Developing a Data

Strategy
By

Heather A. Smith
James D. McKeen

The IT Forum
A focus group of senior IT managers from a variety of different industries
convened regularly by the authors to address key management issues in
IT. This report highlights a recent discussion.
– See back page for details of the IT Forum and other reports.
What’s the Issue?
In many ways organizations seem to be making a full circle from data processing (DP) to
information systems (IS) to information technology (IT) back to a new focus on data
processing. Data, and the information and knowledge that can be derived from it, is
one of the biggest challenges facing organizations these days. There are a number of
reasons for this. First, data is integral to both the work of business and IT, but whereas
business units want data and processes to stay stable and primarily support their
individual functions, IT has an interest in standardizing much of it at an enterprise level
to achieve synergies and create new opportunities. Second, we have come to recognize
that how we manage data in both business and IT will significantly affect our
capabilities, products, and services and thus our business and technology strategy.
Thus, data is no longer considered a subset of a system, but important and valuable in
its own right. Third, we now recognize that data is a both a corporate asset and a
corporate vulnerability that needs to be protected and developed. And finally, data is
booming. New types and amounts of data are already swamping organizations and IT
managers realize that they are only on the tip of a very large data iceberg. For these
reasons, and others, IT managers believe it is now time to develop a data strategy.

This is easier said than done. In an earlier paper on this subject, we concluded:

“All organizations have a generic vision of delivering the right information to the
right person at the right time. However achieving this goal involves careful
consideration of what an organization wants to accomplish with information and
how it proposes to derive business value from it.” (McKeen and Smith 2015).

Similarly, while much has been written about business and technology strategy, data
strategy is very much a green field. Many organizations have found the task of
developing “an intergalactic data model” daunting in the extreme and most have failed
-- and this doesn’t even begin to address data strategy. Today, it is fair to say that
organizations are all over the map when it comes to developing a data strategy, but
most are finding the lack of one to be limiting to what they want or need to do.

Developing an effective data strategy therefore involves first articulating why a

company needs one in the first place and then determining what might be included in
one. As part of this process, it is important to be clear about where a data strategy fits
in relation to all other data and corporate activities, such as master data, big data, and
architecture – to name just a few. Finally, we need to learn from previous failures to
establish realistic and achievable stages towards data strategy maturity and identify
some lessons learned from them.

IT Forum Developing a Data Strategy 2

Why Do We Need a Data Strategy?
A strategy of any type in a business context – be it business, technical, data or
otherwise – is a plan to help an organization succeed. There are two ways it can do this
(Porter 1996, 2001; Smith et al. 2010):

1. Developing new capabilities. With this approach, strategies are designed to help
organizations become more flexible and responsive to changing markets and
competition. Here, the emphasis of strategy is on improvisation, adaptation, and
reinvention throughout the organization. Companies focus on their core
competencies, i.e., key resources and capabilities, and best practices and
benchmarking to keep ahead of or up with their competitors. Their ‘edge’ comes
from continual reinvention and lots of experimentation.

2. Creating a uniquely competitive position. More traditionally, strategies are

designed to perform activities that are different from those of their competitors or
to do them in different ways. These stress tailoring a company’s activities to address
its chosen position in the marketplace and designing them so they reinforce each
other in a coherent
fashion. Competitive
advantage and the ability
to sustain it arises from
integration across all
activities. This approach to
strategy is not to be
confused with performing
similar activities better
than competitors, which is
operational effectiveness
and which should be
considered part of an
organization’s everyday
activities.

As we have noted
elsewhere (McKeen and
Smith 2015), strategies of
all types co-evolve with
Figure 1. Strategy Co-evolves with Technology and new technologies, new
Behavior capabilities, and new
behaviors and structures (see Figure 1). These elements must reinforce each other to
achieve a company’s strategic goal. Although a company’s position relative to its
business environment must always be paramount in all strategy development, business
opportunities can come from any of these elements and business value can only be
derived from all of them working in concert.

