Real-world ranges, what actually drives the number, and how product strategy can cut your build budget by 30–50% before a single developer is hired.
Founders almost always ask this question first, and almost always get a useless answer. "It depends" is technically true and practically worthless. What you really want to know is: what range should I plan for, what makes the number go up, and what can I do about it before I ever talk to a developer?
This essay is the answer I give clients on a Founder Clarity Session, written down so you can use it before you book one.
The honest range
A scoped MVP — meaning a real, usable product that proves a single hypothesis — typically lands between $30,000 and $150,000 in the U.S. and Western Europe market. That's a 5× spread, which feels useless until you understand what moves you across it.
Here is how that range breaks down in practice, based on the quotes I see founders receive every month:
- $30K–$60K: a single-platform app (iOS or web), one core user flow, email/password auth, a basic admin view, no AI, no real-time, simple payments.
- $60K–$100K: two platforms or a responsive web app with a mobile-quality experience, multiple user roles, third-party integrations (Stripe, calendars, maps), basic notifications.
- $100K–$150K: AI features (LLM calls, embeddings, retrieval), real-time collaboration or chat, two-sided marketplaces, compliance work (HIPAA, SOC 2 prep), or anything with a hardware/IoT component.
- $150K+: regulated industries, video processing, custom ML models, anything that needs a dedicated DevOps story from day one.
These are ranges for a competent team that bills between $90 and $180 per hour. Offshore teams can come in 40–60% lower, with a corresponding increase in management overhead and rework risk. Solo "vibe-coded" builds with AI tooling can compress the low end further — but rarely survive contact with a second user, let alone a paying one.
The single biggest cost driver isn't engineering
It's ambiguity.
Every undecided feature gets decided by a developer, in code, while the clock runs. Every "we'll figure that out later" becomes a Slack thread, a meeting, a rework, a regression. None of that shows up on the quote, but all of it shows up on the invoice.
I've watched the same idea quoted at $42,000 by one team and $128,000 by another. The difference wasn't talent. It was the document the founder showed up with. The first team got a PRD with prioritized user stories. The second got a deck and a Figma file with three different navigation patterns.
Strategy is not a tax on your build budget. It's the discount.
What actually drives cost up
If you want to forecast your number, look at these inputs in order. They explain more variance than the tech stack ever will.
1. Scope ambiguity
How many decisions are still open when the engineering team starts? Every open decision is a meeting, a clarification ticket, a rework. Five open decisions cost a week. Twenty cost a month.
2. Number of user roles
A single-user app and a three-role app (consumer, provider, admin) are not 3× the work — they are roughly 2.2×. Each role needs its own onboarding, navigation, permissions, and edge cases.
3. Integrations
Each external system (Stripe, Twilio, Calendly, a CRM, an EHR) adds 1–3 weeks depending on auth complexity, sandbox quality, and how much of their API surface you actually touch. Webhooks are where the budget quietly dies.
4. Real-time and AI
Real-time features (chat, presence, live cursors) and AI features (LLM orchestration, embeddings, evals) both require infrastructure thinking that simple CRUD apps don't. Plan for a 20–40% premium on any feature that involves either.
5. Design polish
There is a meaningful difference between "functional" and "product-quality." Polish is the last 30% of the budget. Many MVPs ship at 70% polish on purpose and re-invest the savings in iteration.
6. Compliance and platform review
HIPAA, SOC 2 prep, App Store review for sensitive permissions, COPPA — these aren't features, they're project-long disciplines. They affect architecture, vendor choice, and timeline.
What drives cost down
The good news is that every cost driver above has a counter-move, and almost all of them happen before code.
- A written PRD with user stories and acceptance criteria — closes the ambiguity loop.
- A prioritized feature matrix that names what's out of scope — kills scope creep before it starts.
- A monetization decision made before kickoff — shapes onboarding, instrumentation, and which flows get priority.
- User journey maps for the surviving features — exposes edge cases at the whiteboard, not at sprint review.
- A defensible technical approach written in plain English — lets your engineering team estimate against architecture instead of guessing.
Founders who walk into a kickoff with these artifacts routinely see quotes 30–50% below founders who walk in with a deck. Same idea. Same engineering team. Different document.
Two budgets you'll actually pay
When you plan, plan two numbers, not one.
The build budget
This is the visible one — the number on the agency or contractor's statement of work. Most founders only plan for this.
The decision budget
This is the invisible one — the cost of every product decision made under pressure during the build, by whoever happens to be in the room. If you don't pre-pay this budget in strategy, you pay it (at 3–5× the rate) in engineering time and rework.
A useful planning ratio: spend roughly 8–15% of your projected build budget on pre-build strategy. On a $50K MVP, that's $4,000–$7,500 of structured product work. The Development Readiness Package sits inside that range on purpose.
A worked example
Two founders, same idea: a marketplace connecting freelance bookkeepers with small e-commerce sellers.
Founder A hands the agency a deck, a Figma file, and a Loom walkthrough. The quote comes back at $95,000 over 14 weeks. The build ships in 21 weeks at $138,000. The product has three flows nobody uses and is missing a refund mechanism that turned out to be load-bearing.
Founder B spends three weeks and $3,200 producing a Development Readiness Package: a PRD, prioritized features, user stories with acceptance criteria, a monetization model, and a one-page architecture brief. They take that package to a trusted offshore dev team and are quoted $8,000 over 11 weeks. The price holds. The build ships in 12 weeks at $8,000. The product has fewer flows, but every one of them is used, and the refund mechanism was in scope from day one.
Same idea. Different approach, different team. Founder A is out $138,000. Founder B is out $11,200. That's a $126,800 difference — and Founder B shipped two months sooner with a product people actually use. That is the real answer to "how much does it cost to build an app."
What to do this week
If you're still in the "getting quotes" stage, do these three things before you sign anything:
- Write a one-sentence hypothesis your MVP must prove. If you can't, you don't have an MVP yet — you have a feature list.
- List every feature, then cut anything not required to test that hypothesis. The cuts are the strategy.
- Get the surviving scope into a document a stranger could quote against in 24 hours. That document is your leverage.
If you'd rather do that with help, a Founder Clarity Session ($197, 45 minutes) will tell you within one call whether your idea is ready for an agency, ready for a Blueprint, or ready to be reshaped before either.