MVP checklist: start without technical debt

This checklist guides you step by step to launch your MVP in four weeks with the right foundations from day one. We’ve used it ourselves and it changes everything. You don’t need to do everything perfectly, but there are things you can’t ignore.

Week 1: Foundations

Tech stack

• Front-end: Choose a simple, modern framework (React, Vue, or even vanilla JS if you’re in a rush)
• Back-end: Lightweight framework (Express, FastAPI, or serverless if possible)
• Database: PostgreSQL or MongoDB depending on your needs, but avoid overly complex solutions
• Hosting: Managed solution (Railway, Render, Vercel) to avoid managing infrastructure

Project structure

• Layer separation: Create clear folders (api/, components/, utils/)
• Environment variables: Use a .env file from the start (and add it to .gitignore!)
• README: Document how to install and run the project

Week 2: Essential features

Core features

• Basic authentication: Email/password with hashing (bcrypt), no need for OAuth in an MVP
• Main CRUD: Basic operations for your main feature
• Validation: Client AND server side (never trust the client)

Minimum testing

• Unit tests: On critical business logic (calculations, validations)
• Integration tests: On main API endpoints
• Goal: 60–70% coverage on critical parts, no need for 100%

Week 3: CI/CD and monitoring

Minimal CI/CD pipeline

• GitHub Actions / GitLab CI: Simple pipeline with 3 steps
  • Run tests
  • Build the application
  • Deploy automatically
• Environments: At least staging and prod

Basic monitoring

• Structured logs: Use JSON format for logs
• Error monitoring: Sentry (free up to 5,000 events/month) or equivalent
• Uptime monitoring: Uptime Robot (free) to check your site is up
• Basic metrics: Response time, error rate (you can use logs for this)

Week 4: Feature flags and documentation

Feature flags

• Simple solution: Environment variables or a free service (LaunchDarkly free tier)
• Usage: Enable/disable new features without redeploying
• Benefit: Test in production with a small group before rolling out

Documentation

• API: Document your endpoints (Swagger/OpenAPI if possible, otherwise a simple README)
• Technical debt: List what's temporary with dated TODO comments
• Deployment guide: How to deploy to staging and prod

Final checklist before launch

Security

• ✅ Passwords hashed (bcrypt, argon2, or scrypt)
• ✅ HTTPS enabled everywhere
• ✅ Server-side validation on all inputs
• ✅ No secrets in code (environment variables)
• ✅ HTTP security headers (CSP, X-Frame-Options, etc.)

Performance

• ✅ Optimised images (WebP, lazy loading)
• ✅ Minified assets in production
• ✅ HTTP cache configured
• ✅ Database indexed for frequent queries

Quality

• ✅ Tests pass in CI
• ✅ No critical errors in logs
• ✅ Monitoring configured and alerts active
• ✅ Database backup configured

What you can postpone (but not forget)

• E2E tests: Not essential for an MVP, but plan for v2
• Advanced performance: CDN, Redis cache, etc. (only if needed)
• Multi-tenancy: If you don't need it now, don't do it
• Microservices: Stay monolithic for the MVP, you can split later

In summary

This checklist helps you launch your MVP in 4 weeks with the right foundations. The goal isn't perfection, but avoiding mistakes that will cost you later. Take the time to do the important things well, and postpone the rest. You can always improve later!

Need help? If you get stuck or want to validate your approach, don't hesitate to contact us. We’ll be happy to help!