MVP stands for minimum viable product, and the word doing the work in that phrase is minimum. It's the smallest version of the thing that real customers can actually use — not a stripped-down demo, not a prototype you'll throw away, but the real app with everything non-essential left out. The cost question only has an answer once you know what's actually in scope, which is why two founders describing "the same idea" can get quotes five times apart.
What actually moves the number
- Sign-in and payments. Both are solved problems with mature tools behind them, but wiring them into your specific app — subscriptions, trials, refunds, team accounts — is real work, and it's the same work whether your app is simple or complex.
- How many places it has to connect to. An app that stands alone is cheaper than one that has to read from your accounting package, write to your CRM, or trigger something in a system you already run.
- How much judgement the AI part needs. "Summarise this document" is an afternoon. "Decide which of these documents needs a human to look at it" is a design problem, not a feature.
- How much design work is actually needed. A functional, well-laid-out interface is included in any serious quote. A fully custom brand identity with original illustration is a separate project.
Why two quotes for the same idea land so far apart
Some of the gap is genuinely different scope — one quote assumed features the other didn't. But a lot of it is structural, and it's worth reading alongside our piece on why agency pricing looks the way it does: a quote built on hours multiplied by a day rate carries a team, a handover chain and a buffer for the unknown that a fixed-price quote from a small team doesn't need to carry.
A quote that's a flat number for a fixed scope and a quote that's "roughly $X, we'll bill as we go" are not the same kind of number, even when the digits look similar. Ask which one you're holding before you compare it to anything else.
What a first version needs, at minimum
- One core workflow, done properly. The thing your customer actually came for, working end to end — not five features half-built.
- Sign-in that's actually secure, not a password field that looks the part.
- Payments, if you're charging from day one — including the boring parts, like what happens when a card fails.
- Somewhere errors get reported to you, so "it broke" doesn't arrive as a customer email three weeks later.
Everything past that list is a second version. The features you're already imagining for month three are usually right to imagine and wrong to build first — they're guesses about what customers will want, and guesses are cheaper to test after launch than to build before it.
A real number to compare against
Ours starts at $1,750, fixed, agreed before anything starts, with a 2–3 week timeline for a first version covering sign-in, payments and the AI features it needs. That's not a floor everyone should expect to match — it's what a lean, integration-light first version costs when there's no team overhead built into the rate. Use it as a reference point, not a target to negotiate a bigger quote down to.
Three questions to ask any quote
- 1What's explicitly out of scope? A good quote lists what it doesn't include as clearly as what it does.
- 2Is the price fixed, or does it move if the work takes longer than expected? Find out who carries that risk before you sign anything.
- 3What happens after launch? Bugs show up in week one whether or not anyone budgeted for them.