Accept that your ICP will change - build the discipline to iterate, not the skill to predict
Every successful founder starts with ICP hypotheses that prove wrong. The winning approach is not better prediction but faster iteration - courage to get narrow, humility to admit mistakes, and persistence to keep testing.
When to use
Throughout early-stage development when evaluating ICP assumptions
Don't do this
Spending months perfecting your ICP definition before testing it with real prospects
3 Founders Who Did This
Every founder featured started with ICP hypotheses that were dead wrong - Retool thought FileMaker developers, Vanta tried voice assistants, Clay served everyone for 5 years
Initially targeted busy founders but shifted to organizations with 50+ employees after discovering founders were price-sensitive and used the product inconsistently due to context-switching.
Described PMF as having two dials (product and market). Kept product vision constant while pivoting messaging and market targeting from FileMaker devs to React developers at Fortune 250 companies