Knowledge Management Case Files

FRAMINGHAM (02/02/2000) - United Technologies Corp., Located in East Hartford, Conn. Employees: 142,000 nationwide; 1998 Revenues: $25.7 billion.

MAJOR DIVISIONS Pratt and Whitney, Otis Elevators, Hamilton, Sundstrand, Carrier, Sikorsky KM Problem How does a company as decentralized as UTC enable knowledge sharing across its disparate businesses?

What can an engineer who designs space suits for NASA possibly learn from one who builds elevators?

Plenty, according to Jean Mayhew. The difficulty lies only in starting the conversation between them. As director of knowledge management at the United Technologies Research Center, the R&D arm of United Technologies Corp., based in East Hartford, Conn., Mayhew is charged with getting a huge-and hugely decentralized-organization to share knowledge centrally. She admits that this is a gargantuan task for a company with about 142,000 employees and five major divisions worldwide, but the possibilities inherent in such knowledge sharing are too good to ignore.

Take the space suits and the elevators, for example. UTC owns Otis, which makes elevators, as well as Hamilton Sundstrand Corp., which is responsible for NASA's haberdashery. "People look at those product lines and say, 'What could they have in common?'" she says. But when broken down into broad disciplines, a whole field of shared concerns emerges. With space suits, Mayhew notes, "You've got materials, you've got controls, you've got understanding the human response to the environment. These issues transfer to both Otis and Carrier [UTC's air conditioning business]."

Mayhew and the KM group at the United Technologies Research Center (UTRC) started the knowledge sharing initiative in November 1998 to support UTC's objective of increasing the company's effective use of information. "We define knowledge management as the process of converting information into value for the corporation," Mayhew says. "Value can be defined as reduced costs, increases in employee productivity and product and service improvements." As UTC grew increasingly global and added new companies through mergers and acquisitions, UTRC saw knowledge management as one avenue for boosting that value through information sharing and brainstorming.

The KM project is still relatively young; the company has spent about a year defining goals and purposes, and is now moving into implementation. Mayhew made an early decision to start small and build the effort incrementally. "One of the first things we did was scope out what's already going on [at UTC] that fits into our KM goals, because there's a lot out there," she says. Preliminary research led the group to concentrate first on UTC's engineering population.

The engineers already had a cross-functional group called the United Technology Engineering Coordination Activities (UTECA), a grassroots organization started in the '70s, dedicated to transfering knowledge across business units.

Expanding UTECA's reach seemed a good place to start.

UTECA was already successful, with 18 different functional disciplines, such as artificial intelligence and engineering operations, trading information informally via channels such as e-mail or conversations. It also holds an annual conference where about 1,500 engineers from around the world meet, greet and talk geek. But 1,500 is a drop in the bucket compared with UTC's total employee base. Mayhew wants knowledge exchange participation to register in the "multiples of thousands, not one thousand," so she aims to formalize UTECA's technology discussion programs as step one of the KM project. (At UTC, the word technology denotes the engineering that goes into a product, not information systems.) She says, "It's not enough for people to just get together and talk about what technology is being developed," as UTECA has traditionally done. "There's got to be a sense of impact in terms of how we are actually putting best practices into play as we go from business unit to business unit. Now we need more structure and accountability."

To build that structure, Mayhew and her group are working with John Cassidy, UTC's senior vice president of technology (in essence, the top-dog engineer) to expand and formalize UTECA's knowledge-sharing programs. The group read studies about how to build communities across an organization and concluded it was vital to get senior- and middle-level managers involved in the effort. "This spreads learning out through the organization more," Mayhew says, and also guarantees a broad support for the project.

Besides, middle managers are the ones who know the experts. "They work where the most experts are and tend to have a good grasp on where learning needs to take place based on where business plans are leading us," she says.

Senior VPs at all UTC divisions tapped their vice presidents of technology, including overseas personnel, to contribute their own knowledge to the project.

First, these executives were charged with making sure that the disciplines currently under study are really the best topics UTC could choose. Second, the VPs had to figure out who knew these disciplines best and get those experts involved.

