Startup idea testing: what actually works.
Use the cheap signals first, then earn the expensive experiments.
Why startup idea testing goes wrong
A flattering score, a few encouraging calls, or a growing waitlist can all be useful. None is a substitute for deciding what must be true before you build.
Five ways to test an idea
Hostile expert simulation is fast. Pre-selling is an honest demand signal. Customer discovery reveals language and workarounds. Landing page tests measure interest. A concierge MVP tests delivery.
The correct sequence
Start with the questions that cost minutes. Move to the experiments that cost money only after the premise can survive a serious attack.