North Star Metric

Origin & evolution

The term was popularised by Sean Ellis around 2010, drawing on his growth work at Dropbox and later through GrowthHackers. Ellis used it to name a practice he had observed in high-growth companies: one metric that captured real value delivered, around which all activation, retention, and referral thinking could align. There is no single canonical inventor of the phrase, and Ellis himself was describing a pattern. He was not inventing a technique from scratch. The modern operational shape of the framework came from Amplitude, particularly through John Cutler's writing and the Amplitude North Star Playbook📒PlaybookCustomer SuccessA standard operating procedure for CSView reference → published in 2019, which codified the relationship between the North Star and its upstream inputs and gave product teams a repeatable method for choosing and using the metric. The specific examples most cited, nights stayed for Airbnb, daily messages sent for Slack, hours watched for YouTube, come from that playbook era.

How it works in practice

The North Star Metric has two components. The first is the anchor metric itself: one number that captures value delivered to customers in a way that correlates with sustainable business growth. The second is a small set of input metrics, typically three to five, each measuring a lever that moves the anchor.

The Ellis test, refined by Amplitude, gives three criteria for evaluating a candidate North Star. It should capture the value the product creates for customers, not just activity on the product. It should move in step with leading indicators of business success. And it should have the quality that if it doubles, the business fundamentally doubles with it.

A worked example. A professional-network product might identify "weekly meaningful connections made" as its North Star: a direct proxy for why members show up at all. The input metrics underneath it might be profile completeness rate (affects relevance of matches), search-to-message conversion (measures whether discovery leads anywhere), and second-message reply rate (measures whether connections survive the opener). Each input has a clear owner and a clear theory of how improving it raises the North Star.

The discipline is in the selection. Teams often land on activity metrics: signups, DAU, page views. These can rise while value delivered falls. The stress test is to ask whether a 2x rise in the candidate metric could happen without any improvement in actual customer outcomes🏆OutcomeStrategyA desired business or user outcomeView reference →. If the answer is yes, the candidate is measuring activity, not value.

When to use it (and when not)

The North Star Metric earns its place when a team has many dashboards but no shared answer to "are we winning?", when growth work optimises for conversions that do not connect to durable value, or when product-led growth strategy requires a single number that aligns the whole team across activation, retention, and referral. It is particularly useful when a new leadership cycle starts and the team needs🫀NeedUserA user need, pain, desire, or constraintView reference → one thing to orient around.

It fails when forced onto products that serve genuinely different customer jobs🔨JobUserJob To Be Done: what the user is trying to accomplishView reference → with no common measure. A platform serving both buyers and sellers may need two metrics held in tension, not one. It also fails when treated as a ceiling: the North Star is a lens for alignment and prioritisation, not a target to be gamed. Teams that optimise directly for the NSM number without maintaining the input model tend to hollow it out.

In the Unified Product Graph

North Star Metric is a collection framework in the metrics category. All slots map to the metric🔢MetricStrategyA unified metric that measures progress, health, or behaviour across the productView reference → entity type, which is the correct distinction: the North Star itself is a metric🔢MetricStrategyA unified metric that measures progress, health, or behaviour across the productView reference →, and each input is also a metric🔢MetricStrategyA unified metric that measures progress, health, or behaviour across the productView reference →. What differentiates them is their relationship and position in the collection, not their type.

• The North Star slot holds a metric🔢MetricStrategyA unified metric that measures progress, health, or behaviour across the productView reference → entity representing the anchor, the number that captures customer value.
• Input Metric 1, Input Metric 2, and Input Metric 3 each hold metric🔢MetricStrategyA unified metric that measures progress, health, or behaviour across the productView reference → entities representing the upstream levers that move the anchor.

Modelling the framework this way in the Unified Product Graph means the North Star metric🔢MetricStrategyA unified metric that measures progress, health, or behaviour across the productView reference → can also appear in OKR key results📈Key ResultStrategyA measurable result tied to an objectiveView reference →, in experiment🧫ExperimentValidationA test designed to validate a hypothesisView reference → success criteria, and in growth team outcome nodes. The input metric🔢MetricStrategyA unified metric that measures progress, health, or behaviour across the productView reference → entities connect to the features⭐FeatureProduct SpecificationA product capability or featureView reference → and experiments that are designed to move them. Nothing is duplicated: one canonical metric record, referenced wherever it is relevant.