Saturday, July 26, 2008

Of Zipf curves and online communities

Over at the Civic and Citizen Journalism Interest Group blog, chairman Jack Rosenberry does some detailed analysis of the Rochester Democrat & Chronicle's online forums and comes away with some numbers:

Among these frequent posters, the top three accounted for nearly a QUARTER of the total postings. Together just three pepole tallied 2,481 postings out of the 10,356, or 24 percent.

Which leaves him wondering:
When we talk about "news as a conversation" does this qualify? Is it a useful exercise for the community just to have the facilities for a conversational commons available, even if they are dominated by a relative handful of users? Is it a useful exercise for the news organization that sponsors it?
It's actually quite a common phenomenon known as a Zipf curve, and it is found not only throughout the Internet, but in many organizational situations. Simply put, it's a manifestation of the Pareto Principle, which essentially says a few people will do most of the work in almost any endeavor.

I don't think this is something to particularly get agitated about. First, in almost any interaction (except elementary school where the teacher steps in), some people will be dominant and some passive. Second, the Zipf curve is so integral to the Internet -- it is the essential basis of Chris Anderson's long tail, for instance -- that I think you have to examine it in that light -- that even if you have dozens or hundreds of people who have made only one or two posts over time, you have accomplished opening up the conversation.

However, I think this heightens the argument for some kind of active moderation as opposed to the largely put it up and leave it alone attitude I see at too many newsrooms. Moderation does not mean you suddenly lose the protection from lawsuits granted to providers of an online forum. Active moderation means you come up with polls and other ideas to encourage people to comment and reassure them it's OK (after all, this is a form of public speaking). And sometimes you literally moderate, suggesting to people it might be time to give someone else a voice.

One of the first systematic studies of this in operation was our first-year report (PDF)on Hartsville Today. I will be doing a two-year review of the site at the NNA/Huck Boyd conference in September. I think this may be the first systematic longitudinal study of such a community news/citizen journalism site. The curve is still there. It is shallower. That may be the best we can hope for.

Labels: , , ,


Post a Comment

Links to this post:

Create a Link

<< Home