Choosing one useful north-star metric
A good north-star metric reflects delivered value. Here is how to test candidates against real product behavior.
A north-star metric is useful when it keeps attention on value delivered to users. It becomes dangerous when it is chosen because the number is large, familiar, or easy to move.
Write the value exchange first
Describe why a user returns and what successful use looks like in plain language. Then list observable behaviors that represent that value—not merely visits, clicks, or time spent.
Test each candidate
A credible candidate should correlate with longer-term retention, respond to meaningful product improvements, resist superficial manipulation, and be understandable across the team. Check these claims with historical cohorts where possible.
Add guardrails
No single metric captures product health. Pair the north star with measures for quality, user harm, reliability, and business viability. Growth that degrades these constraints is not healthy progress.
Document and revisit
Record the exact definition, exclusions, time window, and expected relationship to value. Revisit it as the product and customer base change. The metric is a decision tool, not a permanent identity.