Product Discovery

Figure out whether an idea is worth building before you commit engineering time. The goal is sufficient evidence that the solution will work — not certainty, but enough to bet responsibly.

70-90% of features built without validation fail to deliver business results. Discovery exists to catch the failures early and cheap, so engineering investment goes toward ideas with real evidence behind them.

Core Principle

Every idea tested in discovery should cost at least one order of magnitude less than building the real thing. If engineering a feature takes 4 weeks, discovery should take days. If it takes months, discovery should take weeks.

Process

1. Understand the idea
2. Assess the four product risks
3. Identify what you don't know
4. Design experiments
5. Define evidence gates
6. Write the discovery plan