How we build MVPs in 21 days

Our exact playbook on building MVPs, fast.

Our exact playbook on building MVPs, fast.

In 21 days, our team goes from concept to live MVP for real customers. And we build some pretty complex MVPs - Carvatar Case Study

Here’s our step-by-step playbook.

Days 1-3: Strategy sprint

Clarity from day one enables productivity downstream. With our client, we work to understand the target audience, their pain points and desired outcomes.

  • Distil your vision into a single, testable problem statement. e.g “Team communication is scattered across email, chat, and meetings, making collaboration inefficient and information hard to track”.
  • Identify your audience and recruit them for early input.
  • Prioritize ruthlessly: must-have, nice-to-have, won’t-have.

Tools

  • Pen + paper (❤️ our fave!)

Or, if remote or you simply prefer something digital:

  • Mural
  • Figjam
  • Miro

Days 4-7: Design Sprint

We now turn that problem statement into a real user experience, testing with real users to validate our assumptions.

  • Wireframe UX and data flows.
  • Validate with real users and experts.
  • Add design flair last.

Tools

  • Figma
  • Lovable
  • Bolt
  • Gemini

Days 8-17: Build Sprint

We code, test, and deploy in an automated, continuous loop. Every change is visible on a live URL or TestFlight, so everyone can track progress in real time.

  • Automate builds, tests, and deployments with CI/CD pipelines.
  • Deploy continuously so stakeholders always see the latest functionality.
  • Implement your validated designs in code.

Tools:

  • React
  • Tailwind
  • Supabase
  • Node.js
  • GitHub Actions
  • TestFlight
  • Cloudflare

Days 18-21: Launch Sprint

In the remaining 3 days we deploy everything to the live site and test everything in a live environment - including logins, payments and analytics data.

  • Share the app with real users.
  • Track behavior with analytics.
  • Collect feedback to inform your next iteration.

Tools:

  • Hotjar
  • Mouseflow
  • Mixpanel
  • Typeform
  • Userpilot

Conclusion

We always start by defining the problem statement. We find that this forms a solid foundation for design and code, enabling the team to ship at breakneck speed.

It’s a process we’ve followed for several projects and each time we’re learning better ways to do it faster, better and with more impact for our clients.