Guide
What to decide before building an MVP
Most MVPs go wrong before any code is written. Six decisions to make first: the question you are testing, the first users, the core journey, what you will not build, how you will measure it and when you will decide what is next.
A minimum viable product is supposed to be fast and cheap. Most of the time, when an MVP turns out to be neither, the problem started before anyone wrote code: the team never agreed what the first version was for.
These are the decisions worth making, and writing down, before development starts.
1. The question the MVP must answer
An MVP is an experiment. Its job is to answer one question about your market, such as:
- Will restaurant owners pay monthly for table bookings that sync with their POS?
- Will warehouse staff use a phone app instead of paper pick lists?
- Will clinics book through a link instead of calling?
If you cannot write the question in one sentence, the scope will grow to fill the gap. Every feature request can then be judged against it: does this help answer the question? If not, it waits.
2. Who the first users are
"Everyone who runs a small business" is not a first user. "Five independent pharmacies in one city that we can visit in person" is. A narrow first group makes everything easier: you know which devices they use, which language they work in and which problem hurts most.
Name the group, and if possible, name some of the actual people or businesses. If you cannot think of anyone who would try it in the first month, that is useful information too.
3. The one journey that matters
Most products have one journey that proves the idea works. For a booking product it is "customer finds a slot and books it, and the business sees it". Everything else, such as profiles, settings, reports and notifications, supports that journey or comes later.
Draw the journey as five to ten steps. Those steps are your first version. If a step can be done by hand behind the scenes for now, mark it.
4. What you will deliberately not build
The "not now" list is as important as the feature list, because it is what stops scope creep. Common candidates:
- Payments, if you can invoice the first customers manually
- An admin panel, if a simple database view or spreadsheet export will do
- Native apps for every platform, if a web app or one cross-platform app is enough to learn
- Detailed roles and permissions, if the first customers are small teams
- Integrations, if a CSV import covers the first few weeks
Writing these down matters. When someone asks "can we just add…", you can point to an agreed list instead of negotiating every time.
5. How you will know whether it worked
Decide what you will measure before launch, while you are still neutral. Good MVP measures are about behaviour, not opinions:
- How many invited users actually start using it
- How many complete the core journey
- How many come back the next week without a reminder
- Whether anyone pays, or agrees to pay, when asked
Also decide what result would make you stop or change direction. It is much easier to be honest about this before you have spent months on the product.
6. When you will decide what happens next
Set a date for a decision, typically a few weeks after the first real users start. At that point you choose one of three paths: keep improving the same idea, change direction based on what you learned, or stop and keep the money for something else. All three are good outcomes if they are based on evidence.
A one-page summary
Before you talk to developers, put the answers on one page:
- The question the MVP answers
- The first users, as specifically as possible
- The core journey, in steps
- The "not now" list
- What you will measure, and what result would change your plans
- The decision date
That page is also the best input for an estimate. Our MVP scope planner helps sort features into must-have and later, and the project brief generator turns your answers into a brief you can send to any developer.
- MVP
- Product
Try it yourself
- MVP scope plannerSort your feature list into must, should, could and won’t, size each one, and see what a genuinely minimal first release looks like.
- Project brief generatorAnswer a few questions and get a structured software brief you can share with any developer, including the open questions you still need to answer.
Planning your first release?
Tell us the idea and who it is for. We will help you decide what belongs in version one.
Keep reading
- How to choose a custom software development companyWhat to look for, what to ask and which warning signs to watch for when choosing a company to build your software.
- MVP development cost: what founders should budgetWhat goes into the cost of an MVP, the line items founders forget, and how to plan a budget that leaves room to improve after launch.
- How much does it cost to build an app?The factors that decide what an app costs to build and run, and how to get an estimate you can compare between developers.