A single metric supplies directional consistency
Aliases: one metric that matters · directional alignment · north star
What it is
A north star metric is the one success count an organization orients toward in a given stage. Its first job is alignment: when growth, design, engineering, and operations collide, they share a direction of “better,” instead of cancelling one another with local KPIs. Weekly actives who completed a core task, or creators who produced a shareable result, are typical shapes. Singleness is not a claim that the world has only one important outcome. It is a claim that several parallel “most importants” tear the direction back apart.
Why it happens
Without a shared direction, teams optimize whatever they can move: acquisition wants registrations, content wants time, monetization wants orders, design wants task completion. Those locals sabotage one another in the interface, and each side can produce a chart that proves it is right. A north star recasts the argument from “whose metric matters more” to “does this change move the column we jointly named.” It aligns because it occupies the same slot in planning, experiment ship gates, and external narrative—not because it mathematically summarizes all value. The choice usually sits close to user value and still sensitive to product action: too close to revenue, experience work cannot contribute; too close to attitude, the business side will not treat it as direction. Consistency comes from repeated use, not from a naming ceremony.
Studying it
Compare experiment gates, roadmap items, and cross-team dispute topics before and after a north star is introduced, and see whether conflict shifts from “metrics against metrics” to “how to move the same column.” Run sensitivity checks on candidates: can real product levers move the count inside a reasonable cycle. A count that product action cannot move cannot align action. Coding planning documents also works: if every group still makes trade-offs on a private headline, the north star is a slogan. Separate the effect of “having one direction” from the effect of “having chosen the right one”; the latter is a different claim.
Where it stops holding
A north star replaces several simultaneous headlines, not the rest of observation. In monopolies, heavy regulation, or life-safety products, direction may have to come from a compliance constraint; adding a growth north star then creates a real conflict. Very early ventures change the goal weekly, and locking a count freezes learning. Packaging a composite index as a north star turns alignment back into a shadow fight over weights. Alignment is also not correctness: an error chased by everyone misdirects faster the more consistent it is.
Applying it
- Write the north star in one sentence: numerator, denominator, window, and the user value it sits next to.
- Wire experiment gates, quarterly plans, and cross-team review to that same column; groups may keep local metrics but may not veto the north star direction with them.
- Check that real levers this quarter can move the candidate; if they cannot, choose again rather than keep a slogan.
- Reselect at stage changes, and announce that the old star no longer sets direction, so two stars do not coexist.