What Is a North Star Metric and How to Find Yours for B2B SaaS
Most early-stage SaaS teams I have worked with track revenue, churn, and monthly active users. All three are useful. None of them is a North Star.
A North Star metric is not the same as a KPI. A KPI measures performance against a target. A North Star measures whether the product is doing the core thing it exists to do for users. The distinction sounds subtle. In practice, it determines whether your team is building toward actual product value or toward numbers that can look healthy while the product erodes.
What a North Star Actually Is
The definition I have found most useful comes from Amplitude's product work and Shreyas Doshi's writing on the topic. A North Star metric has three properties: it captures the core value the product delivers to customers, it predicts long-term revenue rather than reflecting it, and it is something the product team can directly influence.
Revenue is not a North Star. It is a lagging indicator of value. If revenue drops, something already went wrong upstream. A North Star catches problems earlier, because it lives closer to the moment of value delivery.
Amplitude describes it as "a metric that represents the core value your product delivers to customers." That framing is useful because it forces the question: what is the core thing we do for users? Not what we sell. What we do for them.
The Three Qualities of a Useful North Star
It represents value delivered, not presence. "Monthly active users" is a presence metric. It tells you someone showed up. It does not tell you whether the product did anything useful for them. A North Star should represent a moment when the product delivered its core value. For a communication tool, that might be conversations that included at least one external contact. For a project management tool, it might be projects moved from in-progress to complete. The difference is the difference between measuring the door and measuring what happened inside.
It leads revenue. When your North Star improves, revenue should follow, with a lag. When it drops, churn should follow, with a lag. If there is no relationship between your North Star and future revenue, you have picked the wrong metric. In my experience, teams that struggle to articulate this relationship usually discover they have chosen a proxy that feels meaningful but is too disconnected from commercial outcome to be useful as an organizational signal.
The product team can influence it. This eliminates net revenue retention (too dependent on sales and CS), NPS (too dependent on support experience and competitive context), and brand metrics (almost entirely outside product's control). A North Star should be something where a product decision made this sprint could move the number within 30 to 90 days.
North Stars by Product Type
Different SaaS categories tend toward different North Stars, and knowing the patterns can accelerate finding yours:
For productivity tools: actions that represent a completed workflow, not a started one. The metric should capture when the product actually saved someone time or effort, not when they logged in hoping it would.
For collaboration tools: the network action, meaning the moment when a second or third person joined a workflow initiated by someone else. Collaboration products deliver value through connection, so the North Star should capture connection, not individual usage.
For analytics and intelligence tools: insights acted on, not reports generated. Any product can generate reports. The moment of value is when a user does something differently because of what the product showed them.
For marketplace or network products: the interaction that creates value for both sides, not just the number of buyers or the number of sellers. Airbnb's North Star is nights booked, not listings available, because a listing without a booking has delivered no value to anyone.
How to Find Your Own
The fastest method I have used is what I call the value delivery interview. Ask your ten most retained users one question: what is the one thing this product does for you that you would miss most if it disappeared tomorrow? Do not prompt with product features. Listen for verbs, not nouns.
The verbs reveal value. "It reminds me," "it tracks for me," "it generates the first draft," "it catches the error before I send." Each verb is a candidate for your North Star.
Then build a simple table. List each candidate. For each, ask: does improving this predict that users stay longer? Can a product decision in the next sprint move this number? If the product hits zero on this metric for a week, would revenue follow downward within 60 days?
The candidate that answers yes to all three is your North Star.
The Most Common Wrong Choices
Sign-ups. Sign-ups measure marketing, not product. A product can have growing sign-ups and declining retention simultaneously. This is exactly the situation where a North Star built on sign-ups becomes dangerous, because the number looks healthy while the product erodes.
Daily active users without definition. "Active" needs to mean something. A user who logs in and immediately leaves is not the same as a user who completes a core workflow. Define active as the minimum meaningful action, then track that. Undefined "active" is almost always too loose to be useful.
Revenue itself. Revenue is what you are trying to predict and grow. It cannot predict itself. A North Star built on revenue gives the team no signal about what is actually working before it shows up in the monthly close.
Using It Weekly
A North Star is only as useful as the conversation it enables. In my experience, the teams that benefit most from having one are teams that reference it weekly, not quarterly.
The weekly question is simple: did this metric go up or down relative to last week, and what product or behavior change explains the movement? The discipline of answering that question every week, with evidence, is what transforms a North Star from a slide into a tool for alignment.
One metric. One question. Every week. That is the practice that makes the North Star worth having.