The North Star Metric: the ultimate guide (and the half most guides skip)
A North Star Metric aligns your team around one number. The half most guides skip: turning it into input metrics and a ranked queue of scored experiments that actually move it.
The North Star Metric: the ultimate guide (and the half most guides skip)
A North Star Metric (NSM) is the single metric that best captures the core value your product delivers to customers, used to align every team around one number. That definition is the easy part, and almost every guide stops there. The hard part, the part that decides whether the metric changes anything, is connecting it to the experiments you actually run each week. This guide covers both: how to choose a North Star, and how to turn it into a ranked queue of experiments instead of a poster on the wall.
Why a North Star Metric matters for growth leads
Without a shared metric, teams optimize their own corner. Product ships features, engineering pays down debt, marketing chases signups, and each is locally rational while the whole stalls. A North Star Metric collapses that into one question every team can answer: did this move the number? It aligns prioritization, sharpens communication, and lets a growth lead defend a roadmap with a metric instead of an opinion.
What makes a good North Star Metric
A usable North Star clears four tests:
- Plain language. If two people read it and picture different things, it is not specific enough.
- A leading indicator. It should predict future success, not report past results. Revenue is a lagging output; the behavior that produces revenue is the leading signal.
- Tied to customer value. It measures the value the customer receives, not the value you extract. The job the customer hired your product to do, in Clayton Christensen's framing.
- Movable. Teams must be able to influence it through the work they actually do.
A common shorthand, attributed to Lenny Rachitsky, is to ask: "which metric, if it increased today, would most accelerate your business flywheel?"
Common pitfalls
- Lagging metrics. Revenue and churn report the past. Use them as guardrails, not as the North Star.
- Vanity metrics. Total registered users only ever goes up and tells you nothing about value delivered.
- Unmovable metrics. If no team can point to the work that shifts it, it cannot guide prioritization.
- Generic metrics. "Engagement" means everything and therefore nothing. The metric should encode your specific strategy.
- Gameable inputs. Once a metric becomes a target, teams find ways to move it that do not create real value (Goodhart's law). If "messages sent" is the North Star, a team can inflate it with notification spam. Pair the North Star with a guardrail metric (retention, satisfaction) so the optimization stays honest, and watch for input metrics that rise while customer value does not.
Commonly cited North Star Metrics
These are the examples most often referenced in public growth write-ups and Lenny Rachitsky's published list of company North Stars. Use them as patterns, not prescriptions: your North Star should reflect your product's specific value.
| Company | Commonly cited North Star | Why it works |
|---|---|---|
| Spotify | Time spent listening | Direct measure of the value delivered (music enjoyed) |
| Airbnb | Nights booked | Captures value for both sides of the marketplace |
| Messages sent | Usage that is the product's core value | |
| Amplitude | Weekly learning users | Encodes the job the product is hired for |
The pattern across all of them: the metric measures a customer doing the thing the product exists to help them do.
The half most guides skip: from North Star to experiments
Here is where most North Star content ends and the real work begins. A North Star you cannot act on weekly is decoration. The bridge from metric to action has two steps.
Step 1: Decompose the North Star into input metrics
A North Star is too broad to experiment against directly. Break it into the two to four input metrics that compound into it. The classic structure is breadth, depth, frequency, and efficiency:
- Breadth - how many users perform the core action
- Depth - how much value each gets per use
- Frequency - how often they return to do it
- Efficiency - how reliably the product delivers the value
If your North Star is "weekly active teams," the inputs might be new teams activated (breadth), actions per team per week (depth), and week-over-week return rate (frequency). Now you have something a team can run an experiment against.
Worked example (illustrative). Imagine a project-management tool whose North Star is "weekly active projects." Decomposed, that becomes three inputs: projects created in the first session (breadth), tasks added per project per week (depth), and the share of projects active two weeks after creation (frequency). Each input now suggests its own experiments. Breadth points at the onboarding flow: does a templated first project lift first-session creation? Depth points at the empty-state: does suggesting the first three tasks raise tasks-per-project? Frequency points at lifecycle: does a day-three nudge improve the two-week active rate? None of these is "improve the North Star," which no one can action. Each is a specific, testable change tied to a metric that compounds upward. That is the difference between a North Star you can run against and one you can only report on.
Step 2: Rank experiments against each input metric
Each input metric generates candidate experiments. The mistake is running them in the order they were suggested. Score them instead. Two scores do the work:
| Score | Question | What it catches |
|---|---|---|
| ICE | How big, how sure, how easy? | The standard prioritization currency |
| ROTI | How much do we learn per unit of time? | Fast tests that ICE buries under slower "high impact" builds |
ICE alone tends to over-rank slow, high-impact builds. Adding ROTI surfaces the painted-door test that answers the question in three days over the build that takes six weeks, which lets you run the next experiment sooner. Over a quarter, that compounding is what actually moves the North Star.
This is the loop most North Star guides leave out: North Star to input metrics to scored experiments to learnings, and the learnings sharpen the next batch. The metric on the wall does nothing; the ranked queue underneath it is the engine.
The original artifact: the North Star decomposition worksheet
Fill this in for your product:
- North Star Metric: the one number that captures core customer value
- Input metric, breadth: how many users do the core action
- Input metric, depth: how much value per use
- Input metric, frequency: how often they return
- Per input metric, list 3-5 candidate experiments
- Score each on ICE and ROTI
- Run the top two by combined score, capture the learning, re-rank
If you keep this ranked rather than static, you have an experimentation program. If it is a slide, you have a North Star and nothing under it.
Where this is weakest
A North Star Metric is a focusing tool, not a strategy. It tells you what to optimize, not whether you are in the right market. A team can move its North Star steadily and still lose if the metric was chosen wrong, which is why the leading-indicator and customer-value tests matter more than the dashboard. And no single metric survives forever; as the product matures, the North Star and its inputs need revisiting.
How to apply this to your experimentation program
Define the North Star, decompose it into two to four input metrics, then attach a scored, ranked queue of experiments to each input. Run by score, capture every learning, and re-rank. The experiment database keeps that queue ranked by ICE and ROTI as you go, and the experiment generator drafts candidates from your business context if you would rather not start from a blank page.
Frequently asked questions
What is a North Star Metric?
A North Star Metric is the single metric that best captures the core value your product delivers to customers. It aligns every team around one number and acts as the leading indicator the whole organization optimizes toward.
What is the difference between a North Star Metric and a KPI?
A KPI is any key performance indicator a team tracks; a company has many. The North Star Metric is the one metric above them all that represents core customer value and aligns the whole organization. Input metrics and KPIs ladder up to it.
What is the difference between a North Star Metric and input metrics?
The North Star is the single top-level value metric. Input metrics are the two to four sub-metrics (typically breadth, depth, and frequency) that compound into it and that teams actually run experiments against. You do not experiment on the North Star directly; you experiment on its inputs.
How do you choose a North Star Metric?
Identify the core job customers hire your product to do, then find the leading, movable, plain-language metric that measures how well you deliver it. Test it against four criteria: simple, leading, tied to customer value, and influenceable by your teams.
How does a North Star Metric connect to experiments?
Decompose the North Star into input metrics, generate candidate experiments for each input, then score and rank them with ICE and ROTI. Run the highest-scoring experiments, capture the learning, and re-rank. The metric guides direction; the ranked experiment queue is what moves it.