IT Forum Developing a Data Strategy 3

Ideally, the “strategy” of an organization is further detailed in specific business and
technology strategies that are consistent and aligned across the enterprise (McKeen
and Smith 2015). Until recently, data issues have been addressed within technology
strategy and in piecemeal ways, such as plans for storage, databases, a data
warehouse, or master data. Developing a separate data strategy that would
complement both business and technology strategy has not been a priority for business
or IT leaders (Dayley et al.2014). Today, several factors are now changing this:

New analytical technologies. These new tools enable large quantities of data from
a variety of sources to be integrated and analyzed in order to accomplish new
activities not possible previously. These include: developing new and more detailed
insights into current company practices (e.g., by particular demographics, or
products); better and more quickly measuring the impact of new strategies on the
market; and using the information derived from it to create new products or
services.

New types of data. Companies are being flooded with data and are therefore
seeking ways to take advantage of it while managing the volume, variety, and
velocity of data involved. Organizing and managing data in the “old days” before
the internet and other forms of external data were available was challenging
enough. Companies have long struggled with data accuracy, integrity, and delivery
issues (Ferguson 2014; Redman 2013). Now, with so many new types of data being
created and made available, much of it external to the organization, such as social
media, sensor, unstructured data to name just a few, companies are recognizing
that they need to “up their game” in utilizing it.

New risk and compliance concerns. Increasingly, governments are enacting laws
to require that companies provide accurate information on their performance in a
number of areas. In addition, high profile security breaches are showing business
leaders first hand some of the market downside of failing to protect their data. This
is driving demands for new policies and practices concerning data usage and
handling, such as retention, security, privacy, and access.

New competition. New products and services are emerging in different industries
that demonstrate the market value of using data for competitive advantage. Stories
of companies that have effectively done this (e.g., UPS) abound, as do stories of
companies that have literally changed the marketspace (e.g., Amazon) for products
and services. The explosion of data intensive devices that are all connected is
creating ever new ways of working, playing, and living that are changing not only
the types of data that are generated but also driving new customer needs for data
didn’t exist previously (e.g., Fitbit).

IT Forum Developing a Data Strategy 4

 New value. Companies are discovering that considerable business value can be
achieved by incorporating data into their strategies. From a productivity point of
view, researchers have found that knowledge workers can spend up to 50% of their
time hunting for data, identifying and correcting errors, and seeking confirmatory
sources for data they do not trust (Redman 2013). Another study found that up to
14% of a company’s revenue is lost when enterprises do not manage and analyze
their data (Bowen and Smith 2014). And data can be used to improve productivity
and effectiveness (McKeen and Smith 2015).

Therefore, many in the focus group believe it is time to take a more comprehensive and
strategic look at how data are used in their organizations. “The demand for data is so
strong in our company,” said one
manager, “that if we’re not out in
front with a data strategy for the
business, it will run ahead of us.”
Another pointed out, “our game
is being upped with the
introduction of analytical power.”
“Because we’re so data
dependent, we are implicitly
driving a data strategy, but this
needs to be better aligned with
our business and technology
strategies,” said a third.

They agreed that a new model of

strategy development should
incorporate business, technology
and data strategies and that each
Figure 2. Strategy Development is Iterative &
of these must not only work in
Collaborative
concert with each other but also
be open to external competitive forces and new possibilities created by new
technologies and new types of data (see Figure 2).

What is a Data Strategy?

Although there is a clear need for companies to strategize about data, there is little
agreement about what actually constitutes a data strategy. Focus group members were
the early stages of thinking about or developing one. They all agreed however that it is
important to conceive of a data strategy as a “living” plan. Although there are many
good reasons to have a data strategy in place, they noted that there is little appetite in
the business for overly constraining approaches that cost a lot and deliver nothing but
data models in a few years’ time. Leaving this comprehensive approach aside, the

IT Forum Developing a Data Strategy 5

companies in the focus group presented a range of philosophies about what a data
strategy should be that trace a maturity path from data chaos to managed agility:

1. Pragmatic. “Our data strategy must be flexible and iterative,” said one manager.
“We’re changing so fast we need pragmatic data solutions that we can use to
enable business results asap.” He compared this approach to feudalism in that it
serves local needs in the lines of business, rather than the enterprise as a whole.
Another explained how this approach functioned in her organization. “There’s no
data that we don’t have but only a few people can tell you where it is.”

