I spent some time talking with Dave Weinstein (hope I spelled that right -- the card is in a pocket somewhere) at GDC and he had a plangent observation: "whenever a professional developer gets a new piece of hardware, s/he creates one of two things for a first test: Pong or Breakout -- they require graphics, player input, detecting collisions, sound can be easily integrated, and they are a complete game".
Start simple. Good advice in almost any endeavor, but particularly in business or programming -- start with the simplest expression of your idea that you can. If it is not compelling, a more comples version may not be either. If you can't build a simple expression of your idea, how are you going to create a larger, more complex one?