Startup Success Metrics

North Star Metric: Why Airbnb Chose Nights Not Revenue

Airbnb's north star metric is nights booked, not revenue. The distinction reveals the logic of north star selection: a good north star metric captures the value that the product creates for users, not just the value it captures for the company. Nights booked measures whether guests are actually staying in homes and whether hosts are actually earning income — the core promise of the platform on both sides. Revenue is a consequence of nights booked, not the cause. When the team debates a product decision, "does this increase nights booked?" is a better question than "does this increase revenue?" because it focuses effort on the mechanism rather than the outcome.

Slack chose messages sent as its north star metric. A team that sends more messages is a team that has built communication habits around Slack — the most predictive signal of whether the subscription will renew. Slack famously found that teams reaching 2,000 messages in Slack retained at dramatically higher rates than those who didn't, which gave the onboarding team a specific activation target to optimise. The process of discovering a north star metric is the process of finding the single user behaviour that is most predictive of long-term retention and revenue in your specific product. For most consumer products it's a frequency signal (messages sent, nights booked, sessions per week); for most B2B products it's a depth signal (features activated, integrations connected, teams onboarded).

Vanity vs Actionable Metrics

Vanity metrics are numbers that feel good to report but don't change a product decision. Total registered users is the most common vanity metric — it only grows, never shrinks, and says nothing about whether current users find value. App downloads measure marketing effectiveness, not product effectiveness; a downloaded app that's never opened costs $0 to create and delivers $0 in value. Twitter followers as a business metric conflates personal brand building with product traction. None of these numbers answer the question "should we change something about the product this week?"

Actionable metrics are numbers where a change demands a response. Weekly active users is actionable: if WAU drops 15% week-over-week, the team investigates immediately. Paid conversion rate from free trial to paid plan is actionable: if it drops from 4% to 2%, something changed in the product or the funnel and the team finds out what. NPS is actionable: an NPS below 20 is a retention risk signal that should trigger a round of customer interviews within the same week. The test for whether a metric is actionable is simple: if the number moves in the wrong direction, does it force a specific team to respond? If the answer is yes, it's actionable. If the answer is "we'd look into it when we have time," it's vanity.

Weekly Metrics Review Ritual

The weekly metrics review is a 30-minute standing meeting that answers three questions for every key metric: What changed relative to last week and relative to expectations? Why did it change — what product change, marketing action, or external event explains the movement? What are we changing in response? These three questions prevent the pattern of reviewing numbers without acting on them, which produces false confidence that the team is data-driven when it's actually just data-reporting.

The review requires a dashboard that's updated before the meeting, not during it. Time spent building a chart in the meeting is time not spent interpreting what the chart means. Amplitude, Mixpanel, and PostHog all provide dashboards that auto-update from instrumentation in the product. The weekly review ritual builds a shared understanding of the business across the whole team — not just the founders — which is one of the most underappreciated benefits of consistent metrics tracking. A designer who knows that activation dropped this week asks different questions during a design review than one who doesn't.

Integrating North Star with OKR

The north star metric and the OKR system serve different purposes at different time scales. The north star is a permanent reference point that doesn't change quarter to quarter — it's the definition of what creating value looks like for your product. An Objective in the OKR system is a quarterly goal: it changes each quarter as the team works on different aspects of the business. The integration point is that the north star should anchor the Objective's ambition — if the Objective doesn't eventually lead to a higher north star metric, it's not worth pursuing.

Key Results, in this integration model, become leading indicators that predict north star movement. For Slack's north star of messages sent, a quarterly Key Result might be "increase the percentage of new teams reaching 100 messages in their first week from 35% to 55%." This is a leading indicator because teams that reach 100 messages in week one are far more likely to reach 2,000 messages total. Building the OKR structure to work backward from the north star — north star defines the Objective's direction, Key Results identify the mechanisms that move the north star — produces a goal system where every quarter's work accumulates toward the same long-term target rather than cycling through disconnected priorities.

Frequently Asked Questions

How do I find my north star metric if I've never tracked it before? Start by listing the five to ten behaviours users take in your product, then identify which behaviour most strongly correlates with retention. Pull cohort data segmented by each behaviour and plot 90-day retention for users who took each action versus those who didn't. The behaviour with the largest retention gap is likely your north star. If you don't have enough data yet, use the theoretical version: which single behaviour would you most want every new user to take in their first week?

How many metrics should I track weekly? Five to eight metrics on a weekly dashboard is the right range. Below five, you lack visibility into important parts of the business. Above ten, the review meeting becomes a reporting exercise rather than a decision-making one. The north star metric, one acquisition metric, one activation metric, one retention metric, one revenue metric, and one product health metric (like API error rate or page load time) covers all critical business dimensions in six numbers.

What is the difference between a leading indicator and a lagging indicator? A lagging indicator measures an outcome that has already occurred — monthly revenue, annual churn. A leading indicator measures a current behaviour that predicts a future outcome — activation rate predicts next month's retention, number of enterprise pilot customers predicts next quarter's revenue. Leading indicators are more actionable because teams can respond to them in time to change the outcome; lagging indicators confirm what already happened. Build your weekly dashboard primarily around leading indicators, with lagging indicators providing context.

How do I get my whole team to care about metrics? Make metrics visible and personal. Post the weekly dashboard in the team Slack channel every Monday morning with a one-paragraph interpretation. Connect each team member's work directly to specific metrics: tell the designer that their new onboarding flow is being measured by the activation rate, tell the engineer that their performance improvements are being measured by session length. Metrics that are owned by a person are updated and discussed; metrics that are company-owned in a spreadsheet are ignored.

When should I change my north star metric? When you complete a pivot that changes what value the product creates, or when you discover — through retention analysis — that a different metric is more predictive of retention than the one you've been using. Changing the north star because the current metric isn't growing is the wrong reason; changing it because you've learned something new about what drives retention in your product is the right reason. A north star that changes quarterly is a company that hasn't found product-market fit yet.