2. Action-oriented. “We are trying to understand what critical business need we

should be addressing with data, and organize our data strategy around it,” said
another manager. This approach could be compared with early industrialization
where design practices were beginning to be introduced.

3. Risk-oriented. Companies begin to move into more enterprise-wide approaches

when they recognize the risks associated with data. “We’re trying to manage our
legal exposure, appropriate uses of data, use of data to make erroneous business
decisions, and operational exposures related to loss of privacy and security,” said
one manager. “We consider ourselves stewards of data and we need to understand
what are the boundaries of control, risk, and regulation that must operate within.
Then, we need to ask ourselves what are we doing within these boundaries to better
manage and use data?”

4. Post-industrial. Although none of the focus group companies is at this point, the
vision for a future data strategy is quite clear. Here, good quality data is available to
appropriate users, along with useful meta-data, such as when it was created and
who created it, and standard definitions, and organized in ways that it can be easily
found, accessed, and utilized. In addition, appropriate tools are available to support
analytics, information creation, and delivery. “This strategy supports a ‘pick what
you need’ model,” explained one manager.

In addition to understanding what an organization wants from its data strategy, it is also
essential to understand what data it wants to include in its strategy. The managers in
the focus group stressed that tackling too much all at once can be a recipe for failure.
“We tried to develop a complete information architecture a decade ago and it wasn’t
successful,” said one manager. “Then a few years later we tried to articulate a
comprehensive strategy but it was too polarizing in the organization. Now we are taking
a more staged and pragmatic approach to strategy development.”

Participants have tried a variety of ways to determine what data to focus on. This
includes limiting their attention to one or all of the following types of data:

Enterprise level data

IT Forum Developing a Data Strategy 6

 Data subject to any sort of compliance or regulatory need
Data needed for competitive purposes
Metadata to identify authoritative sources and content owners
Foundational data, i.e., data that is central to the business, also known as the book
of record
External data to improve business functionality or to develop new opportunities.

Finally, a data strategy must address a number of other business and technical issues
about data once a general direction and scope has been articulated. In an earlier paper
(McKeen and Smith 2015), we suggested there are four major sets of issues related to
data: policy, operations, stewardship and standards. Each of these will guide the
decisions that are made about data and must in turn be informed by a specific strategic
approach and focus that is set at a high level in the organization.

Policy provides the basis for developing the processes, standards and guidelines
needed to manage data assets well throughout the organization. It directs the
development of more detailed practices related to accountabilities, quality,
protection, risk tolerance, compliance, the creation of new assets, acquisition,
retention, integration, analytics, and records management. Because of the large
number of stakeholders affected, it should be developed by a multidisciplinary
team, including IT, key lines of business, and the legal, HR, audit, privacy, and
compliance functions.

Operations includes specific practices that need to be put in place in order for a
strategy to be successful. These practices will vary depending on the strategy
involved but should address: governance, planning, technology and architecture,
education and behavior, and processes for each of areas outlined in the data policy.

Stewardship specifically describes the roles and responsibilities associated with

data management and distinguishes between business and technical
responsibilities. Data management is a boundary-spanning activity. Data owners are
typically business people, although there appears to be a general lack of interest for
this role in the business. Owners define data, determine who should have access to
it, and are accountable for its accuracy, integrity, currency, and meaning.
Technology roles are both custodial -- determining how retained data will be stored
and accessed, mapping which systems use which data -- and strategic -- developing
a technology foundation that supports appropriate data architecture, multi-purpose
access, and protection, as well as the creation of a data dictionary and acquisition of
appropriate analytic tools.

