How to Choose a North Star Metric That Actually Matters

Most teams pick their north star metric the way they pick a logo color: a meeting, a whiteboard, a vote, and then it's frozen on a dashboard nobody questions for two years. The result is a number that goes up while the business quietly goes sideways. A north star metric is supposed to be the one measure that tells you whether you're creating real value for customers — and choosing the wrong one is worse than having none, because a bad north star points a whole org confidently in the wrong direction.

I've watched smart teams optimize signups while churn ate them alive, and watched a five-person startup pick a single honest number and use it to say no to half their roadmap. The difference was never intelligence. It was the metric.

What a North Star Metric Actually Is

The term comes from Sean Ellis, who shaped it watching growth teams at Dropbox, Facebook, and LinkedIn. His framing is the one worth keeping: the north star metric exists to quantify the value customers get from your product over time. As Ellis put it in his conversation with Intercom, "The North Star Metric is really trying to quantify that value over time. If everything you're doing is about trying to expand that, you'll be in good shape for driving long-term growth."

Notice what that definition excludes. It's not revenue. It's not signups. It's not your monthly active users. Those are scoreboards — they tell you what already happened. A north star metric is a behavior, performed by customers, that means they got what they came for. When that behavior expands, value expands, and the money tends to follow.

Amplitude built an entire North Star Playbook around this idea, pairing the headline metric with a set of inputs — the levers teams actually pull day to day. That pairing matters, and we'll come back to it, because the single number alone is a trap.

Customer Value, Not Vanity

Here's the test I apply to any candidate metric: if this number goes up, did a customer's life get better?

Run "total registered users" through that test and it fails instantly. Someone signing up costs them nothing and tells you nothing. Run Airbnb's metric — nights booked — through it and it passes cleanly. A night booked means a guest trusted a stranger's home enough to sleep in it and a host got paid. Both sides won. ProductPlan lists it alongside Facebook's daily active users as a textbook case of a metric that captures the actual exchange of value rather than the appearance of it.

Spotify's choice is just as instructive. Their north star is time spent listening, not accounts created. As the case studies collected by Wudpecker point out, listening time is engagement and satisfaction made measurable — and it correlates tightly with retention. A user who listens twenty hours a week is not churning next month. The metric isn't a proxy for value; it is the value, counted.

The pattern across good north stars is that the number can only move if a customer does something that's genuinely good for them. Vanity metrics let you win without your customer winning. That's the whole distinction.

Leading vs. Lagging — and Why You Need Both

A lot of north star confusion is really a confusion between leading and lagging indicators.

Revenue is a lagging indicator. By the time it drops, the damage is months old. Your north star should sit upstream of revenue — it should be the customer behavior that predicts the money before the money arrives. Time spent listening today is leading; the renewal next quarter is lagging.

But here's where teams get burned. Brian Balfour, who built the growth practice at Reforge, argues bluntly in his essay on the subject that "blindly buying into the concept of the one metric that matters (OMTM) is a fatal oversimplification." His point is that a single north star, watched in isolation, hides problems: "Output metrics can hide customer growth problems percolating under the surface."

His fix is the structure that actually makes a north star useful. Distinguish output metrics (results — the north star itself, revenue, total active users) from input metrics (the actions you can directly influence). Outputs are lagging; inputs are leading. You don't set team goals on the output, because it's too far away and too slow. You break it into a small constellation of inputs, and when those move, the output follows.

So the honest answer to "leading or lagging?" is: your north star is a leading indicator of the business, but a lagging indicator of your inputs. Choose the north star to predict revenue, then choose two to four inputs that predict the north star. That's the whole machine.

The Trap of Optimizing the Wrong Number

Every metric is a target, and every target gets gamed — by your own team, with the best intentions.

Pick "pages viewed per session" and someone will paginate an article into ten clicks. Pick "time in app" and you'll start shipping dark patterns that trap people instead of serving them. The metric you choose silently rewrites your roadmap, because people optimize what they're measured on. This is exactly why Balfour warns against the single number: optimize one output hard enough and you'll strip-mine the inputs that quietly sustain it.

The defense is the customer-value test again, applied as a gut check before you commit: Can this number go up while my customer is worse off? If the answer is yes, you've found a vanity metric or a vulnerable one. Nights booked is hard to game in a way that hurts guests. Engagement-time metrics are easier to corrupt — which is why a team that picks one needs to pair it with a quality or retention input that keeps it honest.

A Method You Can Actually Run

Here's the process I'd give a team starting from scratch:

Name the core value. Finish this sentence: customers hire our product to ______. For Spotify it's "listen to music they love, instantly." Don't skip to numbers yet.
Find the behavior that proves it. What does a customer do when they're getting that value? Not signing up — the recurring action that only happens when the product works. That's your candidate north star.
Run the value test. If this metric climbs, is the customer unambiguously better off? If you can imagine the number rising while customers suffer, reject it.
Check that it leads revenue. Does this behavior, when it grows, reliably precede retention and expansion? If it lags revenue instead of leading it, it's a scoreboard, not a star.
Define two to four inputs. Per Balfour, decompose the star into the levers teams can move this quarter. The north star aligns; the inputs are where the work actually lives.
Write it down precisely, then leave it alone. Amplitude's playbook is right that the metric needs an exact definition — Spotify's "time spent listening" needs a time window and a definition of "listening." Then resist the urge to swap it every planning cycle. A north star you change quarterly was never one.

A good north star metric is less a measurement than a promise: this is the one number we'll defend, the thing we'll say no to other good ideas in order to protect. Most teams never make that promise. The ones that do tend to know exactly what business they're in — and that clarity, far more than any dashboard, is what compounds.

If you want to go deeper on what happens after launch — why the features you ship so often go unused even when the north star looks healthy — I wrote about that here.

Sources

Sean Ellis on charting a path toward sustainable growth — Intercom — Ellis's definition of the north star as quantifying customer value over time, and Airbnb's "nights booked."
The North Star Playbook — Amplitude — the North Star Framework pairing a headline metric with measurable inputs, and the need for precise definitions.
Don't Let Your North Star Metric Deceive You — Brian Balfour — the "fatal oversimplification" warning, output vs. input metrics, and leading vs. lagging indicators.
North Star Metrics — ProductPlan — Airbnb's "nights booked" and Facebook's daily active users as customer-value metrics.
What is a North Star Metric? 7 Best Case Studies in SaaS — Wudpecker — Spotify's "time spent listening" as an engagement-and-satisfaction metric tied to retention.