I stumbled across this article today published by SumoMe.
They studied a gazillion opt-in pop-ups to determine what converted people into mailing list subscribers.
There are a lot of variables, but the broad number is around 1.8% of people that visit a site, if presented with a pop-up, will give their e-mail address and sign up for the newsletter.
Now for a community that that requires collection of more data, that number will probably be a little lower. Based upon what we’ve seen, I think closer to 1% (if it’s clearly sign posted and the form optimised) is probably about realistic. But it gives a decent enough number to ask for at this stage (this assumes you’re using a pop-up for newcomers - we turned ours off and saw about a 10% drop in registrations - @HAWK correct me if that figure is wrong).
Now you can certainly collect more emails by putting together a more compelling offer (and the article shares a lot of tips to do this), but then you will get more people signing up and far less participating. We need people to join because they want to participate in a community, not to download more resources.
Given the broad scope and numbers involved, I thought the benchmark was interesting.