Joe McKendrick on “Closing the Enterprise Data Gap – in Gap Fashion”

Many companies may have enterprise data warehouses and the latest
analytical technologies, but the capabilities are not employed to their
full potential. Often, a change in organizational culture and governance
is required to make the most of these solutions.

Such was the challenge for Gap, Inc. At the recent National Retail
Federation show in New York, Mike Jones, Senior VP for Gap, Inc., discussed how the
retailer re-aligned its corporate data framework to meet the challenges
posed by the recent rough-and-tumble economy. Economic recovery won’t
offer much respite, either – for retailers, things will only get more
competitive.

The company didn’t have a single enterprise view of its data. The
challenge was to bring together data from its five brands – The Gap, Old
Navy, Banana Republic, Athleta, and Piperline – often managed with
separate tools, Jones explained. In addition, the retailer has partners
in 25 countries that need actionable data. “Our challenge was to provide
timely information across the entire enterprise, while at the same time
maintain that brand distinction,” he said. “We needed to be able to
update the data and information where we could measure the results of
business initiatives.”

Key in this model has been the transition from independent data marts to
an integrated logical data model. Corporate data has been systematically
incorporated into a Teradata enterprise data warehouse which facilitates
business intelligence across the enterprise. “We used projects for
specific lines of business to drive what went into our data warehouse,”
Jones explained. “We didn’t have an enterprise view. We had separate
data marts, and had to go back many times to our operational systems for
data, which was a burden.”

The shift for Gap has been to move away from the use of analytical cubes
throughout the enterprise for decision-making, and instead, rely on one
source of the data. Initially, Jones said, there was resistance from
heads of business units who had grown comfortable with doing their own
analysis. “A centralized data model is a key tenet we needed to get our
executives comfortable with.”

Plus, building and generating reports had become a burden to the
company, Jones said. “We didn’t want to be in the business of building
reports, we’re not a report-building organization. We were a little bit
coin-operated. When people needed reports they would come to us.” Now,
with an enterprise data environment, he said, “users across the
different brands are now able to build their own reports and queries.”

In addition, he reported, economies of scale have kicked in as the data
warehouse has grown. “As our enterprise warehouse grew, we needed
incrementally less data and required less effort to enable new
applications and business value,” he said. “The value to effort ratio
increased significantly, and the amount of effort required went down
substantially.”

Key selling points for this centralized approach included the ability to
ask more questions of the data once it was integrated from various
functional areas. For example, Jones illustrated, about 15 questions
could be answered from the sales-related data systems, and 25 from the
inventory systems. When combined as enterprise data within the
warehouse, there were 96 questions that could be answered from the data.

“We’ve shifted to where we’re providing an enterprise view,” Jones said.
“Last year, we spent a lot of time thinking about promotions. The
biggest challenge for us was engaging more with the business and talking
about data, not technology.”

Gap even now has a “data czar” who reports back to the steering
committee, which in turn answers to the CEO. “They meet every quarter
with our CEO and CFO, who make the final decisions.”