How to frame your MVP so you don’t waste time and money

People talk about building an MVP and then end up wasting time and money building something way too big for an experiment.