How to Choose a North Star Metric (and the Inputs That Drive It)

Short answer: a good north star metric measures value delivered to customers, not money taken from them, and rises only when customers are getting what they came for. Choose a count of a meaningful action completed, by active customers, in a period. Then break it into a few input metrics that teams can move directly. Check it with behavior data so it cannot go up while the experience gets worse.

Table of Contents

1. What a North Star Metric Is
2. A Test for Candidates
3. Examples
4. Input Metrics
5. Keeping It Honest
6. Key Takeaways

What a North Star Metric Is

It is one number that the whole company agrees represents customer value. Revenue follows from it, and is not the metric itself.

It serves two purposes: it tells everyone what "better" means, and it settles arguments between teams who would otherwise optimise different things.

It is not a replacement for all other metrics. It is the one at the top.

A Test for Candidates

The third question removes most candidates. "Time spent in app" rises when the product is confusing. "Messages sent" rises with spam. "Signups" rises with poor-quality traffic (vanity metrics).

Examples

Notice the qualifiers. "Orders delivered and not returned" is harder to game than "orders". "Messages that receive a reply" is harder to inflate than "messages sent".

The form is usually: number of [customers or accounts] who [did the valuable thing] in [a period].

Input Metrics

A north star is too slow and too broad for a team to act on week to week. Break it into inputs.

Take "weekly active accounts performing the core action":

Each is something a team can move, and each moves the north star. This is where targets belong (goals for UX and conversion).

Activation deserves its own definition, tied to the moment new users first get value (finding your aha moment).

Track the events that make these measurable:

deepsync("track", "core_action_completed");

Keeping It Honest

Any metric that people are rewarded for will be pushed on, sometimes in ways that harm customers. Three safeguards.

1. Counter-metrics. Pair the north star with numbers that should not get worse.

2. Behavior checks. Watch sessions of the customers behind the number. If "core action completed" is rising and recordings show users repeating the action because it failed the first time, the metric is lying. Rage clicks, errors, and repeated attempts on the path to your north star action are warning signs (frustration signals).

3. Periodic review. Once or twice a year, ask whether the metric still represents value. Products change.

A metric that can be increased by a dark pattern, such as making cancellation hard to improve retention, is measuring the wrong thing. Fix the metric, not the customer's options.

Key Takeaways

• A north star measures customer value, not company revenue.
• Reject any metric that can rise while the product gets worse.
• Break it into inputs that teams own.
• Guard it with counter-metrics and behavior checks.

Conclusion

A north star is useful only if moving it means customers are better off. Choose one that cannot be faked, and keep watching the sessions behind it.

Start free with DeepSync.