In my experience it's best to just get something out there as soon as it's interesting to play with. It'll get content into your system, it'll point out problems and possibilities that haven't occurred to you, and there's nothing like real users to motivate you to focus on getting the most important things done as soon as possible. I've worked at a number of startups that wasted a lot of time working on features that turned out not to matter at all, because they thought they couldn't release without them.
I sometimes hear this idea that you only get once chance to acquire users, that they'll show up, take a look, and if your product isn't good enough, they'll never come back. I think that's just untrue. First off, the chances that your product will cause enough of a stir to reach your entire potential userbase when it's not very good yet seem pretty low to me. And second, there are plenty of examples of companies that had to completely reinvent their product more than once before they found success -- off the top of my head, Flickr is one of those. I worked for another company that did that, and I didn't notice any problems getting people to look at version 2.
I sometimes hear this idea that you only get once chance to acquire users, that they'll show up, take a look, and if your product isn't good enough, they'll never come back. I think that's just untrue. First off, the chances that your product will cause enough of a stir to reach your entire potential userbase when it's not very good yet seem pretty low to me. And second, there are plenty of examples of companies that had to completely reinvent their product more than once before they found success -- off the top of my head, Flickr is one of those. I worked for another company that did that, and I didn't notice any problems getting people to look at version 2.