Case Study 1: Business Intelligence Strategy at Canadian Tire |
||||
Nicole Haggerty and Darren Meister |
||||
“The Retail group has just sent me another ‘quick win’ request,” stated Michael Eubanks, director of marketing information technology (IT), as he walked into his meeting with Andy Wnek, chief information officer (CIO) of Canadian Tire Corporation (CTC). “That’s the second one this week, and I have heard whispers about more. Dealing with these quick wins is going to make it difficult to redevelop the business intelligence (BI) infrastructure. That’s where the real return on investment (ROI) is.” Over the last year, the IT group at CTC had promoted a strategic initiative to deliver real business value from business intelligence (BI) over the next three years. A massive change effort involving infrastructure, organizational structure and business processes across most of the business would be required. Nevertheless, “the plane was still in flight” and current needs could not be completely ignored. As the door swung closed on their meeting, Wnek and Eubanks sat down to discuss how they might keep the plane in the air while rebuilding the engines. |
||||
|
||||
Canadian Tire Corporation |
||||
In 1922, John and
Alfred Billes, two brothers, opened a garage and auto parts store in Toronto,
Canada. By 2003, their enterprise had grown into CTC, a network of businesses
including retail, financial services and petroleum operations. More than
45,000 individuals worked at CTC operations across Canada in more than 1,000
stores and gas bars. CTC businesses were divided into five main groups: |
||||
|
||||
Corporately, CTC had
completed a strategic plan in late 2002 that stated a clear corporate goal
“to become a top quartile performer in our market sector as measured by total
return to shareholders.” This strategic goal was to be accomplished through
four strategic imperatives: |
||||
|
||||
This new CTC
strategic plan guided the development of a CTC IT strategy in February 2003,
an effort led by senior vice president and CIO Wnek, to complete the first IT
strategy document in several years. As CIO, Wnek was responsible for
overseeing the information systems (IS) of the entire enterprise, but not
within the associate dealers’ stores. The BI initiative, while important to
the organization overall, was primarily associated with CTR, and it
represented one of many IS initiatives competing for CTC support and funding
based on the new 2003 IT Strategy 2003. Figure 1 provides an organizational
chart of the senior management at CTC and CTR. |
||||
|
||||
Figure 1: Organizational chart of Canadian Tire Corporation and Canadian Tire Retail |
||||
1 Source: http://www2.canadiantire.ca/CTenglish/h_ourstory.html,
accessed August 22, 2003. |
||||
|
||||
Information Systems at CTC and The New IT Strategy for 2003 to
2005 |
||||
The Web of businesses comprising CTC was accompanied
by a highly complex technical architecture (see Figure 2 for an overview of
their organization). A recent enterprise technology review revealed a
multitude of hardware, software, operating systems, network services,
development tools and applications being utilized across the business.
Source: Company Files. Figure 2: Canadian Tire Corporation enterprise technology overview
For example, CTR ran IBM AS/400 systems at the store level with point-of-sale
(POS) systems and servers that networked to IBM mainframe systems at
the CTR data centre. These systems were funded 50 per cent by CTR and
were supported by the retail systems group at CTC. The systems at Mark’s
Work Wearhouse still remained entirely separate from the rest of the
CTC infrastructure. CTFS in Welland operated on IBM RS6000 with Intel-based
workstations. PartSource and CTP relayed transactions directly into
the corporate network from their POS systems. In fact, the CTC IT group
supported, operated and managed over 100 different mainframe, server
and desktop development and integration tools, 10 different hardware
platforms, 14 operating systems, seven database management systems and
over 450 different production applications and desktop-based applications
and tools. Substantial proportions of the hardware, operating systems,
network services, data services and development and integration tools
were identified as “niche” (exceptions to current standards) and “sunset”
(to be retired) technologies that needed to be addressed in order to
meet the objective of bringing IT spending as a percentage of sales
under industry benchmarks of about two per cent. Currently, CTC’s IT
spending, when measured by accounting for all IT expenditures (including
assets and resources expended on IT in other areas), was around this
benchmark.
The CTC IT strategy document concluded that IS at CTC had evolved into
a highly complex and costly environment that offered substantial opportunities
for consolidation, simplification, integration and cost-cutting. The
results of several IT reviews carried about by consulting firms over
the last seven years supported this conclusion. The results highlighted
several themes faced by the IT group: |
||||
|
||||
In spite of these challenges, IT had
delivered several key initiatives in the past few years, including the
development of a demand forecasting and replenishment system, Y2K upgrades,
the development of www.canadiantire.ca and CustomerLink—a supply chain
management system.
