What an MVP really costs in 2026, and where the money goes

Most founders we talk to have already collected two or three quotes for their first product, and the numbers rarely agree. One agency says $12,000, another says $90,000, and a freelancer says "about three months". None of those numbers is necessarily wrong. They are usually answers to different questions, because each quote assumes a different scope, a different level of polish and a different amount of risk being carried by the builder.

This article is our attempt to make those numbers comparable. We cover realistic 2026 ranges for the four kinds of first product we are asked to build most often, where the money actually goes inside a project, the handful of decisions that move the total most, how to cut scope without cutting the thing that makes your product worth using, and the red flags we look for when a founder asks us to review someone else's quote.

For context: most of the MVPs we scope land between $15,000 and $45,000, delivered in six to twelve weeks. Some are smaller, a few are larger, and we will explain why.

What "MVP" should mean before you price it

An MVP is the smallest product that lets you learn whether a specific group of people will use, and ideally pay for, the thing you believe they need. It is not a cheaper version of the full product. It is a narrower one.

That distinction matters for cost. A "cheap full product" spreads a small budget across twenty features and ships none of them well. A narrow MVP picks the three or four workflows that prove or disprove your core assumption and builds those properly: fast, reliable, secure enough for real user data, and instrumented so you can see what people do.

Before asking anyone for a price, write down in one paragraph:

• who the first 50 users are,
• the one job they will hire the product for,
• what they will do on day one, and
• what number (sign-ups, activation, paid conversions) will tell you it is working.

If a quote cannot be traced back to that paragraph, it is pricing something else.

Realistic ranges by product type

These are ranges we see in 2026 for work done by a small senior team, with design, QA and deployment included. They assume one web platform or one cross-platform mobile codebase, not both.

A few notes on reading the table:

• Marketplaces cost more because they are two products. Buyers and sellers need separate onboarding, separate dashboards and a payment flow that holds money and pays it out, which brings in Stripe Connect or similar, KYC checks and refund logic.
• Mobile adds store overhead. App Store and Play Store review, signing, push notification setup, device testing and release management typically add one to two weeks compared with an equivalent web app.
• The top of each range is usually integrations. A SaaS MVP that syncs with a customer's CRM, accounting system or EHR is often 30–50% more expensive than one that does not, because third-party APIs are where estimates go wrong.

Where the money goes inside a project

Founders often picture an MVP budget as "developer hours". In practice, a healthy project spends roughly a third of its budget on things other than writing feature code, and projects that skip those parts tend to pay for them later.

On a $30,000 SaaS MVP, that works out to around $2,500 of discovery, $4,000 of design, $17,000 of build, $3,500 of QA and $3,000 of infrastructure and launch work.

The phase founders most often ask us to cut is discovery. It is also the cheapest insurance you can buy. A week spent finding out that the payment provider you assumed does not support your country, or that your "simple" import feature has to handle five spreadsheet formats, is a week that saves three later.

Ongoing costs founders forget

The build is not the whole bill. Budget for these from the first month after launch:

• Hosting and managed services: $50–$500 a month for most MVPs on AWS, Vercel, Render or similar. It rises with usage, not with ambition.
• Third-party services: email delivery, error tracking, analytics, maps, SMS, AI model usage. These are individually small and collectively $100–$800 a month.
• Store and developer accounts: $99 a year for Apple, a one-off $25 for Google Play.
• Maintenance and iteration: dependency updates, OS releases, security patches and the features your first users ask for. Plan for either a retained team or 15–20% of the build cost per year as a bare minimum for upkeep.

For founders who want a team that keeps shipping after launch rather than a one-off project, our retained engagements start at $2,699 a month on annual billing. Whether that makes more sense than a fixed-price build depends on how certain your roadmap is, which we come back to below.

What actually drives the cost

When two quotes for the "same" product differ by a factor of three, it is almost always because of these five things.

