Instead, try to think of it in a top-down, breadth-first manner. Pick a market you care about, and the best product in that market. Then think for a week how you could improve that product by an order of magnitude. List ideas. Throw them out. List more. When we're talking about disruptive improvement, thinking about it in terms of "orders of magnitude" is hard, so try to think about it in terms of what huge, meaningful thing the users could do that they can't do now. Design a new product that has a fundamental reason to exist (doing the same thing differently is not a reason).
This way, instead of trusting in your idea, you'll have to put trust into the fact that you can design a great product. I found this to be much easier - in this case all you have to do, is do your best. It's hard for anyone to do anything more than that.
1. Friend: "I have an idea,...., but i'm not sure if it's great or crap. "
2: Me: "Don't worry it IS crap."
3. Friend: "Screw you, i'm gonna build it and show your ass that it's not."
He made that app, and sold it to 4 people. It was more of an (offline) app, so couldn't link it here.