It's sort of a funnel: people you think want it > people who actually want it for free > people who can see themselves wanting it if someone else pays (i.e. future them could pay) > people who are actually willing to pay themselves.
I realized you need to ask a follow-up question. Would you use or pay for product X? If yes, then why are you not using similar services Y and Z that already exist?
If it's truly a good idea, then they already know about Y and Z, because they've been searching for a solution to their problem. They might be paying for one of those existing inferior services too, because they need this problem solved, and they're the only options available. That would be great news. It's bad news when they don't know anything about the existing services.
As a simple example, you could ask pet owners if they would use a social network for pets, where they register their pet, post entertaining photos of their pet, follow other pets, etc. Basically, Facebook where all the profiles and posts are from animals.
You'll get a bunch of yes answers from pet owners, which sounds very promising, and you might start on development. As I said, this is a mistake. If you search on Google, you'll find lots of existing pet social networks. So, ask those same people, if they said yes, why don't they visit Google today, and register on one of those other 10 sites? Why didn't they sign-up on one last month? Usually, once they find out the idea exists (or you develop it), they realize they don't care for, or need the service as much as they originally thought.
As another example, I might want to develop a HN theme with bigger voting icons, larger text and links for improved readability, etc. When I ask myself, I answer yes, it's all positives, and I want and would use this theme. But then I ask, why have I never used any existing HN themes from other users? When did I last search for HN themes, because one probably exists today with some of the features on my list. I then realize that I don't care that much about a theme after all, and I actually prefer running sites as they're originally designed, without third-party themes or add-ons, even if they do add a few beneficial features.
Great way to make this point! So when you ask someone "would you pay for it?" you are really asking "would you use it if someone else paid for it?" Which, really, doesn't tell you very much at all.