1. Number of user roles

Every distinct role (customer, provider, admin, support agent, clinic manager) needs its own permissions, screens and edge cases. Going from one role to three roughly doubles the surface area.

2. Integrations and data you do not control

Payments, calendars, CRMs, identity providers, government APIs and legacy databases each carry unknowns: rate limits, sandbox environments that behave differently from production, and documentation that is out of date. We estimate each integration separately and add contingency to the ones we have not used before.

3. Compliance and data sensitivity

Handling health data, financial data or children's data changes how you store, log, back up and audit everything. A seed-stage healthtech startup we worked with found that meeting HIPAA-aligned requirements (encryption, audit logs, access reviews, a signed BAA with every vendor) added about 20% to the build. That was the right call for them. It would be wasted money for a recipe app.

4. Real-time and offline behaviour

Live chat, collaborative editing, location tracking and offline-first mobile apps require infrastructure and testing that request-response apps do not. If "real-time" is a nice-to-have, a 30-second refresh is often indistinguishable to users and far cheaper.

5. Design ambition

A clean interface built on a well-maintained component library costs a fraction of custom illustration, bespoke animation and a unique design language. Early users forgive plain. They do not forgive slow or confusing.

How to cut scope without cutting value

Scope is the only lever that reduces cost without reducing quality. These are the cuts we most often recommend.

• Replace admin panels with existing tools. For the first few months, an internal tool built with Retool, a well-structured Airtable or even direct database access for one trusted person is often enough. That can save $3,000–$8,000.
• Use hosted auth and payments. Clerk, Auth0, Supabase Auth or Cognito for sign-in; Stripe Checkout and the Stripe customer portal for billing. Building these yourself is expensive and risky.
• Do things manually behind the scenes. Matching, onboarding, content moderation and reporting can be done by a founder for the first 100 users. Automate once you know the process.
• Ship one platform first. Web first if your users are at desks; one cross-platform mobile app if they are not. Rarely both.
• Defer settings, roles and customisation. Every "users can configure…" feature multiplies testing. Pick sensible defaults.

A B2B scheduling startup came to us with a 40-feature spec and a $95,000 quote from another vendor. After a week of discovery we agreed on 11 features, a Retool admin and Stripe Checkout, and delivered in nine weeks for $38,000. They added four of the deferred features after launch. The other 25 never came up again.

Red flags in a quote

When founders ask us to review another vendor's proposal, these are the things we look for.

• A single number with no breakdown. You should see phases, features or at least a rough allocation of effort. A lump sum makes it impossible to discuss trade-offs.
• No discovery or design phase. This usually means the vendor intends to start coding from your document, and every ambiguity becomes a change request.
• No mention of testing, deployment or hosting. Someone will have to do this work. If it is not in the quote, it is either missing or coming later as an extra.
• Unclear ownership of code and accounts. You should own the repository, the cloud accounts, the domain and the app store listings from day one, not receive them at the end.
• Fixed price with an unlimited feature list. Either the vendor has not read the list carefully, or they plan to deliver thin versions of everything.
• Estimates much lower than everyone else's. A quote at a third of the market rate usually reflects a misunderstanding of scope, not a secret efficiency.
• No demos until the end. You should see working software every week or two.

Summary and checklist

A realistic 2026 MVP from a senior team costs $15,000–$45,000 for most web SaaS products, more for mobile apps and marketplaces, and less for a marketing site. The total is driven far more by scope, roles, integrations and compliance than by hourly rates.

Before you sign anything, check that you can answer yes to each of these:

• I can describe my first 50 users and the one job the product does for them.
• The quote breaks cost down by phase or feature.
• Discovery, design, QA and deployment are explicitly included.
• Each third-party integration is listed and estimated separately.
• I will own the code, cloud accounts and store listings from the start.
• I will see working software at least every two weeks.
• I have budgeted for hosting, services and maintenance after launch.
• I know which features I would cut first if the budget tightens.

If you want a second opinion on a quote, or a scoped estimate for your own idea, we are happy to spend an hour on it with you.