The big questions I've learned to ask is : What happens if the idea doesn't mature to its 100% ? Is it useful/worth doing/helpful to the product and direction at 50% ? 20%? Whats the first version and will it be something we can make use of in itself ? Is the "prototype" a product in itself ?
This isn't an execution issue. Its just practical. From what I've observed in the software world, rarely is an idea executed to its fullest ideal. There are engineering issues, time-to-market issues, other priorities, reactions to real need that create directional changes, etc, that all lead to a smaller subset or version of the idea being the end-product for the longest time possible. This is also a good thing - because it leads to practical, usable R&D rather than early, limited "lab prototypes" that don't really fit anywhere yet, but "we're working on ver 2". It also gives a feature/product a shot at collecting market feedback early, and helps optimize resources on it. Most importantly, it reduces wasted effort on half-bakery stuff that never got released. Sure, we still have a lot of those, but thats where the learning happened ;)