Six steps
- Problem and user. Who has the problem, how they solve it today and what they would pay for. One conversation with three potential customers beats a week of guessing.
- One action. The main thing a user must be able to do in the first version — book, buy, upload, track.
- Written scope. Screens, rules and a fixed price, agreed before any code.
- Design and build. Simple, clean screens; accounts, data and payments if needed; progress you can see every few days.
- Launch. On your domain, with analytics to see what users do and error tracking to catch problems.
- Learn. Talk to the first users and decide what to do next.
What goes into the written scope
- The list of screens and what each one does.
- User roles — for example customer and admin — and what each can see and change.
- Rules: prices, limits, notifications, what happens on errors.
- Integrations: payments, email, messengers, spreadsheets.
- What is explicitly not included in this stage.
- Price, deadline and payment stages.
A written scope is the cheapest insurance in software: it turns “I thought you meant…” into a five-minute conversation instead of a paid rework.
Timeline and price
| Stage | Time | Price |
|---|---|---|
| Pre-sales MVP | 1–2 weeks | from $300 |
| Core MVP | 2–4 weeks | from $800 |
| First-customers MVP | 4–8 weeks | from $1,500 |
Technology choices
For most MVPs the right choice is a web app: it works on every phone and computer, needs no app store and is cheap to change. For audiences that live in Telegram, a Telegram Mini App gives the same benefits plus sign-in without passwords. Proven, common technologies are better than exotic ones: they are cheaper to host and easier for the next developer. Hosting usually costs $5–20 a month.
Launch checklist
- Your own domain with HTTPS.
- Analytics and error tracking.
- Automatic database backups.
- A real test payment, then a refund.
- A privacy policy and terms, if you collect personal data or take payments.
- A clear way for users to contact you — a chat, a messenger link or email.
Mistakes that make an MVP expensive
- Features for users you don't have yet.
- Polishing the design before anyone has used the product.
- Working without a written scope — and paying for every misunderstanding later.
- Building native apps for two platforms before the idea is proven.
- Accounts, servers and code registered to the developer instead of you.
Building an MVP with AI tools
AI coding assistants make development faster, and AI app builders can produce a clickable prototype in an afternoon. For a product that takes real payments and stores real data, someone still has to own the architecture, security, payments and hosting. A sensible split: use AI tools to explore the idea, then build the MVP properly once you know what it should do.
Questions
Do I need a technical co-founder?
Not for an MVP. A developer can build it, and the code and accounts stay with you.
Can I pay by stages?
Yes — each stage is agreed and paid separately: 50% before, 50% after.
Should I build a web app or a mobile app first?
A web app or a Telegram Mini App in most cases: cheaper, faster and no app store. Native apps make sense once the idea is proven.
How do I know the MVP worked?
Decide on one signal before launch — for example ten paying customers in the first month — and check it honestly.
How long does MVP development take?
From one to two weeks for a pre-sales MVP to four to eight weeks for a product with accounts and payments.