The MVP and product-market fit

Ask five founders what an MVP is and you'll get five answers, most of them wrong. "It's the cheap first version." "It's the product with fewer features." "It's what we ship before the real thing." All of these miss the point, and building an MVP on the wrong definition wastes exactly the time the MVP was meant to save.

What an MVP actually is

The minimum viable product is the smallest thing you can build to learn whether you're on the path to product-market fit. The keyword is learn. An MVP isn't a smaller product, it's an experiment. Its output isn't revenue, it's validated learning: the maximum amount you can learn about real customers for the minimum amount you have to build.

The concept was coined by Frank Robinson in 2001 as the "right-size product, big enough to cause adoption and sales, but not so big as to be bloated and risky." Eric Ries then made it famous in The Lean Startup (2011), framing it as the fastest way through the build-measure-learn loop.

What an MVP is not

• Not a low-quality product. "Viable" matters. It has to actually deliver the core value to be worth learning from.
• Not a feature-reduced version 1. An MVP is scoped around a hypothesis to test, not a roadmap trimmed for time.
• Not necessarily software. A concierge MVP, delivering the outcome by hand, is often the fastest learning of all.

How the MVP drives toward product-market fit

The MVP is how you climb the fit ladder without betting the company on a single guess:

1. Build the smallest thing that delivers the core value. Enough to be genuinely useful, nothing more.
2. Put it in front of real users and watch behavior, not opinions.
3. Reach problem-solution fit: confirm the MVP solves a real problem for early users who keep coming back.
4. Iterate toward product-market fit, measured by the share of engaged users who would be very disappointed to lose it.

Each loop is cheap because the MVP is small. That's the entire economic argument: the earlier you learn you're wrong, the less it costs to be wrong.

The most common MVP mistake

The classic failure is building an MVP and then never measuring what it taught you, shipping it, feeling productive, and moving on to the next feature. An MVP with no measurement isn't an experiment, it's just a small product. The whole value is in reading the result: does it retain, do users come back, and would they be very disappointed to lose it? That last question is the Sean Ellis test, and it's the cleanest signal your MVP can give you about whether you're on the road to fit.