It's true that the only way to be sure of what people will pay is to see what they pay. But by the time you have an actual product to sell, you will have asked people for advice and feedback about 8 million times. It's very natural at that point to go for 8 million + 1 and say, "What would you pay for this?"
Pricing is one of the few things in business that's traditionally amenable to gathering hard data. Which is why I'm so excited about the Lean Startup approach; there one tries to apply the same data-driven spirit to everything.
The traditional and very successful approach to getting past the brittleness of "What would you pay for this?" is to sell vaporware. It's how Micro-Soft (before the S was lowercase and the hyphen was dropped) and many others got started.
You say to yourself: "I have a pretty good idea of how much work and expense it will take me to build this", then you go off and try to sell it as though it was already done or just getting polished for release. Once you have orders (or even better, cash advance!), you (privately) commit to either finish the product or cancel/refund orders if you fail^Whave a great learning experience.
It's dishonest, but it works, and if you structure your offer fairly (don't induce reliance on the vaporware, only induce desire), it's not evil.
Successful established authors do this before they write books. It's a standard technique in publishing.
It's true that the only way to be sure of what people will pay is to see what they pay. But by the time you have an actual product to sell, you will have asked people for advice and feedback about 8 million times. It's very natural at that point to go for 8 million + 1 and say, "What would you pay for this?"
Pricing is one of the few things in business that's traditionally amenable to gathering hard data. Which is why I'm so excited about the Lean Startup approach; there one tries to apply the same data-driven spirit to everything.