North Star Metric: How to Pick the One Number That Tracks Real Value
A north star metric is the one number that best captures how much value your product delivers to its users, chosen so that when it rises over time, the business is likely to be healthier too. It measures an outcome people got, such as notes revisited or invoices sent, not a count of visits or a bank balance. A small product should pick one and write down the reason.
What makes a number a north star?
Three tests decide it. First, it reflects value received by the user, not value captured by you. Second, it moves when the product gets better and stays put when it does not, so it is a leading indicator rather than a lagging one. Third, you can influence it with work you do this month. Revenue fails the first test on its own, because a customer can pay and get nothing, and page views fail the second, because they rise with a viral post and fall back again whatever the product does.
The metric is a choice about what your product is for. A team that picks “signups” is saying the product exists to collect names. A team that picks “reports shared with a second person” is saying the product exists to help people communicate results. The same code could carry either choice, and the numbers on the dashboard would push the work in different directions.
What does a good one look like for different products?
| Product | Candidate metric | Why it reflects value |
|---|---|---|
| A recipe-scaling app | Recipes scaled and then marked as cooked | Cooking is the goal, not scaling |
| A link shortener with stats | Links clicked by someone other than the creator | The product worked when others used the link |
| A habit tracker | People who logged a habit on at least four days of the week | Repeated use shows the habit is forming |
| A small marketplace | Transactions completed | Both sides got what they came for |
| A music-practice tool | Minutes of practice logged | Practice is the outcome the player wants |
Notice that each includes a verb of success, not of presence. “Opened the app” is the type of metric to avoid. If you are unsure, finish the sentence “the product worked when the user ...” and see what comes next. The answer, measured across all users in a period, is a draft north star.
How do you break it into inputs you can act on?
A single number is too coarse to steer by, so decompose it into a few drivers. For the recipe app, “recipes cooked this week” equals new users who scale a first recipe, plus returning users who scale another, times how often each of them follows through. Each driver maps to a different piece of work: onboarding for the first, reminders for the second and a clearer cooking view for the third. That is where the earlier metrics in this glossary fit: activation rate feeds the first driver and retention rate the second.
- Write the north star as a count over a period, such as the count in a calendar week.
- List three to five inputs that multiply or add up to it.
- Pick the weakest input and run one change aimed at it.
- Review all inputs together at a set time each week, so the review is a habit.
Do you need one before you have users?
You need a guess, held lightly. Before launch, write the sentence “the product worked when ...” and decide how you would count it. After the first fifty users, check whether the guess matches what your happiest people actually do, and change it if it does not. A north star chosen on day one and never revisited is a risk, because the early users may show you the real value was somewhere else.
Keep it modest at first. With ten users the number will be tiny and noisy, so record the raw count in a spreadsheet each week and read the people behind it. When you launch your product free, note the metric at launch week and again a month later. The comparison is the first evidence of whether you are building something that delivers value. The launch metrics to track list shows the supporting numbers that sit around it, and the music and audio ideas on the board offer a sense of how makers phrase outcomes for a niche audience.
What goes wrong when picking one?
- Choosing a vanity number. Total registered users only goes up, and page views follow campaigns. Neither shows that anyone got value.
- Choosing revenue and nothing else. Revenue matters, and MRR is the right number for it. As a north star it tells you about the past, not about value that is not yet paying.
- Changing it every month. A metric that moves with each new idea cannot show a trend. Revisit it when your understanding of the user changes, not when a month looks bad.
- Gaming it. If you reward the number directly, behaviour bends toward the number. A weekly count of “notes created” can be inflated by splitting one note into five. Pair it with a quality check, such as notes revisited.
- Having five north stars. The point of the name is that you steer by one. Supporting metrics are fine, but only one is first.
Note that the metric measures users’ results and not your effort. Hours spent, features shipped and posts written are inputs to your work, and they belong in a separate log. For a first measured week, see how to measure a launch.
Frequently asked questions
What is a north star metric?
It is the single number that best captures the value users get from your product, chosen so that growth in it signals a healthier business. Examples are messages answered, recipes cooked or transactions completed. It should measure an outcome for the user, change when the product improves and be something your work can influence.
Is revenue a north star metric?
Usually not on its own. Revenue shows that customers paid, but it lags behind value and says nothing about users who are getting benefit without paying yet. Track revenue as MRR or paying customers, and choose a north star that counts the outcome users came for, since that tends to lead revenue.
How many north star metrics should I have?
One. Supporting metrics can sit beneath it as inputs, such as activation, repeat use or sharing, but a team steers by a single number. If you find yourself defending two, you probably have two kinds of user or two products, and you should decide which to serve first.
Can a north star metric change?
Yes, and it should when you learn that your first guess did not match what users value. Change it deliberately, write down why and the date, and avoid changing it because a month looked poor. A metric that moves often cannot show a trend, so keep it for at least a quarter.
Written by Post your build ideas with AI assistance and edited for this site. Where a page names another product, the facts came from that product’s own pages on the date shown above; terms change, so check them before you rely on them. About the publisher.