I follow the Lean Startup Circle group and came across this article:
Minimum Desirable Product
What’s a minimum “test” of your product? And what are you testing?
A hypothesis-driven approach to product development dictates that you build as much as you need to test our your product, but not more and not less. But what are you “testing” your product for?
It seems like there is a lot being written and tried out in this area, which got formally rolling with Crossing the Chasm, but with a lot of focus now on how to get to the chasm in the first place and how to find a really small chasm you can jump on a bike instead of a motorcycle.
Having been through this process a few times, and seen the results of what happens when this kind of process isn't followed, I'm considering coming up with a mantra for startups. Repeat it everyday and you can avoid the stupid costly problems.