Andrew Crossley
    STAGE 3-5 · 6 WEEKS

    MVP development

    MVP development is building the smallest deployed version of a product that delivers real value to a real user and produces learning you can act on. It is not a prototype and not a demo: it is live, instrumented, and someone can pay for it. With AI build tooling a properly scoped MVP takes four to six weeks and costs £12,000 to £25,000 rather than the three to six months an agency will quote.

    WHO THIS IS FOR

    Where MVP builds go wrong

    • The scope grew from one user journey to four because each addition seemed small.
    • You are three months in with nothing deployed, because everything is nearly ready.
    • An agency quoted six figures and nine months for something that should take six weeks.
    • You have a beautiful prototype in Figma and no way for a customer to actually use it.
    • Admin panels, settings screens and edge cases got built before the core journey worked.
    • There is no analytics, so launch will teach you nothing beyond whether the servers stayed up.

    THE FRAMEWORK

    The Crossley Method: idea to first revenue in seven stages

    See the full method
    1. STAGE 1DiscoverWeek 1
    2. STAGE 2ValidateWeek 2
    3. STAGE 3PrototypeWeek 3
    4. STAGE 4Build MVPWeeks 3-4
    5. STAGE 5LaunchWeek 5
    6. STAGE 6First RevenueWeek 6
    7. STAGE 7ScaleOngoing

    WHAT THE WORK COVERS

    MVP development across stages three to five

    This service assumes the problem is validated. If it is not, start with product validation — building first is the most expensive way to discover nobody wanted it.

    Prototype

    The core journey made clickable and put in front of five real users before a line of production code exists. Cheap to change, and it always changes.

    Build MVP

    One journey, built and deployed daily with AI tooling. Auth, data, the core loop and nothing else. Everything not on the critical path is deliberately deferred.

    Instrument

    Analytics, activation funnel and — where AI features exist — an evaluation set. Shipping without measurement means launch produces opinions rather than data.

    Launch

    Landing page, pricing, onboarding and a payment or waitlist path, so the product can convert rather than just exist.

    HOW IT RUNS

    The six-week build

    1. Week 1

      Scope and prototype

      Define the single journey. Clickable prototype tested with five users. The hardest work of the project is deciding what is not in it.

    2. Week 2

      Foundations

      Data model, auth, deployment pipeline and the first thin slice of the core journey live on a real URL. Deployed from day one, never local-only.

    3. Weeks 3-4

      Core build

      Daily deploys. The journey working end to end, including the AI components with evaluation attached where relevant.

    4. Week 5

      Launch surface

      Landing page, pricing, onboarding sequence, analytics, and payment or waitlist depending on your model.

    5. Week 6

      Launch and handover

      Public launch, first users, full code and infrastructure handover, and a next-six-weeks roadmap based on the real usage data.

    PRICING

    MVP development pricing

    Fixed price against a fixed scope, paid in two milestones. Scope changes are priced separately and honestly rather than absorbed and resented.

    Prototype only

    £3,000

    One week. Clickable prototype of the core journey, tested with five users. Enough to raise against or to decide against building.

    Six-week MVP

    £12,000-£25,000

    Validated idea to launched product: build, launch surface, analytics and handover. Range depends on integrations and AI complexity.

    Build with your team

    £4,000-£8,000/mo

    You have engineers; I bring the product leadership and AI build practice. Ongoing, one to two days a week.

    Building rather than hiring? Get a number in 60 seconds with the MVP cost calculator.

    COMPARISON

    Fractional-led build vs a development agency

    Agencies are good at building specifications. They are poor at deciding what the specification should be, and at seed stage that decision is the whole game.

    FactorProduct-leader-led buildDevelopment agency
    Typical cost£12,000-£25,000£60,000-£150,000
    TimelineFour to six weeksThree to nine months
    Scope decisionsActively cut, because scope is the riskExpanded, because scope is the revenue
    Who you deal withOne person, throughoutAccount manager, then a rotating team
    Analytics and evaluationBuilt in from week twoUsually a line item you have to ask for
    After launchRoadmap based on real usage, optional retainerChange-request pricing

    PROOF

    Products shipped this way

    Wocal — £2.7M pre-money

    Founder and CPO 2020-2025. Scaled from a single venue to 300+ by shipping narrow and expanding only where usage justified it.

