The assumption you are testing
Write the belief that would end the business if wrong: that a segment has the problem, will pay, will switch from what they use today or that suppliers will join. The MVP exists to test that belief.
Guide · Minimum viable product
A minimum viable product is an experiment with real users, built to answer the question most likely to sink the business. This guide shows how to choose that question, scope the smallest product that answers it and run the learning loop that follows.
Riskiest assumption named before scope
One complete journey, not many partial ones
Success and failure criteria agreed up front
Before you build
Write the belief that would end the business if wrong: that a segment has the problem, will pay, will switch from what they use today or that suppliers will join. The MVP exists to test that belief.
Some assumptions can be tested with a landing page, a concierge service run by hand or a clickable prototype. Build software when the question needs real usage over time.
Choose one user type and one end-to-end path, from first visit to the moment they receive value. Everything outside that path is a candidate for manual work or later releases.
Agree before launch which numbers or behaviours mean continue, change direction or stop. Deciding after seeing results invites rationalisation.
Security of personal data, payments handled through a proper provider, accessibility basics and legal terms are not optional just because the product is small.
Structure
The minimal screens and logic for one user to reach value: sign-up, the main action, the outcome and a way to come back.
Admin screens, spreadsheets or team members doing work by hand behind the scenes, so automation is built only once the process is proven.
Event tracking for each step of the journey, a simple dashboard and a way to talk to users, such as in-app feedback or scheduled interviews.
Authentication, payments, email, analytics and hosting bought rather than built, keeping engineering effort on what is unique.
A sensible data model, version control, automated deployment and basic tests, so a successful MVP can evolve rather than be thrown away.
Planning sequence
Talk to potential users about how they handle the problem today and what it costs them. Look for behaviour and spending, not compliments about the idea.
State the assumption, the metric that tests it and the threshold that counts as a pass. Keep it to one page that everyone signs off.
List every feature idea, then remove anything not needed for the chosen user to complete the journey once. Move the rest to a later list.
Line up the people who will try it at launch, ideally from the target segment, so early data reflects real demand rather than friends being polite.
At the agreed checkpoint, compare results with the threshold, read the qualitative feedback and decide whether to continue, adjust the proposition or stop.
Mistakes
Teams add feature after feature to feel ready, and the experiment becomes a full product launched late. Each addition delays the evidence you need.
A prototype that cannot take real data, payments or repeat use cannot answer questions about retention or willingness to pay.
Launching without event tracking or a plan to talk to users leaves you with opinions instead of evidence.
Feedback from colleagues, investors or a broad audience can point in the wrong direction. Test with the specific segment the business depends on.
The term is often used for a cheap first version, but its useful meaning is narrower: the smallest product that lets you learn whether a key assumption holds. Its output is evidence, and the software is the instrument that produces it.
Framing it this way changes decisions. Features are judged by whether they help answer the question, and launch is judged by whether you learned something, not by how polished the product looked.
List the beliefs your business depends on: the problem exists, this segment feels it, they will pay this much, they will switch, suppliers will join, unit costs work. Rank them by how uncertain they are and how fatal being wrong would be.
The top one is your MVP’s target. Often it is about demand or willingness to pay rather than technical feasibility. If feasibility is the real risk, a technical spike may come before any customer-facing product.
Draw the path from discovering the product to receiving value. Each step needs to work, but each can be crude. Matching can be done by a person, reports can be emailed manually and admin can live in a spreadsheet.
This concierge approach keeps engineering focused on what users touch. It also teaches you how the work is really done before you automate it, which avoids building the wrong workflow into code.
Instrument every step in the journey so you can see where people drop out. Combine numbers with conversation: a short call with users who left early explains more than any funnel chart.
Set thresholds in advance. For example, decide what share of trial users must return in their second week for the idea to be worth continuing. Without a threshold, any result can be read as encouraging.
Small does not mean careless. Real users trust you with their data, and a breach or lost payment can end the business faster than a missing feature. Use proven services for authentication and payments, encrypt data in transit and write basic terms and a privacy notice.
Accessibility basics such as readable contrast, labelled forms and keyboard navigation cost little at the start and exclude fewer early adopters. Retrofitting them later is more expensive.
Most MVPs produce a mixed answer: some users love it, most drift away. Look for the segment that behaves differently and ask why. Narrowing to that group is a common next step.
When the core assumption holds, move to the next one on your list and invest in the foundation: tests, monitoring, data model fixes and automating the manual steps that now limit growth. Effort grows with the number of user roles, integrations and compliance needs you add.
FAQ
Minimal enough that you can launch and learn quickly, but complete enough that a real user can get real value from it. If users cannot finish the main journey, you learn about your bugs rather than your market.
No. A prototype tests usability or appeal with simulated content; an MVP is used for real, with real data and often real money, so it can test behaviour such as repeat use and payment.
No-code and low-code tools are a good fit when the journey is standard and speed matters most. They become limiting with complex logic, heavy data or strict security needs. Plan how you would migrate if the experiment succeeds.
Those tied to your hypothesis: activation (reaching value for the first time), retention (coming back), conversion to paid, or supply-side sign-ups. Avoid vanity numbers such as page views that rise without proving anything.
If the evidence supports the hypothesis, the next step is usually to strengthen the foundation, automate manual steps and test the next riskiest assumption. If it does not, change the proposition or audience and test again.
Next
Where to go when the hypothesis is set and you need the product built.
Open registerStarting an MVP from a proven model when the risk is market, not mechanics.
Open registerAdding someone to own discovery, prioritisation and the learning loop.
Open registerHow larger organisations run MVPs and pilots inside existing businesses.
Open registerWhy most early products ship faster as one deployable unit, and the signals that justify splitting later.
Open registerPrimary sources
Dated official documentation, standards, and research that support the factual claims on this page.
How to read this page
Scope, cost and timeline figures are planning guidance until confirmed in a written proposal. Diagrams and screens are illustrations, not client deliverables, and source-code rights and handover follow the signed agreement. See the delivery process and the full disclaimer.
Next decision
Define outcomes, constraints, evidence, rights and handover before delivery begins.
Commercial rights, repositories, environments, documentation, acceptance and handover remain contract-defined.