Standards enable the organization to develop enterprise-wide solutions, better

integrate evolving new technologies, implement partnerships, and create innovative
solutions (Palem 2014). As noted above, there is ongoing tension in many
organizations between IT, which wants to promote the adoption of standards, and

IT Forum Developing a Data Strategy 7

 the lines of business, which tend to prefer to use data in their individualized context.
Many focus group organizations have made considerable progress in this area. Most
have established master data for key customer and corporate records. Several are
using compliance requirements to move forward with standardization of key data
assets. Nevertheless, standards continue to be a divisive issue that needs to be
pursued with caution.

Where Does Data Strategy Fit?

A significant challenge for any organization is determining where data strategy fits in
the organization structure. Is it a business or an IT function? At what level in the
organization should it report? On one hand, it is the business that must decide which
data are important and how to use it to deliver business value. On the other, IT has the
responsibility for the data lifecycle from acquisition through organization, processing,
maintenance and retention. The majority of focus group companies now have a senior
data role in their IT organizations, with titles such as Director Data Services, AVP Data
Strategy, or Chief Data Officer. Only one had positioned this role outside of IT. As with
other boundary-spanning initiatives, IT is often called upon to bring its enterprise
perspective to the table. However, in doing so, there is a danger that the business
could see data strategy as “just another technical initiative”.

The focus group stressed that technology should not be the sole driver of data strategy
or it will be doomed to failure. Multi-disciplinary governance is essential to provide the
sponsorship and guidance to make it effective (Laney and Beyer 2013; Daley et al.
2014). Most of the group’s participants had therefore created data governance councils
to provide the necessary direction to both IT and business regarding data strategy and
its implementation. It is this council that ensures data strategy is consistent with and
complementary to business and technology strategy, focuses on key data assets, and
that new opportunities for leveraging data are addressed. It also determines how fast to
move with implementing data architecture, standards, roles, and activities and acts as
mediator in the event of disagreements over the management of shared assets.

A second challenge for data strategists is: Where does it fit with the rest of the “data
ecosystem”? Today, there is considerable confusion among both business and IT
leaders about both the variety of data an organization must cope with and they can
typically have a number of different ongoing initiatives associated with some type of
data. This problem is further exacerbated by the fact that the data ecosystem is also
changing rapidly with new types and sources of data being developed on a regular
basis and new forms of competition evolving through the use of data analytics. The
result is a rather fuzzy picture where some pieces overlap and with significant gaps,
making it difficult to articulate a clear data strategy.

IT Forum Developing a Data Strategy 8

A key prerequisite to developing an effective data strategy is therefore understanding
and clarifying an organization’s data ecosystem and putting in place ongoing activities
to further clarify it and fill in gaps. A company’s data ecosystem consists of both data
and activities that utilize it. As Figure 3 shows, a good starting point for data strategists

Figure 3. A Data Ecosystem

to inventory all types of data currently available in their organizations, ongoing projects
to use data or create new information, current data delivery and custody practices, and
technology strategy, architecture, and tools. “It’s unlikely that we will be addressing the
full ecosystem with our new data strategy initiative but we want to make informed
decisions based on our knowledge of it,” said a manager.

The Data Journey

The focus group stressed that developing an effective data strategy is a journey, not a
one-time activity. “We’re maturing in our data management practices,” explained a
manager, “and hoping that as we demonstrate the value of data, we can enhance and
extend our data strategy.” Thus, in addition to understanding the big picture of data
strategy, there are several practices that will develop organizational readiness to
become more strategic about data. These include:

Improving data quality. Data quality problems “plague every department, every
industry, at every level for every type of information” and employees routinely work
around or correct errors in their daily work (Redman, 2013). Poor data quality costs
the organization – in unnecessary work, in driving poor decisions, and generating

IT Forum Developing a Data Strategy 9

 distrust in data based decisions. High quality data enhances trust in what it’s saying
and encourages people to rely more fully on it (Ferguson 2014; Bowen and Smith
2014) and is a key factor in improving the information orientation of an organization
which has been shown to have a strong relationship with performance (Davenport
2007; Marchand and Kettinger 2000). Focus group members stressed that data
quality problems should not inhibit the development of a data strategy and efforts
to improve quality should address the root causes of poor data, rather than “fixing”
it at a later data. To do this, best practices include improving communication
between the creators and users of the data and developing data quality metrics to
monitor improvements (Redman 2013). In some cases, the growing size of the data
sets available may make errors less significant to correct as well.

