The most expensive mistake in software is building the wrong thing. How to scope an MVP that proves the idea for a fraction of the full-vision budget — without burning the runway.
