Bootstrap Institute logo Doug Engelbart's
   Colloquium at Stanford
An In-Depth Look at "The Unfinished Revolution"
Session 3
Knowledge management: Key findings and lessons
Marcelo Hoffmann1.*
- unedited transcript -

Okay, thank you, Doug. Okay. So what I will do today is give you a very, very short run-down of knowledge management in the business environment. It will be a very short-twenty minute, or so-presentation. And, take it with a grain of salt--it's meant to be for business because those are our clients.

But the meta-lesson from this, if you will take is, is how does this relate to bootstrapping? How does this relate to the government, or NGO's, and so on? We'll have Peter Yim speak afterwards about an NGO and what they do, which in a sense are a type of knowledge management. But my bias will be directly related only to companies because that's the environment where I work. 

Okay, I have to move this to get-sorry, presentation on the left there? Okay, thank you.

So, the concept is nothing new. It's been around for a long time. Different authors have taken different perspectives on this. 
 

 Concept of KM is not new
  • Vannevar Bush described machines cross-indexing human knowledge and following links from one source to another (Atlantic Monthly, 1945) 
  • Douglas C. Engelbart and his team developed hypertext, groupware, windowing, and other tools to facilitate online collaborations and knowledge work in the 1960s and 1970s 
  • Peter Drucker described the "knowledge worker" in the early 1980s

Doug has presented his views. Peter Drucker has written extensively about knowledge workers. But what is new now is the Internet. It is the networks; it is the greater computing power and the like.
 

 Status of KM in business
  • Current state similar to that of quality movement thirty+ years ago:

  • - Confusion/differences in semantics
    - Evolving methodologies 
    - Great variation among practitioners
    - General consensus regarding its long-term value
  • KM is unlikely to go away as just another fad, instead, will likely evolve and become more of a science, as it matures over time...

We can say that the status is pretty much equivalent to the quality movements of 20-30 years ago when you talk about knowledge management in the corporate settings. Everybody speaks of knowledge management as if it were something consistent, unique, individualizable, et cetera; it is not. And I will give some descriptions of what different folks mean by knowledge management and then talk about some cases, some suggestions for implementation, and then what we see in the future. And then, perhaps if you will keep in your mind-how does this relate to Doug's ideas, theories, strategies, and the like. One thing I would like to emphasize-that this knowledge management stuff is not a fad. It's definitely something that will remain for a while, but it will evolve.

The popularity has been only in the last five years. It probably might change in terms of methodology, syntax, semantics, and the like, but it's not going to go away as some of the other business consulting fad have come and gone. Perhaps, something to keep in mind, also, is what methodologies are applicable to be transferred from the knowledge management to bootstrapping? I believe some will, some will not; but perhaps it's something we should look at collectively. For lack of a definition, there are many descriptions, but one thing that is very consistent is something about value-either creating value, or re-using value, or conversion into or out of value, reuse of knowledge and the like.
 

KM implies value, purpose and appropriate methods
  • "The art of creating value by leveraging intangible assets" (Sveiby, others)
  • Driven by a growing understanding that knowledge is valuable, mostly intangible, difficult to hold on to, and usually product of collective thought
  • Corporate KM is the evolutionary result of many business trends, enabled by information technology
  • KM is primarily about management!

It's a result of many evolutionary processes and trends. But, definitely, it's mostly about management. When you really get down to it, if you cannot manage a corporation, you're not going to be able to do knowledge management particularly well. And, I'll get into that in greater detail later on. 
 

 KM involves a range of concerns and practices
  • Valuing knowledge: recognition and evaluation
  • Exploiting intellectual property
  • Managing knowledge workers
  • Capturing/sharing/distribution of work-based learning (from projects, business processes, workflow, clients, competitors ...)
Practices overlap tacit and explicit types
of knowledge ...

The practices vary. And, these I would to emphasize. Different writers, different authors and consultants speak as if knowledge management were any one of these four components--and sometimes a mix, combination, hybrids, and the like-but all these are included by some anyway. Valuing knowledge has been around for a while, getting increasing importance in terms of recognition and evaluation. Evaluation is very problematic. There are some attempts to go beyond accounting, but it is a matter of judgment. There are no ways to measure and value knowledge in a unique, repeatable way. But there's some ways in which you can compare greater or lesser value. 

