Every founder asks this question, and almost every answer is useless. "It depends" is true and unhelpful. "€50,000" is specific and probably wrong. The problem is that MVP means something different to the person asking than to the person quoting.
So let's break the number into the parts it's actually made of.
The three things you're really buying
Any quote you receive is some combination of these, whether or not the breakdown is shown to you.
- Decisions. What gets built, what gets cut, what the thing is for. This is the expensive part and it is almost never itemised.
- Construction. Turning those decisions into working software. This is the part everyone thinks they're paying for.
- Risk. The buffer the supplier adds because they don't know your codebase, your stakeholders, or how many times you'll change your mind.
A quote that looks expensive is usually heavy on item three. A quote that looks cheap has usually skipped item one, and you will pay for it later in change requests.
Rough ranges, honestly stated
These are European rates for senior work in 2026. US agency rates typically run 1.5–2× higher for equivalent seniority; offshore rates run lower but shift cost from the invoice to your own management time, which is rarely accounted for.
- A single validated flow — one core journey, real backend, shippable to a test group: €7,000–15,000.
- A credible consumer MVP — auth, a core loop, payments, push, store release: €25,000–60,000.
- A B2B product with real integrations — SSO, roles, an external API you don't control, an admin surface: €60,000–150,000.
- Anything touching regulated data — health, finance, identity — add 30–50% for compliance work that is invisible in the demo.
If a supplier gives you a firm number before asking what the product is for, they are pricing item two and hoping items one and three don't bite. They will.
What actually moves the number
In roughly the order that surprises people:
- Integrations with systems you don't control. A payment provider, a legacy ERP, someone else's undocumented API. This is the single most common source of overrun, because the effort is unknowable until you're inside it.
- The number of people who can say no. Every additional approver adds calendar time, and calendar time is money on a day rate. A three-stakeholder project is not 50% more expensive than a two-stakeholder one — it's often double.
- Design that hasn't happened yet. "We'll figure out the screens as we go" is a decision to pay for the same screen three times.
- Native vs cross-platform. Real, but usually smaller than people expect — see Native vs Flutter vs React Native.
- Your own responsiveness. A team waiting two days for an answer bills those two days to something.
The cheapest MVP is the one you don't build twice
The most expensive projects I've seen were not the ones with big budgets. They were the ones that shipped something nobody wanted, cheaply, and then had to start again — this time with less money and less patience from the board.
Before you compare quotes, be able to answer: what single thing must be true for this to be worth continuing? If the MVP doesn't test that, its price is irrelevant.
How to make the quotes comparable
Send every supplier the same three things and the numbers you get back will finally mean something:
- The one user journey that matters most, described end to end in plain language.
- The systems it must talk to, named, with a note on whether you control them.
- The date something has to be in front of real users, and what happens if it isn't.
Suppliers who come back with questions rather than a number are the ones worth talking to.
A worked example
A founder came to me last year wanting a marketplace app. Two sides, payments, chat, reviews, ratings, a web admin panel. Three agencies had quoted between €90,000 and €140,000, and all three were quoting honestly for what had been described.
The uncertainty that mattered was whether suppliers would list at all. Everything else — the payments, the chat, the ratings — was only worth building if that turned out to be true, and none of it tested it.
What we actually built, in one two-week sprint: a listing flow, a browse view, and a contact button that opened WhatsApp. No payments, no chat, no ratings, no admin panel. Around €8,000.
Thirty suppliers listed in the first fortnight, and almost all of them tried to move the conversation off-platform immediately. That's a finding worth six figures, and it arrived for the price of a rounding error. The full build happened later, deliberately, with the disintermediation problem designed for rather than discovered in month five.
The €90k quote wasn't wrong. It was a correct price for the wrong scope, and no amount of negotiating the rate would have fixed that.
Five questions to ask before accepting any quote
- What did you assume that I didn't tell you? Every estimate rests on assumptions. Getting them written down converts them from future excuses into present risks.
- Which part of this are you least certain about? An honest supplier names one immediately. A supplier who says "none of it" hasn't thought about it, and you'll pay for that thinking later.
- What would you cut if the budget were half? The answer reveals what they think is essential — and whether they understand what the product is for.
- Who writes the code, and what percentage of their week do I get? The second half of that question is the one that matters. "A senior lead" at one day a week is not a senior lead.
- What happens on the day we disagree about whether something is in scope? There will be such a day. Better to know the mechanism now, while everyone is still friendly.
Why cheap quotes get expensive
The pattern is consistent enough to be predictable. A low quote wins on price, then meets reality:
- Month one: things move fast, because the easy 60% is genuinely easy and everyone is motivated.
- Month two: the integration nobody scoped turns out to need an account, a contract, and a sandbox that takes three weeks to provision.
- Month three: change requests begin. Each is individually reasonable and collectively equal to the difference between the cheap quote and the expensive one.
- Month four: you're choosing between paying more and shipping something you'd be embarrassed by.
You didn't save money — you deferred discovering the price. That deferral has a cost too, paid in runway and in the credibility you spent telling your board the number.
A note on fixed prices
A fixed price is not a discount — it's a transfer of risk. The supplier prices the risk of being wrong, and you pay for that certainty. Whether it's worth it depends on how well-defined the work is, which is exactly the subject of fixed price vs time and materials.
What I'd say is this: if a supplier will fix a price without understanding the work, the fixed price protects them, not you. The scope will simply shrink until it fits.