I would say that a lot depends upon the type of product that you're building. For example, if it's a web-based social networking app, send your beta announcement mail to, say, 10 people or 50 people and track views and signups for 2-3 weeks. See what kind of network effect you get. Pick your initial invitees wisely. If you want users, then you want them to be the so-called influencers.

I also have this question. I am a little reticent, though, because I'm concerned about scalability. In my case, I am writing the front end in typical database-backed fashion, but the backend is several custom C++ components. So while I want users, I'm not sure how much I want them on day one. Everything I read says, "get users, figure out your weaknesses, and scale as needed." That's reassuring. Nevertheless, the unknowns that we face can still be daunting.

Larger companies and people who have been around the block will often use a PR firm or will simply file a press release themselves. It costs about $1000, from what I understand.

I think from the standpoint of your average bootstrapper, guerilla tactics are in order. I plan on using Google Analytics. I haven't tried it yet, but I plan to. It will, for example, show hits based on geographic location. You can use that to target local advertising in the markets where you're getting some initial traction. I don't think there is really a set approach. When you're a guerilla, anything is possible.