A North Star Metric is the single number a company optimizes that best captures the value the product delivers to customers and that, over time, predicts revenue. It is not the same as revenue itself: revenue is the outcome, the North Star is the leading indicator. Picked well, a North Star aligns product, growth, and sales on a shared cause-and-effect chain. Picked badly, it drifts the company.
What good North Star Metrics look like
| Company | North Star | Why it works |
|---|---|---|
| Slack | Daily messages sent in active workspaces | Use of the product = value delivered |
| Airbnb | Nights booked | Captures supply, demand, and pricing |
| Notion | Weekly active editors | Editing = real use, viewing alone is shallow |
| Spotify | Time spent listening | Direct measure of value to user |
Each example shares three traits: it is a count of a customer action that delivers value, it is measured per-customer (not aggregated revenue), and it precedes revenue rather than equaling it.
Three tests for a candidate metric
Before adopting a North Star, run it through:
- The pause test. If the metric goes up while revenue goes down, what would that mean? If the answer is “nothing concerning,” the metric is wrong.
- The lever test. Can a product team, today, identify three concrete features or experiments that would move it? If not, it is too abstract.
- The customer test. Does the metric correspond to something a customer would describe as value if you asked them? If you have to translate, it is too internal.
A common failure is picking a metric that fails the customer test, like “users created” instead of “users who completed onboarding.” Account creation does not deliver value.
Why a single metric
Companies that pick five “primary metrics” pick none. The point of a North Star is to force trade-offs: when you must choose between two roadmap items, you pick the one that moves the North Star more. With five metrics there is always a justification for any choice.
The North Star is supplemented with input metrics (a tree underneath) and guardrail metrics (things that should not get worse, like churn or NPS). But there is one North Star.
How to operationalize
- Pick the metric. A 4-week exercise with the CEO, head of product, and CRO.
- Build the tree. Decompose into 3-5 input metrics. Notion’s weekly active editors decomposes into signups, activation rate, and reactivation rate.
- Set a target. Annual or 3-year. Make it ambitious; a target that requires a 10% improvement does not change behavior.
- Wire it to OKRs. Every team’s quarterly OKRs ladder up to either the North Star or one of its inputs.
- Review weekly. Single dashboard, single number on top.
Common pitfalls
- North Star = revenue. Revenue is the outcome. The metric should be the cause.
- Vanity North Star. “Users” or “downloads” without a quality dimension. Always add an action filter.
- Changing it too often. A North Star that changes annually is not a North Star. Commit for at least 3 years.
- Metric without a tree. A single number with no inputs cannot be moved by a team. Build the input tree.
Related
- PLG (Product-Led Growth) — where North Stars are most explicit
- Unit economics — the financial cousin
- NRR vs GRR — common guardrail metrics
- Notion — common home for the metric tree doc