Exploiting intellectual property--that's something that's very much in vogue--particularly by large companies that have been able to reduce the cost of maintaining intellectual property, selling what they have that they're not using. Companies like Dow and Motorola--there are a number of cases where they have used this very effectively both to reduce their cost, and to increase their profitability. 

Managing knowledge workers-that's been around forever, perhaps since the middle ages. But obviously it's getting more and more important now. And I will emphasize this later on in my presentation--that if there is a meta-message behind this knowledge management, the fad or movement, is that what is being used rather effectively is what's considered explicit knowledge-stuff that falls into databases, knowledge bases, manuals, you know, all the stuff that's obviously very valuable, and useful, and repeatable, and reusable, and so on. What isn't well used, or not as well used, is what's tacit, implicit in people's heads, that sort of thing. We haven't quite learned how to reuse what's in people's heads. But I'll give some suggestions as to technologies that might help reduce the problem of reusing implicit knowledge. The last one is the one that perhaps has the most potential, the greatest value, and arguably, the most difficult to capture.

This whole issue of capturing-I will read it directly-sharing/distribution of work-based learning from the day-today work of knowledge workers. And this implies from within the firm, outside, competitors, suppliers, allies, and the like. This is very problematic, and for this you would need to have something like what Doug calls a Dynamic Knowledge Repository. What we have seen in the companies we have studied is that this is not done particularly well. It's done very ad-hoc. Some companies will bet on others, but perhaps by chance, or by cultural reason, there is no great methodology that can be transferred. And the practices of the four components that you see above here do overlap, and at times, keep expanding. Now when looking at the approaches, the formal approaches used by companies who are very heavily involved with knowledge management evolve. The first movement, if you would, of keeping track of repositories has been pretty much done by most large companies. It's fairly straightforward; it's not that difficult to do. 
 

Approaches evolve
  • Repositories (to hold "knowledge as object") are often first and tend to span the organization: 

  • - Yellow/white pages, directories
    - Structured internal knowledge (data, manuals, code, other documentation)
    - Knowledge maps
    - Archives of online dialogues, threads of discourse
    - All supported by search engines, data-mining, visualization

And you see all the different kinds of collections of knowledge, or information, or data, if you would. But the latter component, which we think is very important, is relatively new-the whole thing about search engines, data-mining, visualization-making stuff available more easily, making stuff available when people need it, ideally as they need it. That's a little bit of a stretch as of now, but perhaps in time. Technology solutions are coming. The second set of activities; the second set of approaches is this whole arena of networks. And that's for the purpose of creating, holding, transferring tacit, or implicit, knowledge. This is much more problematic. It's much more about managing people, deciding organizations, evolving organizations and the people who are within them.
 

Approaches evolve (cont.)
  • Networks that create and hold tacit knowledge derived from work processes and interactions, tend to follow:

  • - Connections between individuals, teams, groups
    - Staff rotation among functions, groups, businesses
    - Communities of learning and practice support ad-hoc groups, creativity, innovation
  • Face-to-face remains crucial and combines with electronic means as systems tend to blend content and interaction over time

Different companies use different approaches. The classic ways are connecting people, teams, and ad-hoc get-togethers. Some of the companies locally you can see they design their offices and their buildings to increase the likelihood that people will meet with others. Several folks here who have worked for Sun will smile when I mention that at the new Sun facility in Menlo Park there are like mini-cafeterias every so often as a way to have folks gathers there, exchange information, knowledge. And, surprise, surprise-there are whiteboards right across there. Well, it is to get this exchange, which cannot be done in other, more formal ways. 

And, the last point about communities of learning and practice is something that has been going on now for a while, but increasingly there is more attention to this. And, I think that as the technology to connect on the networks, which has become increasingly more powerful, this will become even more popular. However, face-to-face will not go away any time soon. If there's a question about this, perhaps there might be competition with very high definition screens, and so on--if you can get the dynamics that are better than being there. But I'm not sure that, that will be the case-not yet, not for a while anyway. 

Fig. 1

If we look at the enablers and resources that are used for knowledge management-you can go from the left side, going from noise, data, information. I spoke about knowledge as object, then going on from knowledge as process. The higher, right-hand corner will be knowledge as emergent properties. And this is really tricky to go to. That's something that I believe is still a research question. Perhaps by using bootstrap in evolving organizations beyond the current state, this will work out. But there is no methodology approach that I know of for all the companies I've interviewed, and visited with, and spoken to. But, in a sense, there is this transition from the low-end stuff to the higher-end, increasingly more tacit, and more complex, and much more difficult to transfer. 
 

