Rick Dove, Paradigm Shift International, www.parshift.com,

Companies have managed knowledge,
and people have collaborated,
since time began.

If you build it they will come. Not! The experts are back again saying that they've learned from experience in their own company, and with others - that there isn't a natural pent up demand for this new infrastructure - after the ribbon cutting ceremony a few people are eager to exploit the system but not many. So the experts are ready to help with the next phase: cultural reprogramming. They understand this phase well from first hand experience - they've lived through their own indoctrination period successfully. Of course their people are all MBAs who naturally make a living by communicating and collaborating.
   But they'll show you that it works in industry, too. British Petroleum is living proof of that. BP began its $12 million dollar pilot-program in 1995, and "about a third of the money was spent on behavioral scientists who helped the people in the pilot programs learn how to work effectively in a virtual environment [Harvard Business Review, Sep-Oct '97, p. 153]." BP has good bottom line results to show for it: oil drillers who ran into really expensive problems got on the network and found others who had solved these very same or similar problems before. It's been quite effective for BP, and made good times of idle time for the drillers isolated on these drilling platforms - kind of like ham radio with a video link.
   Companies have been managing knowledge and people have been collaborating in communities of practice as long as there have been companies and people. This is not new stuff, this is not unnatural stuff. What is new, and useful, is the technology that can broaden the communities to include people who may never shake hands as well as people who are simply somewhere else. What is unnatural is the new abstraction of knowledge management into something artificial, something that can be dreamed up anew from logic and ideal visions. Something that can be purchased like a box of Cheerios - just add milk to nourish and reinvigorate the corporation.
   But they point with pride to what they've come to think is a real knowledge management system. After all, most of the experts say that this is what has to be done. And the experts know because they've already done it for themselves and for others, and they'll help get it started for you, too. The results are refreshingly tangible: you can count the bytes of captured "knowledge", you can count the numbers of accesses to this so-called knowledge base, and you can count the man-hours and dollars involved in creating and maintaining this resource. Numbers that all measure some form of performance and prove that the company is taking knowledge management seriously.
   And there's more to do - more man-hours to measure and more information technology to buy. Those same experts and a few new ones can now show that it is at least equally important to develop and support communities of practice - one of the most effective forms of collaborative learning. This makes a lot of sense and has a nice clear path to follow: you buy stuff from computer vendors and from software vendors. Things like network servers and Lotus notes, maybe even collaborative project notebooks and whiteboard services on the Internet. Once the infrastructure is up and running it's time to get some users, and create those communities of practice - you know - people who have common work-interests even though they reside in different places throughout the company, people who want to work together on the corporate intranet to swap knowledge and help each other do better faster.
  

©1999 RKDove - Attributed Copies Permitted
Essay #053 - Published 5/99 in Automotive Manufacturing & Production, Gardner Publications

========= Reply =========================
From: m.crawford@deuk.co.uk (Mark Crawford), Date: Tue, 1 Jun 1999
What about measuring your performance. Models are good at showing you what to do, but how to implement it.
How to know how well you are managing. With direct measures it is much easier to keep motivated and implement it. Indirect measures such as cost savings, lead time reduction does not provide adequate evaluation.

========= Reply =========================
From:(Rick Dove), Date: Tue, 1 Jun 1999
Good point, Mark. Indirect measures are often like statistics - you can use them to make any point you want, especially when it is difficult to show direct and unequivocal causal effect. But direct measures also present difficulties. They often measure the wrong things just because they are the only (or the easiest) things to measure directly.

Like measuring the hit count on a knowledge-repository web site, for instance. I'll pick on this issue to make my point, but any issue would do.

A hit on a web page could simply mean someone referenced it and found it inapplicable - often the case when running through a list of search engine results. This is actually worse than it sounds. It is a performance count in the positive direction that actually should be negative as it wasted someone's time.

Far worse, however, are the hits that a web page counts that are done by robots and not people. Search engine robots account for a large number of page hits as the engines attempt to update their indices frequently. These days the robotic "readers" are inflating page-hit numbers even higher by downloading an entire site for presumed off-line reading later by a human.

I am suspicious that I have some pages on my web site that have a higher percentage of robotic readers than human ones - but if I implemented a page-hit-count meter it would indiscriminately count them all. Friends have suggested that I place a hit count meter on the "home" page to impress visitors with the magnitude of the company they keep. Unfortunately more visitors enter the site through other portals than ever refer to the home page - as they follow a link to a specific piece of information. This is a growing phenomenon on our web site as an increasing number of university courses employ selected direct-linked pages as mandatory course reading.

I am pleased that some of our material is deemed useful in university courses. But counting these genuine human readers is also a misleading performance metric - for us - as our objective for this knowledge repository is to affect they way people approach the solution of real problems in business.

There are more examples of misleading metrics here, but the point should be clear. Trying to measure the value of knowledge management activities is bound to distort the process as people evaluated on these metrics begin to manage the metrics. It is a simple and seductive matter for a webmaster to run up a page meter count, for instance, and it can even be done quasi-legitimately.

Better to design a KM practice that "pays as it goes" - meaning the acquisition/creation/cataloging activities have a direct and (relatively) immediate pay-off for the people and organizations involved. Creating a priesthood of knowledge managers will undoubtedly get lots of knowledge managed, but what we really want is to see useful knowledge applied effectively. If some KM activities require dedicated staff-work, the useful approach is to rotate real people through these positions. And the real measure of usefulness will be in their subsequent productivity and innovation when they return to line work.

This web-site knowledge repository pays as it goes. In the extreme, generating its content is reward enough. It is our way of attending to our own personal continual learning needs - all of which is eventually applied by us. Thank you for the opportunity to work through this line of thought.

========= Reply =========================