Minimum Viable Product (MVP)

Details
Full Name

Minimum Viable Product

Also known as

MVP, Lean Startup MVP

Core Concepts:

Smallest thing that tests a hypothesis

An MVP is the minimum product that allows a team to learn whether a specific hypothesis about user needs is valid — with the least amount of effort.

Validated learning is the output

The defining output is not a feature set, not revenue, not users — it is learning. "Did we learn whether our hypothesis is true?" is the only success criterion.

One hypothesis at a time

An MVP targets a single, falsifiable hypothesis ("users will pay for X", "users need feature Y more than feature Z"). Bundling hypotheses into one MVP makes the learning ambiguous.

Minimum means minimum

If a hand-drawn mockup, a landing page, or a concierge service can test the hypothesis, that is the MVP. Code is often not required. The "P" stands for product, but the product can be an illusion.

Build-measure-learn loop

The MVP is the first turn of a tight feedback cycle: build a minimal thing, measure how real users respond, learn from the data, then decide whether to pivot or persevere.

Not a small v1

A common misuse is calling the first release of a polished product "the MVP". A real MVP would be embarrassing to ship in production — its job is learning, not market entry.

Pivot or persevere

The MVP gives you the evidence to choose: continue in the same direction (persevere) or change course based on what you learned (pivot).

Distinct from Walking Skeleton

A Walking Skeleton validates architecture; an MVP validates market demand. You can have one without the other.

Key Proponents

Eric Ries ("The Lean Startup", 2011); Frank Robinson coined the term in 2001

When to Use:

  • Validating whether a new product or feature solves a real user problem before investing in full development

  • Testing pricing, positioning, or target-segment hypotheses with minimal risk

  • Early-stage startups with limited runway who cannot afford to build the wrong thing

  • Internal product experimentation where feature ideas need evidence before prioritization

  • LLM prompting: "design an MVP for X" signals test the hypothesis, don’t ship a complete product

  • Jobs to be Done - Framework for articulating the user need an MVP is trying to validate

  • Impact Mapping - Technique for aligning MVP scope with business outcomes

  • User Story Mapping - Visual tool for identifying the minimum set of stories for an MVP

  • Walking Skeleton - Architectural counterpart; validates structure rather than market demand

Criticism:

  • The term is diluted in practice: Nielsen Norman Group (Sara Paul, "Minimum Viable Product (MVP): Definition") notes that "'MVP' often means different things to different people" — from paper prototype to released product — and advises defining what you are testing and what you expect to learn before using the word

  • Henrik Kniberg’s widely cited corrective "Making sense of MVP — and why I prefer Earliest Testable/Usable/Lovable" (Crisp, 2016 — the skateboard→car drawing) argues "Minimum" and "Product" mislead teams into shipping useless fragments; each increment should be a smaller complete solution to the user’s need, not a piece of the final one

  • Disambiguation advice: state which sense you mean — Ries’s learning vehicle ("maximum validated learning for least effort") or Kniberg’s Earliest Testable/Usable/Lovable staging — rather than relying on the bare acronym