What is early warning signals?

Early warning signals is a concept in personal epistemology: Measure things that predict outcomes rather than waiting for outcomes themselves.

Example: A startup tracks monthly revenue to judge product-market fit — but revenue is a lagging indicator that arrives months after the decisions that shaped it. Sean Ellis's product-market fit survey asks one question ('How would you feel if you could no longer use this product?') and counts the percentage who say 'very disappointed.' That single leading indicator predicts sustainable growth 6 to 12 months before revenue metrics show it. The startup that measures only revenue is steering by the wake of the boat. The startup that measures the Ellis score is steering by what's ahead.

This concept is part of Phase 24 (Feedback Loops) in the How to Think curriculum, which builds the epistemic infrastructure for feedback loops.