Working on data definitions. Common definitions of key data elements, including

their attributes, are a fundamental component of data integrity, that is, data that
consistently means the same thing everywhere in the organization (Bowen and
Smith 2014; Lewis 2012). “We suffer from different data definitions,” said one
manager. “Everyone agrees that we should have a common definition but wants to
use their own version.” Appointing data stewards to “own” key data elements and
work with stakeholders to define them appears to be the best practice for making
headway in this area, but members warned that it can be a frustrating and time-
consuming activity unless there’s a direct value motivation. One company is using
advanced tools to assist this work such as, a semantic data base that links similar
concepts with different names, and automated introspection searches that pull out
linked data to make it easier to develop common definitions. “Tools can make this
problem more tractable,” said the manager involved, “but in the end the only way
to achieve common definitions and ensure these aren’t changed is through having a
data owner.”

Developing engagement. Many companies are seeking to demonstrate the

potential value of a data strategy by engaging people with data. “The more we can
provide self-serve analytics, the greater the value we can deliver to the business,”
said one manager. “We are trying to deliver pragmatic data solutions that will
enable business results asap and show how data can make us competitive,” said
another. Although many business managers are beginning to recognize data as a
business asset to be protected, most are still constrained by their experience and
need more insights into what new opportunities might be possible with data.
Generating excitement and engagement around small data initiatives is a therefore
a key way to obtain funding for more significant data initiatives.

IT Forum Developing a Data Strategy 10

 Setting priorities. All companies in the focus group had a number of clear data
priorities, even if they did not have a formal data strategy. These tended to be more
technical than business-oriented, reflecting the areas in which IT has control.
“Wherever possible, we do foundational work with data,” said a manager. This
includes: decoupling data from tools and applications to make it more accessible;
storing data in its rawest form so that it can be used for a variety of purposes; and
developing meta-data to ensure consistency.

Advice to Managers
Moving forward with a data strategy can be challenging, said the focus group. Funding,
skills, and resources are in short supply. Working with different business units with
differing business strategies, priorities, and timelines can mean that progress is patchy
at best. And generating business interest and appetite for foundational data work can
be a struggle. Focus group managers had several recommendations for others who are
finding it difficult to make progress with their data strategy:

Put someone in charge. It is more likely that progress will be made if someone is
paying attention to developing a data strategy and making it a priority.

Have a vision and set priorities. All members of the focus group knew what their
top data priorities were and why. Some were working on obtaining an enterprise
mandate for data; others on data architecture and standards; and many were trying
to provide easier access to data. Having a big picture vision enables IT to undertake
smaller projects of more practical value to business that still make a contribution to
achieving a broader data strategy.

Keep it pragmatic. Even with a mandate, it is important to continually show real

business value from data. “We are more likely to get support for and understanding
of data as a business asset, if we can deliver business value with it,” said a
manager. A corollary to this is that no organization should attempt to implement a
data strategy in a “big bang”. “This approach has consistently proven ineffective,”
said a manager.

Educate the business. Small proof of concept initiatives can help the business see
the opportunities for using data more effectively and highlight the value of how a
clear data strategy can support and enhance a business strategy.

Measure progress. People pay attention to what is measured. Therefore, several

participants have added data metrics to their enterprise scorecards. “We have
scorecard metrics around data quality and data governance,” said a manager. It is
also important to know what success looks like, especially in the early days of
strategy development. “We believe that success is any result that leads to more
demand for data delivery work that can, in turn, fund foundational data work,” a
manager explained.

