Strange for this to be such an early post in the blog--you'd think epiphanies like this would come later--but it's certainly useful early on.
Solve a problem that affects you.
Look at open source software; the developers are driven by a common problem, not by money (but perhaps recognition). Define success as making your life easier and gaining experience. You'll gain more--in both experience and recognition--through the completion of a profitless startup than by abandoning a potentially profitable one. Yes, a business model is great, but don't focus too much on it. A business model doesn't define a good idea; it's just a criterion for a potentially profitable idea.