    CallFlow AI, Journly, Vynl

    Three AI products taken from idea to deployed and instrumented, each scoped to a single core journey before build began. Live demos available.

    Co-Ride

    Community carpooling platform, founded November 2025, pre-seed, currently running the same six-week structure described above.

    Enterprise grounding

    Just Eat 2021-2025 and Sage. Useful for MVPs that will eventually need to survive procurement, integrations and scale rather than just a demo day.

    The scoping decision that determines everything

    An MVP has one job: get a real user through one valuable journey and learn from it. Every feature that is not on that journey is a delay to learning, and delayed learning is the only genuine risk at this stage.

    The practical test I use is simple. Write down the single sentence describing what a user does and what they get. Anything that does not directly enable that sentence goes on a list for later. Admin panels, settings, team management, onboarding wizards, notification preferences — all of it can wait, and most of it turns out to be unnecessary once real usage arrives.

    This is unpopular in the room, because every deferred item has an advocate and each one individually seems small. The compounding is what kills projects: four small additions is a doubled timeline, and a doubled timeline is where MVPs die.

    How AI tooling actually changes MVP delivery

    The build itself is roughly three to five times faster. Tools like Lovable, Cursor and Claude Code turn a well-specified screen into working code in minutes rather than hours, and the boilerplate that used to consume the first two weeks of a project largely disappears.

    What has not changed is everything around the code: deciding what to build, judging whether the output is right, understanding the user, and getting someone to pay. Those remain human, and they remain the bottleneck. The result is that product judgement is now a larger share of the total effort than it was five years ago, not a smaller one.

    There is also a specific trap. AI tools make it cheap to build, which makes it cheap to build the wrong thing at speed. A team without scope discipline will produce four half-working features in the time it takes a disciplined one to ship a single journey that works, and only the second of those can be launched.

    What ships with every build

    • A live URL from week two, deployed continuously, never a local-only demo.
    • The core user journey working end to end for a real, non-technical user.
    • Analytics with an activation funnel defined before launch, not after.
    • Evaluation sets for any AI feature, so quality is measurable across releases.
    • A landing page with real pricing and a working payment or waitlist path.
    • Full ownership: your repository, your infrastructure, your accounts, no lock-in.
    • A written next-six-weeks roadmap based on what the usage data actually shows.

    FREQUENTLY ASKED

    MVP development questions

    How much does MVP development cost?
    A properly scoped MVP built with AI tooling costs £12,000 to £25,000 and takes four to six weeks. A clickable prototype alone is around £3,000. A traditional development agency will typically quote £60,000 to £150,000 over three to nine months for comparable scope.
    How long does it take to build an MVP?
    Four to six weeks for a validated idea with one core user journey. Week one is scoping and prototyping, weeks two to four are the build, week five is the launch surface and week six is launch and handover. Timelines beyond eight weeks almost always indicate a scope problem.
    What should an MVP include?
    One core user journey working end to end, authentication, the data model behind it, analytics with an activation funnel, and a way to pay or join a waitlist. It should not include admin panels, settings screens, team management or edge-case handling until real usage justifies them.
    Can you build an MVP without a technical co-founder?
    Yes. AI build tooling means a product-led build with a single experienced person can deliver a production-grade web MVP. You will eventually need engineering depth to scale, but you do not need it to prove demand or to reach first revenue.
    Do I own the code?
    Entirely. Your repository, your hosting, your database, your third-party accounts, handed over at the end of week six. There is no proprietary platform and no lock-in of any kind.
    What is the difference between a prototype and an MVP?
    A prototype simulates the experience so you can test the concept — it is not deployed and nobody can genuinely use it. An MVP is live, handles real data, and a real customer can complete a real task and pay for the result.
    What happens after the MVP launches?
    You get the live product, the analytics, the code and a next-six-weeks roadmap grounded in real usage. Founders who want continued product leadership move onto a fractional CPO retainer at one to two days a week.
    What if my idea is not validated yet?
    Start with a two-week validation sprint instead. It costs around £3,500 and either produces the behavioural evidence that justifies a build or saves you the £12,000 to £25,000 the build would have cost.

    GET IN TOUCH

    Tell me what you want to build

    The idea, the user, and any evidence you already have. I reply within 48 hours with a scoped quote and a timeline.

    Your details are used only to reply to you and are never sold or shared. See the privacy policy.