North Star Metric.

What a North Star Metric is

When a product has dozens of indicators, it's easy to get lost in them. Conversion, retention, activation, average order value, cost of acquisition - everything matters, and that is exactly why it's unclear what to look at first. The North Star Metric, a “guiding metric”, solves precisely this problem.

It is one key indicator that reflects two things at once: the value the user gets and the growth of the business. If it grows, people are getting what they came to the product for more often, and the company earns more. If it stands still, there is no movement, however much the other numbers grow.

An important clarification: such a metric is not a line in a report for the owner. It is a reference point for the whole team at once: marketing, product, development and support. Everyone understands what they influence, and they argue not about whose task is more important but about which hypothesis will move the shared indicator faster.

Let's look at what it looks like in practice, how it differs from ordinary indicators and how to choose your own.

How it differs from ordinary indicators

Any business has indicators of three types, and confusing them is the most common mistake.

Activity indicators answer the question “how much did we do”: the number of posts, reach, the number of ads launched. They speak about the team's work, not the state of the product.

Financial indicators answer the question “how much did we earn”: revenue, profit, average order value. They matter, but they react with a delay and are poor at suggesting what exactly to do today.

The guiding metric sits between them. It shows whether people are getting value right now, and at the same time it is tied to money: if people use the product more often and more deeply, revenue follows.

The check is simple. Ask yourself: if this indicator doubled, would the business be noticeably better off? For reach the answer is usually “not necessarily”. For a correctly chosen guiding metric - “yes, almost certainly”.

What it looks like in examples

For an audio streaming service, the guiding metric is minutes of listening: if a person listens more, they are satisfied and come back. For a short-term accommodation rental service - the number of nights booked: not views and not sign-ups, but the fact of use. For a corporate messenger - the number of messages sent within teams.

From Russian examples: for a music service it is the number of tracks listened to per user per month, for a foreign language school - the number of lessons held per student.

Note what they have in common: in every case the metric describes the action the person came for. Not registration, not installing the app, not subscribing to a newsletter, but the use itself.

For retail and services the logic is the same. For a food delivery service it is the number of orders per customer per month, for a chain of car repair shops - the number of cars serviced with a repeat visit, for a construction company - the number of completed projects for which the client is willing to recommend the contractor.

How to choose your own

Selecting a guiding metric usually takes several attempts. It helps to move step by step.

1. Formulate which customer task you solve. Not what you sell, but what changes for the person after the purchase.
2. Find the action in which that value shows up. Usually there is one, and it is easy to recognise: those who perform it come back more often than the rest.
3. Check whether it can be measured from a single data source. If the indicator has to be assembled by hand from three spreadsheets, it won't become a daily reference point.
4. Check the link to money. Take the data for a past period and see whether revenue grows together with this indicator.
5. Make sure the team influences it. The exchange rate also affects revenue, but you can't control it.

To get through this path quickly, you need data on customers and enquiries collected in one place. How to set that up is covered in our piece on CRM systems, and ways to link promotion spend to returns - in our article on data analysis.

What this metric should not contain

The main trap is indicators that are pleasant to show but change nothing. Likes, followers, views, the number of app installs. They grow, the report looks upbeat, and there are no sales.

Such numbers aren't useless in themselves - they help assess individual pieces of work. But in the role of the main reference point they are harmful: the team starts optimising precisely them, and so doing what increases the number of likes rather than the value.

The second trap is a metric influenced by only one department. If the indicator sits entirely in marketing's zone, product and support stop considering it their own, and the point of a shared reference is lost.

The third is a formula that is too complex. An indicator calculated from five coefficients is one nobody keeps in their head, which means it doesn't work as a reference in daily decisions.

What to do after choosing

The chosen metric is not a monument. It has to be worked with.

First: break it down into its components. The number of orders per customer per month is made up of the number of active customers and purchase frequency - and each part can be worked on separately. This breakdown turns an abstract goal into a list of specific tasks.

Second: make it visible. The indicator should be on a shared screen or in a regular team mailing, not in a file on the analyst's computer. A reference point nobody sees is not a reference point.

Third: test hypotheses rather than argue about them. Every change is an experiment with a clear expectation: what should move and by how much. The method is described in our article on A/B testing.

Fourth: review it once a year or when the business model changes. A metric chosen for the launch stage rarely suits the retention stage.

Does a small business need this

The approach came from product companies, but for a small business it is even more useful: there are fewer resources, and the price of spreading yourself thin is higher.

The practical benefit is that arguments about priorities disappear. When the team has one shared reference point, the question “launch ads or redo the website” is decided not by a vote but by assessing which will move the main indicator more.

The second benefit is a clear link between work and result. An employee sees how their task affects the overall number, and that works better than an abstract call to work harder.

You can start without complex systems: choose an indicator, learn to calculate it once a week and watch how decisions affect it. The role of analytics in this process is covered in a separate piece, and how to retain customers when the indicator comes down to repeat purchases - in our article on bringing customers back.

Think about it: which number in your business shows that the product really solves the buyer's task? That is your guiding metric. And if you need to link it to advertising campaigns and budget, that is a job for advertising and contextual ads setup.