How to Launch an MVP in 8–12 Weeks: A Practical Guide
A software MVP with a genuinely disciplined scope is realistic in 8–12 weeks and typically costs £30,000–£80,000 in the UK. The timeline depends almost entirely on how ruthlessly the scope is cut to a single core user journey — not on how fast the engineering team works.
An MVP is not a smaller version of your product idea — it is the smallest thing you can put in front of real users that tests your riskiest assumption. Most MVPs that overrun their timeline do so for the same reason: they were scoped as a cut-down full product, not as an experiment. Get the scope right and 8-12 weeks is a realistic, achievable timeline for most software MVPs. Get it wrong and no timeline will save it.
What “minimum” actually means
Minimum does not mean low-quality — a broken MVP tells you nothing useful. It means minimum surface area: the smallest set of features that lets a real user complete the core action your product exists for, end to end, without you having to explain away five things that “aren’t built yet.” Everything else — admin niceties, edge-case handling for scenarios that might not even occur, a second user role you think you’ll need eventually — waits.
Ruthless scope prioritisation
A simple, effective filter for MVP scope: for every feature on your list, ask “if this were missing, could a user still complete the core journey and would we still learn what we need to learn?” If yes, it is not MVP scope. Common categories that fail this test and should be cut for v1:
- Admin dashboards beyond the bare minimum to operate the product manually behind the scenes.
- Account settings and preferences beyond what is legally or functionally required.
- A second or third user role, unless the core journey is genuinely impossible without it.
- Notification systems beyond the one or two that are core to the product working at all.
- Any integration that is a “nice to have” rather than a hard blocker to the core journey.
A realistic 8-12 week breakdown
- Week 1: Discovery — lock the single core user journey, the tech stack, and what “done” means for launch. This is the week most rushed MVPs skip, and the one that determines whether the other 11 go smoothly.
- Weeks 2-3: Foundations — authentication, data model, core infrastructure, deployment pipeline. Unglamorous, and the thing that makes weeks 4 onward fast rather than firefighting.
- Weeks 4-8: Core feature build, in weekly or two-weekly cycles with a working demo at the end of each. This is where scope discipline matters most — new ideas surface constantly at this stage, and almost none of them belong in this build.
- Weeks 9-10: Payment integration (if applicable), edge-case handling for the core journey only, and the unglamorous but essential legal basics — privacy policy, terms, cookie consent.
- Weeks 11-12: QA, real-device testing, soft launch to a small user group, and fixing what that group actually finds rather than what you assumed they would.
Twelve weeks is realistic for most software MVPs with a single core journey and no unusual compliance burden. Anything involving payments infrastructure from scratch, regulated data, or a genuinely novel technical problem should budget toward the longer end, or accept that week 1 will reveal a longer honest timeline.
Tech choices that keep you fast without a rebuild in a year
- Favour managed infrastructure (managed databases, managed hosting, managed auth) over self-hosted equivalents — the ops overhead is not worth carrying for an unproven product.
- Choose “boring”, well-supported technology over the newest framework — an MVP’s job is to prove the product idea, not the tech choice.
- Build the data model to be extended, not rebuilt — this is the one area worth a bit of extra upfront thought, since migrating data later is consistently more expensive than migrating code.
- Avoid premature abstraction — code written for a feature you might build in six months slows down the feature you are building this week.
The mistakes that most often blow the timeline
- Scope creep from stakeholders who were not in the discovery week. Lock scope with everyone who has a vote before week 2, not during week 6.
- Designing for scale you do not have yet. Handling ten million users is not the MVP’s job — handling the first hundred real ones properly is.
- Treating the MVP as the final product. The plan should assume rework after real user feedback, not treat v1 as the finished architecture.
- No plan for what happens after launch. An MVP without a follow-up iteration budget just becomes an abandoned experiment — decide the post-launch plan before you launch, not after.
Budget expectations
Most software MVPs with a genuinely disciplined scope land between £30,000 and £80,000 in the UK market for an 8-12 week build, depending on complexity of the core journey and whether payments or third-party integrations are involved. Anything quoted significantly below that range for a genuine software MVP is usually a signal that scope, quality, or seniority is being cut somewhere — worth asking exactly where.
After launch
The MVP’s job ends the moment you have real usage data. What comes next — iterate, pivot, or stop — should be decided from that data, not from how attached anyone is to the original idea. Most successful products bear only a rough resemblance to their MVP six months later, which is the system working as intended, not a failure of the original build.
Validating before you build a single line of code
The fastest, cheapest MVP is the one you do not have to build because a lighter test already answered the question. Before committing 8-12 weeks of engineering, it is worth ruling out:
- A landing page test: describe the product, collect intent (sign-ups, a waitlist, even card details for a pre-order) before building anything, to check demand exists at all.
- A manual “concierge” version: deliver the core value by hand — spreadsheets, manual matching, a person doing what the software will eventually automate — for a handful of real users, to learn what actually matters before automating it.
- A clickable prototype: useful for validating usability and flow, but be honest that it does not validate whether people will actually use or pay for the real thing — that only comes from something real.
None of these replace a real MVP forever, but skipping straight to a 12-week build without any of them is the single most common way founders spend real money learning something a £500 landing page test would have told them for free.
What a sensible post-MVP roadmap looks like
The months after launch matter as much as the twelve weeks before it, and are where most of the original scope cuts get revisited with real evidence behind them:
- Weeks 1-4 post-launch: fix what real users actually hit, not what you assumed they would — these are rarely the same list.
- Month 2-3: add the highest-friction item from real usage data, usually one of the features deliberately cut from v1.
- Month 3-6: revisit the original roadmap against what you now know — expect to drop things you were previously certain about.
- Ongoing: a standing monthly or retainer-based capacity for iteration, rather than re-scoping a new project from scratch every time — this is usually the point a team moves from project pricing to a retainer model.
Frequently asked questions
Should the MVP be built on the same tech stack as the eventual full product?
Usually yes, if you already have conviction in the core idea — rewriting a validated MVP onto new infrastructure is wasted effort. If you are still genuinely testing whether the idea works at all, optimising for build speed over long-term architecture is the more defensible choice.
How many features is too many for an MVP?
If you cannot describe the core user journey in three sentences without using the word “and” more than twice, the scope is already too broad.
What if we run out of budget before week 12?
This is almost always a scope problem, not a budget problem — it means the MVP was not actually minimum. Re-cutting scope to the true core journey, even mid-build, is usually a better outcome than shipping something half-finished on the original plan.
If you have a product idea and want a straight opinion on whether 8-12 weeks is realistic for it, send us the outline — we will tell you honestly what fits in that window and what does not.