Skip to contentSkip to navigation

Some Principles of Knowledge Management

More than ever, companies are realizing that their real advantage lies in what they know. But how do you manage knowledge?

(originally published by Booz & Company)

We do not know one millionth of one percent about anything,” said Thomas Alva Edison who probably wouldn’t have appreciated Professor Davenport’s observation that the most valuable asset companies have is the knowledge of their employees. More than ever, companies are realizing that their real advantage lies in what they know. But how do you manage knowledge?

Enunciating 10 principles of knowledge management, Professor Davenport establishes a framework which senior company executives can use to govern their approach to knowledge. He also uses the Hewlett-Packard Company, in an accompanying case study, to illustrate how rapidly the knowledge management concept can spread throughout a company even without a top-down mandate.

Systems such as the World Wide Web and Lotus Notes are giving managers real opportunities to capture and distribute knowledge. Hewlett-Packard’s Product Processes Organization (P.P.O.), for example, has formed a Knowledge Management Group and is developing a Web-based knowledge management system called Knowledge Links. The group intends to develop a variety of services with applications not only for P.P.O. but for Hewlett-Packard as a whole.

Professor Davenport warns that managing knowledge in organizations will lead to new problems and issues, even resistance by what he calls an 'anti-intellectual' orientation in the United States. But he points out that as free natural resources and cheap labor are exhausted, the last untapped source of commercial advantage is the knowledge of people in organizations.

Many companies are beginning to understand that the knowledge of their employees is their most valuable asset. Even so, few have actually begun to actively manage that asset on a broad scale. Thus far, they have addressed the issue at a philosophical or a technological level, with little discussion about how knowledge can be used more effectively on a daily basis.

For the past two years, I have been working with organizations in this area. The first order of business has been to establish the framework a company will use to govern its approach to knowledge. From that base can flow detailed approaches and plans.

Even at this early stage of knowledge management, enough lessons have been learned to offer some basic rules of thumb. What follows then are 10 principles of knowledge management, along with a sense of their real-world implications.

1. Knowledge management is expensive (but so is stupidity!)

Knowledge is an asset, but its effective management requires the investment of other assets, namely money and labor. Employees need to be trained in the ways of capturing, packaging and categorizing knowledge. And they will need the right hardware and software tools to do that.

While few companies have calculated the cost of knowledge management, there are some quantified estimates. Robert Buckman of Buckman Laboratories estimates that his firm spends 3.5 percent of its revenues on knowledge management. McKinsey & Company has long had an objective of spending 10 percent of its revenues on developing and managing intellectual capital.

But while knowledge management is expensive, the obvious retort is that not managing knowledge is even more so. What is the cost of ignorance and stupidity? How much does it cost an organization to forget what key employees know, to be unable to answer customer questions quickly or at all, or to make poor decisions based on faulty knowledge?

Just as organizations concerned about the value of quality figured out the cost of poor quality products and services, companies trying to assess the worth of knowledge can attempt to measure the cost of not knowing. Of course, such an assessment could lead to political problems within the organization, but that is the province of another principle.

2. Effective management of knowledge requires hybrid solutions involving both people and technology.

"Computers that think are almost here," a Business Week article recently announced, adding that "the ultimate goal of artificial intelligence--human-like reasoning--is within reach." Such a prediction may produce a feeling of déjà vu for managers and professionals, who have been hearing about machine-based knowledge since the 1950's. The fact is that companies wishing to effectively manage knowledge still need a heavy dose of human labor.

We need to construct a hybrid environment that uses humans and machines.

People may be expensive and cantankerous, but they are quite accomplished at certain knowledge skills. When we seek to understand knowledge, to interpret it within a broader context, to combine it with other types of information or to synthesize various unstructured forms of knowledge, humans are the recommended tool.

Computers and communications systems, on the other hand, are good at capturing, transforming and distributing highly structured knowledge that changes rapidly. They are increasingly useful--though still a bit awkward--at performing these tasks on less structured textual and visual material. But it remains the case that most people do not turn to computers when they want a rich picture of what is going on in a particular knowledge domain.

Given this mixture of skills, we need to construct hybrid environments for knowledge management that use humans and machines in complementary ways. Just as sophisticated manufacturers have realized that "lights out" factories are not necessarily the most effective or flexible, it is clear that knowledge factories should offer not only the bits and bytes of a computer but also someone to talk to.

