How do you pick a north star metric?

Your north star metric is the single number that best captures whether your product is delivering real value to real users — not just activity, but genuine progress toward the outcome your customers care about. Pick it wrong and you'll optimize your way into a local maximum that looks healthy on a dashboard while the business quietly stalls. Pick it right and every team, every sprint, and every hire aligns toward the same underlying truth.

Start with the value moment, not the funnel

Most founders default to a metric that's easy to measure — signups, page views, revenue — rather than one that actually captures the moment their product worked for someone. The right place to start is to ask: what does a user do or experience when they get the core value from my product? For a collaboration tool, that might be the moment a second team member joins a shared workspace and takes action. For a marketplace, it might be the first completed transaction where both parties were satisfied. For a SaaS product, it's often a usage event that strongly predicts whether someone will still be paying ninety days from now.

This 'value moment' framing forces you to think like your best customers, not like a growth analyst. Once you've identified that moment, your north star is usually a measure of how frequently or how broadly it's happening across your user base. The metric should be something that, if it went up meaningfully, you'd be genuinely confident the business is healthier — not just bigger.

A common mistake is picking a metric that's a leading indicator of something else rather than the thing itself. DAU is often this kind of proxy. It tells you people showed up, but not whether they got value. 'Tasks completed per active user per week' is much closer to a north star because it's harder to fake and more directly tied to retention and word of mouth.

The metric must be simultaneously leading and actionable

A good north star metric has two properties that are somewhat in tension: it has to predict long-term business outcomes (revenue, retention, referral), but it also has to move on a timescale short enough that your team can act on it. If your metric only moves quarterly, your team can't course-correct fast enough to matter. If it moves daily but has no relationship to whether users actually succeed, you're optimizing noise.

The practical test is whether you can trace a credible, falsifiable chain from the metric to a business outcome. 'If weekly active writers on our platform goes up by 20%, our 90-day retention historically improves by 8 points, which drives expansion revenue.' That chain doesn't have to be perfect — early stage companies rarely have the data for precise correlation — but it should be directionally defensible and something your team believes in enough to make hard tradeoffs for.

One useful exercise: write down three metrics that might be your north star, then ask for each one — 'If this number doubled but nothing else changed, would we actually be winning?' If doubling your signups while leaving engagement flat doesn't excite you, signups isn't your north star. If doubling weekly activated users would genuinely feel like a breakthrough regardless of what revenue looked like in the short term, you're closer to the right thing.

Avoid vanity metrics and composite indexes

Vanity metrics are numbers that trend up and to the right by default just because a company is alive and spending on acquisition. Total registered users, cumulative downloads, and raw page view counts all have this property. They can mask severe underlying problems — high churn, low activation, or a product that works for one narrow cohort but nobody else. A north star metric should be capable of going down. If it can't decline, it's not giving you honest signal.

Composite scores — weighted blends of multiple metrics into a single index — feel sophisticated but usually create more confusion than clarity. When the number moves, nobody knows which underlying variable drove it, so nobody knows what to work on. The simplicity of a single, unambiguous metric is a feature, not a limitation. Your team should be able to explain it in one sentence to a new hire on their first day.

Paul Graham's observation about startups and growth is relevant here: he has emphasized that the cleanest signal a startup can track is whether the thing users actually want is happening more often. That principle cuts through the temptation to track elaborate dashboards and keeps teams focused on whether the product is working in the world.

How to validate your choice before committing

Before you declare a north star and build dashboards around it, run a three-week validation sprint. Pull your best users — the ones who renew, refer others, and expand their usage — and look at what they have in common behaviorally in their first thirty days. Which events did they all hit? Which frequency or depth thresholds did they cross? That cohort analysis often surfaces the real value moment more clearly than any top-down theorizing about what your product is 'for.'

Then compare that to your worst cohorts — users who churned early or never converted from free to paid. The north star metric candidate should sharply differentiate these two groups. If your best users hit it early and your churned users never did, you've found something real. If the distributions overlap heavily, the metric isn't predictive enough to be useful.

Once validated, stress-test the metric against a perverse incentive question: 'How could a team game this number without actually improving the product?' If the answer is easy — like 'send more notifications to drive opens' — your metric is vulnerable to being optimized in ways that harm the user experience. The best north stars are hard to game because they require genuine user success, not just user action.

When to change your north star metric

Many founders treat their north star as permanent infrastructure once chosen, but the right metric often changes as the company matures. At seed stage, the right north star might be a raw engagement signal — 'users who completed their first meaningful workflow.' At Series A, when you have enough data to measure retention accurately, it might shift to something like 'users who returned and completed that workflow at least three times in 30 days.' At Series B and beyond, it might evolve toward a metric that connects more directly to monetization or expansion revenue.

The trigger to reconsider your north star is when you've been moving it consistently but some other important business signal — retention, NPS, revenue per user — isn't responding. That divergence usually means the metric has drifted from the real value moment, either because the product changed, the user base shifted, or you got so good at moving the number that you've started gaming it accidentally.

Changing the metric is not an admission of failure. The discipline is in being explicit about why you're changing it, documenting the old series alongside the new one, and making sure your team understands what the new metric is measuring and why it's a better proxy for user success. Metric changes done silently or too frequently destroy organizational trust in the measurement system entirely.

“Growth is the form a startup has to take to be a startup.”

— Paul Graham, source

The one thing to do

This week, identify your single most important user value moment, check whether it differentiates your best users from churned ones, and make that event-frequency your north star — everything else is a supporting metric.

Frequently asked questions

Can revenue be a north star metric?

Revenue can work at later stages when your product-market fit is stable, but early on it's usually too downstream to give fast feedback. A usage or engagement metric that strongly predicts revenue is more actionable for a seed-stage team because it responds faster to product changes.

Should every team in the company track the same north star?

Yes — that's the point. The north star is a single shared direction, not a departmental KPI. Each team will have sub-metrics (inputs) that they believe move the north star, but the whole company should be oriented toward the same destination so tradeoffs between teams can be made rationally.

How is a north star metric different from an OKR key result?

OKR key results are quarterly targets — they're inherently time-boxed and can change each cycle. A north star metric is persistent and describes the fundamental value your product creates, not a milestone. OKRs often ladder up to moving the north star, but they're not the same thing.

What if I have a two-sided marketplace — whose value do I measure?

In marketplaces, the north star usually captures the moment both sides get value simultaneously — a completed transaction, a successful booking, a matched connection. Metrics that only measure one side (e.g., supply listings) can grow while the other side suffers, so a 'matched' metric is almost always more predictive than a unilateral one.

Sources

More playbook answers · Growth Prophet home