The North Star Framework
The North Star Framework is the cleanest answer I know to a question every growing product faces: what is the one number this whole team should orient around? It was formalized by John Cutler and the team at Amplitude (building on the "North Star Metric," a term coined by Sean Ellis), and it's earned a permanent place in how I think about outcome ownership. This note is the framework as its authors define it, plus where I push back.
The North Star Framework as described here is the work of John Cutler and Amplitude. The underlying "North Star Metric" term was coined by Sean Ellis. I'm presenting their model, not mine.
#The core idea: one value metric, plus its inputs
Most metric programs fail in one of two ways: a single top-line number (revenue, DAU) that's too lagging and too gameable to guide daily work, or a dashboard of fifty metrics that guides nothing. The North Star Framework threads the needle with a two-layer structure:
- The North Star Metric (NSM) is the single metric that best captures the core value your product delivers to customers. Crucially, it's a leading indicator of sustainable business results: value delivered to users that predicts revenue, rather than revenue itself. Amplitude's canonical examples: Spotify's "time spent listening," Airbnb's "nights booked."
- Input Metrics are typically three to five metrics that the team can directly influence and that together drive the North Star. They're the levers; the NSM is the dashboard.
The relationship is the whole point: if you move the inputs, the North Star should move, and if the North Star moves, the business should follow. It's a causal chain from daily work → user value → business result.
#A worked shape
Amplitude often frames the inputs along dimensions like breadth, depth, frequency, and efficiency. A simplified media example (my illustration, in the spirit of the framework):
| Layer | Metric | Who owns the lever |
|---|---|---|
| North Star | Weekly engaged readers | The whole product org |
| Input: breadth | New readers activated / week | Growth team |
| Input: depth | Articles read per session | Content + UX |
| Input: frequency | Return visits / reader / week | Retention / notifications |
| Input: efficiency | Time-to-first-article | Onboarding |
Each team can see their lever and how it ladders up. That's the framework's real gift: it connects an individual PM's metrics to a value metric the exec team cares about, so strategy and daily execution share one number.
#Why I rate it
The thing the North Star does better than OKRs is durability. OKRs reset quarterly; a North Star is meant to persist for years, giving the org a stable orientation while the OKRs underneath it change. It also enforces my favorite discipline, which is picking a metric that's user value, not just business extraction. A team optimizing "nights booked" is forced to make the product genuinely better; a team optimizing raw revenue can degrade the experience to hit a quarter. The North Star is Outcomes Over Outputs crystallized into a single, shared, leading number.
#Where I push back
A North Star is a simplification, and simplifications can be weaponized. Optimize any single metric hard enough and you'll find the perverse path, because Goodhart's law always wins eventually. "Time spent" can mean addictive dark patterns; "nights booked" can mean ignoring host quality. A North Star needs guardrail metrics (things you refuse to harm, like churn, trust, unit cost) sitting alongside it, or you'll optimize the company into a corner.
Two more cautions. First, picking the NSM is hard and high-stakes. Choose the wrong metric and you point the whole org at the wrong hill, so it's worth real discovery before committing. Second, the causal links between inputs and the North Star are hypotheses, not facts. They need validating with experiments, or you end up religiously moving an input that turns out not to drive the star at all.
#North Star vs OKRs vs the metrics you already have
They're complementary, not competing, and each layer answers a different question. The North Star points the durable direction; the input metrics name the levers; OKRs are the quarterly bets on which levers to push; and RICE picks which features might move an input.
#How AI is changing it
AI doesn't change the framework, but it makes living inside it far easier. Self-serve, natural-language analytics (increasingly built into platforms like Amplitude) let a PM interrogate the input-to-North-Star relationship conversationally and spot when an input is decoupling from the star far sooner. What stays human is the choice of the metric and its guardrails: a model can compute correlations between candidate inputs and the star, but deciding which value metric represents the company's soul, and which harms are unacceptable in pursuit of it, is a strategic and ethical call the org must own.
#Continue Reading
- OKRs and Goal-Setting for the quarterly bets that push the North Star's input metrics.
- Outcomes Over Outputs is the belief the North Star operationalizes.
- Product Analytics and Metrics is the measurement foundation the framework sits on.
- Business Outcome Ownership for connecting your work to a value metric that predicts the business.
- Fluency with Data for validating that the input-to-star causal links are real.