Knowledge management is a subject that cuts across many
disciplines, each providing a slightly different perspective
on the field. A recent roundtable discussion,
moderated by editor Rowan Dordick, brought together
four of Research's leaders in knowledge management. Cathy Lasser
is business information executive and director of information systems
at Research and leads the division's
efforts in implementing knowledge management. Bob
Easton, manager of collaborative computing,
is overseeing the rollout of Expert Network at the Thomas J.
Watson Research Center. Marshall Schor, manager of knowledge systems,
works on many topics related to knowledge management, including systems
for the support of group collaboration. John Thomas, a research staff member
, is interested in developing tools to capture and summarize semistructured
information such as stories and pattern languages.
DORDICK: Is knowledge management something new or a new name for something old? And are people who are using this term talking about the same thing, or is there a spectrum of meanings?
EASTON: In his book Working Knowledge, Larry Prusak compares managers to Molière's Monsieur Jourdain, who was delighted to discover that he had been speaking prose all his life. Knowledge is something we've been doing all along, but it's just taken on new importance.
THOMAS: I agree that a lot of the activities now given the label of "knowledge management" have gone on before, and a lot of the technologies are extensions of existing ones. The difference is that now people are thinking of knowledge management from a systems perspective and in terms of the entire knowledge cycle.
LASSER: But there's also the new problem of the explosion of information, which makes it harder to get a handle on what is actually useful.
EASTON: Also, as people become more specialized, and as tasks become more complex and require a greater range of knowledge, there is a greater need for people to be able to find information possessed by others in the organization. For that to happen, the information has to be made available.
LASSER: And that's what knowledge management entails: managing and sharing all of an enterprise's information assets, from databases and documents to policies and procedures, as well as previously unarticulated expertise and experience resident in individual workers.
THOMAS: It is an illusion, however, to think that someone or some group has the ability to direct the creation and dissemination of knowledge. What one can do is foster a culture and an ecology of knowledge that promote those goals.
DORDICK: Then why use the term "management"? Is there something qualitatively different about knowledge management that distinguishes it from the traditional practices of disseminating knowledge?
THOMAS: Certainly, as Cathy said, more and more information is available, and so being able to find what's relevant in that sea of possibilities is a challenge that's only going to increase and require new approaches to sharing.
LASSER: And those companies that can make the best use of their knowledge will enjoy a competitive advantage, so there is a tremendous motivation.
DORDICK: Can you give an example?
SCHOR: The sharing of expertise is one. For instance, British Petroleum, it has been reported, found the costs to sink a well differed widely from team to team. They also found that the expertise of a highly productive team could be transferred to other teams simply by using videoconferencing and language translators.
DORDICK: That raises the question, to what extent is knowledge management a cultural issue -- a way of interacting within an organization -- as opposed to a strictly technical one?
THOMAS: They are interrelated, but the cultural
aspect is very important. In particular, knowledge management can be seen as a movement to enable us to understand how information has been created, the contexts in which it is applicable and even how to apply it.
DORDICK: But to benefit from knowledge management, do corporations have to provide specific tools, or is the first step to change the way employees think about everything they do?
EASTON: Most people who write about this topic say that it's much more about culture and people than it is about technology. And in fact, some writers say that if you're spending more than 30 percent of your budget on technology, you're spending in the wrong area. You really ought to be working on the culture and the people. Bacon's claim that knowledge is power might lead some people to think one benefits by hoarding one's knowledge, but when you're working as part of a team, you begin to see the value of collaboration and sharing.
DORDICK: What's the best cultural model to achieve that kind of sharing?
EASTON: There may not be an ideal model, but there may be some very general characteristics, even ones that transcend species boundaries. In a Harvard Business Review article last year [March-April 1997], the author contrasted the abilities of different species of birds earlier in the century to learn how to pierce the aluminum foil tops of milk bottles and drink the cream. It turned out that titmice, which flock together, learned, whereas the territorial robins never did. So, it was the flocking behavior, or the collaboration, that enabled individuals to learn and then share their knowledge.
DORDICK: Since it's improbable that everyone in a large organization will "flock together," how do you ensure that knowledge diffuses throughout the corporation, and not just among local groups?
THOMAS: An interesting approach to that problem took place in
Sweden. Karl-Henrik Robert, a leading oncologist and social thinker,
eventually got the entire country to
move toward a sustainable economy through a community of communities. While everyone in Sweden had the same simple systems view of ecology and a common set of principles, the details of implementation were left to a number of specific and fairly independent communities. That work may be a blueprint for how corporations might attack complex problems that require large-scale cooperation.
LASSER: It's also true that different levels of an organization need different kinds of knowledge, and even different tools with which to analyze them. So, while the knowledge base may be the same, different aspects are used by different groups.
DORDICK: Is it necessary, then, for someone to winnow out what's valuable for various groups, or is everyone responsible for finding what they need?
THOMAS: That's another one of these tensions. If Marshall [Schor], say, is knowledgeable in expert systems, he may have some idea of what's valuable for the rest of the corporation and make that available. But there may be some other aspect of what he's doing, such as the methodology of knowledge engineering, that he doesn't realize is relevant to me because he's not doing my job.
SCHOR: The process of packaging valuable knowledge is essential to knowledge management. That's why principals in IBM Global Services are encouraged to devote time to what they call "harvesting" their consulting experiences.
DORDICK: What incentives do most employees have to participate in such knowledge sharing?
EASTON: There are three basic motivations for sharing one's knowledge: reciprocity, reputation and altruism.
THOMAS: Yes, and a particularly valuable form of reciprocity is feedback: we all have particular blind spots -- we forget things, we overlook things and sometimes we're plain wrong. So, by sharing what we know, the quality of our own knowledge can actually become better.