Emphases vary
  • MECHANISTIC 
  • Short Term, Tactical 
  • Efficiency/Effectiveness 
  • Automation 
  • Content 
  • Military/Hierarchical 
  • Small-scale
  • ORGANIC 
  • Long Term, Strategic 
  • Creativity/innovation 
  • Human-centered 
  • Interaction-oriented 
  • Biological/Adaptive 
  • Large-scale

The emphases, by various companies, vary dramatically. If you look here, from the left side, the more mechanistic approaches to the organic. If there's one message that we can convey it's that, in general, most modern companies are going from the left side to the right side, but not all evenly and not evenly down this list of characteristics. One company might be very interested in short-term tactical ideas, but very involved with creativity and innovation, and so on. So, you would see somewhat of a zigzag of organizations if you were to take some sort of measurement across a large number. But, in general, we have seen a transition from the left side to the right side, towards more biological approaches-and also larger-scale, and more chaotic, perhaps going along with, the Santa Fe Institute talks about that, chaos and chaos theory. Another way of looking at this knowledge management and the processes underneath is that they tend to go from the simple/tangible on to the less tangible, but potentially the greater the benefit.

Fig. 2

And, when I've given these presentations in different companies and different companies, what tends to come back from the dialog is most companies are at the left-hand corner, bottom left-hand corner. Trying to avoid repeating the invention of the wheel. And they keep doing that, over time, more efficiently. Some are going to the concept of how do they make a better wheel-but not that many, and not often on purpose. If this were drawn to scale, there would be a major chasm between that and the fellow on the hand-glider, because that's changing paradigms. That's very difficult-going from there to thinking about thinking. The meta-approaches that Doug proposes, I have yet to see any organization, any other speaker, besides Doug and the group that's promoting and trying to bring this bootstrapping to happen, suggesting this.

So, if there were a chart describing where the companies are that we have observed, it would be largely at the bottom left, with a few going up, and the, nobody at the top, right-hand side. And, there might be some hypotheses about countries, and size of organization, and so on. Going back to, say, some visits in Japan some time back-Japanese companies, for example, are very good at not re-inventing the wheel. They transfer that kind of knowledge relatively well. They have a very hard time with creativity and innovation, and so on. Different cultures will work differently-better for some things, less well for others. Organizational models are something very important to look at for considering knowledge management.
 

Organizational models set context
  • Command-and-control ["military"]: rigid, hierarchical, moderate scalability, effective in narrow domains
  • Exchange economy ["marketplace"]: scales up easily, useful for explicit transactions and exchanges among many
  • Gift society ["family"]: difficult to scale up, enhances deeper, long-term relationships; enables serendipity and high-value sharing; implies complex interactions (shared values and high emotional quotient are crucial...)

The three kinds of organizations, for lack of a better taxonomy that we have seen-command-and-control, exchange economy, and gift society. What's very common is that many speakers present the idea for gift society as if that were the one that they would get to first. I have yet to see that happen-if anything, you can get into exchange economy where there's a trade off: I will give you something if you will give me something in return that's equally valuable or better. And one of the suggestions we make is to go for that, to go for the exchange economy. Don't assume that you will generate a gift society or a gift group just by itself. You have to develop the trust-go from command-and-control to exchange economy and then to a gift society. And, in fact, most organizations will probably oscillate between an exchange economy environment and a gift society environment. You know, depending if the trust goes up and down. But, it's still better than working in command-and-control where folks will only contribute what they have to contribute, and no more. So, this is something to consider if one were to analyze what is the situation in a particular company and we suggest to also consider how to ferment the evolution from an exchange economy to a gift society. What we have found among the cases that we have analyzed in greater detail is that they're extremely contextual to be successful. There's no generic approach that will work in different companies. 
 

