How a north star metric works
A north star metric is the single measure a business agrees best represents the value it delivers to customers. It sits above the usual reporting: not revenue, which is the result, and not traffic, which is an input, but the moment where a customer actually gets what they came for. A delivery business might choose orders delivered on time; a training provider, courses completed; a booking site, stays taken rather than bookings made.
The point is that it stands in for value received, not money collected. Revenue can be lifted for a quarter by discounting, or by pushing sales that will later be refunded. A well-chosen north star metric usually cannot be pushed that way, which is what makes it useful as a tiebreaker when teams disagree.
Underneath it sit the inputs that move it: the ordinary key performance indicators each team owns. The north star does not replace them. It explains what they are all supposed to add up to.
Why a north star metric matters
Most disagreements inside a business are really disagreements between metrics. Marketing wants more enquiries, sales wants better ones, operations wants fewer awkward jobs. A shared measure of delivered value gives all three a way to settle the argument without appealing to seniority.
It also guards against short-term wins that cost more than they bring. When the agreed number rewards customers being served properly, tactics that inflate one stage of the funnel while making the experience worse stop looking clever, because their damage shows up in the same place everyone is watching.
Common mistakes with north star metrics
Choosing a number that can only rise is the first. A cumulative total, such as registrations to date or total followers, can never fall, so it never warns you that anything is wrong. Choosing revenue is the second: it is the outcome you want, but it says nothing about whether customers got value, and it can be propped up in ways that hurt next year.
The rest are about discipline. A metric picked in a workshop and never mentioned again changes nothing. One that no team can visibly influence breeds fatalism instead of focus. And a business carrying several competing north stars effectively has none — at which point the label is decoration on an ordinary vanity metric.
How to act on it
Ask what a genuinely satisfied customer has actually done, then count that. Test the candidate against three questions: would it fall if service got worse, can teams influence it through work they control, and would you be comfortable if it were the only number reported this quarter. If the candidate fails any of them, keep looking.
Then break it into the inputs each team owns, and show both on the same report so the connection is visible rather than assumed. Review the metric itself rarely, since changing it often defeats the purpose, but do change it when the business changes what it sells. And keep the ordinary safeguards running beside it: a single number is a focus, not a substitute for looking at the business.