I've always been of the opinion that you need a very strong auteur at the head of a company driving product to make something awesome. Tivo, Sling, Flip, Apple, and others are clearly true to themselves and not full of compromise. Posterous, Skitch, Balsamiq, and Google are examples of the same.

But that doesn't mean you don't heed feedback. Apple wasn't against cut-and-paste; they just didn't want to have an offering that wasn't up to their standards. Google tests the hell out of everything.

I still think you have to put your first visionary offering out there - IMVU had their client that Eric had to throw away - and see what happens. Too much CD "work" is just an excuse to not build something and is just the same wasted "effort" as premature optimization.