Skip to content
All guides

· 7 min read

PoC vs prototype vs MVP: which should you build first?

PoC vs prototype vs MVP: what each one proves, who it is for, what it costs in time, and a simple way to decide which one your startup should build first.

Blueprint-style cover reading: PoC vs prototype vs MVP

A proof of concept, a prototype and an MVP are three different things that answer three different questions. Founders often use the words as if they meant the same, and then pay for the wrong one: a full build when a two-week test would have answered the question, or a demo when they needed paying users.

In short: a proof of concept (PoC) answers "can this be built?". A prototype answers "is this the right design?". A minimum viable product (MVP) answers "will people use it and pay for it?". Build the one that answers the question you are least sure about. Most products need a prototype and then an MVP, and only some need a proof of concept first.

PoC vs prototype vs MVP: side by side

Proof of conceptPrototypeMVP
The question it answersCan this be built?Is this the right design?Will people use it and pay?
What it isA rough technical test of the one risky partScreens you can click through, with nothing behind themA small, real product that works end to end
Who uses itYour own teamYou, a few future users, investorsReal users
Real data and paymentsNoNoYes
Typical timeDays to a couple of weeksTwo weeksEight to twelve weeks, design included
What you keep afterwardsAn answer, and notesThe designsThe product, the code and the users
What it cannot tell youWhether anyone wants itWhether it can be built, or whether people will payNothing it was scoped to test

The prototype and MVP times are from our own process. A proof of concept takes as long as the doubt it is testing, and a larger or more regulated product takes longer at every stage.

What a proof of concept is for

A proof of concept tests one technical doubt, as cheaply as possible. It has no design, no sign-in and no polish, and nobody outside the team sees it. When it has given its answer, most of it is thrown away.

You need one when the product depends on something that might not work:

  • A technique that may not be accurate enough. Reading data out of scanned documents, matching people to each other, or an AI feature that has to be right most of the time.
  • An integration you do not control. Another company's system that your product cannot live without, and whose access rules you have not tested.
  • A hardware or device limit. Bluetooth devices, background location or offline use on older phones.
  • Speed or volume. A calculation that has to finish in a second, or a report over a very large set of records.

You do not need one when the product is made of things that are known to work: accounts, lists, forms, search, payments, chat and maps. Most marketplaces, booking apps, stores and internal tools are in this group. Their risk is not whether they can be built. It is whether anyone wants them.

Write down, before the test starts, what result counts as a pass. "The parser reads the totals correctly on 9 of 10 sample files" is a pass mark. "It seems to work" is not.

What a prototype is for

A prototype is the product's screens, linked together so that you can tap through them on a phone as if it were real. Nothing is saved and nothing is sent. It exists to test the design before anyone pays for code.

A prototype is good for three things:

  • Finding the confusing parts. Put it in front of five people who could be users and watch where they stop. Changing a drawing takes an hour. Changing built software takes days.
  • Agreeing what is being built. A feature list means different things to different people. A screen does not.
  • Showing investors and early customers. It makes the idea concrete in a meeting, long before there is a product.

A prototype cannot tell you whether people will pay. People are polite about a demo. Only a real product, with a real price, tests that.

In our own process the prototype is weeks two and three: every screen designed for phone and laptop and linked into something you can click through, before the build starts. It is a step inside the MVP, not a separate project. Our week-by-week guide to how long it takes to build an MVP shows where it sits.

What an MVP is for

An MVP is the smallest version of the product that a real user can sign up to, use for its main job and, if you charge, pay for. It is real software on real servers. It is small because it does one job, not because it is unfinished.

An MVP tests the things the other two cannot: whether people come back, whether they pay, and what they ask for next. That is why it is the only one of the three that ends with users.

The usual mistake is to make it too big. Every feature added "just in case" delays the day you learn something. Our guide on how to scope an MVP covers how to choose the first user and the one job, and what to leave for version two.

Which one should you build first?

Ask which of these you are least sure about.

  1. "I am not sure it can be built." Start with a proof of concept of that one part. Then go on to the prototype and the MVP.
  2. "I am not sure the design is right," or "I need to show it to investors next month." Start with a prototype.
  3. "I am not sure people will use it or pay." This is most products. Design a prototype, then build the MVP.
  4. "I am not sure there is any demand at all." Before any of the three, test demand with a landing page, a waiting list or a no-code version, and talk to the people who sign up.

Common paths

  • A marketplace, booking app or store. No proof of concept. Prototype, then MVP. The parts are known. The risk is demand and trust.
  • A product built around an AI feature or document processing. A short proof of concept on real sample data first, with a pass mark agreed in advance. Then prototype and MVP.
  • A product that depends on another company's system. A proof of concept of the connection first. If access is refused or limited, you have found out in a week instead of in month three.
  • An internal tool replacing spreadsheets. Often a prototype is enough to agree the screens with the people who will use it, and then the tool is built in one go.

Mistakes to avoid

  • Calling a prototype an MVP. If nobody can sign up and nothing is saved, it is a prototype. Tell investors and customers which one they are looking at.
  • Growing a proof of concept into the product. It was written to answer a question quickly, with none of the care a product needs. Keep what you learned and build the product properly.
  • Building a proof of concept for something that is not in doubt. Sign-in and payments work. Testing them proves nothing.
  • Skipping the prototype to save two weeks. The design questions do not go away. They come back during the build, where each one costs more.
  • Treating the MVP as the whole product. If the first version takes six months, it is not testing an idea any more.

Questions founders ask

Is a proof of concept the same as a prototype?

No. A proof of concept tests whether a technical part works, and usually has no screens worth showing. A prototype shows what the product will look and feel like, and has no working parts behind the screens.

Can a prototype become the MVP?

The designs can. The screens you approved in the prototype are the ones that get built, so none of that work is lost. A clickable prototype has no code behind it, so there is nothing to "upgrade": the MVP is built from the designs.

Do investors want a prototype or an MVP?

It depends on the stage. A prototype can be enough for a first conversation, because it shows what you intend to build. Investors who want evidence that people use the product will ask for numbers, and only an MVP gives you those.

How much smaller is an MVP than the full product?

Small enough to launch in weeks, large enough to do its one job properly. A useful test: remove a feature, and ask whether the first user could still get the job done. If yes, it can wait.

What comes after the MVP?

Real use. The first month after launch tells you which of your guesses were right, and that decides version two. Nothing on the version-two list should be built until then.

Where we fit

We build MVPs: a blueprint week, a clickable prototype in weeks two and three, then the build, live in 8–12 weeks for a fixed price. If the product has a technical doubt, say so in the first call: that part should be tested before anyone quotes for the rest. The details are on our MVP development page, and this post is part of our MVP guide.

If you are not sure which of the three you need, tell us what you are building. We will tell you which question we would answer first.

  • MVP
  • Startups
  • Product strategy

Building a first version of your product?

We scope it in a week and launch it in 8–12, for a fixed price. Tell us what you want to build.