It took me seven years and two failed startups (one of them through YC) before I finally had a hit (the Ruby on Rails Tutorial). I think part of what made it work was intentionally taking a medium risk for a medium reward. My first two startups were all-or-nothing (which turned out to be nothing), whereas the Rails Tutorial, even had it "failed", would still have produced a five-figure annual income stream. (It succeeded, so it's six.)