~/learn/Product Metrics · Metric foundations
North Star metrics & metric trees
Pick one number that captures the value customers get, then break it into parts teams can move.
why this matters
'What metric would you use to measure the success of X?' opens most product data science interviews.
A North Star metric is the single number that best captures the value customers get from a product — and that predicts long-term business success. Spotify might use 'time spent listening', Airbnb 'nights booked', Slack 'messages sent by teams that stay'. It is not revenue itself: revenue is the result of delivering value, and it moves too slowly to steer by.
On its own a North Star is too big for any one team to move. A metric tree breaks it into parts that multiply or add together. Each branch becomes something a specific team owns. When the top number moves, you walk down the tree to find which branch moved it.
Metric tree — Airbnb-style
Nights booked ├── Active guests × bookings per guest × nights per booking │ ├── new guests (marketing) │ └── returning guests (retention) └── constrained by: available listings (supply team)
Revenue = Users × Conversion rate × Orders per buyer × Average order value
Interview trap
Don't answer 'what's the North Star for X?' with a list of five metrics. Pick one, say why it reflects customer value, then show how it breaks down. Naming 'revenue' or 'DAU' with no reasoning is a weak answer.
Pro tip
Good test for a North Star: if this number grows, are customers clearly getting more value? If you can grow it while making users unhappy (e.g. notifications sent), it's the wrong number.
key takeaways
- Pick one North Star tied to customer value.
- Break it into a tree of team-owned parts.
- Walk the tree to explain movements.
check your understanding
Which is the best North Star for a music streaming app?