A good app can still fail if the business direction is unclear.

Start with the problem.
Then build the smallest version that proves people care.