Then figure out, in terms of expenses and whatnot, what you'll need to get to demo and beta and beyond (should this prove feasible).

From there, it gets trickier, but if you've built a good product that relieves some need without causing too much additional pain, you'll (possibly/probably) be fine.

Shouldn't data exist in the form of direct competitors? If there aren't any, then does this market exist? (It's entirely possible that it does, and no one but you sees it yet).

Data should also exist in the form of knowing roughly how many people could benefit from your product (and who might consider it a step up). If you're basing this on a market need, then that population should be at least a somewhat knowable quantity.