What is a Minimum Viable Product (MVP)?

Definition

A Minimum Viable Product is the smallest thing you can ship that solves a real problem enough that someone pays or uses it seriously and generates learning data.

Why it matters

The MVP is where most founders lose six months. They add "just one more feature" instead of shipping. Meanwhile, the market changes, the team burns out, and no learning happens. Drew Houston's original Dropbox MVP was a 5-minute video demo of a product that did not exist - 75,000 people signed up for the waitlist in a weekend, proving demand before a single line of sync code was written. The lesson is not that every MVP should be a video; the lesson is that every MVP should answer a specific question faster than any other version.

How it applies

You are building an AI tool for lawyers to summarize contracts. A full product would need OCR, a UI, a document library, version control, team collaboration, and integrations. MVP instead: a Google Form where a lawyer pastes a contract and hits submit, you run it through Claude overnight, and email back a summary by 9 AM. No UI, no backend. First 10 customers get it free in exchange for feedback. Within two weeks, you know whether the summary is actually useful, which 2 sections matter most, and what the lawyer would pay. That is enough to decide whether to build a real product.

Common mistakes

• Confusing MVP with beta - an MVP is the first thing that creates value, not the polished beta.
• Shipping an MVP with no instrumentation - you spent weeks building it and learn nothing.
• Adding features that "everyone expects" before testing - users expect many things they do not actually need.
• Treating the MVP as permanent - an MVP is disposable; it exists to teach you, not to scale.