I had a similar discussion with my friend about this; we're both students, looking to launch web businesses. I'm of the philosophy 'launch a crappy version one

I had a similar discussion with my friend about this; we're both students, looking to launch web businesses. I'm of the philosophy 'launch a crappy version one and iterate' (I read a lot of pg's essays), he's worried that if his site isn't really good when he launches, he'll put off a lot of potential customers. I countered that in the early stages it's simply not going to encounter enough users to put off any significant slice of his market. To be fair to him, though, he needs a critical mass of users to get his site off the ground much more than I do; in such a case, a big 'launch' can be helpful. Has anyone tried the tactic of: crappy beta -> feedback from a few thousand users -> get decent -> big marketing launch? Could that work?