Building Your MVP

What to build, what to skip, and how fast to move.

What Is an MVP (And What It's Not)

As Dalton Caldwell, Managing Partner at YC, explains: "The keyword here is viable. A product that doesn't work at all and is useless to everyone is not viable. It has to be useful enough to serve some kind of purpose for the customer."

An MVP is not a prototype. It's not a demo. It's the smallest version of your product that delivers real value to real users. Minimum refers to scope (few features), not quality (poor execution).

The 90/10 Solution

Paul Buchheit's framework: the first version is not going to be the final version, and it will very likely — a lot of the code — be rewritten, and that's okay.

DoorDash Example

Built "in one afternoon" with HTML/CSS, PDF menus, Google Forms for the back end, and Find My Friends for driver tracking. They constrained to Palo Alto only — which helped them get delivery unit economics right before expanding.

How Long Should an MVP Take?

Validation

Time: 1–2 weeks

Output: 10+ customer conversations

Scoping

Time: 2–3 days

Output: v1 feature list

Build (non-AI)

Time: 2–4 weeks

Output: Live product

Build (AI-native)

Time: 4–8 weeks

Output: Live product + AI

Soft launch

Time: Day 1

Output: First users, first feedback

Technical Decisions

Choose tech you already know. Use third-party services for auth, payments, infrastructure. As Diana Hu explains: "Choose the tech for iteration speed. Keep it simple. Don't just choose a cool new programming language — choose what you're comfortable enough to launch quickly."

The Tools to Use

Auth: Auth0 or Clerk. Payments: Stripe. Cross-platform: React Native. Cloud: AWS or GCP. Landing pages: Webflow. Back-end: serverless (Lambda, Firebase). Don't build from scratch.

Common MVP Mistakes

Mistake: Building too many features

70% of MVP features go unused. If you can remove it and users can still complete the core workflow, remove it.

Mistake: Skipping validation

Build something to show users first, get their reaction, then decide what to build.

Mistake: Over-engineering

Pokemon Go launched with technical issues — users couldn't log in — and it didn't kill the company. They made over a billion dollars in revenue.

Mistake: Getting attached to bad ideas

Optimizely started with a Twitter referral widget. When feedback showed it wasn't working, they pivoted quickly to A/B testing.

Next: Getting Your First Users →