North Star Metric: How to Choose One for a Startup, With Examples
What a north star metric is, a five-question test for choosing one, examples by business type, the input metrics that move it, and what to use when you are too early to have the data.
By Suren, Founder & Growth Strategist at Gufy · Published
A north star metric is the single number that best shows your customers are getting the value your product promises. Pick the count of the moment a customer gets what they came for, not the count that flatters the company.
This guide gives a test for choosing one, examples by business type, the supporting numbers that move it, and what to do if you are too early to have the data.
Why one number
A startup can measure hundreds of things. When every team reports its own, nobody can say whether the company as a whole is getting better. One shared number gives a common answer to "is this working?" and a way to choose between pieces of work: which one moves the number more?
We use the term "focus metric" with clients for the same idea. If the whole business is working to move one or two numbers, and those numbers are rising, the startup is on a good track.
A five-question test
Run any candidate through these.
- Does it count value the customer received? "Quotes sent" does. "Accounts created" does not.
- Does it lead to revenue? If it rises for a year, would income follow?
- Can the team move it? Product, marketing and support should each be able to name something they can do about it.
- Can you measure it now, and often? A number you can only see once a year cannot steer weekly decisions.
- Would it be hard to cheat? If you could raise it in a way that leaves customers worse off, it is the wrong number.
Examples by business type
| Business | A reasonable north star | Why |
|---|---|---|
| B2B SaaS, used by a team | Weekly active teams doing the core action | Value is delivered when the team uses it, not when one person signs up |
| Quoting or invoicing tool | Quotes or invoices sent per week | The action is the value |
| Marketplace | Completed transactions | Both sides got what they came for |
| Consumer subscription app | Subscribers active this week | Paying and using together |
| Social or dating app | Users who made a connection this week | Opening the app is not the value; the connection is |
| Agency or service | Outcomes delivered across all clients, such as meetings booked | Totals across clients show whether the method works |
Two notes on that table.
Match the period to how the product is naturally used. Some products are used daily, some weekly, some a few times a year. A founder may want people in the app every day, but customers decide the pattern, and you find it by observing what your best users do. Daily active users is the wrong measure for a product people need once a week.
For an agency example, the total matters more than any one client. One client might get five meetings and another fifty. Each needs attention on its own, but if the total keeps rising, the method is working.
Vanity numbers to avoid
- Sign-ups, installs, registered users. They only go up and say nothing about value. An app can report tens of thousands of installs while a small fraction of those people open it in a month.
- Page views and followers.
- Leads, when the business needs customers. A team steering by leads will produce more leads, including poor ones.
- Revenue alone, for the reason in the questions below.
The inputs that move it
A north star tells you whether things are working. It does not tell you what to do on Monday. For that, break it into the inputs that produce it.
For "weekly active teams sending quotes":
| Input | What it measures |
|---|---|
| New accounts | How many teams start |
| Activation rate | The share that send a first quote (see the aha moment) |
| Return rate | The share of active teams still active the next week |
| Depth | Quotes sent per active team |
Each input belongs to someone, and each suggests experiments. ICE and RICE scoring is a simple way to choose which experiment to run first.
Leading and lagging indicators
A north star is often a lagging number: by the time it moves, the cause is weeks old. Inputs are leading indicators.
This matters most when the true result arrives late. Take a tax product whose real value is tax saved, visible once a year. You cannot steer by that. You can steer by the things that produce it: customers signed up, transactions processed correctly, transactions approved by the user. If those are healthy all year, the annual figure is very likely to be healthy too.
The same logic applies to quarterly targets. A target checked only at the end of the quarter tells you that you missed it. Set a checkpoint halfway, with a rule agreed in advance for what changes if you are behind, and watch the leading indicators weekly.
If you are too early to have data
Most guidance on this topic assumes an analytics team. An early startup often cannot say how many of its users came back last week. That is the first thing to fix.
- Set up product analytics before you spend on acquisition. You cannot improve what you cannot measure, and ad platforms report installs and clicks, not whether anyone stayed. Best Product Analytics Tools for Startups covers the choices.
- Track a short list of events: signed up, did the core action for the first time, did it again. That is enough to see activation and retention.
- Choose a provisional north star from the five questions and write down why. Expect to change it once.
- Before launch, count learning. Customer conversations held, people who paid or pre-ordered. Those are the right numbers for a product that does not exist yet.
Why this comes before ad spend
Suppose an app spends $600 in a month on ads and gains 600 users, at $1 each. That sounds efficient. If 1% of them are still active a month later, six people remain, and the real cost is $100 per retained user. Whether that is acceptable depends on what a retained user is worth, which is a unit economics question. Without retention tracking the founder sees "600 users for $600" and keeps spending.
A bucket with a hole in it does not fill faster because you pour faster. A north star built on active, retained users makes the hole visible.
Common mistakes
- Choosing what is easy to count.
- Choosing revenue and nothing else.
- Changing it every quarter. Revise it when the product or customer changes, not when the number disappoints.
- No inputs. A north star without the numbers that drive it is a scoreboard nobody can affect.
- One number for two different customers. A marketplace may need one per side.
- Tracking it without acting on it. Review it weekly, alongside what you tested.
Common questions
What is a north star metric?
A north star metric is the single number that best shows your customers are getting the value your product promises. When it rises, the business should be healthier over time. It gives a team one shared answer to the question 'is this working?'
What is a good example of a north star metric?
For a quoting tool, quotes sent per week. For a marketplace, completed bookings. For a team chat product, active teams sending messages. For a meeting-booking service, meetings booked across all clients. Each counts the moment a customer receives value, not a moment that only matters to the company.
Is revenue a good north star metric?
Usually not on its own. Revenue tells you what customers paid, which follows from the value they received and arrives later. A team steering by revenue alone can raise it for a while in ways that hurt customers. Choose the measure of value delivered, and track revenue alongside it.
What is the difference between a north star metric and a KPI?
A north star metric is one number for the whole company. KPIs are the many measures individual teams use to manage their own work. The useful KPIs are the inputs that move the north star, such as new sign-ups, activation rate and how often active users return.
Can a startup have more than one north star metric?
One is the aim, because the point is focus. Some products need two, for example a marketplace with buyers and sellers. More than two or three and nobody is steering by anything.
Where Gufy fits
Gufy sets up the measurement that makes a north star usable: product analytics with a clear event plan, dashboards a founder can read, and acquisition reported against retained users. We then work on the inputs, from onboarding to lifecycle messaging.
If you cannot say how many of last month's new users are still active, start there.