To build on their strengths and the recently developed business strategy,
IT Strategy 2003 laid out a strategic vision to be “an agile IT team,
aligned to business priorities, operating a simpler technical environment
with the appropriate standardized processes.” Consistent with this vision,
three strategic IT imperatives were identified: |
||||
|
||||
These three imperatives were laid out to
guide action and prioritizing within the IT group and the CTC network of
business as a whole.
Four programs were developed to enact the IT vision for the period of
2003 to 2005. The first program was to implement a CIO governance council
that would meet quarterly and assume responsibility for developing enterprise
standards, monitoring IT spending, undertaking annual IT planning, monitoring
the IT strategy and providing opportunities for sharing and collaborating
across the enterprise to realize synergies.
The second program, organizational and people capabilities, specified
key capabilities and services the IT group would need to be able to
offer to the organization. Business-consulting, solutions integration,
end-user services and support, platform operations and management, and
enterprise IT planning and architecture were just some of the capabilities
outlined in the IT strategy document. |
||||
The third program, process improvements, included coordinating an annual
IT strategy planning process based on the corporate strategic planning
cycle, the development of an opportunity management process that would
standardize and regulate the identification, analysis and approval of
IT project requests and a monthly IT scorecard reporting on key performance
indicators including IT spending as a percentage of sales.
The final program was entitled technological direction, and it laid
the foundation for re-architecting the organization. To support the
IT imperatives outlined above, this program identified five areas for
attention: business intelligence and data management, application deployment,
integration and messaging, standardization and simplification, and security
deployment. Business intelligence and data management activities were
identified as high priorities. Consequently, BI quick win projects were
prioritized first with development and implementation of a BI strategy
and structure to follow over the life of the IT strategic plan. |
||||
Business Intelligence and the Retail Environment |
||||
Historically, retail organizations have
invested significantly less in information technology than other industries
(two per cent of revenue compared with about eight per cent of revenue across
other industries). This spending has focused mainly on POS and supply chain
management systems. One of the important exceptions to this rule is Wal-Mart.
However, due to increasing competition, retailers have turned to BI to
improve sales and better serve customers. A recent Forrester Research report
indicated that IT executives at 286 North American companies with over $1
billion in sales intended to make business intelligence analytics their
second largest IT investment after their Web portal investments.2
BI is the consolidation and analysis of internal data (e.g., transactional
POS data) and/or external data (e.g., purchased consumer demographics)
for the purpose of effective decision making. Assembling and merging
data from various sources is a complex task, and analysis requires the
use of highly sophisticated skills. At the core of all BI initiatives
is a data warehouse to hold the data and analytics software. The data
warehouse stores data from operational systems in the organization (e.g.,
inventory, POS, accounting, marketing, etc.) and restructures it to
enable queries and models to extract decision support reports. With
no clear dominant players in the business intelligence market place,
many niche players have emerged to serve data warehouse and BI analytics
markets—an industry that is expected to grow from US$30 billion to US$75
billion by 2005 in North America alone.3 Figure 3 provides partial information
on players in this marketplace. |
||||
|
||||
For example, in order to improve profit and customer service, Hudson’s
Bay Company (HBC), a major Canadian department store chain, is currently
in the process of a multi-year effort to upgrade its IS. Two major data
warehouses (one for department store operations and one for discount
store operations) have been merged into a single data warehouse enabling
executives, store managers and business analysts access and interpret
data about store sales, category sales, financial performance and suppliers.
Ace Hardware (American hardware retailer similar to CTR in that stores
dealer-owned) launched a BI initiative with its 5,000 stores whereby
store owners and Ace executives could view and analyse information to
aid in category management and promotion decisions. With over 65,000
products from 3,000 vendors, these activities would be impossible for
a single store to undertake. A price-setting model has been particularly
effective in allowing Ace Hardware store owners to see the implications
of setting prices above or below those recommended by head office. While
the current application focuses mainly on analysing historical trends,
Ace is also developing an application to use and view real-time POS
data in real time to see the status of pricing, product and promotion
decisions. Numerous other examples of the effective application of BI
in the retail industry have also been observed.4
Despite the benefits from BI and data warehousing investments, implementation
of these projects consumed huge organizational resources and was difficult.
An IDC report revealed many challenges associated with the iterative
nature of over 400 BI implementations reported by over 1,300 respondents.
