Definition of the product MVP
The acronym MVP stands for minimum viable product. In common usage, a product MVP is the version of a new product reduced to what is needed to test a hypothesis with real users, before investing further.
Eric Ries gives a definition in a post of 3 August 2009, “Minimum Viable Product: a guide” (startuplessonslearned.com, read on 12 November 2026): the version of a new product that allows a team to collect, with the least effort, the maximum amount of validated learning about customers. He adds two warnings. Despite its name, the MVP is not about creating minimal products. And the definition is no formula: it requires judgement, depending on the context.
The glossary of the SAFe framework (framework.scaledagile.com, read on the same day) gives the word a meaning of its own: an early and minimal version of a new solution, sufficient to prove or disprove an epic hypothesis. The epic is the portfolio initiative presented in our lean portfolio management entry.
Hypothesis, complete journey, measurement: what makes an MVP
The two definitions agree on one point: the MVP serves to learn. The three elements that follow are our synthesis, not a quotation.
- A hypothesis written down before development. For example: a given audience will adopt a given service in place of a given habit. Without it, there is nothing to prove or disprove.
- A complete journey, even if narrow. The user completes a real task. In a story mapping workshop, the top row of the map describes this smallest system that works end to end.
- A measurement, then a decision. The criterion is set in advance, and the reading leads to continuing, pivoting or stopping.
What goes into the MVP and what waits can be read in the order of the Product Backlog, for which the product owner is accountable in Scrum.
Product MVP, proof of concept, prototype, pilot: not to be confused
These four words are a matter of usage, and their boundaries vary from one organisation to another.
| Object | Question asked | Who uses it | What follows |
|---|---|---|---|
| Proof of concept (POC) | Is it feasible? | The team, outside production | A technical decision |
| Prototype | What would it look like? | Users, in a demonstration | A revised design |
| MVP | Does the usage hypothesis hold? | Real users, for a real task | Continue, pivot or stop |
| Pilot | Does the chosen solution hold up in real conditions? | One site or team, before general rollout | The rollout |
The pilot assumes the solution has already been chosen: it tests its rollout, whereas the MVP tests the value of the product.
Drifts in usage also blur the word.
- The renamed first release. A reduced scope, with no hypothesis or measurement, is a first version, not an MVP.
- The minimum that is no longer viable. Without security, or without the step that completes the journey, the product cannot be entrusted to real users.
- The temporary kept on. Code written quickly to learn becomes technical debt if it stays in service without being reworked.
Example: booking aerial work platforms at an equipment rental company
Take a construction equipment rental company. Its hypothesis: site managers will book from their phone rather than calling the branch. The MVP is limited to aerial work platforms and two branches. The journey is complete, from search to confirmation; behind the scenes, an employee still enters each booking in the branch management tool. A feature flag opens the service only to customers of these two branches.
The criterion is written in advance: the share of aerial platform bookings made through the app, and the number of customers who return to it. Depending on the reading, the company extends the service to other equipment, changes target or stops.
Problem to solve, users, scope of the first version, hypotheses to test: this framing, which precedes development, is presented on our custom software development page.