The smallest build that answers a real question — and why most so-called MVPs answer none.
A minimum viable product is the smallest version of a product that can test the assumption the business is betting on. It is not a cheap version of the full product and not a prototype. The distinguishing question is: what would we have to see for this to be worth continuing, and what is the least we can build to find out?
Most MVPs fail in one of two directions. Either they become a full product with the word minimum attached, taking a year and testing nothing; or they are so thin that nobody can use them and the result tells you nothing about whether people would. Both failures come from skipping the first question — which assumption is this testing?
What has to be true for this to work? Usually it is about demand or willingness to pay, not about whether the software can be built.
What would you need to observe to continue? Decide before building, because afterwards the bar moves.
One user type, one path through the product, end to end. Breadth is what makes an MVP late.
Minimal in scope, not in quality. A crashing MVP tests whether people tolerate crashes, not whether they want the product.
Against the signal defined up front. An MVP that ships and is simply extended has skipped the decision it existed for.
A business plans an MVP for a field service app. The initial list has 34 features. Everyone agrees to cut it down, and after two workshops it has 29 — because every feature has an advocate and no one wants to be the person who removed theirs.
A better cut asks a harder question: what single workflow, end to end, proves this is worth building? The answer is that a technician receives a job, completes it on site, captures a signature, and the office sees it as done. That is perhaps six features. Scheduling optimisation, parts ordering, customer notifications, reporting and the other twenty-three are all plausible and none of them tests the premise.
Built as six features, the app ships in weeks and immediately teaches things no workshop would have: technicians need the customer's phone number more than the job description, signatures fail in sunlight, and the office wants a photo rather than a tick. Two of those were not on the list of 34 at all. Built as 29 features, it ships in months and teaches the same things later, after the effort has been spent in the wrong places.
| Term | What it means |
|---|---|
| Minimum viable product | The smallest build that delivers real value and produces real learning. |
| Scope creep | Additions that turn an MVP into a full product before anything has been validated. |
| Validated learning | Evidence from actual usage, as distinct from opinions gathered in a meeting. |
| Technical debt | Shortcuts taken deliberately for speed. Acceptable when recorded; corrosive when forgotten. |
| Pivot | Changing direction based on what the MVP revealed. The outcome an MVP exists to make affordable. |
Every stakeholder's must-have arrives, and the MVP becomes version one by another name.
Without one, any outcome can be read as encouraging, and the project continues by default.
Users judge a broken product as a bad product, not as an early one.
One segment, done well, answers the question. Three segments triples the work and blurs the result.
| Metric | What it tells you | What distorts it |
|---|---|---|
| Features in MVP scope | Smaller is usually better. A list that only grows during planning is a warning sign. | |
| Time to first real user | The clock that matters, because learning does not start until someone uses it. | |
| Usage of each shipped feature | The honest verdict on what was worth building. Routinely surprising. | |
| Assumptions tested | Whether the build is producing evidence or just producing software. |
We build and support the systems this page describes. Start with a free working proof of concept — no cost, no commitment.