Understanding Chief Data Officer
Understanding Chief Data Officer
strataconf.com
Presented by OReilly and Cloudera,
Strata + Hadoop World is where
cutting-edge data science and new
business fundamentals intersect
and merge.
Job # 15420
Understanding the Chief
Data Officer
How Leading Businesses Are
Transforming Themselves with Data
Julie Steele
Understanding the Chief Data Officer
by Julie Steele
Copyright 2015 OReilly Media, Inc. All rights reserved.
Printed in the United States of America.
Published by OReilly Media, Inc., 1005 Gravenstein Highway North, Sebastopol, CA
95472.
OReilly books may be purchased for educational, business, or sales promotional use.
Online editions are also available for most titles (http://safaribooksonline.com). For
more information, contact our corporate/institutional sales department:
800-998-9938 or [email protected].
While the publisher and the author have used good faith efforts to ensure that the
information and instructions contained in this work are accurate, the publisher and
the author disclaim all responsibility for errors or omissions, including without limi
tation responsibility for damages resulting from the use of or reliance on this work.
Use of the information and instructions contained in this work is at your own risk. If
any code samples or other technology this work contains or describes is subject to
open source licenses or the intellectual property rights of others, it is your responsi
bility to ensure that your use thereof complies with such licenses and/or rights.
978-1-491-92058-9
[LSI]
Table of Contents
v
Understanding the
Chief Data Officer
Introduction
Its been hard to miss the swelling tide of big data over recent
years, but just what that term means to business and how it should
be managed within an organization is still evolving. With the
increasingly vital role technology has come to play over the last sev
eral decades, the Chief Technology Officer (CTO) and Chief Infor
mation Officer (CIO) were introduced and became familiar roles in
many organizations. Less well understood is the nascent but rapidly
spreading Chief Data Officer (CDO) position.
Many companies understand that data, when used correctly, can
yield tremendous valueor even change entire industries. Just look
at what Amazon or Netflix has done with recommendations, to take
a new school web example, or what Walmart has done with supply
chain optimization, to take an old school retail example. What
company wouldnt want to achieve such results through data?
If those examples constitute the carrot, then there is also a data stick.
Industries such as finance and healthcare, which regularly deal with
sensitive personal information, have become more heavily regulated
in terms of how they must handle and protect their data. Even
without regulation, some recent hacks have made the prospect of a
large data breach a very scary possibility for anyone handling credit
card transactions.
Whether tempted by the carrot of new products and efficiencies, or
harrowed by the stick of privacy and security concerns, many
1
companies have elected to address these issues by appointing a
CDO. But the territory is still relatively uncharted.
Right now, there are as many implementations of the role of CDO as
there are organizations implementing it. Everything about the job,
from reporting structure to primary responsibilities to required skill
sets, can vary with the industry, company, and individual. But there
are some very distinct patterns and categories, and some common
threads that can yield insight for those considering a CDO of their
own.
This report presents a picture of the current landscape, as well as
some guidelines and best practices for those considering adding a
CDO to their own organization. I spoke with a dozen professionals
who have performed the role in various settings including health
care, telecom, finance, marketing, insurance, and government at the
municipal, state, and national levels. Their collected wisdom shows
how the right data leadership can make companies more customer-
focused, competitive, and influential.
Centralization
Enterprise-scale companies may consist of dozens or even hundreds
of smaller companies, divisions, and other components. And each of
these produces data. The CDO is in charge of gathering the data
from across these different silos and bringing it into one central
Evangelization
The CDOs team is not a new department that is simply appended
onto the old way of doing things, like a third arm that adds incre
mental capability. It is more like developing a nerve system: it works
with all the other parts of the organism, collecting information and
passing signals back and forth in a way that allows for better collec
tive action and decision-making. A nervous system is not made of
muscle; its job is to inform, not act all by itself.
The CDO team is typically quite small compared to the rest of the
enterprise, so convincing others in the organization that this kind of
work is worth investing in is critical to success. The CDO must be
an advocate for data-driven approaches, and achieve buy-in from
Facilitation
Of course, once youve won everyone over to the importance of
working with data in new ways, you then have to remove some bar
riers and free up the resources to make it feasible. The ideal CDO is
one who makes better, more efficient action possible for the rest of
the organization.
Part of the job was to represent the resource needs for data practi
ces within the priorities of the agency, said former FCC Wireless
Telecommunications Bureau CDO Rob Alderfer. A lot of the stuff I
Reporting Structures
The responsibilities outlined here are the central themes that have
emerged over a period of time. The reality is that, for many early
CDOs, job responsibility number one was to figure out where they
fit and what their other duties should be. The role has sometimes
been created without a very specific idea of what the organization
hopes to accomplish with data, or how the CDO role should be
positioned relative to the existing hierarchy.
Bonaguro experienced this in San Francisco, where the position was
mandated by legislation but not well outlined: Defining and under
standing where the role sits in the existing structure was something
that had to be done.
