The North Star Metric Trap

A fintech client of mine had one number on every screen: weekly active users. It climbed for eleven straight months. The board loved it. Then renewals came in, and a third of those active users left. When we finally watched the sessions, the pattern was plain. People were logging in over and over because they couldn't finish what they came to do. The North Star wasn't measuring success. It was measuring struggle, dressed up as engagement.

How a North Star becomes a vanity metric

The North Star Metric was supposed to be a discipline. One number that captures the value customers receive, so a growing company doesn't scatter its attention. In practice, most teams skip the hard part — defining customer value — and go straight to picking a number that already looks good. Sessions. Sign-ups. Messages sent. Time in app.

These numbers share a convenient property: they go up when the company does almost anything. Ship a feature, send a campaign, add a notification — the line moves. That is exactly what disqualifies them. A North Star that rises regardless of whether the customer succeeded isn't guiding anything. It's applause.

The test is simple. Can this number go up while a customer is failing? For most North Stars I've audited, the answer is yes. Engagement rises when users are lost. Deflection rises when customers give up. Time in app rises when the task takes too long.

If your North Star can climb while your customers struggle, it isn't a star. It's a mirror.

Lagging indicators wearing leading badges

The second failure is quieter. Teams pick a real outcome — net revenue retention, renewal rate, lifetime value — and treat it as a steering instrument. But those numbers are verdicts, not signals. By the time retention drops, the moments that caused it are months old. Customers don't leave suddenly. They drift away. Revenue metrics record the drift after it has finished.

A useful North Star sits upstream of the money. It measures the moment value is delivered, not the moment value is billed. For a logistics platform, that might be shipments delivered on time without a support contact. For a B2B tool, the weekly report a customer actually forwards to their boss. When that moment happens reliably, retention follows. When it stops, you get months of warning instead of a post-mortem.

When making the number replaces serving the customer

The third failure is cultural. Once a number carries bonuses and board slides, teams stop asking what the customer needs and start asking what moves the metric. The two questions can look identical for a quarter. Then they diverge.

I've watched this happen across banks, hospitals, and software companies. Notification schedules tuned to inflate daily actives. Onboarding steps cut, not because customers didn't need them, but because they slowed activation. Support deflection celebrated while repeat contacts climbed. Nobody was lying. Everyone was optimizing. Optimization is not understanding.

The tell is in the meetings. When a metric review spends an hour on how to move the number and no time on what the number means, the North Star has already inverted. It no longer represents the customer to the company. It represents the company to itself.

Build it bottom-up, from real moments of success

A North Star worth steering by isn't chosen in a strategy offsite. It's found in the field. Start with moments where a customer demonstrably got what they came for, then work backward to a number that counts those moments and nothing else. The sequence I use:

• Collect twenty real success stories. Not survey scores — specific accounts, specific tasks completed, in the customer's own words.
• Find the shared moment. What actually happened when it worked? First report shared. First payment reconciled. First shift covered without a phone call.
• Define the metric as a count of that moment: how many customers experienced it this week.
• Stress-test it. Can it rise while a customer struggles? If yes, tighten the definition until it can't.
• Pair it with a why-layer — session evidence, verbatims, friction signals — so the number never travels alone.

The result is rarely elegant. 'Weekly accounts that reconcile in under ten minutes' looks worse on a slide than 'weekly active users.' It's also nearly impossible to fake, which is the point. A metric built from the customer's moment of success can only be moved one way: by producing more of those moments.

That fintech client replaced their engagement star with 'users who complete a transfer on the first attempt.' The number was embarrassing at first. Six months later it was the most trusted line in the company — because when it moved, everyone knew a customer had actually won.