What an MVP Actually Is
A minimum viable product is the smallest version of a product that delivers real value to real users and lets you start learning and earning from it immediately. Not a cheaper draft of the full thing — the smallest complete thing that does one real job well.
That word “viable” is doing all the work, and it’s the word everyone drops. An MVP isn’t a half-built product you’re embarrassed by. It’s a finished, narrow one. It does less, not worse.
When an MVP actually matters
It matters the moment you’re tempted to build the whole thing before anyone has paid you a penny or used a single feature.
The mistake an MVP exists to prevent is the most expensive one in software: building the impressive, complete version of something nobody wanted in the shape you built it. Every feature you ship before you have a real user is a bet placed blind. The MVP move is to place the smallest bet that still tells you something true — ship the one thing your product genuinely has to do, get it in front of real people, and let what they actually do correct your aim before you’ve spent the budget.
So it matters most when the stakes of being wrong are high and your certainty is low — which, for anything new, is always. It matters less when you already know exactly what’s needed because you’ve built it ten times before and you’re just executing. Even then I’d usually still start narrow, because “I’m certain” is the sentence that precedes most wasted budgets.
The part most people get wrong
The MVP is not a cost-cutting exercise. It’s a sequencing exercise. You’re not building a worse product to save money — you’re building the right slice first so the money you do spend is spent on something proven.
Here’s how I actually think about it, and it’s where the worldview comes in. Take whatever you were about to spend renting a SaaS tool to solve this, or building the full custom version. The lean MVP — the 20% that does your actual job — is usually a few weeks of focused work, and it very often costs about what year one of the alternative would have. Year one washes. The difference is everything after. With the MVP you own, year two starts with the build already banked and you compound: you extend it into the next thing you need, connect it to the system next door, shorten your ROI every month. Build it fast, bank the money early, and grow from the MVP instead of starting over each year. That’s the whole game.
When it’s overhyped
“MVP” has been worn smooth by overuse, and there are two ways it goes wrong.
One: people use “MVP” as cover for shipping something broken. A product that crashes isn’t an MVP, it’s a bug. Viable means it works at the one thing it claims to do. Cut scope, never quality, inside the scope you kept.
Two: people MVP the wrong layer. They build a flimsy throwaway prototype and then have to rebuild it properly the moment it works — so they pay twice. A good MVP is small but real: built to own and extend, not to bin. The slice is narrow; the foundation under it is sound.
Where to go next
This is one of the terms I explain plainly, because the misunderstanding of it costs people real money. It’s tied directly to build vs buy — the MVP is usually the “build” answer done sanely — and to technical debt, because a lean MVP is how you avoid the worst debt of all: a big system built for needs you guessed at.
If you’ve decided the answer is to build, how I’d build the MVP is the practical next step. And if you just want a straight read on whether your idea wants an MVP or something else entirely, tell me about it — that’s a short conversation worth having before you spend anything.