Skip to main content
← All guides
Tool + guide · September 2026 · 6 min read

The north star metric

One number that says whether customers are getting what they came for. Bring your candidate below and have it tested — including against the question that breaks most of them.

Short answer

What is a north star metric?

A single number capturing the value your product delivers to customers, used to focus everyone on one thing. The operative words are “to customers” — revenue and signups measure what you received, which is why neither qualifies however important they are.

The short version

  1. Three requirements that cannot be traded off: it measures customer value, it moves before revenue, and a team can influence it this quarter.
  2. Revenue cannot be a north star: it is a lagging indicator, and by the time it moves the decision is a quarter old.
  3. The test that breaks most candidates: name a way the number could rise while customers are worse off.
  4. One. Two north stars restore exactly the ambiguity the idea exists to remove.
Test your candidate ↓

Free · No signup · Nothing leaves your browser

Bring a candidate, and we will try to break it

This does not suggest a metric — nobody can do that without knowing your product. It takes the one you were going to put on the wall and tests it against the criteria, including the one most teams skip.

  1. 1.Does it measure value the customer receives — not value you extract?

    Critical

    Revenue, seats sold and page views measure what you got. A north star measures what they got, which is the only thing that keeps producing revenue.

  2. 2.Does it move before revenue does?

    Critical

    A lagging metric tells you about a decision made last quarter. The point of a north star is to be the earliest honest signal you have.

  3. 3.Can a team change it this quarter by doing their job well?

    Critical

    If nobody can influence it, it is a weather report. Teams disengage from numbers they cannot move.

  4. 4.Can anyone in the company explain it in one sentence, without a formula?

    A metric people cannot restate is a metric people cannot use. Composite indices score well on paper and are ignored in practice.

  5. 5.Can you measure it today, weekly, without a new data project?

    A north star you will be able to compute next quarter is not steering anything this quarter.

  6. 6.Is it one number, rather than a set of them?

    The value of a north star is that it forces trade-offs to be made explicitly. Three north stars is the same as none.

The question most teams skip

Can you name a way this number could go up while your customers are worse off?

Think hard rather than fast. Time in app rises when people cannot find things. Sessions rise when the job takes several attempts. Tickets resolved rises when the product creates more problems.

0 of 6 answered

Metrics that pass every checklist and still fail this one

  • Time spent in the apprises when people cannot find what they came for
  • Number of sessionsrises when the job takes several attempts instead of one
  • Support tickets resolvedrises when the product creates more problems to resolve
  • Signupsrises when you buy traffic that never comes back
  • Features shippedrises when nobody is checking whether anyone wanted them
  • Revenuerises when you raise prices on customers who are already leaving

None of these is banned — each is the right metric somewhere. The point is that the checklist alone would have waved them all through.

What this cannot know. Whether your candidate actually leads revenue at your company — only your own history answers that, and most startups have too little of it to tell. A north star is also a focusing device rather than a strategy: it says what to watch, never what to do about it.

If the product is new enough that no metric has history yet, the question underneath is still whether anyone wants it — which the Go/No-Go rubric asks directly.

Test the idea behind it →

A perfect metric on a product nobody wants measures nothing beautifully. A 15-minute session ends with a written GO / WAIT / NO-GO and the reasoning behind it. Free tier, no card.

The question that breaks most candidates

Every north star article gives you a checklist, and most candidate metrics pass it. Here is the question the checklists leave out:

Name a way this number could go up while your customers are worse off.

Time spent in the app rises when people cannot find what they came for. Sessions per week rises when the job takes three attempts instead of one. Support tickets resolved rises when the product generates more things to resolve. Each of those sails through a checklist: they measure customer behaviour, they lead revenue, teams can move them, they are simple and measurable and singular.

And each of them goes up on your worst quarter — and the team, watching the wall, will be pleased. That is not a metric problem you can manage around. It is a metric that will actively mislead you at the exact moment you most need to be told the truth.

If you can name such a way, the candidate is out, whatever else it scores. If you cannot, write down that you tried — because the next person to propose a metric will not, and the record is what stops the conversation restarting.

Four that hold up

Not because they are famous — because each one is hard to grow while making the product worse.

Airbnb

Nights booked

A night booked is a guest housed and a host paid. Both sides got what they came for.

Spotify

Time spent listening

Unusually, time here is the value rather than a symptom of confusion — people came to listen.

Slack

Messages sent within a team

Measured within a team, not across the product, because the value is a team communicating.

Amplitude

Weekly querying users

Not logins. Someone asking a question of their data is someone getting the thing they bought.

What a north star cannot do

It cannot tell you what to do. It says what to watch. Teams that confuse the two end up optimising the number rather than the thing the number was standing in for — which is the same failure as gaming, arrived at sincerely.

It cannot survive without input metrics. A single number a quarter away from anyone's daily work is decoration. The useful structure is the star plus the two or three things that visibly drive it.

It cannot be validated early. Whether your candidate really leads revenue is a claim about your own history, and most startups have far too little history to check it. Early on, a north star is a hypothesis you are agreeing to act on, and it is more honest to say so.

And it cannot fix a product nobody wants. A perfect metric on a product with no demand measures nothing beautifully.

Before the metric, the demand

Choosing a north star assumes there is value being delivered and the question is how to see it. If that assumption is the shaky part, the metric will be immaculate and flat.

A 15-minute session works through seven criteria — is the problem real, does a market exist, will they pay — and ends with a written verdict and the reasoning.

Test the idea behind the metric →

15 min · free tier, no card

Frequently asked questions

What is a north star metric?+
A single number that captures the core value your product delivers to its customers, used to focus a company on one thing. The important half of the definition is "to its customers": revenue and signups measure what you received, not what they got, which is why neither works as a north star even though both are important numbers.
What makes a good north star metric?+
It measures value the customer receives, it moves before revenue does, and a team can influence it within a quarter. Those three are not tradeable — a number that misses any one of them cannot steer anything. Beyond them: anyone should be able to explain it in a sentence, you should be able to measure it weekly today, and it should be one number rather than an index of several.
Can revenue be a north star metric?+
No, and it is the most common wrong answer. Revenue is a lagging indicator: by the time it moves, the cause happened a quarter earlier and the decision that would have changed it is long past. A north star has to be the earliest honest signal you have, which means it sits upstream of money rather than being money.
What is the gaming test?+
Name a way the metric could go up while your customers are worse off. Time spent in the app rises when people cannot find what they came for. Number of sessions rises when the job takes several attempts instead of one. Support tickets resolved rises when the product creates more problems to resolve. If such a way exists, the metric will rise on your worst quarter and the team will be pleased about it, which is disqualifying however well it scores on everything else.
How many north star metrics should a company have?+
One. The entire value of the idea is that it forces trade-offs to be made explicitly — when two teams want opposite things, the north star decides. Two north stars restore the ambiguity the exercise was meant to remove, and three is the same as none.
Is this checker free, and does my data leave the browser?+
It is free and requires no account. Everything is computed in your browser — nothing you enter is sent to us or stored anywhere.

Related guides