Importantly, the study findings indicated that 35 per cent of all BI
implementations were unsuccessful, 35 per cent were adequate and 30
per cent were described as successful: the larger the organization and
the more complex the BI implementation, then the lower the likelihood
of launching BI successfully, including being on time and on budget.
These organizations also prioritized the 10 biggest challenges to achieving
success in implementing business intelligence, in descending order:
budget constraints, data quality, understanding and managing user expectations,
culture change, time required to implement, data integration, education
and training, ROI justification, business rules analysis, and management
sponsorship. However, two groups of respondents reported different priorities
for these challenges. IT managers more frequently mentioned data quality
and cultural change as their biggest hurdles. Business managers placed
higher priority on education and training for end-users. The study concluded
that BI initiatives were iterative projects whereby user expectations
and training needs expanded as they gained access to and experience
with analysing data. Additionally, variation in the viewpoints of IT
and business managers could be pressure points for implementation. Given
that most organizations expect their data warehouses and BI investments
to continue to grow, these variations in opinions, especially with respect
to data quality and end-user training and education, pointed to important
areas of BI project management. |
||||
2Business intelligence becomes a hot commodity,
especially for retailers. |
||||
3Ibid. |
||||
|
||||
Business Intelligence and Data Warehousing at Canadian Tire |
||||
BI analytics started at CTC in 1994 with the
development of the information warehouse (IW), which was implemented by the
CTC IT group at the request and funding of CTR. Around that time, the CTC
chief executive officer (CEO) began trying to change CTR’s image and role
from that of a wholesaler to that of a retailer. This led to the realization
that more data was required in order to begin analysing data like a
retailer—going beyond the store level to examine product, store and margin
trends. To facilitate this new logic, the IT group built the IW into which
they extracted, transformed and loaded data from a variety of sources,
including POS data downloaded from the stores. |
||||
At that time, Wnek was chief financial officer (CFO) for CTR. His efforts
to provide better information for business decision-making led to the
creation of retail analytics in the CTR finance group. This ultimately
led to a separate department—the Finance Retail Analytics Group (FRAG)—that
performed the bulk of the analysis and prepared reports for the various
marketing departments at CTR. Between 1994 and 1998, the IW grew dramatically
as more educated end-users in CTR Marketing and analysis within FRAG
demanded more data and CPU time to conduct analysis to support business
decisions.
During this time period, BI efforts fragmented—a situation that persisted
and then accelerated from 1998 onwards. The IT group gradually took
on a more technical focus for the IW, focusing on loading data and transforming
it into more and more summary tables to balance the need for CPU time
for user queries against capacity constraints. The goal of summary tables
was to cut down on the duplication of queries. During this period, due
to a lack of resources which were being used for other projects), the
IW was evolving on old infrastructure and a poorly defined data model.
Further, a lack of standard data definitions meant that several versions
of the truth could be extracted from the IW; depending on the way you
defined it, you could end up with six different numbers for inventory
levels. Also, some data was just simply not available—a marketing analyst
in the sports segment for instance, could not evaluate the results of
a weekly promotional effort on golf clubs nor evaluate the performance
of various brands against each other e.g., how Titleist performed against
Nike products). In its current state, the data model in the IW did not
reflect the data requirements of the business.
Throughout CTC, user groups gradually undertook more responsibility
for IW data management activities so they could perform their own analytic
tasks. User groups developed applications and hired business analysts
who extracted data from the IW, then cleaned it, integrated additional
data and transformed it into their own reports. While this division
of labor and IT resources enabled better business decisions because
it facilitated better analysis of the available data, it also distributed
BI and IS responsibilities and resources across the organization. For
example, as many as 100 people were being employed in end-user communities
in CTR finance and supply chain in positions that were largely IT responsibilities
(technology acquisition and management, application development, database
management, technical support, etc.) but who worked outside of governance
of the IT function. These shadow IT groups provided an alternative source
of IT resources to the user groups but at unknown cost and security
risk to the CTC IT infrastructure. By 2003, CTC IT was largely seen
as a hardware provider and manager but not as a strategic business partner.
CTR marketing maintained its own analysts, developers and end-user support
for BI efforts, and CTR FRAG provided most of the retail analytics needed
by marketing. |
||||
Figure 4 provides a depiction of the current architecture of the IW.
