Folksonomies In The Enterprise

In one of my last posts, I talked about how the old world of UDDI taxonomies is being replaced with the new world of Web 2.0 folksonomies - this includes things like ratings and tagging vs. the rigidly structured categorization provided by "old school" technologies like UDDI.

The real question, though, is whether these techniques can be translated to work within the enterprise?  Said another way, will things which work well when a community has 1M+ members as well when a community has 10 or 100 or 1000 members?  Do ratings and tagging need a critical mass of users before they become useful?

When I started writing this post, I had imagined coming up with examples of how tagging and rating just won't work when there are only 10 or 100 people in the community.  But, as I struggled to come up with these examples, it hit me -- a taxonomy is actually just a special case of a tag-based folksonomy.

A taxonomy is a case where the only people ordained to define tags are the "experts" (and, yes, the quotes around experts are on purpose here).  In their infinite wisdom, these experts define very subtle differentiation and very sophisticated relationships among their tags - in an effort to build the perfect taxonomy. In many cases, of course, the "experts" have lost sight of the reason for their taxonomy in the first place - to help the non-experts get their jobs done.  So, given this target audience, "subtle" and "sophisticated" are very poor design goals in reality.

There's nothing to stop an organization by allowing their "experts" to first define the tags they think are appropriate - let them build their shrine to the taxonomy gods - but then allow the non-experts to add additional tags, ratings, etc.  And, most importantly, let the folksonomy evolve over time.  Nothing is worse than a taxonomy that can't be changed after it's created.  The world just isn't a static place.