What an MVP Is, and What It Isn't
By CodexierPublished 6 min read
Minimum viable product is one of the most used and least agreed-on terms in software. Investors mean one thing, developers another, and founders often mean a smaller version of everything they eventually want. This guide gives a working definition, separates an MVP from a prototype and a beta, explains how small is small enough, and shows the shapes an MVP can take, so you can plan one that actually teaches you something.
The purpose: learning with real users
The reason to build an MVP rather than the full product is uncertainty. You do not yet know whether the problem is painful enough, whether your solution fits, whether people will pay, or what they will do the second time they log in. A finished product answers those questions expensively. An MVP answers them cheaply, but only if real people use it in real situations. Internal demos and friendly feedback do not count.
Not a prototype, not a beta
| Thing | Who uses it | What it answers | Built to last? |
|---|---|---|---|
| Prototype | You, test participants | Is the flow understandable? Does the idea make sense? | No, thrown away |
| MVP | Real users with a real problem | Will they use it, pay and come back? | Core yes, rest no |
| Beta | Early customers of a near-finished product | What is broken or missing before general release? | Yes |
| Full product | Everyone | Can it scale and compete? | Yes |
A clickable prototype is the right tool when the question is about the design or the concept, and it costs a fraction of code; our product design sprint produces exactly that. A beta is what you run when the product exists and you want to catch defects. An MVP sits between them: enough engineering to be trusted, little enough to be cheap to change.
How small is small enough
Start from the one job the user hires the product to do, and build only the path that completes that job. Everything that helps the user after the job is done, or that serves a second kind of user, waits. A practical test: write the user's story from sign-up to result in five sentences. If a feature does not appear in those sentences, it is not in the MVP.
- One user type. A second role doubles the product; add it when the first role pays.
- One core flow, end to end, working well. Not three flows working badly.
- Manual behind the scenes is fine. If you can fulfil something by hand for the first fifty customers, do not automate it yet.
- Standard components for everything that is not your idea: logins, payments, email, admin.
- A way to talk to users built in: a feedback form, a support email, a short call after the first week.
What must still be solid
Minimum does not mean careless. Users judge trust in the first minute, and a product that loses their data or exposes it to others is not viable at any size. The list of things an MVP must get right is short but not negotiable: the core flow works every time; user data is stored securely, backed up and separated between customers; logins and password resets work; if you take payment, receipts and refunds work; and you meet the basics of GDPR, meaning a privacy notice, a way to delete an account and a record of what you process.
What may be rough: the visual design beyond clarity, the settings pages, reporting, integrations you can replace with a manual export, and any feature for a second user type. Roughness there is a sign of discipline, not neglect.
Examples of MVP shapes
Concierge MVP
You deliver the service by hand behind a simple interface. A booking form and a person doing the work. Learns whether the service is wanted before any automation exists.
Single-feature MVP
One flow done properly, such as uploading a document and getting it analysed. No dashboard, no team features, no integrations.
Spreadsheet-plus MVP
The product's logic lives in a spreadsheet or a no-code tool; the code is only the part users touch. Cheap to change while the rules are still moving.
Internal-first MVP
Built for your own team or one pilot customer with a written agreement. Learns in a controlled setting before the public sees it.
When you should not build an MVP at all: if you have not talked to potential customers, a landing page and conversations come first and cost almost nothing. If the idea is a copy of an existing product with a small twist, the question is not whether it can be built but why anyone would switch, and no MVP answers that. And if the budget cannot cover a solid core, build a smaller MVP rather than a fragile bigger one. Our MVP planning blueprint exists to find the smallest viable shape before code is written; what the build then costs is covered in what an MVP costs, and a free 15-minute call is the fastest way to find out which shape fits your idea.
Frequently asked questions
How long should it take to build an MVP?
Weeks, not quarters, once the scope is defined. If the plan says six months, the scope is too big for the questions you are trying to answer. Cutting to one user type and one core flow is what brings the timeline down.
Should an MVP be free for users?
Usually not. Willingness to pay is one of the main things you are trying to learn, and free users answer a different question. A low price, a pilot fee or a paid trial tells you more than a large number of free sign-ups.
Can an MVP be built with no-code tools?
Often, and it is a good way to keep the first version cheap. The limits are custom logic, integrations the tool does not support, and buyers who ask security questions. Many products start no-code and are rebuilt in code once the model is proven.
What happens after the MVP works?
You will know what users actually do, which is rarely what the plan assumed. The next version is built on that knowledge, often replacing parts of the MVP that were deliberately rough. That is the MVP doing its job, not a failure of the first build.
Not sure what the smallest viable version of your idea is?
Describe the job your product does for its first user. In 15 minutes we sketch the MVP shape with you, name what must be solid and what can wait, and say whether a build or more conversations should come first.
Book a free 15-minute call