Profiting from Mistakes

Some big blunders bring great opportunities; others just waste time and energy. At Plano, Texas-based Frito-Lay Inc., one error led to the creation of an impressive knowledge management system; a second mistake undermined users' acceptance of it.

The story begins in the late 1990s, when one of Frito-Lay's biggest customers adopted a more centralized decision-making structure. Frito-Lay's regional sales teams, designed to deal with regional customer offices, found themselves struggling to work effectively with this huge account.

Moreover, the customer -- a sophisticated, multibillion-dollar supermarket chain -- began demanding more from the Frito-Lay sales teams. When a Frito-Lay salesperson suggested a new way to merchandise a product, the supermarket wanted the facts and figures to back it up. "They were pushing us to support [our plans] with quantitative and qualitative research," recalls Mike Marino, Frito-Lay's vice president for category and customer development.

Sensing a trend, Frito-Lay created a handful of national sales teams to focus on top customers, such as the supermarket chain. Theoretically, this would bring more of Frito-Lay's resources to bear on its customers' needs, particularly for information. But the company seemed to have forgotten the needs of its sales teams.

The teams, used to working regionally, found nationwide collaboration difficult. Although Frito-Lay had rich stores of market research and other pertinent customer information housed in databases at its headquarters, there was no easy way for a team member in, say, Peoria, Ill., to find what he needed. Frustration rose, performance suffered and sales team turnover reached 25%.

Then, in early 1999, Marino engaged Dallas-based Navigator Systems Inc. to help. Frito-Lay and Navigator envisioned a Web-based portal that would combine tools for knowledge management and collaboration, enabling the team to better serve the customer while helping reduce frustration and turnover. Navigator is a consulting firm that specializes in the development of custom business intelligence, enterprise collaboration and e-commerce applications.

The project team would pilot the portal with the supermarket team because it had the most centralized and demanding customer. "We knew if we could deliver there, we could satisfy any customer," Marino says.

The pilot sales team told the project team what kind of knowledge they needed. The request ranged from simple information, such as why Frito-Lay merchandises Lays and Ruffles products in one part of a store and Doritos in another, to more complex research on what motivates shoppers as they move through a store.

Then the project team had to find the knowledge. The team went prospecting in Frito-Lay's databases in departments such as marketing, sales and operations. They scoured the Web for external sources such as trade publications and industry organizations. They identified in-house subject matter experts and noted their areas of expertise in an online database.

The KM Bandito

In October 1999, the project team presented a working prototype to a core of beta users from the pilot sales team only to find that in the quest for speed, a classic and crippling error had been made: The project team had neglected to involve the sales teaam in the design of the tool. The prototype the project team had built could be marginally useful to any sales team, but it wasn't specific enough to be transformational for the pilot team.

"Conceptually, it was a great idea," says Frito-Lay sales team leader Joe Ackerman. "But when folks are not on the front line, their view of what is valuable is different from those running 100 miles an hour in the field."

The project team needed to backtrack and plug in the missing features, but it also had to win back the sales force, who suspected that even a revised tool would be a waste of time.

"We roll out lots of new things at the company," says Ackerman, but for teams in the field, "those things can create additional work."

This kind of mistrust isn't unfounded. Tom Davenport, director of the Accenture Institute for Strategic Change in Cambridge, Mass., says many knowledge management initiatives end up complicating work. "The tough part about knowledge management in general is that it gets added on to what people already do," he says.

To build a knowledge management system that streamlines work rather than creating more takes a lot of predesign brainstorming with users, says Davenport. "You have to understand and even redesign work processes, so it's baked in as part of their day-to-day work."

The project team then spent the next four months working with salespeople to evolve the prototype into a system they would embrace. Some of the changes were "baked in," like a call-reporting feature that everyone is required to use.

"So many people want to know what happened on a sales call, the account manager can be on the phone for days," Ackerman explains. "Now, we're able to post that to a Web site. It frees up the account manager to document the call once and move on."

Other changes included enabling users to analyze and manipulate data rather than just viewing it and developing reports tailored to customers' needs. "The [original] reports were very general," Ackerman says, so users would have had to spend lots of time reformatting them for customer presentations. "Now they can use them in more of a turnkey fashion."

With the changes in place, the project team wisely enlisted Ackerman for the official rollout. "If it comes from the field, it's really better-received than if it's from headquarters," he says. "So we made sure it was embraced by the team leader -- me."

The Upshot

Ackerman says better collaboration has helped to significantly reduce turnover, while improved access to knowledge-base data has enabled account managers to present themselves as consultants with important data to share.

The system has continued to evolve since its inauguration. Today, Ackerman says, it's used for daily communication, call reporting, weekly cross-country meetings, training, document sharing and access to data and industry news.

But the pilot team still isn't using the tool as regularly or as effectively as it could. "They're still getting good at it," Marino says. "There are two or three things that they take advantage of."

This isn't necessarily a bad thing, says Peter Novins, vice president of knowledge community and portal services at Cap Gemini America LLC in New York. "In knowledge management, early losers are the people who have the most to offer," he says. The richer the content, the more work it takes to become familiar with it, Novins adds.

Marino claims that the portal has been a big success. The pilot team, whose inaugural year as a customer-based team in 1999 was a bad one from a financial standpoint, exceeded its sales plan for last year and grew its business at a rate almost twice that of Frito-Lay's other customer teams. There was no turnover except for promotions. The concept is now being tailored to three other Frito-Lay sales teams and departments, and other divisions of Purchase, N.Y.-based PepsiCo Inc., Frito-Lay's parent company, have expressed interest in it.

Frito-Lay's sales management team reaped valuable lessons from the experience. For starters, it discovered that you can't redesign a sales team without realigning its support structures. In addition, early involvement of users is essential for buy-in.

"You have to get more input from the end users," says Michele Fraijo, a category manager on the pilot team who is helping to adapt the system for another team. "That gives people a sense that it's theirs."

Finally, don't create a plain-vanilla tool; give it some flavor. "If I would do anything differently, it would be to be more customer-relevant from the outset," Marino says. "I'd go deeper on more things that really would have been relevant to that first team. They have to have that feeling that 'hey, this is really helping me do my job better.' "

Kathleen Melymuka is Computerworld's senior editor, management.