Re:Review others websites and they review yours
by terrym on Thursday October 21, 2010 @01:23PM
(#1816489)
I think your intuition is correct - that user contributions generally follow a power law, so when you randomize 1-to-1 interaction among all members, the likelihood of a person in the long-tail getting feedback from another user in the long-tail is high. A good essay on this: http://www.shirky.com/writings/powerlaw_weblog.html
Proving or disproving these assumptions would be pretty important for strategic direction at a QA site, incidentally, since the world if I'm right looks very different from the world if I'm wrong. Since StackOverflow's data set is public, you could do it as a weekend project. In general, though, this is the kind of stupidly powerful thing you can go with the data you already have in your startup.