Many-to-Many: Tagging's power law

A lot of people have considered folksonomies (informal classification systems defined around group consensus) as classification systems and found them wanting. I think even considering folksonomy from that perspective is wrong-headed. Instead, folksonomies might be a good vehicle for observing the uptake of new memes (pivotal ideas) within groups.

Sections: Emerging Practice

Topics: folksonomy

Many-to-Many: Tagging's power law:

Ben Hyde looks at four popular bookmarks at del.icio.us and plots how many times each is tagged with the same word. E.g, BoingBoing is tagged as “blog” 200 times and as “news” 90 times. The curve is that of a classic power law: The most frequently used tags are used waaaay more frequently than lesser-used tags.

Ben stresses that four bookmarks don’t constitute a significant sample, but wouldn’t we expect a folksonomy to assume the shape of a power law distribution?

If you think of tags as representing memes (pivotal ideas we have all agreed on) and URLs as items to be understood within that system of memes, this result is not surprising at all. Folksonomies (see here for a definition) just reflect the group consensus on how things should be classified. With things that are well understood within a group, that consensus tends to be pretty solid.

I think the interesting implications come in thinking how you should act based on this result. If you are trying to get a new idea accepted (say you are a marketer) then you should not be too discouraged if your new meme is fairing poorly initially. The question would be trend in uptake. You would also want to see how adoption was going within potentially more influential groups.

Bud posted this on February 3, 2005

Trackback Pings

TrackBack URL for this entry:
http://thecommunityengine.com/cgi-sys/cgiwrap/fpgibson/thecommunityengine.com/mt/mt-tb.cgi/331

Comments

Post a comment

Name:

Email Address:

URL:
Remember Me? YesNo

Comments: