The Short Answer
What are the stages of MVP development?
Six stages: discover the problem, define the hypothesis and success metric, design the core flow and scope, build the smallest working product, launch to a small group of real users, and learn from the results before deciding to persevere, pivot or stop. Most of the risk is removed in the first three stages, before any code is written.
The stages are not a waterfall. Each one exists to answer a question cheaply before the next, more expensive stage begins. Eric Ries defined the MVP as the version of a product that gets the most validated learning with the least effort; the process is how you keep the effort low.
The Six Stages of MVP Development
Discover
Understand the problem and who has it, from conversations rather than assumptions.
Define
Write the hypothesis and the one metric that will prove or disprove it.
Design
Map the core flow, cut scope to the Must Haves, and test the design with users.
Build
Build the smallest production-quality product, with weekly demos.
Launch
Release to a small group of real users and track the metric from day one.
Learn
Compare results with the target, measure fit, and decide what happens next.
| Stage | Question it answers | Output | Go deeper |
|---|---|---|---|
| 1. Discover | Is this a real problem for a reachable group of people? | Problem statement and interview notes | MVP vs prototype vs PoC |
| 2. Define | What exactly are we testing, and how will we know? | Hypothesis, success metric, MVP type | MVP examples |
| 3. Design | What is the smallest flow that tests it? | Core flow, prioritised features, scope document | How to scope an MVP |
| 4. Build | Can we ship it, properly, in weeks? | Working product, repository, handover | How long an MVP takes |
| 5. Launch | Will real users use it? | First users, usage data | MVP launch strategy |
| 6. Learn | Do we keep going? | Fit score and a decision | After the MVP |
The UK government's Service Manual runs a similar sequence for public services: discovery, then alpha (testing prototypes), then beta (building it for real, first for a limited group). Its phases are sized for services used by millions of people; a startup can usually move through the same questions far faster.
An MVP Roadmap, Week by Week
What does an MVP development roadmap look like?
A roadmap for a focused MVP with one core flow can run about ten to twelve weeks end to end: one to two weeks of discovery and definition, one to two weeks of design and scoping, three to six weeks of build, and two to four weeks of launch and learning before the first decision point.
| Weeks | Stage | Milestone at the end |
|---|---|---|
| 1–2 | Discover and define | Hypothesis, metric and MVP type agreed |
| 3–4 | Design | Core flow tested with users; scope document signed off |
| 5–10 | Build | Production release with analytics on the core flow |
| 11–12 | Launch | First cohort of users onboarded personally |
| 13+ | Learn | Fit survey, results against target, persevere/pivot/stop decision |
These are planning figures for a single-flow product, not data. The build length depends on hours and team size; use the calculator in how long it takes to build an MVP to size your own.
MVP Development Strategy: Five Decisions
What is an MVP development strategy?
An MVP development strategy is the set of choices that decide what the MVP tests and how: the hypothesis, the success metric, the type of MVP, the build approach, and who you launch to first. Getting these right matters more than any technical choice.
- The hypothesis. One falsifiable sentence. If the team cannot agree on it, stop and do more discovery.
- The metric. Behaviour, not opinion: payment, repeat use, a referral. Set the target and date before launch.
- The MVP type. The cheapest thing that can test the hypothesis, from a landing page to working software. See seven MVP examples.
- The build approach. No-code, an AI app builder, a freelancer or a studio, depending on what the product must prove and how long the code must last. See MVP development company vs AI app builders.
- The first users. A specific group you can reach personally. See MVP launch strategy.
MVP Development Checklist
Work through it stage by stage. Progress is kept only while the page is open.
Interactive checklist
MVP development checklist, stage by stage
Eighteen checks across the six stages. The bar shows where you are.
0 of 18 done. Current stage: Discover.
Stage 1 · Discover
Stage 2 · Define
Stage 3 · Design
Stage 4 · Build
Stage 5 · Launch
Stage 6 · Learn
Frameworks Worth Borrowing
| Framework | Use it in | What it gives you |
|---|---|---|
| Build–Measure–Learn (Lean Startup) | Define, Learn | The loop: every build is an experiment with a metric |
| MoSCoW (DSDM) | Design | Must, Should, Could, Won’t; Must Haves under about 60% of effort |
| Usability testing with five users (Nielsen Norman Group) | Design | About 85% of usability problems found per round of five |
| Sean Ellis survey (40% test) | Learn | A product-market fit score from active users |
| GOV.UK phases | Whole process | Discovery, alpha, beta, live, with clear exit questions |
Sources: Agile Business Consortium, Nielsen Norman Group, First Round Review, GOV.UK Service Manual.
Where the Process Breaks
- Skipping define. Building before the hypothesis and metric are written down produces a product that cannot fail, and so cannot teach you anything.
- Scope growing in build. The Standish Group found half of software features are hardly ever used. New ideas go on the Won't list, not into the sprint.
- Launching to friends. Polite feedback from people who know you is not data.
- No decision date. Without one, a weak result turns into months of waiting for it to improve.
For the full picture, see the complete guide to MVP development. Loomstrat runs stages three to five as a fixed-price build in 3 to 6 weeks; see MVP development.
Frequently Asked Questions
What are the stages of MVP development?
Discover the problem, define the hypothesis and success metric, design the core flow and scope, build the smallest working product, launch to a small group of real users, and learn from the results before deciding to persevere, pivot or stop.
What is the MVP development process?
A sequence of cheap tests before expensive ones: research the problem, write a testable hypothesis, design and scope the smallest flow that tests it, build it properly, launch it to real users and measure one metric decided in advance.
What does an MVP roadmap include?
Milestones for each stage: hypothesis and metric agreed, scope signed off, production release, first users onboarded, and a decision point after results and a product-market fit survey. For a focused MVP with one core flow, that can be about ten to twelve weeks end to end.
What is an MVP development strategy?
The choices that decide what the MVP tests and how: the hypothesis, the success metric, the type of MVP, the build approach and the first users. These matter more than technical decisions.
What should be on an MVP development checklist?
A one-sentence problem and hypothesis, a success metric with a target and date, a mapped core flow, MoSCoW-sorted features, a signed scope with IP assignment, a deploy pipeline from week one, analytics on the core flow, a small first launch group and a decision date.
Which framework is best for MVP development?
No single framework covers everything. Build–Measure–Learn sets the overall loop, MoSCoW handles scope, five-user usability testing checks the design, and the Sean Ellis survey measures fit after launch.
Treat each stage as a question with a cheap answer, and the expensive stages get smaller, shorter and far more likely to work.