The major challenges were the multiple independent data sources not
included in the IW, lack of standard data definitions and consequent
inaccuracies in the data, the strained resources associated with storage
and querying the IW, and the increasing delays and denial of access
to information required by the end-users. |
||||
Source: Company files Figure 4: Current business intelligence environment |
||||
When Wnek assumed the role of CTC CIO, he understood the barriers to
BI success. His work in the mid-90s had lead to the initial development
of the IW and FRAG. This gave him a good perspective as to what was
needed to realize the value of BI. To get the ball rolling, four major
activities were undertaken: |
||||
|
||||
4Business intelligence buy-in. Source: http://www. informationweek.com/story/showArticle.jhtml?articleID=970006, accessed August 24, 2003. 5Business Analytics Implementation Challenges: Top 10 Considerations for 2003 and Beyond, January 2003, IDC Report #28728. |
||||
Business Intelligence Environmental Assessment and Quick Wins |
||||
In their assessment, the consultants from
Cap Gemini found BI to be a crucial element to the long-term success of
CTR—hence its placement as a major IT program in the IT strategy.
The vision established for BI at CTC was to provide “the right information
for the right decisions at the right time, enabling proactive, accurate
business decisions.” |
||||
|
||||
Four guiding principles were established to support these goals: to
be business driven, to support the IT strategy (including making technical
changes in line with the technology plan), keep the learning in-house
(even if external expertise was used), and make changes sustainable.
Effective execution of the BI strategy over the next three years would
result in a new BI environment (see Figure 5).
Figure 5: Future business intelligence environment
In this new world, data would be sourced and consistently managed and
integrated from across the company, historical data would be organized
according to standard data formats and housed in the central data warehouse,
and there would be simple and easy to update access to metadata6 (data
about the data). This would result in various data views or data marts.7
Decisions about implementing physical data marts versus special views
of the data for different areas had not yet been made. As an example,
the financial data mart (or view) would provide consistent access to
standard financial data that would be the basis for enterprise performance
management. Similarly other areas would have their own data marts/views
to assist in their own decisions while masking the complexities associated
with access to the full corporate data structure. BI specialists would
assist in organizing the data marts/views, retail analytic specialists
would have access to BI tools and the data warehouse to perform sophisticated
analysis and predictive modeling, and end-users would have instant access
to information they needed to make relevant business decisions.
Given the current state of BI at Canadian Tire, several steps were planned
for the short term–namely dedicating resources to implementing quick
win projects and finalizing a detailed BI strategy-and-planning document
to serve as a guide for prioritizing actions over the next three years.
Quick win opportunity assessments happened in early 2003. These projects
consisted of short-term actions that IT could take to improve BI capabilities
and to provide users with new information. These included opportunities
such as providing access to daily promotional sales data; market basket
analysis capabilities; forecasting and model simulation of incremental
sales; pricing optimization reports by region; price competitiveness
analytics and brand analysis such as comparisons by brand, brand manager,
margin, shipments, cannibalization, etc. Quick win projects were selected
based on offering the highest potential value at the lowest cost to
IT resources.
The development of the BI strategy document had also commenced with
a series of user meetings, surveys, benchmarking studies and workshops. 6Meta-data is data about data. Source: http://www.techweb.com/ encyclopedia/defineterm?term=meta-data, accessed August 22, 2003. 7A data mart is a subset of a data warehouse for a single department or function. Source: http://www.techweb.com/ encyclopedia/defineterm?term=data+mart, accessed August 22, 2003. |
||||
|
||||
Going Forward: Creating a BI Mindset and Realizing Value |
||||
As Eubanks and Wnek’s meeting progressed,
they reflected on the difficulties faced in the next several months. First,
they had to determine what to do about new quick win requests while they
finalized the BI strategy and implemented the program. On the one hand, what
was the point of the BI strategy and program plan if they kept reacting to
new quick win requests, rather than using the plan to prioritize them? However,
on the other hand, they had to think about the business implications of not
delivering new quick win requests that could provide real value to the
business today. They wondered whether rejecting new requests would diminish
the end-users’ enthusiasm for the BI initiative and whether the work would be
picked up by the shadow IT groups.
Second, several elements of the BI strategy still needed to be finalized,
and programs, with their ensuing priority and timeline, needed to be
developed. Eubanks and Wnek wondered what technological, organizational
and people implementation plans would need to be included on the timeline
in order to successfully deliver on the BI strategy over the next 2.5
years, given that the organization was expecting the program to be completed
by the end of 2005. |
||||
Case Study Questions |
||||
|