Knowledge Management: Four Obstacles to Overcome

Knowledge management (KM) continues to be trendy and appealing — a recent conference in San Antonio, Tex., drew some 250 people from 109 companies — but so far it's a concept with a checkered career. A handful of organizations, such as the World Bank and Xerox Corp., have developed robust programs for sharing knowledge that are by now well integrated into everyday operations. But many others have launched initiatives that start with a bang and end with a whimper. According to Houston-based American Productivity & Quality Center (APQC), which has conducted several studies of KM, most programs are in danger of getting stuck in the pilot phase.

Why so? Experts at APQC and elsewhere point to several potholes on the road to successful knowledge management. Among them:

Too many companies treat KM as an end in itself, argues Nancy M. Dixon, a professor at George Washington University and author of the new book Common Knowledge: How Companies Thrive by Sharing What They Know. They create a program because they think it will pay off later — or just because they think an effective organization should be able to share knowledge across departments.

	The number-one reason KM initiatives may not function … is that the 'evangelists' fail to connect with the real business issues.
	— Carla O'Dell, president of APQC

That's exactly backward. "The goal is not to make knowledge management happen," says Dixon, "the goal is to deal with the organization's most pressing issues and to use KM where it's appropriate to do so." APQC president Carla O'Dell underscores the point. "The number-one reason KM initiatives may not function in an organization is that the 'evangelists' fail to connect with the real business issues," she says. "That's the pitfall of just trying to create vehicles for people to communicate and share — it's a solution looking for a problem."

Successful programs start with business problems that KM can help solve. Ernst & Young developed its "PowerPack" document collections — information tailored to specific industries — with the express goal of helping consultants develop proposals faster. Ford Motor Co.'s so-called Best Practice Replication system helps plant managers achieve the companywide objective of 5% annual productivity improvements. At Anadarko Petroleum, a Houston-based oil exploration and production company, "everybody wants to bring in a new well," says planning analyst Tiffany Tyler. "That's such a buzz for people. And if we can show them we can cut cycle time by a week and make a better decision about where to drill the hole" — goals of the company's KM program — "it's, 'Well, don't you want to do that?' The answer is almost universally yes."

Many companies focus their attention on the KM pilot project and forget about the rollout, says Richard McDermott, president of McDermott Consulting in Boulder, Colo., and a collaborator on some of the APQC studies. "Pilots, because they're new, get lots of attention and support. Senior managers see the pilot and think it's great — they'll lead the charge." Over time, though, key managers get transferred, the marketplace shifts, and attention is focused elsewhere. Pretty soon, resources have dried up.

To avoid this, McDermott advises planning the rollout at the same time you're launching the pilot. That allows managers to acknowledge (and plan for) the fact that "the real work comes after the pilot is done." It's important, too, to make sure that the company is prepared to commit real resources to the KM effort. "If you're going to start three pilots and you invest half a million to a million dollars in that, you ought to be thinking of a considerably larger investment in the overall rollout," says McDermott.

Best-practice companies don't hesitate to spend such sums. Many of the benchmark companies studied by APQC laid out more than $1 million to get their programs started and spend more than that amount annually on ongoing development and maintenance. Ernst & Young, Dixon points out, spends a whopping 6% of revenue on knowledge management, though that figure includes technology investments that benefit KM but weren't designed exclusively for that purpose.

Knowledge-management initiatives are likely to peter out unless responsibility for them is somebody's job. Big organizations may need several such somebodies. "There's a general rule of thumb: if you're a large corporation and you're trying to get started in this, you probably need four core people dedicated to it in addition to the IT people," says Carla O'Dell, who in addition to her APQC role is also coauthor of If Only We Knew What We Know: The Transfer of Internal Knowledge and Best Practice. Smaller companies can get by with one dedicated person, such as Anadarko's Tyler. But if no one's accountable for KM, it won't last.

The larger the initiative, the more people and job descriptions are required. Like many big professional-services firms, for example, Arthur Andersen has an extensive KM system, dubbed Knowledge Base. One thing that makes it work, says Andersen consultant W. Todd Huskinson, is the fact that literally hundreds of the firm's employees work on it — helping project managers elicit the lessons from their latest engagement, editing the content and keeping it fresh, even staffing a hotline to help consultants navigate the KM system. What's more, the firm's end-of-engagement checklist includes a question as to whether the deliverables and lessons were added to Knowledge Base. "So the partners and managers need to sign off on that: 'Yes, we've made a contribution based on this engagement.'"

Knowledge management is not a one-size-fits-all program. On the contrary: It works best when individual programs are tailored to the needs of individual users. Ernst & Young's PowerPacks are designed for consultants in particular industries, points out Nancy Dixon. Chevron's Project Resources Group, a team of internal consultants with special areas of expertise, is a resource for managers of capital-intensive products. Knowledge-transfer systems, Dixon says, "are less useful and less effective when they are designed for just 'anyone' in the organization."

There's another way in which KM needs to be tailored: it needs to fit an organization's culture. A major requirement, according to one APQC study, is that it must be "tightly linked to a pre-existing core value of the organization." Two examples:

— At IBM's Lotus Development Corp. unit, there's a widespread feeling that Lotus "invented" collaborative technology — so "contributing ideas to a Lotus database and checking on the insights of others is simply how people do business at Lotus." Moreover, the Lotus culture is forgiving: projects don't have to be perfect the first time; "employees feel free to 'try things out' and modify what does not work." That's reflected in the way knowledge is shared.

— At Ford, the emphasis is on avoiding risk and getting it right. "Because Ford's development teams use an intranet Web site to communicate and to share knowledge within and among teams, their original documents, analyses, and discussions are available to senior managers. The opportunity to look at original work, rather than 'massaged' reports, builds on Ford's values of careful, complete, detailed analysis."

KM is still a young field; it's not surprising that so many organizations have only begun to implement it, nor that so many have bogged down. Still, they can learn from the pioneers. "Those that have been at it for a long time had to learn what they know by trial and error," says O'Dell. "But those that are coming into it are able to leapfrog over a lot of those early problems and make much faster progress."

One sticky point about knowledge management (KM): it's tough to assess how you're doing. Tough, but not impossible. "Two things need to be tracked," says George Washington University professor Nancy M. Dixon: "an outcome measure and a process measure." The outcome measure depends on the objective you're trying to reach through better management of knowledge — increased productivity, faster turnaround time, whatever. "So long as KM is aimed toward a particular business goal," says Dixon, "it's possible to track it."

The process measure depends on the nature of your knowledge-management system. You want to know whether people are doing what you'd like them to do — tapping into the database, taking advantage of peer-assist consulting capabilities, putting designs into the design bank. It's hard to know when KM is "implemented," says Dixon. But you can know if people are doing the activities that you think of as knowledge management, and you can assess whether they're reaching the goals you're seeking.

· · · ·

Reprinted with permission from the Harvard Management Update, August 2000.

See the latest issue of Harvard Management Update.

All images © Eyewire unless otherwise indicated.