Quick answer
A North Star Metric is a primary measure representing the value customers receive from a product and leading sustainable business outcomes. A good candidate describes a meaningful value event, uses the correct customer or account unit, can be influenced by the product and predicts longer-term health without merely restating revenue. Teams pair it with a small set of actionable inputs and non-negotiable guardrails. Selection requires strategy, customer research and historical validation; adoption requires a versioned definition, ownership, decision rituals and regular revalidation as the product or market changes.
What is a North Star Metric?
A North Star Metric is a shared measure of value delivered to customers. It gives teams a common direction between day-to-day inputs and slower outcomes such as retention, expansion and sustainable revenue. It should embody product strategy rather than summarize activity.
One number cannot explain a business. A North Star works with inputs, financial outcomes and guardrails. Its advantage is focus: teams can debate the value mechanism and coordinate decisions without turning a KPI catalogue into competing definitions of success.
Why the North Star framework emerged
Digital products generate abundant events, but abundance can weaken decisions. Page views, clicks and active-user counts are easy to retrieve without showing whether people achieved the promised result. Teams needed a measure connecting use to customer value and strategy.
Amplitude's North Star Playbook formalized the framework around a metric and inputs. The emphasis is alignment and learning, not a permanent slogan. Its account of changing the metric shows that a North Star can evolve with strategy and intended customer impact.
Criteria for a strong North Star Metric
A strong candidate represents value received, not internal activity. Product and go-to-market teams can influence it, it occurs early enough to guide decisions and plausibly relates to durable customer and business health. It must be understandable and measurable.
Specify the actor, action, quality threshold, frequency and window. In B2B, account collaboration may represent value better than individual logins. In marketplaces, use a balanced transaction or successful match rather than demand activity that ignores supply welfare.
Value
State the customer problem and observable moment of received value.
- What changes for the customer?
- Which behavior demonstrates it?
Candidate
Create measures with explicit units, events, windows and quality thresholds.
- Who or what is counted?
- Can the definition be audited?
Validate
Test whether candidates lead durable customer and business health.
- Is the relationship stable?
- What alternative explains it?
Decompose
Connect the North Star to actionable inputs and protective guardrails.
- What can teams change?
- What must not worsen?
Operate
Use the metric in decisions while reviewing its validity and incentives.
- Which decision changes?
- How could the measure be gamed?
Build a North Star input tree
Inputs describe how the organization may influence the North Star. They often include breadth, such as the number of activated accounts; frequency, such as valuable cycles per account; depth, such as collaborators or completed work; and quality, such as successful or accepted outcomes.
A mathematical decomposition can expose double counting and clarify ownership, but a driver tree is still a hypothesis. Historical relationships may reflect selection, seasonality or a common cause. Test interventions at the input level and observe their effect on the North Star and downstream outcomes.
How to choose a North Star Metric
Begin with strategy and customer research. State whose problem the product solves, what value looks like, its frequency and long-term success. Generate several candidates before discussing dashboard fields so instrumentation does not dictate the concept.
Score candidates against value, influence, leading relationship, measurability, durability and gaming risk. Analyze historical cohorts and lags, inspect segment differences, and run a cross-functional workshop to surface assumptions. Pilot the definition in decisions before making it an incentive target.
- Customer and problem explicit
- Value event observable
- Unit and identity rule defined
- Quality threshold included
- Time window matches use
- Candidate alternatives compared
- Historical cohorts examined
- Lag to durable outcome tested
- Input tree documented
- Guardrails assigned
- Gaming risks reviewed
- Revalidation date scheduled
North Star Metric example
Hearthline's hypothetical process begins with household value rather than the events already easiest to count. App opens could rise because of confusing reminders, while recipe saves may represent aspiration without completion. A completed feasible plan is closer to the proposed outcome but still requires validation.
The candidate avoids tracking every meal. Privacy, burden and accessibility constrain measurement. Inputs help teams investigate onboarding and repeat planning, while guardrails discourage activity created through pressure or waste.
Hearthline is a hypothetical household meal-planning app. Teams currently celebrate app opens, recipe saves and grocery-list exports separately, but none is a shared expression of whether a household successfully plans and prepares useful meals.
