Essay

The MVP Definition Is Broken

2 min read
The Origin

The Rule That Used to Work

The MVP was invented to save you from yourself. The rule was simple — do not build too much, ship the minimum, and learn fast. Building was slow and expensive, so cutting scope cut risk. The idea was to lose six weeks testing a bad idea, not six months building the wrong product.

That was solid math until AI showed up and made building almost free.

The New Reality

The Gap Nobody Warned You About

Your MVP now takes days, and the whole post-ship checklist that used to keep your team busy while feedback trickled in gets done by Tuesday. But your users still take six weeks to tell you what they think. That window between done-building and received-feedback — that gap — is something nobody warned you about.

So you are stuck with the product live, the inbox empty, and the team staring at the ceiling — and inside that gap, you only have two options and both of them are bad.

  • You React Too Early

    The Panic Trap

    If you act on every signal that trickles in, you are not learning — you are panicking, and one unhappy user ends up rewriting your roadmap.

  • You Wait Too Long

    The Idle Trap

    If you wait for enough data, your team goes idle, and idle teams do not stay quiet. An empty mind is the devil's workshop. They start inventing problems and solving feedback nobody gave.

Historical Parallel

Marketers Solved This Years Ago

This is not a new problem. Marketers solved it years ago when digital media made content creation almost free but audience attention still took time to earn.

They never send one message and wait. They send three, with the same budget, testing different bets, and they let the market pick the winner.

You can do exactly that with your product, but you stagger them.

The Strategy

Ship MVPs, Not MVP

You never have just one assumption — you have several, each worth testing. So build for all of them.

  1. Ship MVP 1 — Start Collecting

    You ship MVP 1 and start collecting feedback. While it bakes, you build MVP 2 testing a completely different assumption.

  2. Ship MVP 2 — Keep Moving

    While MVP 2 ships, MVP 3 is already done. Run them as variants, the same way you would run an A/B test, and let real usage tell you which one wins.

  3. Never Be in the Gap

    You are never in the gap because you are always moving. If assumption 1 is wrong, assumption 2 is already live and assumption 3 is already built.

One rule: each MVP must test a different assumption, not a different button color — a different customer, a different problem, or a different promise.

The Real Value

AI Compressed the Cost of Being Wrong

This is the real value of AI — not speed, but recovery. Earlier, a failed assumption meant six weeks of waiting for feedback plus another six weeks of rebuilding. Now your next bet is already live before the first one fails. AI did not just make building faster, it compressed the cost of being wrong. And when the cost of being wrong goes down, you can afford to make more bets.

The old MVP is dead. It was built for the old world. That world is gone.

Ship more, wait less, and bet often. Build MVPs, not MVP.

Pankaj Sarda

Engineering Leader

Building systems in bits, atoms, and books. Bridging high-scale software infrastructure and real-world operations.

Read full profile