The VPs-about eight in all-spent time analyzing the core technology areas that are critical to both products and customers, as well as defining who the appropriate experts are. They talked to UTECA members and picked the brains of their staff and that of other business units, and recommended topics and topic experts at a meeting with Mayhew and Cassidy in December 1999. The VPs are currently fine-tuning topics and recruiting the identified experts, Mayhew says.

While the VPs identify the who and the what of knowledge sharing, it's up to Mayhew and her group to organize the how-the format of the program. She wants to build it into something more than a series of classes and take it one step further than such standard knowledge management techniques as maintaining a list of experts. "We've done classes, listed standard best practices and lessons learned already," she points out, "and we've had a directory of experts in place since 1989."

Again, the group is drawing on the UTECA veterans' knowledge to design the way the new, structured knowledge sharing program will run. "We're in the organization phase," she says. "We see us putting together projects and classes where knowledgeable people can share what they know. We see us partnering with customers and people outside the corporation, too, including academic institutions."

Mayhew isn't quite sure of the role that information technology will play in the UTECA knowledge sharing program. "One of our challenges is understanding why many of the IT applications haven't been as successful as we think they should have been," she says, and she hopes to address that by inverting standard systems thinking. "Often when we create an application, we devise the app and then impose it on people and expect them to adapt," she says. "If you take a look at how people are naturally sharing information and then build a system that supports that, you're more likely to be successful."

Mayhew sees this program as an important goal for the United Technologies Research Center to attain. As the host of the project, Mayhew says, "It's important that we provide a curriculum that attracts really good minds to cooperate and provides opportunities to expand what we know."

E-mail Senior Editor Carol Hildebrand with your KM ideas at cjh@cio.com.

EXPERT CRITIQUE BY THOMAS H. DAVENPORT MAKE SHARING PAY The United Technologies case study illustrates some difficult issues in the second wave of knowledge management. UTC has the distinct challenge of having begun some of the early stages of knowledge management in the late '70s. It's no longer content with practices many other companies are now adopting as first steps. Its situation reminds us that knowledge management isn't new, only rediscovered. But once communities of practice are in place, how do we get greater value from them? There are approaches that UTC has identified, and perhaps one promising tack that it's not pursuing.

UTC manager Mayhew realized that a key source of value in an organization of independent business units is to transfer ideas across unit boundaries. In large, diverse companies, individuals are usually rewarded for advancing their own business units, not for contributing ideas to other groups. But sharing knowledge across boundaries is a low-cost way to boost the value of the entire corporation. That knowledge transfer will probably require not only occasional face-to-face gatherings among engineers with similar concerns but also some adjustment in how managers evaluate and reward good sharers.

Another means UTC has pursued for advancing technical knowledge is focusing on the knowledge that really counts. A key aspect of any knowledge strategy should be deciding what knowledge types are most critical to company objectives.

Mayhew and her team decided to focus on engineering knowledge as most important. Now they're focusing even more tightly by asking business unit engineering VPs to agree on what engineering discipline knowledge is key to success and to identify experts in those fields. This ensures that limited resources for knowledge management will be devoted to knowledge that really matters.

One angle that UTC apparently hasn't yet pursued could help expand the population of knowledge sharers. Mayhew notes that 1,500 engineers already enthusiastically share knowledge but that they represent a small percentage of the total. Such knowledge "heat seekers" go out of their way to pursue knowledge management and are willing to add it on top of their existing jobs.

Most knowledge workers, however, already feel that they've got too much to do and adding knowledge management on top of their existing roles won't be popular. I'd suggest that UTC consider redesigning the jobs of some key engineers in order to "bake in" the KM duties. This will make knowledge behaviors part of the day-to-day work rather than something extra.

Thomas H. Davenport is a regular contributor to CIO; see his column.

• Bookmark this page
• Share this article
  •  
  • facebook
  • slashdot
  • digg
  • Reddit
  • stumbleupon
  • linkedin
  • twitter
• Got more on this story? Email Computerworld
• Follow Computerworld on twitter

• Share
  Share this article

  •  google+
  • facebookfacebook
  • slashdotslashdot
  • diggdigg
  • RedditReddit
  • stumbleuponstumbleupon
  • linkedinlinkedin
  • twittertwitter
• print
• email
• Bookmark