%20scilla.studio.svg)
A North Star Metric (NSM) is the single number that best captures the value your product delivers to customers, the one metric the whole company can rally behind because moving it up means users are genuinely getting more of what they came for. A good North Star sits one step upstream of revenue: it measures delivered value, which then turns into money. The classic examples (Airbnb's nights booked, Spotify's time spent listening, Slack's messages sent within a team) all share that shape. Revenue itself is almost never a good North Star, and I'll explain why below.
Here's the part most "ultimate guides" skip: choosing a North Star is less a measurement exercise than a strategy exercise. The number is easy. Agreeing on what value actually means for your product, and resisting the pull toward whatever is easiest to count, is the hard part.
North Star Metric: the one metric that most accurately reflects the core value your product delivers to customers, chosen so that growing it reliably grows the business.
There's no formula, which trips people up. A North Star isn't computed; it's selected. But a good one is recognisable by a clear test: if this number goes up, are customers better off and is the business healthier? If you can answer yes to both with a straight face, you've probably found it. If moving the number could make customers worse off (more ads shown, say) while the business looks better for a quarter, you've found a vanity trap instead.
A single worked example makes the shape concrete:
Airbnb. Value moment = a guest completes a stay → North Star = nights booked → inputs = listing quality, search-to-book conversion, repeat-booking rate.
The North Star also anchors that small set of input metrics, the three or four levers your teams can actually move week to week that feed the star. Nights booked is the star; new-listing quality, search-to-book conversion, and repeat-booking rate are the inputs. You don't ship features against the North Star directly. You ship against the inputs and watch the star respond.
The selection comes down to four questions, in order.
1. What is the core value moment in your product? Not the signup, not the purchase, but the moment a user gets the thing they came for. For a video tool it's a video watched by someone other than the creator. For a project tool it's a task completed by a team. Write that moment down in plain language first. The metric is just a count of that moment happening.
2. Does it lead revenue rather than lag it? A North Star should be a leading indicator. Revenue, MRR, and total signups are lagging; by the time they move, the value (or the lack of it) already happened weeks ago. Pick the upstream behaviour that causes the revenue, so the metric gives your teams time to react.
3. Can the whole company influence it? A good North Star is broad enough that product, growth, marketing, and support can each see how their work feeds it, but specific enough that it isn't just "revenue with extra steps." If only the growth team can touch it, it's a team KPI, not a North Star.
4. Does it survive the gaming test? Ask: what's the laziest way to make this number go up without helping a single customer? If that lazy path is easy and tempting (think "send more notifications" to lift DAU), the metric will get gamed under pressure. The best North Stars are hard to inflate without delivering real value.
Run a candidate through all four and most pretenders fall away fast.
The canonical examples all share the same shape: each counts a value moment, not a transaction, and each goes up only when customers are actually getting what they came for.
| Company | Value moment | North Star | Why it works |
|---|---|---|---|
| Airbnb | A guest completes a stay | Nights booked | Counts delivered value (a stay), not just bookings that might cancel. |
| Spotify | A listener finds music worth their time | Time spent listening | Hard to fake; you can't inflate it without people genuinely listening. |
| Slack | A team communicates in the tool | Messages sent within a team | A message is the value, and the team frame stops single-user vanity. |
| A person reaches someone they care about | Messages sent | Tracks the core job directly; growth means more real conversations. | |
| Facebook (early) | A user connects with friends | 7 friends in 10 days | A leading behaviour that predicted long-term retention, rather than a lagging total. |
Notice none of them is revenue, and none is a cumulative total that can only rise. Each can fall on a bad week, which is exactly what makes it useful.
These get muddled constantly, so to be concrete:
Put simply: the North Star is the destination, the input metrics are the steering wheel, and OKRs are this quarter's route.
This is the section I'd read first, because the failure modes are predictable and expensive.
Mistake 1: making revenue the North Star. Revenue is the result of delivered value, not a measure of it. Optimise revenue directly and you'll find short-term levers (aggressive upsells, dark-pattern pricing, squeezing existing customers) that lift the number this quarter and quietly erode the product. It also tells your teams nothing actionable: "make more money" isn't a brief. (And if you're tempted because your LTV:CAC looks great, read why a great ratio can mean you're underinvesting; high ratios hide their own problems.)
Mistake 2: picking a pure vanity metric. Total registered users, page views, app downloads, cumulative anything: these only go up, so they always look like progress. A metric that can't go down can't tell you when something's broken. If your North Star never falls on a bad week, it's decoration.
Mistake 3: choosing what's easy to measure over what matters. DAU is easy. "Weekly active teams that completed a meaningful action" is harder to instrument but far closer to real value. Teams default to the easy number and then spend two years optimising the wrong thing. Measure the right thing badly before you measure the wrong thing precisely.