For example, when compiling computerized databases of organizational knowledge, it makes sense to include "pointers to people." That is what the G.M. Hughes Electronics Corporation did, capturing its best-process re-engineering practices in a database that combined human and computerized knowledge. Each entry was submitted to an editor, who screened it for usefulness and relevance. Entries recorded just enough about the practice to pique the reader's interest, and included the name and phone number of a person who could describe it in detail. Use of the database is solid and growing.

3. Knowledge management is highly political.

It is no secret that "knowledge is power," and thus it should not surprise anyone that knowledge management is a highly political undertaking. If knowledge is associated with power, money and success, then it is also associated with lobbying, intrigue and backroom deals. If politics plays no part in a knowledge management initiative, it is a safe bet that the organization perceives nothing of value is at issue.

What does knowledge politics mean for effective knowledge management? Some managers will argue that politics only gets in the way. But astute managers will acknowledge the value of politics and cultivate its use, lobbying on behalf of knowledge. They will broker deals between those who have knowledge and those who use it. They will cultivate influential "opinion leaders" as early adopters of knowledge management approaches. At the highest level, they will try to shape the governance of knowledge to better utilize it across the organization.

4. Knowledge management requires knowledge managers.

Key business resources like labor and capital have substantial organizational functions devoted to their management. Knowledge will not be well-managed until some group within a company has clear responsibility for the job. Among the tasks that such a group might perform are collecting and categorizing knowledge, establishing a knowledge-oriented technology infrastructure and monitoring the use of knowledge.

The goal should be to facilitate the creation, distribution and use of knowledge by others.

Several professional services firms already have knowledge management roles in place. Booz-Allen & Hamilton, McKinsey, Andersen Consulting, Ernst & Young, Price Waterhouse and A.T. Kearney all have chief knowledge officers. Buckman Laboratories reoriented its information systems organization to become managers of knowledge, and it now calls the group the knowledge transfer department. The Hewlett-Packard Company created one knowledge management group within its corporate products and processes organization, and another within its computer systems marketing group. (See the article at the end of this piece.)

A knowledge management function could inspire resentment and concern within the organization if it seeks to collect and control all knowledge. The goal should merely be to facilitate the creation, distribution and use of knowledge by others. Furthermore, the knowledge managers themselves should not imply by their words or actions that they are more "knowledgeable" than anyone else. In fact, one knowledge manager at Hewlett-Packard argues that the most important qualification for such a role is being "egoless."

5. Knowledge management benefits more from maps than models, more from markets than hierarchies.

It is tempting when managing knowledge to create a hierarchical model or architecture, similar to the Encyclopedia Britannica's Propaedia, that would govern collection and categorization. But most organizations are better off letting the knowledge market work, and simply providing and mapping the knowledge that its consumers seem to want. The dispersal of knowledge as described in a map may be illogical, but is still more helpful to a user than a hypothetical knowledge model that is best understood by its creators, and rarely fully implemented. Mapping organizational knowledge is the single activity most likely to yield better access.

Knowledge managers can learn from the experience of data managers, whose complex models of how data would be structured in the future were seldom realized. Companies rarely created maps of the data, so they never had any guides to where the information was in the present.

Letting the market work means that knowledge managers try to make knowledge as attractive and accessible as possible, and then observe what knowledge gets requested using what specific terms. For example, at Teltech, a Minneapolis company that manages a knowledge network of external experts, clients who call for referrals are unlikely to always use the same terms as the experts use in describing their work. The function of connecting client needs to available expertise is performed by Teltech's online search and retrieval system. Called the Knowledgescope, the system is effectively a map or thesaurus of more than 30,000 technical terms. It is maintained by several full-time knowledge engineers, who add 500 to 1,200 new concepts per month to the database while removing outdated ones.

Each technical term has a preferred usage and several possible synonyms. Teltech's goal is to have the terms in the database that are used by clients. Therefore, each day the knowledge engineers receive a list of terms sought unsuccessfully in the database by Teltech's knowledge analysts or by clients accessing the database directly. Many of the unsuccessful searches are misspellings, but valid misses are added to the database.

Until recently, Teltech's approach to structuring knowledge had been hierarchical, rather than thesaurus-based. Its previous database was called the Tech Tree and it had several key knowledge branches, including scientific/technical, medical and chemical. However, both the clients and the knowledge analysts found it difficult to navigate through the tree, and new terms tended to be added at inappropriate places.

