you might push them away -> This is a testable claim about empirical reality.

My feeling is that you need to get a user to invest a bit of time -> This is also a testable claim about empirical reality. I actually come down totally differently on this one: I think getting a user to "success" quickly trumps all other considerations for retention, even if getting them to "success" quickly causes their first use of your service to be much shorter on average than it would otherwise have been.

But, hey, what is the point about making testable claims about empirical reality if you don't actually test the suckers. So I'm going to dig into my data this weekend and figure out which one of us is right, at least for my app/users, and blog about it. Please mock me mercilessly if I don't have this done by Sunday.