SCHOR: That's standard practice in research: in the course of presenting papers at meetings and submitting them for publication, you receive a stream of comments that can help you clarify your argument.
DORDICK: How might something like that be adopted within a corporation?
THOMAS: Lotus Notes® databases on various topics exist within Research, and to some extent could serve as a prepublication repository of ideas.
LASSER: Yes, and there are also specialized discussion databases in Lotus Notes called Teamrooms, as well as personal Web sites. But as these resources proliferate, we are faced with the problem of managing them; at the very least, we need to find a simple way to let people know what's available.
EASTON: This is where knowledge management takes a slightly different direction from Marshall's example of principals aggregating and harvesting knowledge from within their community. One of the huge challenges is to improve searching technology, taxonomy, categorization and text analysis so that one can find the results one needs easily and quickly. (See stories, pages 26-27.)
DORDICK: Those are undeniably important technologies, but they seem to have less to do with sharing or making tacit knowledge explicit -- the well-worn themes of knowledge management -- than with information retrieval.
EASTON: The word "management" may be part of the problem here. If you rephrase the issue as "how do I find the information that I need at a particular moment," knowledge management can be seen as part of our lives already. The challenge is to make it simpler and more encompassing.
DORDICK: Is there a difference between information management and knowledge management?
SCHOR: It's true that a lot of examples of so-called knowledge management are really instances of information retrieval. I think knowledge management is trying to go beyond that, but we need to address the fundamentals, as well.
DORDICK: We've discussed the benefits, but are there also risks in sharing information that need to be considered? For example, might erroneous information spread because people uncritically reuse flawed
results?
SCHOR: That's one possibility. More serious is the case in which I make available a presentation that I have given to a customer and somebody else sees that I recommended a particular component without knowing the reasons I recommended it. For example, it may have been a requirement for a configuration the customer already had. But it may be that I would not generally recommend that component.
DORDICK: So, if we start to rely on shared knowledge and don't individually go through the rigor of thought necessary to test assumptions or test knowledge, how do we know that the thought processes that led to those conclusions are correct?
SCHOR: I think the problem and, in part, the solution exists today. For example, when people present new project proposals to management, it's customary to test them by asking: have you thought about this, what did you assume here, and so on. So that's already part of the culture, and that will continue.
LASSER: The biggest problem, however, is that in order to really be successful you need to rely on outside information, often on the Web where there may be only indirect evidence of the quality of the information -- for example, how many people link to a particular site.
DORDICK: In effect, then one puts one's trust in a virtual community.
LASSER: That's right.
DORDICK: Aside from strictly technological projects, what is being done to promote knowledge management within Research?
LASSER: Research has decided that knowledge management is a key area and will be one of its goals for 1999. Its importance lies not only in the opportunity it provides to improve the way we do our own jobs, but also in the opportunity it provides for IBM to be a leading-edge supplier of knowledge management technology. I think management is actively trying to promote the benefits of sharing and the importance of collaboration, across the division and, indeed, across the company as a whole. Also, I think the redesign of the Research intranet to allow easier searching and page authoring is a key part of the knowledge management environment.
DORDICK: Has any thought been given to such simple things as spaces for informal gatherings and "watercooler" chat?
LASSER: That probably falls in the category of quality-of-life issues, which is being addressed, and there is a team of people looking at what can be done to make us more creative. It certainly has connections to knowledge management: a centrally placed Ping Pong table, for example, as in our China Lab, could prove to be a meeting ground where ideas get generated.
SCHOR: So you could look at those cultural things the same way that you'd look at certain technology: they aren't knowledge management per se, but they enable it.
EASTON: And the sense of community they engender is very valuable. People will exchange information more easily with others whom they know, and they are more likely to trust the accuracy of such information. As companies build internal Web sites, one of the things they need to do is make them function as communal meeting places. And we're struggling with that here in Research, as well.
LASSER: It's all part of making the environment here more creative, more dynamic, more interactive and more collaborative.
DORDICK: As these projects move forward, how does one measure how well specific technologies and cultural changes are doing in realizing the goals of knowledge management?
EASTON: This is one of those areas where measures have been weak and subjective. Of course, frequency of use is a good measure, but anecdotal information -- feedback from people who have used it -- is also valuable because it tells us how to improve it. To a certain degree, we've provided the IS Web site today with some feedback mechanisms to collect comments, by means of a "thumbs up" and a "thumbs down" button on each of our Web pages. And human nature being what it is, people press the "thumbs down" button much more frequently, so we ask people what it is they like about this and what they dislike. And I think we're going to have to do that with every new project.
THOMAS: That is a good reminder that the organization as a whole, not just its individual members, can learn. Feedback loops allow an organization to assess its own performance at various levels. It can then use that knowledge to improve its policies, its procedures, its practices, its product mix -- even its culture -- over time.
EASTON: Yes, and a good example of how critical that can be is found in pharmaceutical companies. Because of the extensive testing and regulatory concerns involved in taking a new pharmaceutical product to market, the cost of not getting to market is more than a million dollars a day. Clearly, learning how to reduce that time saves a lot of money. One way is to learn how to find your way through the bureaucratic process, and another is to learn how to accelerate the tests, and to collect that information and put it in a place where the next team can use it to advantage.
LASSER: And there are similar, if less dramatic, benefits to be gained in every organization. I think the projects under way in Research to build knowledge management tools are very valuable, but I think we will really begin to see the utility of knowledge management when everyone begins to make it a part of his or her job, and the many improvements that result spread through -- and benefit -- every level of the organization.
