This is a follow up to Shane's early question on site promotion. In one respect, things are going fairly well. Our traffic is steadily increasing: from an early number of around 2500 visits/day, we are now at 3200 visits/day. We also have nearly 1500 registered users (part of this is no doubt some reddit love we got for a post on functional data structures).
And yet, I think our numbers overestimate the true penetration of the site within the theoryCS community at large. Here's an experiment I did:
Estimate the fraction of users with a reputation in the range $(1, 101) \cup > [102, \infty)$, and compare this to the fraction of MO users in the range $(1, \infty)$
The rationale for this is that merely signing up gets you a reputation score of 1, or 101 if you come in via the SE system. So that score doesn't reflect any actual involvement. MO on the other hand is an SE 1.0 site, and so they don't have the 101 bonus.
While in one respect this is an unfair comparison, because MO has been around a lot longer, it's not entirely unfair because we're only comparing ratios here.
TCS: roughly 20%
MO: roughly 40%
(I did a crude approximation by looking at number of pages of users, because I have a life :))
It seems to me that this might suggest that our relatively large user numbers (relative to this stage in the beta) is an illusion, and so we need to work more at getting people in the community to participate. I know for a fact that many of my colleagues even in the narrow realm of computational geometry aren't here.
I don't have too many concrete ideas apart from some of the suggestions from the last post. But maybe it's a matter of doing more in-community - mentioning the site in talks, citing the site generously when appropriate, and encouraging co-workers and students to post questions (note that I'm not mentioning things like twitter which only touch a small fraction of the theory community)