I'll tell you what not to do: spend $10-30,000 on letting a MVP company build and support your MVP for you. (Enterprise variant: spending $x00,000's on a proof-of-concept or "MVP" of some product without any consumer data or analytics backing it up.) Fool's errand. I haven't met a company that succeeded this way. (I never used such a company, but I've known one or two.)

MVPs should be "easy", dirt cheap and highly accessible. If shit breaks because too many people want in, then that might be an idea worth building on. If the app is too hard to build and isn't getting traction quickly, it might be too hard or the wrong thing.