MVP in 6 weeks: the schedule that actually works

The obsession with the "perfect MVP" is the #1 reason Costa Rican startups don't ship. They spend 6 months polishing features nobody asked for, while a competitor launches something uglier but already getting paid. This is the schedule we use to take an idea to production in 6 weeks, with the hard rule that keeps the project focused.

💡 Does your idea fit in 6 weeks? The interactive quote builder gives you a USD range + estimated duration in 30 seconds.

💰 What budget does an MVP need? See the complete Costa Rica pricing guide — USD 3,000–5,000 ranges for MVPs, broken down per feature and per vertical.

The hard rule of the 6-week MVP

A 6-week MVP is ugly, breaks on edge cases, and has one great feature — but it's alive and getting paid.

If your MVP at week 6 can't charge a real customer, you didn't finish an MVP — you finished a demo. And demos are not business.

Week-by-week schedule

Weeks 1–2: Discovery + Architecture

No production code yet. Decide what to build and how.

What happens:

• Interviews with 8–12 prospects. Client team + us. Question: "what problem costs you most to solve in X?" — don't show your idea.
• Main feature design. Low-fidelity wireframes. Critical flow only.
• Stack decisions. Next.js or Astro, Stripe or LemonSqueezy, Supabase or RDS. (See our note on picking the stack.)
• Minimum technical setup. Repo, Vercel preview, domain, Stripe sandbox, analytics.

What does NOT happen: production backend, nice-to-have integrations, polished design.

Week 2 output: a 2-page doc with the main feature defined + a Figma with the critical flow + the repo set up.

Weeks 3–5: Build in weekly sprints

3 one-week sprints. Each sprint ends with a client demo and, where applicable, 1–2 beta users.

Sprint 3 — Skeleton:

• Auth working (Magic link or Google OAuth — no email + password yet).
• Main feature in its crudest version. Unstyled. Functional.
• Stripe in sandbox connected but not in the flow yet.

Sprint 4 — Connect the flow:

• Main feature polished in its 2–3 main states (loading, success, error).
• Real Stripe checkout (sandbox payments but full flow).
• Transactional email via Resend (welcome, payment confirmation).

Sprint 5 — What's missing to sell:

• Simple landing (1 page) with value prop + button to Stripe.
• Basic onboarding so a new user can reach the feature in 60 seconds.
• Account page + cancel subscription (Stripe Customer Portal).

What does NOT happen in these 3 weeks:

• Heavy automated tests (a minimal E2E set is enough).
• Internationalisation (unless critical day-one).
• Polished admin panel (admins use the database or Retool).
• "Wow" design — clean design is enough.

Week 6: Hardening + Deploy + First users

The week where the MVP goes from project to live product.

Monday–Tuesday: critical bugs and edge cases that surfaced in demos. Brutal triage: only what blocks use, not the pretty stuff.

Wednesday: deploy to production. Stripe live mode. Final domain. SSL, analytics (PostHog), Sentry or equivalent for errors.

Thursday–Friday: first 5–10 real users (the ones you pre-sold in the early weeks). Direct WhatsApp or Slack support. Every reported bug fixed in under 24 hours.

Week 6 output: product in production, charging real cards, with 5+ active users. A 2-minute Loom of the recorded demo to send to investors or more prospects.

Why pre-selling beats a perfect MVP

Most founders want "build first, sell later". It's exactly the reverse. The correct plan:

1. Week 0 (before the 6 weeks): pre-sell. Talk to 10 prospects. Sell the problem, not the product. Ask for commitment (deposit, letter of intent, waitlist with card).
2. If nobody pre-pays: the problem isn't big enough. Don't build the MVP. Iterate the offer.
3. If 3+ pre-pay: you have signal. Start the 6-week schedule.

This changes everything:

• MVP budget is partly covered before starting.
• You have 3 guaranteed users on launch day.
• Week 6 feedback comes from people with skin in the game.
• The team builds with confidence knowing the market validated.

Pragmatic stack for an MVP

Total production stack cost for 100 users: USD 0–20/month. Yes, zero. Free tiers of Vercel, Supabase, Resend and PostHog cover the first month easily.

Mistakes that kill the 6-week schedule

1. "While we're at it, let's add X feature". No. Feature X goes to v2. Any deviation breaks the rhythm.
2. Designing before validating. If you didn't pre-sell, what you're designing is probably the wrong solution.
3. Ambitious stack. Microservices, Kubernetes, async messaging — not for an MVP. PostgreSQL + Next.js + 1 external service is enough.
4. Not deploying in week 1. The deploy must be working by day 5. If you wait until week 6, you'll discover 20 problems on the last day.
5. 5+ person team. Max 2–3 people for an MVP. More team = more coordination = less speed.

In summary

If you have an idea and want to know if it fits in 6 weeks, let's talk. We don't charge for that conversation. In 30 minutes we tell you whether it's worth starting or whether you still need pre-sales.