"MVP" has become a word for everything: a landing page with a waitlist, a prototype exported from a design tool, a half-finished platform that ran out of money. Ask five agencies to build you one and you'll get five quotes for five different things.
Here's the definition we've worked with since 2017: an MVP is one core journey, fully built, in front of real users. Not a little of everything — all of one thing.
This piece covers what that costs, which shortcuts come back to bite, and how you'll know your MVP is done. With team sizes and weeks as we plan them, and the route to a fixed price.
All of one thing
The most common MVP mistake isn't too much thrift — it's the wrong kind of thrift. Building ten features to 60 percent instead of one to 100. The result demos well and can be given to nobody.
An example from our portfolio: HappyBee shows parents what their children eat at school — weekly menus, nutrition, per child. One purpose, one user group. The case study is here.
That's the shape most ventures should start in. Assuming it has to be built at all — Anjan has worked through when an off-the-shelf product is the better call. A product that does one thing right. Extending comes later, on a clean base — the other way round works too, it's just more expensive.
What an MVP costs
Here is a lean MVP with us, in team and weeks:
| Project shape | Team | Duration |
|---|---|---|
| Lean MVP — one core journey, in front of real users fast | 2–3 engineers | 8–14 weeks |
Everything a first production-ready release needs is in there: architecture, design, development, QA, release. Not a click-through prototype — a product real users use with real data.
Why no amount? Because an MVP with a single connection to a legacy system can take much longer than one without, and a number that doesn't know that won't help you. Swiss clients get their quote in francs, at a fixed price — our Swiss cost guide has typical market ranges, and our cost calculator works out team and weeks for your specific shape in two minutes.
And as always: indicative 2026 figures, not a quote. What moves the number is complexity — integrations, compliance, data migration and the reliability bar you need on day one. After a free 4-hour workshop, you get a written scope and a quote within three working days; for a clearly defined scope, that quote is a fixed price. If the scope is still open, a short paid discovery sprint comes first, credited in full against the build, and the fixed price follows it.
The shortcuts that come back to bite
An MVP lives by leaving things out. But there are two kinds of leaving out, and the bill arrives with the wrong one.
What you should never cut:
- Login, permissions, data-protection basics. "We'll harden it later" means, in practice, you build it twice. We build privacy-first, with GDPR and the nDSG considered from day one — retrofitting is always dearer than building in.
- An admin surface. Without one, your developer answers every support question with a database query. A simple one is fine — but it has to exist.
- The architecture decisions that can't be retrofitted. Multi-tenancy is the classic — why "we'll add it later" is the most expensive sentence in SaaS is here.
- Measurement. An MVP is an experiment. An experiment without metrics is an opinion with a budget.
What you can safely cut:
- The polished design system — a clean, plain UI serves the first cohort fine.
- Integrations two through ten. One integration that holds beats five halves.
- Separate native apps per platform — cross-platform is the default road now, it's how we build mobile products.
- Any feature whose sentence starts with "And later users could also…"
How you'll know it's done
An MVP isn't done when nothing is missing. It's done when a stranger completes the core journey without your help — and you can measure what they did.
Which is why one sentence belongs before the first sprint, and many projects never write it: "This MVP succeeds if number X goes from A to B by date C." When that sentence is missing, the gap isn't technical. It's a decision nobody has made.
And after? Version 2 gets built from usage data, not from the original ideas list. Annual support typically runs at 15–20% of the original build cost — that goes into our quote from day one too, because the second bill shouldn't be a surprise.
The sober way in
If you're seriously weighing an MVP: the first workshop takes four hours and costs nothing. Within three working days of it you have a written scope, an explicit "won't build" list and a quote, at a fixed price when the scope is clearly defined. If the scope is still open, a short paid discovery sprint comes first, credited in full. The document is yours to keep, even if you build with someone else afterwards.
First know what's being built. Then a number that holds.
By the way: what AI changes about the building itself — and what stays human — we've sorted out here.