The team states its value theory: households benefit when they choose feasible meals together and complete those planned meals with less coordination burden. It interviews households to understand substitutions, skipped meals and accessibility needs.
A candidate North Star is weekly households completing at least a minimum number of self-selected planned meals, with a confirmation rule designed to avoid demanding or intrusive tracking. Raw opens and saves remain diagnostics.
Hearthline checks whether the candidate precedes continued household use and appropriate paid retention across household types. It does not assume that correlation establishes the value event as the cause of those outcomes.
Potential inputs include eligible households reaching a first feasible plan, plan completion rate, active household members and repeat planning. Guardrails cover notification complaints, food-waste signals, accessibility failures and support load.
Teams use the tree to form hypotheses, but experiments retain their own primary metrics. The North Star is reviewed when strategy, meal workflows or measurement quality materially changes.
Hearthline and its metric candidates are hypothetical. A real North Star requires customer research, lawful measurement and organization-specific validation.
Validate predictive and causal claims
Compare cohorts at different candidate levels while accounting for tenure and eligibility. Check relevant segments and whether the North Star precedes retention or expansion. Look for nonlinear thresholds and high activity that signals difficulty rather than value.
Predictive association is not a causal guarantee. Motivated customers may both perform the value event and retain. Use experiments or appropriate causal designs to estimate the effect of specific interventions. Revalidate after major product, pricing, audience or instrumentation changes.
Use the North Star in planning and experiments
Use the metric tree to connect strategy, team problems and experiments. A team should explain which input it will change, for whom, through what mechanism and how it could contribute to the North Star. This creates alignment without making every test wait for a distant outcome.
Experiments still need sensitive primary measures and guardrails. Review North Star movement by mature cohort and interpret delays. Discuss customer evidence and mechanism alongside the number so metric change does not substitute for understanding.
Govern definitions, incentives and access
Create a semantic contract for events, formula, windows, identity, exclusions, latency and owners. Version changes and maintain a bridge where possible. Audit tracking because a stable dashboard can conceal broken events or changing eligibility.
Avoid attaching individual compensation directly to an untested North Star. Goodhart's law warns that a measure can degrade when it becomes a target. Pair it with quality, trust, financial and customer guardrails, and give teams a safe route to report gaming or adverse effects.
Limitations and common North Star mistakes
Some businesses serve multiple sides or have distinct value propositions that resist one measure. A portfolio may need a primary North Star with explicit balanced measures, or separate product-level stars beneath a shared strategic outcome. Forcing false unity is worse than acknowledging the structure.
Common mistakes include choosing revenue, active users or engagement time without a value theory; creating a composite no one understands; changing definitions silently; and treating correlation as causation. A North Star is a decision tool grounded in strategy. It cannot repair weak product-market fit or replace customer judgment.
Choose the measure that best represents customers receiving the promised value, then surround it with actionable inputs and protective guardrails.
Frequently asked questions
Can revenue be a North Star Metric?
Revenue is usually a business outcome rather than a direct measure of customer value. It belongs beside the North Star, though value-based transactions may form part of a carefully defined metric.
How many North Star Metrics should a company have?
The framework aims for one primary company or product measure. Complex portfolios may need nested product metrics or balanced measures, with their relationship stated clearly.
What is the difference between a North Star and a KPI?
The North Star expresses the central customer-value theory. KPIs monitor supporting performance, financial results, risks and operations. A company usually has many KPIs but one primary North Star.
How often should a North Star Metric change?
Only when strategy, customer value, product structure or evidence materially changes. Review it on a planned cadence, but avoid cosmetic changes that destroy comparability.
Should every experiment use the North Star as its primary metric?
No. Many tests need a nearer, more sensitive outcome. State the proposed link to the North Star and monitor downstream effects when the sample and time horizon allow.
Sources and further reading
- Amplitude: The North Star Playbook ↗Publisher framework for selecting a North Star and connecting it to actionable inputs
- Amplitude: Introducing the North Star Playbook ↗Concise publisher definition and implementation context
- Amplitude: Evolving Our Product North Star Metric ↗First-party case showing how a value metric changes with product strategy
- Royal Statistical Society: Goodhart's Law and Performance Measurement ↗Publisher discussion of how targets can distort the measures used to manage performance