Solve the Right Problem
You should constantly move from idea to idea when you're working on a new startup... until you get to something that is scalable. Then stop the ideation and focus on execution.
Rarely will the first idea be the one that should be scaled. My friend Scott Clark (BuzzMaven) talks a lot about adjacent possibilities as a way of exploring new ideas which I think applies here. You need to find a good solid idea to start the process and then riff on it. This takes time so you should make sure you have the runway to go through this process.
I've been "riffing" on my new startup Punndit for about a year and a half now. The original idea has matured and we now have a much more solid understanding of where we can create real value and build a business. That was unclear 18 months ago. We went from what would have been solving a local maximum to what we now think could be a global maximum. Not sure you can get there without going through that process.
Your original idea is rarely the best answer so being comfortable with an iterative design process is critical to solving the right problem.