Mistake 4: confusing an engagement count with delivered value. "Messages sent" works for Slack because a message is the value. "Minutes in app" for a productivity tool is the opposite: more time spent often means the tool is slower or more confusing, not better. Always ask whether more of the metric means the customer won or lost.
Mistake 5: setting a North Star before product-market fit. If you don't yet have a repeatable group of users getting durable value, you don't have a North Star to optimise; you have a hypothesis to test. Premature North Star worship sends teams chasing a number before there's value underneath it. (If you're not sure you're there yet, start with whether you actually have product-market fit.)
Here's what I actually think: for most early-stage B2B SaaS tools, the honest North Star is rarely a clever custom metric. It's retention of the activated user, because retention is the one number that proves the value moment keeps happening. A flashy "engagement score" is usually retention wearing a costume.
Why retention? Because it's the metric that lies the least. Acquisition can be bought. Activation can be juiced with a slick onboarding. But users only come back, week after week, if the product is genuinely doing a job for them. Retention is delivered value, measured honestly, over time. That's the whole point of a North Star.
And here's where benchmarks help you sanity-check the input to your North Star. For B2B SaaS, healthy retention from first use looks roughly like this:
| Metric | B2B SaaS (avg) | Consumer apps (avg) | Source |
|---|---|---|---|
| Day-1 retention | 50 to 70% | 20 to 30% | Pendo, Amplitude (B2B); Adjust, Statista (consumer) |
| Day-7 retention | 40 to 60% | 8 to 15% | Pendo, Userpilot (B2B); AppsFlyer, Amplitude (consumer) |
| Day-14 retention | 35 to 55% | 4 to 8% | Amplitude, Mixpanel (B2B); Mixpanel, Amplitude (consumer) |
| Day-90 retention | 25 to 35% | 1 to 4% | Pendo Product Benchmarks (B2B); AppsFlyer, Adjust (consumer) |
B2B and consumer follow completely different mechanics, so never compare them directly. These are the same benchmarked bands set out in full in retention rate benchmarks.
These are directional context, not targets, and that distinction is the whole point. A B2B tool sitting at the low end of the Day-7 band isn't "failing"; it might serve a deliberate, infrequent workflow (think tax software) where monthly retention is the honest unit. The benchmark tells you which questions to ask, not what your number should be. (For the full retention picture, see retention rate benchmarks.)
So which metric matters for your tool? Work backwards from the value moment, pick the leading behaviour that proves it's happening, and, for most B2B tools, anchor it on retained, returning users rather than a vanity total or revenue. Then watch the curve flatten (the point where each cohort's retention stops sliding and settles into a stable floor of users who keep coming back), not just the headline number climb. That flattening is what durable value looks like in a chart.
A North Star without inputs is a poster on a wall. Once you've chosen the star, build the small constellation around it:
The discipline is in step 3: input metrics must be things teams can actually influence with the work in front of them. If you can't draw a line from a sprint's output to an input metric to the star, the system is decorative.
If you want to pressure-test the input that anchors most B2B North Stars (retention) against real industry bands, the free benchmark.scilla.studio tool charts your retention curve, K-factor, and unit economics against B2B and consumer ranges in a couple of minutes. It won't pick your North Star for you (nobody can; that's your strategy), but it'll tell you fast whether the value underneath it is actually holding. Benchmarks are context, not targets, and the tool is built to keep you honest about both.
What is a North Star Metric in simple terms? It's the single number that best captures the value your product delivers to customers, chosen so that when it goes up, both customers and the business are better off. Examples: nights booked (Airbnb), time spent listening (Spotify), messages sent (Slack).
Can a company have more than one North Star Metric? Generally no, the point is focus. You have one North Star supported by three to four input metrics. Multiple North Stars usually means the company hasn't agreed on what value means yet. Large multi-product companies sometimes run one North Star per product line, but each product still gets a single star.
Should revenue be my North Star Metric? Almost never. Revenue is a lagging result of delivered value, and optimising it directly invites short-term levers that erode the product. Pick the upstream behaviour that causes revenue instead, so the metric gives teams time to act.
What's the difference between a North Star Metric and a KPI? A KPI is any metric you track for health (retention, CAC, churn, NPS). The North Star is one specific metric (usually one of your KPIs) elevated above the rest as the company's strategic focus. All North Stars are KPIs; very few KPIs are North Stars.
How often should I change my North Star Metric? Rarely. A North Star should be durable across years, not quarters. Re-validate it annually against your strategy, but if you're changing it every few months, you're probably still searching for product-market fit rather than optimising a found one.
What North Star Metric should a B2B SaaS tool use? For most early-stage B2B tools, the honest answer is retention of the activated user, the number that proves the value moment keeps happening. Custom "engagement scores" are usually retention in disguise. Anchor on returning, active users rather than a cumulative vanity total or revenue.
