The concrete and yet elusive goal to waging a successful KM campaign is the quest to determine two things really:
(1) Who created the content you're identifying; and,
(2) Who was it intended for
If your KM system can solve these two mysteries you are well on your way to getting KM right. Because if your colleagues know this they can reconstruct not only topics, formats, and date ranges but the actual motives of the person or teams that created the work and thus how in-step it is with their own tasks, priorities, and personal investment in the material itself.
The other 20 percent? That falls outside the domain of KM central. That's up to the domain leaders. Ultimately though the remaining 20% answers to the role of knowledge contributor. Most of us are members of the content producing and consuming camps. It's a goal no more abstract than garbage-in, garbage-out. Your keynoter is also part of the clean-up crew. This reciprocal arrangement bodes well for tackling the 80/20 rule.
Under any practical arrangement the reach of the KM team does not extend beyond the 80 percent. Designing an experience is one thing. Endorsing and classifying content is quite another. But how does the designer impress upon their content producers the need for direct engagement in the mechanics of sound KM execution.
I put this question to a panel of intranet managers, information architects, and leading practitioners in a 2005 roundtable in Searcher Magazine entitled: Under Budget, on Time, and in Sync: How to Stage Successful Rollouts. One of the participants, Christie Confetti-Higgins of the internal portal team for SunLibrary within Sun Microsystems focused on the breadth versus depth issue that lies at the core of content production (and producer participation in it). Christie focused on the need to shape an enterprise taxonomy around the routines and priorities of her users -- not around the completeness of a subject or discipline it addresses:
There was a fine balance between how many levels of the taxonomy hierarchy to go down before it was no longer valuable to the user. Do you drill down to 10 levels in a competitive taxonomy or will 5 levels do? Some of the questions we asked ourselves were based on the content. Was there a lot of content around a particular area of [industry] competition that would benefit from further categorization? If so, we then investigated that option. We continually looked at how users would find the information. At some point, it is too detailed and then the return diminishes for us in terms of time to develop and the end user in terms of time to now find that information. There is a thing as too much detail :).
Three years since its publication the excessive detail Christie refers to here could be buried in the long tail of your search log results or the tags too specific to be reused in future business efforts. In an SLA conference last fall she positioned it as embedding a client-centric perspective in your KM system. This is an approach that values timely sharing over exhaustive collecting and shows a direct pay-off to producers -- even when their participation falls outside the scope of their performance reviews.