The general advice is "do things that don't scale" so rather than trying to post about it and get some analytical data, just go talk to 20 people in your target market for an hour each and see what resonates with them, what problems they have etc -- all before building anything (except maybe a simple mockup if you feel it will get your point across better.)

Definitely both but likely starting by talking to your target demographic and then building an MVP if it seems like people would pay money for it.