Successful approaches are necessarily contextual
  • Chaparral Steel 

  • - Small, highly focused mini-mill
    - People-centered, mix work-shifts, other "soft" methods
    - Tacit knowledge ("apprenticeships")
    - Individual and small-group learning
  • Boeing 777 Program

  • - Very large, aerospace firm
    - Technology-centered: CAD, high-speed networks
    - Explicit knowledge ("engineering")
    - Large-scale, shared understanding across time and space

There are two cases that I have found that are classics: Chaparral Steel, a small mini-mill, some couple thousand people, and Boeing, the triple seven-I just couldn't think of Boeing as a company because it's just too large. But if you can compare Chaparral Steel, they work essentially as if they were a company working in the Middle Ages with apprenticeships. And they have their R&B on the production floor, so newer folks learn from people who have been there longer and who have done steel-making longer, and so on-very tacit knowledge. Very little use of computers. Once they sell the steel, it's gone; they don't have to guarantee it anymore. If you contrast that to Boeing where they have to keep track of everything they've done for years and years, maybe for 40 or 50 years after the airplane will fly, it's a very different environment. It's a different approach-and for good reasons. So if I were to take another metaphor-it's sort of right-brain/left-brain, but the whole concept here is to be specific. What kind of organization do you have? What do you want to get out of it? And then, what do you decide for it? Perhaps, what sort of folks do you hire, train, support, that sort of thing? One of the suggestions we have found is very useful is, go for the early benefits. There's nothing better, to build up the bandwagon effect, than being successful early on. And what we have found within that is the idea of going after improved knowledge repositories.
 

Early benefits are common and desirable
  • Improved knowledge repositories and means to share ("to know collectively what we already know individually")
  • Improved efficiency and effectiveness
  • Improved communications, reduced cycle times
  • Ultimate benefits are indeterminate: "users build the road as they travel..."

It's much easier to prove the benefit of these than to prove the benefit of a network that might take six months, a year, two years to come up with something interesting and exciting. Also, going for effectiveness and efficiency, and even though these may be lower-level gains, they're still useful to create the environment that, in a sense, will take you to a higher-level. But one thing to keep in mind-you cannot predict what will happen in the longer run. Literally, users build the road as they travel it; and it's a co-discovery between the developers, the designers, participants, and the like. It's an interesting travel with all kinds of detours. Some of the risks that we have found in organizations that try to do this fast is shifts in power, the changes in politics and relationships and so on, were very problematic. And, often times, the senior manager was not aware of what the effects would be until after the fact, until it was a major crash.
 

Associated risks
  • Unexpected changes to organizations-shifts in power, relationships, value proposition(s)
  • Security issues and info-glut now, privacy in the future
  • Large-scale efforts create problem of dividing costs in a distributed environment: costs and benefits do not correlate easily-they spread across time, space, and organizational structures)

And, the problem of implementing this things is also difficult enough that we suggest: start small, evolve, grow, see what works in your organization, and then, as long as there are successes, things will evolve nicely. Otherwise, if you try to do something big, there will be major problems in allocating costs because those who pay are not necessarily those who benefit. Then, there will be a major fight for who pays, who gains, where, when. In addition, these things spread over time, which is very problematic because there is still-as I mentioned before--not very good ways to measure what the value of improved knowledge management will be over time. We have found that there are at least two kinds of resources that are absolutely crucial. The organizational part, in Doug's terminology, would be the human systems, the agreements, the excitement, and he commotion of the right kinds of activities, of behaviors, that sort of thing. We would add mentoring, training, support to the point of being pedantic.
 

Resources needed: Organization 
  • Belief in sharing, but also in delegation, use of judgment, and selective ignorance! 
  • Time to meet/listen/learn from peers, clients, providers, competitors, other stakeholders; travel...
  • Promotion of excitement in the process
  • Willingness and ability to foment group coordination: high-level champion(s), consistent leadership, congruent incentives
  • Staff mentoring, accessible training and support

Definitely leadership from senior executives is very important.
 

Resources needed: Infrastructure
  • Appropriate places to meet/listen/learn with and from others...
  • Connected workstations and/or personal computers
  • Adequate network architecture, database, and infrastructure to search, exchange, share
  • Ongoing mentoring, training and support in dynamic environments

Besides that infrastructure, I will go back to what I mentioned earlier: even the architecture of the buildings where collaborators work is important-for folks to meet in an environment that's conducive for them to do what they want to do, better. Obviously, computers, networks, all that, they are nice, nothing new there but important to have ongoing mentoring, training, and support. We have had cases of that, even at SRI, where as Microsoft keeps changing the software, we need support where we didn't think we would need support. So, I've had several cases of that myself.
 

