North star metrics

A north star metric is the single number a company is trying to grow. The shorthand: if everything else went away and you could only optimise one metric, this is it. Input metrics are the operational metrics that compound into the north star.

A few real-world examples:

• Airbnb: nights booked (north star). Listings, bookings per listing, repeat rate (inputs).
• Spotify: time spent listening (north star). Active users, songs per session, sessions per week (inputs).
• Slack: messages sent within an organisation (north star). Active users, channels created, integrations active (inputs).

The pattern: north stars measure outcome, inputs measure the leverage points. A test moves an input metric. The bet is that moving the input compounds into the north star over time.

Why this structure matters for CRO

CRO tests almost always optimise input metrics. Conversion rate, add-to-cart, sign-up rate. The north star is too far downstream to move in a test window. The test’s primary metric is the input. The north star is the outcome the input is meant to compound into.

Without an explicit input-to-north-star mapping, you’re optimising disconnected metrics and hoping they’re the right ones. The map should answer:

• Which input metrics actually drive the north star? Tested via historical correlation or holdouts, not assumed.
• What’s the conversion path from input to north star? Sign-up doesn’t help if the new users don’t activate.
• Which inputs have the most leverage? Some inputs barely move the north star even when they shift dramatically. Others have outsized effect.

What makes a good north star

A few properties of a healthy north star metric:

• Tied to value delivered, not value extracted. Time-on-site is a vanity north star - it can rise because the product got worse and users took longer to complete tasks. “Customer outcomes achieved” is a healthier framing.
• Measurable at high frequency. Quarterly revenue isn’t a great north star because the feedback loop is too slow.
• Resistant to gaming. A north star that incentivises behaviour the company doesn’t want (clickbait headlines, dark patterns) corrupts the team’s decisions over time.
• Aligned with retention. Acquisition north stars look great until churn catches up. Retention-flavoured north stars age better.

Where this falls down

• Picking a vanity north star. Companies that choose “monthly active users” without defining “active” end up optimising for low-signal engagement. The team works hard and the underlying business doesn’t grow.
• Treating the north star as the only metric. A north star is the one number you’d save if you had to pick one. It doesn’t mean you ignore everything else. Guardrails still matter.
• Letting the north star drift. Companies often pick a north star at a particular stage of growth and never revisit it. The original north star (e.g. sign-ups) stops being the right outcome metric at scale.
• Confusing input metrics with the north star. A team that’s optimising 14 input metrics without an explicit hierarchy effectively has no north star.