Teltech has found the thesaurus approach to be much more satisfactory. It has mapped the knowledge world rather than modeled it.

6. Sharing and using knowledge are often unnatural acts.

If my knowledge is a valuable resource, why should I share it? If my job is to create knowledge, why should I put my job at risk by using your knowledge instead of mine?

We sometimes act surprised when knowledge is not shared or used, but we would be better off assuming that the natural tendency is to hoard our own knowledge and look suspiciously on knowledge that comes from others. To enter our knowledge into a system and to seek out knowledge from others is not only threatening, but also requires much effort -- so we have to be highly motivated to undertake such work.

If the knowledge manager adopted this principle, we would not assume that the installation of Lotus Notes will lead to widespread sharing, or that making information available will lead to its use. We would realize that sharing and usage have to be motivated through time-honored techniques--performance evaluation, compensation, example.

Some companies are beginning to employ those techniques. The Lotus Development Corporation, now a division of I.B.M., devotes 25 percent of the total performance evaluation of its customer support workers to knowledge sharing. Buckman Laboratories recognizes its 100 top knowledge sharers with an annual conference at a resort. ABB ASEA Brown Boveri Ltd., the Swiss-Swedish conglomerate, evaluates managers not only on the results of their decisions, but also on the knowledge and information applied in the decision-making process.

7. Knowledge management means improving knowledge work processes.

It is important to address and improve the generic knowledge management process, but knowledge is generated, used and shared intensively through a few specific work processes. While the details vary by company and industry, these include market research and product design and development, and even more transactional processes like order configuration and pricing. If real improvements are to be made in knowledge management, gains must occur in these key business processes.

Knowledge work processes are only rarely addressed in process improvement initiatives.

Two colleagues and I recently carried out research involving more than 25 companies that had attempted to improve knowledge work processes. We found processes oriented to creating (e.g., research), packaging (publishing) and applying (system development) knowledge. In general, the most effective improvement approaches struck a middle ground between top-down "re-engineering" of the process and bottom-up design by autonomous knowledge workers. Creative knowledge work required less top-down intervention, and knowledge application processes a bit more. However, surveys of companies on their re-engineering efforts have confirmed that knowledge work processes of any type are only rarely addressed in process improvement initiatives.

8. Access to knowledge is only the beginning.

If simple access to knowledge were sufficient, then there would be long lines outside the nation's libraries. Access is important, but successful knowledge management also requires attention and engagement. It has been said that attention is the currency of the information age.

More active involvement with knowledge can be achieved through summarizing and reporting it to others, through role-playing and games based on usage of the knowledge and through close interaction with providers. This is particularly important when the knowledge to be received is tacit, as Ikujiro Nonaka has long noted.

Some companies have already begun to help their managers and employees become more engaged with knowledge. Jane Linder, an information (and market research and strategic planning) manager for a division of the Polaroid Corporation, worked with a supportive division president to create a "war games" exercise for managers and professionals. Participants digested market research and then played roles as competitors or as Polaroid in making sales presentations to customers. The marketing-oriented exercises were a big success, and now Polaroid is assessing the use of engagement approaches for other types of knowledge. Toyota and Nissan have both sent car designers to the United States to receive tacit knowledge by fraternizing with particular customer segments.

9. Knowledge management never ends.

Managers may feel that if they could only get their organization's knowledge under control, their work would be done. However, the tasks of knowledge management are never-ending. Like human resource management or financial management, there is never a time when knowledge has been fully managed.

One reason is that the categories of required knowledge are always changing. New technologies, management approaches, regulatory issues and customer concerns are always emerging. Companies change their strategies, organizational structures and product and service emphases. New managers and professionals have new needs for knowledge.

This rapid change in knowledge environments means that companies should not spend considerable amounts of time in mapping or modeling a particular environment. By the time they finished, the environment would no longer exist. Instead, descriptions of environments should be "quick and dirty," and only as extensive as usage warrants.

10. Knowledge management requires a knowledge contract.

It is not clear in most organizations who owns or has usage rights to employee knowledge. Is the knowledge of employees owned or rented? Is all of the knowledge in employee heads the property of the employer? How about the knowledge in file cabinets or computer disk drives? What about the knowledge of consultants while they are consulting? Outsourced employees? Few companies have policies to deal with these issues.  

