The central line is, "Does it immediately interest our most likely customers? And if it doesn’t, there’s a problem. And it might not be worth working on it at all, or starting again."<p>But "starting again," isn't exactly "giving up." You don't always know your most likely customers (look at boomers on Facebook), or what your product is going to be (consider Google's humble origins).<p>And just because you're not on a trajectory to be the next Facebook or Google (or even Dogster) doesn't mean you don't have a potent service for an important niche market.<p>At each level of your StartUp's development feedback is important - but feedback is just for steering.
I'm all for the 'parallel tracks' method of choosing what to do.<p>Start any number of projects, make it ones that you feel good about but you are not sure which one is the 'best', or even if there is a winner amongst them. Then try them out in front of real people, on the web. The ones that gain traction fastest are the ones that you should put more time and effort in.<p>In the unlikely case that you have multiple 'winners' you can always divest the runner up and use the proceeds to boost your real winner.