Most MVPs fail because the scope is wrong, not because the team was slow. Here is the structure we use to ship real products in eight weeks — and the tradeoffs worth making.
What an MVP actually is
We use a working definition that has held up across every consumer and B2B product we've shipped:
An MVP is the smallest thing you can put in real users' hands that produces a real, decision-grade signal.
That definition rules out a lot. It rules out feature checklists copied from competitors. It rules out internal demos. It rules out anything you would not be willing to send to a stranger.
The shape of a good 8-week MVP
We scope our MVP Sprint package against this shape:
- Weeks 1–2: Discovery and design. Real conversations with real users. A core flow on paper.
- Weeks 3–6: Build. One or two flows, end to end, on production-grade infrastructure. Auth, database, payments if needed.
- Week 7: Polish, instrumentation, real-data testing.
- Week 8: Launch and the first round of feedback.
What is not in scope: nice-to-haves, admin tooling that would be lovely to have, integrations with everything, custom analytics dashboards. All of those are post-MVP problems if the MVP is ever worth them.
The five tradeoffs worth making
- Choose one user, one flow. If there are two equally important users, you are scoping two MVPs.
- Choose boring foundations. Postgres, Next.js, Stripe. Not "interesting" choices. The MVP is not the place to learn a new stack.
- Choose to launch with a sharp opinion. Vague apps are easy to build and impossible to learn from.
- Choose to skip admin UI. Use the database directly for the first month. You don't have admin users yet.
- Choose to instrument from day one. If you can't see what people do, you'll guess instead of learn.
What you get on the other side
A real product, in front of real users, with a real signal — not a deck about a product that might exist. That signal is the foundation everything else gets built on: roadmap, fundraising, hiring, even pricing.
How we run them
This is exactly the playbook behind our MVP Sprint package. Two operators, eight weeks, fixed scope, production deploy, post-launch support. No open-ended hourly billing.
If you have an MVP you've been circling for months, send us the one paragraph version. We'll come back with a proposed scope you can argue with.