How The Gap is reshaping I/S for a better business fit
- By Sandra Taylor
- July 24, 2001
Standing from left:
David Bergen, Phil Wilkerson
and Mick Connors (seated) of The Gap Inc.
This article chronicles one company's evolution and the use of information systems
(I/S) as an integral and crucial component of its business strategy.
Reshaping the business vision
To understand The Gap's Information Systems, San Francisco, is first to understand
the intense, results-oriented business environment created by one Millard S.
"Mickey" Drexler -- Gap's CEO since 1995. Drexler took the reins at a precarious
time in The Gap's history.
Sales at stores that have been open for over one year -- a common barometer of
success (or failure) in the retail industry -- had dropped to an all-time low.
From a respectable 12% average growth during 1986 to 1991, the rate had fallen
to an ominous 1% in 1993 and 1994. In 1995, sales went flat.
With the storm clouds overhead, Drexler ascended as CEO.
With an emphasis on aggressiveness and innovation, Drexler began reshaping, clearly defining and differentiating the company's lines of business -- its core Gap chain, as well as Banana Republic, Old Navy Clothing Co., GapKids and Baby Gap. Changes were intense. Banana Republic, for example, was totally transformed. Instead of selling product lines that actually competed with The Gap stores, Banana Republic discarded the old and brought in a totally new line of products designed to appeal to an up-scale market. The repositioning has been so successful that the company is now planning to augment store sales with a catalog business.
Why do we need client/server testing?
* Working systems are profit generators
* Business systems provide competitive advantage
* Only one chance to start it right
* Industry experts estimate $30 billion a year is spent on application
Source: The Gap Inc.
Far from being content with repositioning, Drexler is also an avid believer
in expansion. One aspect of expansion is the number of stores, on both the domestic
and international fronts. Currently, more than 1,600 stores exist in the U.S.
and over 200 do business abroad. With more than $400 million slated for additional
space in the immediate future, Drexler is not yet finished expanding The Gap
The I/S connection
Sitting squarely in the middle of this business repositioning, growth and
change is the corporate information systems group -- certainly not a place for
the faint-hearted or those fearful of responsibility, challenge and/or innovation.
Under the leadership of CIO Mick Connors, a seven-year veteran at The Gap, I/S
and its strategies are evolving at a pace mirroring that of the business.
Said Connors, "Our role, even the environment, has changed dramatically. We play a much larger role in the company's competitive business strategies. As such, change is constant." In many instances, those changes mean I/S will provide the key enabling technologies underlying a series of competitive business moves. For example, I/S is currently in the process of evaluating both Intranet and Internet-enabled applications to streamline and accelerate not only internal business processes, but the company's ties to the external world as well.
Reshaping the I/S function
Internet, Intranets -- this company obviously knows to exploit the newest technologies. What makes this aspect of the story interesting is the metamorphosis of the I/S function -- and indeed, of The Gap's I/S management team itself.
"Our role has changed
dramatically. We play a much larger role in the competitive business strategies."
In the early '90s, The Gap was a traditional glass-housed, mainframe-centric organization, run by the then newly-hired Mick Connors. In those days, when The Gap implemented its purchase order system, MVS, CICS and Cobol were the technologies of choice, and waterfall methodology was used to model the business processes. It would be hard to find a more classic example of the mainframe system development environment. Approximately one year later, Connors brought in David Bergen to manage the development side of the house, and a year after that, Tom Stromberg to manage technical operations.
In the intervening years, the management troika has learned the lessons only experience can teach -- the difficulty of maintaining aging mainframe applications, the performance problems of some of the early RAD-oriented application development tools and the growing need to view I/S as a strategic business weapon. But perhaps the most important lesson of all is the need to deliver value to the company. Said Connors, "Building credibility is essential, and the only way to build that credibility is to deliver early and often." In a discipline plagued with cost and time overruns, The Gap's management team's ability to deliver on its promises is unique.
The new order
With the appointment of Drexler as CEO and the roll-out of his aggressive business initiatives, the need to tighten the linkage between I/S and business strategies became increasingly apparent. I/S staff changes, as well as the technological changes, were soon forthcoming. (See "Technical Architect -- the newest title on the I/S organizational chart.")
Organizationally, Connors added a new function -- Technical Architect -- and filled the slot with one Phil Wilkerson who, for all intents and purposes, is a walking technology encyclopedia.
The Gap Inc.
The team at The Gap which implemented the use of I/S as an
integral and crucial component of business strategies.
Source: The Gap Inc.
Together, Connors, Bergen, Stromberg and Wilkerson are responsible for charting
the course of development and ensuring the course coincides with The Gap's business
strategies. Each brings his own unique strength and perspective to the process.
From Wilkerson came the state of myriad technologies and their applicability to
the needs of The Gap. Bergen contributed the ability of the development group
to internalize the technologies, then develop and deliver applications based on
those technologies. Stromberg provided the ability to physically and feasibly
deploy, administer and manage the resulting systems. And Connors supplied the
bridge between I/S and the business strategies -- and the ultimate decision-making
responsibility. (See Fig. 1
Blueprint for the future
Connors and company also recognized their existing computing infrastructure could not weather the rising demands from the business side of the house. Thus began the ordeal of developing a new computing architecture. Organizations that have gone through this process know it to be laborious and painful. Given the time pressures of the full-court press deployed by CEO Drexler and the self-imposed "deliver early and often" philosophy espoused by Connors and Bergen, the process was exceptionally complex. Complex and difficult -- yes. But a process that resulted in an architecture as modern and aggressive as The Gap's business strategy.
