How to find a problem for your solution

All developers with entrepreneurial ambitions are guilty of this - we create a great solution with no users, since we were so sure of it's usefulness that we didn't actually do any research. This is my (oh I don't know..) 7th time doing this. And I still don't learn from past mistakes. I get closer to how it's really supposed to be done, but not actually get there.

Also I haven't got an answer for the question just yet, but I do have a plan. Since my product got a lot of people interested in it (but very few paying users), then it's worth getting into further.

So plan B is just having a methodical approach to developing and marketing. Meaning I'm going the hypothesis->experiment->learn way of doing things.

Before I develop I clearly write down why I'm developing it, why I'm hoping it'll work, what criteria means it's successful and then learning from the result. The marketing (which is actually more important at this point) will also go down this route.

My wishful thinking is that at least I'm going to have an eye on whether this product of mine has any chance of rising up from the mud, or it'll forever stay in the ground.

Also I'm reading Running Lean, which is great and I'm getting a lot of ideas from it.