MVP and Validated Learning

A chef dreams of opening a restaurant. Location, interior, forty-seat dining room, a full menu — two years of savings, one enormous bet. Instead, she parks a food truck outside an office district for a month. Same signature dishes, a fraction of the cost. The truck isn't her dream. The truck is a question: will strangers pay for this food, again and again? Four weeks later she has an answer no business plan could have given her.

By the end of this video, you'll be able to define an MVP the way its inventor meant it, spot the two ways people get it wrong, and name the thing an MVP actually produces — which isn't revenue.

MVP stands for minimum viable product, and the definition that matters comes from Eric Ries, who coined the modern usage: the MVP is the smallest version of an idea that lets you learn the most, for the least effort.

Read that carefully, because the goal in that sentence is not shipping. It's learning. An MVP is last lesson's slicing taken to its logical extreme — the thinnest slice you'd build not primarily to deliver value, but to answer a question. Will they use it? Will they pay? Will they care? Will they come back?

The food truck is exactly this. So is a one-page website describing a course before the course exists, counting who clicks "notify me." So is a pilot workshop for eight people before building a full curriculum for eight hundred. Each one is a question wearing work clothes.

And what an MVP produces has a name: validated learning — evidence from real behavior that your idea works, or doesn't. Not opinions from a survey. Not applause from a focus group. People doing things: ordering, clicking, returning, paying.

The word "validated" is doing the work. Everyone has beliefs; a plan is a stack of them. Validated learning is a belief that reality has countersigned. Notice how this closes a loop you've been building all module: Lesson 1 taught you that outputs are hypotheses — "we believe X moves Y." The MVP is the cheapest possible way to test exactly that sentence. And it's Module 2's empiricism — decide from observation, not prediction — pointed at the riskiest belief you hold.

Now the two failure modes, and both hide inside the name.

Failure one ignores the M: building too much. If your "MVP" takes a year, it's not a minimum anything — it's the restaurant with extra steps.

Failure two ignores the V: shipping junk and calling it lean. If the food truck serves cold, sloppy food, the chef learns nothing about her restaurant idea — people rejected the execution, not the concept. The test only works if the slice is genuinely good at the one thing it does. So hold the guardrail as a slogan: minimum in scope, viable in quality. Small, and excellent at its small job.

You might think MVP means "version one, but cheaper" — the real product, just worse, shipped early to save money. That's understandable; it's how the term gets used in half the meetings where it appears. Actually, an MVP isn't an early version of the answer. It's an instrument for checking the question. Some MVPs — the landing page, the pilot workshop — are thrown away entirely once they've taught what they came to teach. The product was never the point. The learning was.

Three things to keep. One: an MVP is the smallest version that maximizes learning for the least effort — a question wearing work clothes. Two: what it produces is validated learning — beliefs countersigned by real behavior. Three: minimum in scope, viable in quality — small, and excellent at its small job.

Pause and look at what Module 3 gave you: outcomes as the destination, stories as the language, order as the strategy, the MVP as the probe. One question remains, and it's the whole next module: once work starts… why does everything take so long? Module 4 has the answer, and it's sixty years old.