What is measurement decay?

Measurement decay is a concept in personal epistemology: Regularly check that your feedback loops are still connected to meaningful outcomes.

Example: A startup tracks 'daily active users' as its north-star metric. Six months later, the product team has optimized onboarding flows that inflate DAU by re-engaging churned users through dark patterns — push notification spam, confusing unsubscribe flows, auto-opening the app. DAU is up 40%. Actual user satisfaction, measured by NPS, has cratered. The feedback loop that once told the team 'people find this useful' now tells them 'people can't figure out how to leave.' The metric didn't break overnight. It decayed one optimization at a time.

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