Micheline Casey also encountered this at the Federal Reserve Board:
Theyd never seen a CDO, and they werent sure at all what a CDO
was supposed to do. They were sure something was needed, but they
werent sure what that looked like, smelled like, tasted like.
So what you find right now is that the reporting structures vary
every bit as much as the job responsibilities: some CDOs report
directly to the Chief Executive Officer (CEO), while others report to
the CTO, CIO, or even the Chief Financial Officer (CFO). Some
larger enterprises such as AIG have multiple CDOs, one for each
major group, and sometimes also one for the overall conglomerate.
One of the reasons we are seeing such a variability with CDOs is
because you may have one business whose definition is that the
CDO is really the data steward, and they report to a team within the
CIOs office (so there may be one or two levels down from the CIO).
On the other end of the spectrum, we have a Chief Data Officer who
is peer to the Chief Marketing Officer and the Chief Financial Offi
cer, and who is really changing the direction of where the organiza
tion is going around data and analytics, said Mark Ramsey of
Samsung.
This latter scenario seems to be the ideal one. In most cases, it
doesnt make sense for the CDO to be nested under the CIO in the
long run: their areas of responsibility certainly converge occasion
ally, but the CDO is responsible for organization-wide policy and
Reporting Structures | 9
data management. To go back to the earlier analogy of the nervous
system: such a system works best when it is wired directly into the
brain of the place.
Business Challenges
The most important part of the CDOs job is to align technology
with business goals. Much can be done with data, as the press
delight in showing us every day, but the key is to do things with data
that will directly support business objectives. The CDO is in charge,
in large part, of making sure that the paths an organization pursues
with data are pursued for the right reasons. And typically, those are
business reasons.
Data, while supported by technology, is not fundamentally a tech
nology problem. Your information systems can house the data, but
your questionslike, Should we be running this program?those
are business questions, said Joy Bonaguro of San Francisco. And
On data strategy
In order to develop good business questions and to answer them,
one needs to have a data strategy. The ideal data strategy is written
in collaboration with all the business stakeholders, so it is well
understood and agreed upon. It outlines the things that matter to
the company, lays out a roadmap for how data will be used to help to
achieve those goals, and provides actionable plans for how to get
started.
That last part, action, is critical. Without the ability to act, a data
strategy is just another document that will moulder away. It is great
to have a strategy, but we actually have to deliver results, and we
have to make a difference for the business, said Floyd Yager of
Allstate.
Kolker, of Seattle Childrens, explained that in the healthcare busi
ness especially, the ability to take action is imperative. One of the
lessons we learned the hard way is that data analytics, data science,
data modeling is not enough. Its necessary, but we wanted not just
to get data and do analytics on it but to get something that is action
ableactionable insights. We wanted to change the business, and
change is very tough, especially in an industry like ours where were
talking about peoples lives and health. Its not like somebodys trying
to optimize clicks.
In addition to being actionable, a good data strategy must also be
flexible. It should be a living document that can adjust as business
priorities change and technology evolves. It can lay out a new road
map when data indicates that your current tactics are no longer
working: when your investment strategy stops yielding a competi
tive advantage or your supply chain is no longer optimal, for
instance, or when another concern altogether eclipses those issues.
On data governance
Of course, how the data gets used, particularly in certain settings
like hospitals and financial institutions, can be very sensitive. As dis
cussed at the top of this report, part of the initial genesis of the CDO
role was the need for regulatory compliance and oversight of whats
called data governance: ensuring that data is handled according to
strict standards and guidelines.
In newer businesses such as car-sharing or online dating, there
arent yet regulations regarding customer privacybut there proba
bly should be. In the absence of any clear guidelines, companies in
these fledgling industries must guess at how to avoid violating their
users trust. To cross that line is not only disastrous for public rela
tions, but also a potential legal problem. Conscientious data gover
nance is almost more imperative in the absence of regulation, since
the boundaries between fair data handling and mistreatment can be
much less clear.
While it seems like a no-brainer to address these issues before any
thing else, not everything can be done all at once. The ability to
organize and prioritize is crucial, and a CDO can bring a lot to the
table here.
On the governance and policy side of things, said Jennifer Ippoliti
of Raymond James, I have created a central point of escalation for
data issues so that we can look at them together: we can prioritize
them independently in terms of what is best for the enterprise as a
wholeand not just one particular user or group of usersand
Conclusion
Just as the duties and purview of the CIO used to vary from com
pany to company, so are the duties and the purview of the CDO a bit
muddy right now. Over time, it may stabilize as the CIO role did:
possibly even into multiple distinct positions. But the common
threads of centralization, evangelization, and facilitation are likely to
Conclusion | 19
About the Author
Julie Steele is Director of Communications at Silicon Valley Data
Science. She thinks in metaphors and finds beauty in the clear com
munication of ideas. She is particularly drawn to visual media as a
way to understand and transmit information, and is co-author of
Beautiful Visualization (OReilly, 2010) and Designing Data Visuali
zations (OReilly, 2012). She teaches an evening course on Data Met
rics and Visualization at the School for Visual Arts in New York
City, and can be found tweeting about all things data-related as
@JSteeleEditor.