Many organizations have held that employee knowledge--at least that developed between 9 and 5--is the property of the corporation. However, several societal changes make such an approach problematic. Employees are moving more often to new jobs and new organizations; the distinction between work life and home life is more ephemeral, and there are more contingent workers. In any case, few companies have done a good job of extracting and documenting the knowledge of employees.

If knowledge is really becoming a more valued resource in organizations, we can expect to see more attention to the legalities of knowledge management. Perhaps the greatest problem with increased knowledge management is the increased population of lawyers it will engender. Intellectual property law is already the fastest-growing legal field, and it will only grow faster.


As can be easily deduced from these principles, managing knowledge in organizations will lead to a variety of new problems and issues. Moreover, this type of initiative will face resistance. Knowledge management runs counter to the direction of American society; we prefer television to books, intuition to research, pragmatists to theorists. The serious pursuit of knowledge in organizations will be challenged by an anti-intellectual orientation in the United States that has been present since the days of the frontier.

But now the new frontier is in our minds. As free natural resources and cheap labor are exhausted, the last untapped source of commercial advantage is the knowledge of people in organizations.

It is the very early days for knowledge management, and even the principles and rules of thumb described above will engender considerable disagreement. The good news is that almost anything a company does in managing knowledge will be a step forward.

Case Study: Knowledge Management at Hewlett-Packard

Knowledge management is exploding at the Hewlett-Packard Company. While there has been no top-down mandate to manage knowledge at this highly decentralized computer and electronics manufacturer, many divisions and departments are undertaking specific efforts to do just that.

A corporate "knowledge czar" would not fit with HP's culture, but many managers are attempting to capture and distribute the knowledge residing in their own business units and departments.

The efforts are springing up quickly, and it is difficult even to identify and track all of them. The Computer Systems Marketing organization, for example, has put a large amount of marketing knowledge into a World Wide Web-based system that can be accessed around the world. It contains product information, competitive intelligence, white papers and ready-to-deliver marketing presentations.

HP Laboratories, meanwhile, is developing approaches to facilitate access to both internal and external knowledge. And Corporate Information Systems is putting document-based knowledge of procedures, personnel and other information into Web and Lotus Notes systems; its managers argue for a "pull" approach to knowledge distribution (accessed when needed), rather than the typical "push" strategy. Information Systems is also attempting to map the various sources of knowledge about information systems development and management around the company.

Perhaps the most focused, intensive approach to knowledge management is in the Product Processes Organization, or P.P.O., which provides the company's product divisions with such services as purchasing, engineering, market intelligence, change management and environmental and safety consulting.

Bill Kay, the director of Product Processes, believes that information and knowledge management should be a core competence, and he even put his unit's Information Systems group at the center of the P.P.O. organization chart. Product Processes has adopted many approaches to knowledge transfer in the past, including catalogues of documents, video and audiotapes of meetings, best practice databases and the Work Innovation Network, a series of meetings and ongoing discussions on change management topics.

Until recently, however, there had been no formal responsibility for knowledge management in Product Processes. But in 1995, Mr. Kay formed a Knowledge Management Group within P.P.O.'s Information Systems Group. Its initial charter was to capture and leverage knowledge of the product-generation process in the various HP divisions.

The group quickly developed a prototype of a Web-based knowledge management system called Knowledge Links. Its primary content is knowledge about the product-generation process; the knowledge may come from a variety of functional perspectives, including marketing, R.&D., engineering and manufacturing. The knowledge going into Knowledge Links comes from outside the Knowledge Management Group, but group members add value by identifying, editing and formatting the material, making it easier to access and use.

The P.P.O. Knowledge Management Group intends to develop a variety of other services with broader applications for P.P.O. and for HP as a whole.

It plans to develop for internal HP clients versions of Knowledge Links for other types of knowledge. It has developed an assessment tool for weighing the levels of knowledge management capability in P.P.O. and other company groups. It is beginning to map key knowledge domains within P.P.O. Finally, within the context of a Knowledge Links development activity, the group offers advice about how best to foster knowledge-creating and sharing behaviors, without which the technology is of little value.

Hewlett-Packard has been a knowledge-oriented company since its founding in 1938. Now, however, many managers throughout the company are realizing that the knowledge management concept must be taken to a higher level.

Illustration by Evangelia Philippidis

Reprint No. 96105

Get s+b's award-winning newsletter delivered to your inbox. Sign up No, thanks
Illustration of flying birds delivering information
Get the newsletter

Sign up now to get our top insights on business strategy and management trends, delivered straight to your inbox twice a week.