ISD technical architecture DB2 access model
The new computing strategy utilizes an open systems architecture
and an object/component-based development model.
Source: The Gap Inc.
The new computing strategy utilizes an open systems architecture. The major components,
shown in the Client/Server Architecture diagram, blend many of the computing industry's
premier and most well-known technologies -- Forté, Tuxedo, Informix, Solaris
and NT, as well as MVS, CICS, etc. (See Fig. 2
Perhaps the most significant fundamental change for The Gap is the introduction of object/component-based development; as witnessed by the company's selection of Forté as the primary development tool suite, complemented by C++ code when and as necessary. This transition is not without a cost. Said Wilkerson, "The move to object-orientation will require a significant investment in our internal resources and in the external resources we employ to assist us with business-related issues. However, we cannot lose sight of what this technology offers from a competitive viewpoint. Object technology will enable the development community at The Gap to respond to changing business requirements in a more timely fashion."
"We cannot lose
sight of what this technology offers from a competitive viewpoint. "
The major technical and cultural impact of this choice falls squarely on the shoulders of VP of Development David Bergen and his development group. Said Bergen, "Object technology extends our development portfolio. With the aggressive current and near-term business initiatives we have on our plate, we need all the ammunition we can muster if we are to deliver the features and functionality needed by our business partners." The Gap's technological arsenal is indeed growing. In addition to the information-rich legacy systems, The Gap now has investments in OO and client/server technology, as well as various packaged systems and a growing data warehouse initiative. The intergration of these technologies alters much more than the technical development process -- it ripples through the very structure of the development organization and the job skills required to populate that organizational structure. Like the introduction of object/component development, the organization itself is an evolving entity
Throughout this evolution, Connors remains committed to a tight link with the I/S business partners -- from "business architects" involved in the overall planning phase to "business experts" actively participating in the development process. Such user involvement has proven time and time again to be a critical success factor.
The mainframe lives on
The preceding architecture diagram depicts another increasing common occurrence, not only at The Gap, but across the industry in general. Far from being rolled into mothballs, the mainframe is playing an increasingly active role in today's client/server systems. In fact, in SPG's annual market research survey, nearly 50% of the respondents cited "Build Systems That Interoperate With Legacy Data/Processes" as a critical strategy. An additional measure of the importance of the mainframe is another metric from the same survey. For the 49.5% of respondents who cited "Build Systems That Interoperate With Legacy Data/Processes" as a critical strategy, their importance ranking of 8.2 (out of a possible 10) indicates that this strategy is extremely important. For these respondents and The Gap, client/server systems that embrace the mainframe (and the wealth of data and processes that reside on that mainframe) is the order of the day.
"We need all the
ammunition we can muster if we are to deliver the features and functionality
The commitment to quality
As if The Gap did not have enough on its plate, another key aspect to the architecture exists -- automated software testing. As veterans in the software development discipline, both CIO Connors and VP of Development Bergen have a working appreciation of the need for quality and perhaps, a greater appreciation of the process needed to fix errors in production-level software. While both of these managers "cut their QA teeth" with a manual testing process and an army of testers all trying to hit the ENTER key at the same time, Connors and Bergen are now firmly in the automated testing camp. Said Connors, "In today's environment, we simply cannot afford to deliver the software twice. We have to do it right the first time." Echoing Connors sentiments, Bergen noted, "I've learned the hard way, and those lessons never leave you. We have to institute a testing methodology that will deal with these complex systems we're building."
Bergen's philosophy for ensuring a quality development effort is relatively simple. His "you build it, you own it" approach extends to all his development teams. While the team may in fact have a development arm and a maintenance arm, the team as a whole is responsible for both the initial and the ongoing maintenance of the software it develops. Bergen's challenge now is to integrate support for The Gap's chosen test software from Mercury Interactive into his organization. While the tool suite may be new, it is doubtful if Bergen's build-it/own-it philosophy will change.
Technical architecture for TP
The Gap's development teams had prototyped and tested Java-enabled
access to existing business rules and were accessing data on DB2
and Informix databases.
Source: The Gap Inc.
I/S, like its business partners, is on a fast track. In the space of less than
14 months, various architectural and technological approaches have been evaluated
and discarded or accepted. Due diligence on major product sets, such as the application
development tool suite, has been conducted and a consensus reached. Organizational
issues have been and continue to be addressed. Business partners (both internal
and external) have been apprised of the new I/S directions. And a prototype planning
and distribution system that both utilizes the new system architecture components
and validates the architecture is well under way.
Hot on the heels of the initial planning and distribution system came the first uncompromising test of the new architecture -- Java-enable one of the applications. In the space of less than a week, The Gap's development teams had prototyped and tested Java-enabled access to existing business rules and were accessing data on the DB2 as well as the Informix databases. (See Fig. 3.) While certainly not a production-quality system, the architecture shown in the Web-enabled diagram is proving it can live up to Connors and Bergen's deliver-early-and-often rule of thumb.
The development of a new computing architecture involving new technologies and new applications, especially under the given time pressures, is nothing short of awesome. Given the stakes, one had to wonder if the pressure ever reached the flash point. The final question posed to Connors, Bergen, Stromberg and Wilkerson dealt with organizational and personal dynamics: "What is the secret to getting the three potentially competing functions to work in concert?" While openly admitting that the process was not always a smooth one, Connors replied, "First of all, the four of us share the same fundamental vision. We're all pulling on the same rope, in the same direction. The other key ingredients are the trust and respect we all have for each other. That's not to say we always agree with each other. It does say we can work together to arrive at a conclusion."