Information technology enablers in use now
  • Major investment in electronic networks, tools, Internet, Web 
  • World Wide Web browsers, authoring tools, and other Internet standards help to develop, share, exchange, and present
  • information/knowledge (HTML, XML, KQML, Java, others..)
  • &Search technology to find sources, content, contacts
  • Developments from artificial intelligence: speech recognition, natural language, knowledge bases...

Several things that we see as enablers now that also I mentioned earlier-about the networks, the tools, the internet-a lot of value will be gained from the transition from HTML to XML, perhaps KQML, Java, others. I think we'll have some speakers later--next week and the following week-getting into this in much greater detail. So, I will not dwell onto this one. Search technology is going to be very important because soon there will be the ability to capture, as Jim Spohrer said last week. You know, if you can have a very small form factor hard drive that can store all the audio for your whole life, well that's nice, but then how do you find anything. So, if we have better search techniques and approaches, particularly things from artificial intelligence-speech recognition, natural language, knowledge bases, and so on-in a somewhat automated fashion. That'll be quite nice and useful. These things are here now; they can be used better, but they're there. Okay, one thing that's coming up, that's not quite there yet, the idea of cross-referencing multi-media-actually SRI has some efforts on this where you can cross-correlate different media streams from video, sound, documents, and so on. And, if you can do what amounts to a multi-media search, there's a great potential there-both for good and for spying, and for privacy violations, and for security problems. But still, it's something worth looking at because it's got very good potential. This whole idea of automating some of the activities that people do now also has huge potential, although it will never, or not for a long time, become completely automated. I don't think people will be pulled out of the loop; but, perhaps the lower-level activities, thee lower-level tasks, can be automated. I believe Doug goes along with that in terms of co-evolution and so on. The suggestions that we have made that seem to be consistent with most of the success cases that we've seen follow here.
 

Information technology enablers in future
  • Possibility to "store everything" digitally to categorize and analyze when needed
  • Cross-referenced multimedia (telephone conversations, e-mail, videotaped meetings, streaming video....)
  • Integration of artificial intelligence techniques to access, understand, and perhaps extend large corpus (neural nets, software agents, genetic algorithms...)
Methodology for new initiatives
  • Identify needs given organizational strategy and objectives 
  • Assess and address human and organizational factors (culture, structure, incentives, practices...)
  • Define the "knowledge architecture(s)": repositories, networks, organizing principles...
  • Choose and deploy technology infrastructure
  • Implement in modules, test, improve
  • Repeat above steps as necessary..

The idea is to start looking at what are the needs. Take a very gook look at the human organizational elements, factors, environment, culture, all that. Then start the design process with the knowledge architecture. What should be the ideal combination of things so that users would get benefits fairly early on? Only then think of the technology infrastructure, and then stop. Start doing this in modular fashion: test, improve and repeat the steps. This is a highly intricate process. I have yet to see any successful company deploy a formal, large knowledge management system in one fell swoop. It just doesn't happen; it's too complicated.
 

Possibilities ahead
  • Create greater value from knowledge resources:

  • - Improve processes, globally: potential for large-scale implementations in collaboration and knowledge exchanges
    - Enhance individual/group creativity and innovation; adaptability to new environments
    - Develop seamless web of resources: combine people, technology, knowledge, and processes in continuous cycle(s) of learning and synergy

Some of the possibilities ahead include: the processes in a much more global fashion--large scale implementations, what Doug talks about, you know, meta-meta levels--large organizations collaborating in a closely coupled fashion, maybe across countries or large enterprises. Another great opportunity is the one to enhance individual and group creativity. That's something we don't have much of a handle yet, but perhaps by reducing the lower-level functions then perhaps folks will have the time and ability to do the higher-level activities.

And then, this whole notion of this seamless web of resources combining them in a fashion that is useful at the time that has needed, in the best way. That is a little bit of a pie in the sky for now, but perhaps not in a few years. It is something that we will have to evolve into. And, I want to finish with some comments on a slide presented by Jim Spohrer last week about learning that although when it comes to knowledge management it has taken, to some extent, the first three of these components, performance support, practice skills, habits, and training. It has yet to, within the corporate setting, look at education, research, the finding of new answers and the finding of new questions. There's still a lot of opportunity there, but I'm not sure that we quite have it yet. And, with that, I finish my presentation and cede the floor back to Doug. Thank you.

[<] principal lecture]


Top


 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

 ---
 
 

















Above space serves to put hyperlinked targets at the top of the window
Copyright 2000 Bootstrap Institute. All rights reserved.