of the centralized analytic team or competency
center itself.
Skilled
Successful practitioners establish a reputation for being highly-skilled specialists who
are committed to staying abreast of emerging tricks and techniques and incubating
those capabilities for the enterprise.
This does not mean, however, that all development occurs on a centralized team or in
a center of excellence. Highly functional teams realize when their expertise is and isn’t
necessary and strive to engage only in a value-added manner. This might run the gamut
from delivering a project or program, advising on it or acting as a consultant.
If analytics is new to the organization or function the analytic center may play a central
role in project execution. As the analytic maturity of different functions or the organiza-
tion at large improve, the ability to act as a trusted adviser and consultant can be more
important.
Results Oriented
Last but not least, highly performing organizations focus on outcomes. Action, not just
insight, is the name of the game.
Analytic
innovators
embrace their
role as change
agents.

17
Cultivating an analytic culture and embracing data-driven decision making is no small
undertaking. It’s easy to get lost in the hype or become discouraged, especially when
dealing with often inflated expectations for immediate outcomes associated with big
data. Or while that so-called little data is still proving to be a big challenge.
But while the journey forward may at times feel arduous, it is achievable – even if your
enterprise wasn’t born with analytic DNA.
Organizations that make the turn engage the mind, connect with the heart and develop
the analytic muscle of their organizations to achieve their goals. These elements are criti-
cal for organizations to transcend traditional mindsets and create an analytic culture that
embraces and drives innovation.
CONCLUSION
MIND
ALIGNS
analytics to
business
strategy.
CULTIVATES
a willingness to
experiment and
the will to
execute.
INVESTS
in and nurtures
required skills,
processes and
solutions.
figure 5. the analytic enterprise

SAS Institute Inc.
100 SAS Campus Drive
Cary, NC 27513-2414
USA
Phone: 919-677-8000
Fax: 919-677-4444
about the author
KIMBERLY NEVALA is responsible for industry education, key client
strategies, and market analysis in the areas of business intelligence
and analytics, data governance, and master data management. She
has over 15 years’ experience advising clients on the development
and implementation of strategic customer and information manage-
ment programs and managing mission-critical projects.

A frequent speaker and writer, Kimberly is often consulted on the
topics of business strategy and alignment. She is the co-author of
the first eBook on data governance, The Data Governance eBook:
Morals, Maps and Mechanics, as well as Planning Your BI Program: A
Portfolio Management Approach and Top 10 Mistakes to Avoid When
Launching a DG Program.
SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA
and other countries. ® indicates USA registration. Other brand and product names are trademarks of their respective companies.
10 _S1 .0 14

�ݺ�ߣ

Ast 0127927 anatomy-of_an_analytic_enterprise_107111_0514

More Related Content

Ast 0127927 anatomy-of_an_analytic_enterprise_107111_0514