North star metric

A north star metric is the single number that counts how often users take the action that shows your product delivered its core value. It is not a count of visits or logins. It tracks the behavior that matches your value proposition, so the whole team can steer onboarding, product and marketing work toward one shared outcome.

How it is measured

Start from the action that proves a user got value. Then count it.

North star metric = number of users (or actions) completing the core value action in a set period, usually a week.

Amplitude's product benchmark report draws the line clearly: Just because a user returns to your site doesn’t mean they’re active. The report describes the North Star as an action that aligns with your value proposition.

A concrete case

Take a travel app. The report's example is the difference between “opening the app” vs. “booking a trip.” App opens look healthy on a dashboard and say little about value. Trips booked per week is the number that moves when the product works.

To find your equivalent, the same report suggests you Use cohort analysis to uncover common behaviors, paths, and milestones that correlate with activation. Compare what your retained users did in their first week with what everyone else did. The action that separates the two groups is your candidate.

How to use it this week

Once it is defined, it gives your experiments a target. The report notes that Understanding your North Star will help you develop strategies to nudge users toward it. In practice, that means onboarding changes, emails and in-app prompts all get judged by whether more users reach the core action, and sooner.

Log each idea in a growth experiment backlog with the north star as the success measure. Check it against your activation rate.

Usual mistakes

• Picking a traffic or engagement count, such as visits or sessions, that rises without users getting value.
• Choosing revenue. It lags behind user behavior and is hard to move with a single experiment.
• Tracking three "north stars" at once, which removes the point of having one.
• Defining it once and never testing whether it actually predicts retention.

It matters most when your team is small and every hour of work needs one clear target.