IT Forum Developing a Data Strategy 11

Conclusion
Developing a data strategy is not for the faint of heart. It can mean stirring up a hornet’s
nest of territorialism and politics. It can also feel like a thankless task with very little
support or appreciation from business. Moving down this path involves knowing where
you are at present and where you want to be and then developing a business appetite
for improving data strategy through such activities as education and proof of concept
initiatives. As with many other organizational changes over the years, IT is anticipating
the future, and it’s all about data. Now, IT’s challenge to implement an effective data
strategy that will not only deliver immediate value to the business, but also put the
pieces in place to more fully enable new business strategies in the future.

References
Bowen, R. and A. Smith. “Developing an Enterprise Data Strategy”, Healthcare
Financial Management, V. 68:4 (April 2014): pp. 86-89.

Dayley, A., D. Logan, and G. Landers. “Best Practices for Data Retention and
Policy Creation will Lower Costs and Reduce Risks”, Garner Research:
G00262827 (April 2014).

Davenport, T. and J. Harris. Competing on Analytics: The New Science of

Winning, Harvard Business School Press, Boston (2007).

Ferguson, R. “Elevating Data, Analytics to the C-Suite”, MIT Sloan Management

Review, Reprint no. 55401 (March 25, 2014).

Laney, D. And M. Beyer. “Big Data Essentials for Business and IT”, Financial
Times, August 2, 2013.

Lewis, N. “Six Ways Healthcare can Assess Data Strategy”, Information Week,
April 16, 2012.

Marchand, D., W. Kettinger, and J. Rollins. “Information Orientation: People,

Technology and the Bottom Line”, Sloan Management Review, (Summer, 2000):
pp: 69-81.

McKeen, J. and H. Smith. IT Strategy: Issues and Practices (3rd Ed.). Pearson
Prentice Hall: Upper Saddle River, N.J., 2015.

McKeen, J. and H. Smith. IT Strategy: Issues and Practices (3rd Ed.). Pearson
Prentice Hall: Upper Saddle River, N.J., 2012.

Palem, G. “Formulating an Executive Strategy for Big Data Analytics”,

Technology Innovation Management Review, March 2014: pp:25-34.

IT Forum Developing a Data Strategy 12

 Porter, M. “What is Strategy?”, Harvard Business Review, (November-December
1996).

Porter, M. “Strategy and the Internet”, Harvard Business Review, (March 2001):
pp. 62-78.

Redman, T. “Data’s Credibility Problem”, Harvard Business Review, December

2013.

Smith, H. and J. McKeen. “Developing and Aligning a KM Strategy”, Journal of

Information Science and Technology, Volume 7, Number 1, June 2010, pp. 40-
60.

Smith, H., McKeen, J.D. and Singh, S., "Developing and Aligning a KM Strategy,"
Journal of Information Science & Technology, V. 7, No. 1 (June 2010): pp. 40-60.

IT Forum Developing a Data Strategy 13

Concept
The purpose is to bring senior IT managers together to examine topics that are of
critical concern to them and their organizations. Via the Forum, members share
experiences, learn from their peers, establish valuable networks, and develop
practical strategies for creating, implementing, and managing IT solutions.

Recent Papers
Managing IT-based Risk Enhancing Development Productivity
Application Portfolio Management Managing Maintenance
Building a Strong Relationship with Managing IT Demand
the Business IT in 2015
Effective IT-Business Business Intelligence
Communication Improving Customer Experience
Total Cost of Ownership Mobile Technology
Enabling Collaboration with IT Redefining IT
Identity Management Innovation with Technology
IT Shared Services Emerging Technology Management

Participating Organizations
Bell Canada Empire Financial Group
BMO Financial Group LCBO
CAA Ontario University Application Centre
Canadian Tire Parmalat
CIBC Sun Life
e-Health Ontario

Membership
Membership in the IT Forum is limited and by invitation only. The annual fee is $3,000.
Please direct inquiries to Dr. James McKeen at jmckeen@business.queensu.ca.

IT Forum Developing a Data Strategy 14