How to Choose Your North Star Metric (Most Companies Pick Wrong)
A practical framework - three tests and three real business models - for figuring out which number actually deserves to run your company.
Revenue is up 18% this quarter. Everyone in the Monday standup is smiling. And yet three people in that room already know something is wrong - they just don't have a number to point at yet. That gap, between the thing you're measuring and the thing that's actually happening in your business, is exactly why most companies get this wrong when they're figuring out how to choose a north star metric for their business.
Here's the mistake almost everyone makes: they treat the North Star Metric like a trophy. The single most important number, the one that goes on the wall, the one the CEO quotes in the all-hands. That framing is backwards. A North Star isn't the number that makes you look good at the end of the quarter - it's the number that would have warned you three months earlier that the quarter was going to be bad.
Why the metric you picked is probably the wrong one
Most teams default to a lagging indicator - revenue, total users, total orders - because it's easy to explain to a board and it's already sitting in the dashboard. The problem is that lagging indicators are confirmations, not warnings. By the time revenue dips, the damage was done 60-90 days ago, in churn, in a broken onboarding flow, in a KOL campaign that brought in the wrong audience. You can't steer a ship by looking at its wake.
A North Star Metric's job isn't to tell you how you're doing. It's to tell you, early, that you're about to stop doing well.
The three tests - run your current metric through all of them
Before you commit to a number, stress-test it. If it fails any one of these, drop it and keep looking. This is the actual framework we use when a founder asks us how to choose a north star metric for their business - not a formula, a filter.
- Causal test: does moving this number reliably cause revenue or retention to move later, or does it just move alongside them by coincidence? 'App opens per week' feels meaningful but often fails this test completely.
- Company-wide test: can product, marketing, ops, and customer service all meaningfully influence this number through their own work? If only one team can move it, it's a team KPI, not a company North Star.
- Leading test: does this number change before revenue changes, or after? If you only find out it dropped the same month revenue drops, it's too late to be useful.
Most 'North Stars' we see in the wild pass one test, maybe two. Total customers passes the company-wide test but fails leading and often causal - you can grow total customers while the business is quietly dying underneath. That's the trap.
What this actually looks like, by business model
There's no universal answer, because the right North Star depends entirely on what causes what in your specific business. That said, three patterns show up constantly:
- Ecommerce: repeat purchase rate within a defined window, usually 60-90 days. First purchases are bought by ads; repeat purchases are earned by the product and experience. This is the number that tells you whether you're building a brand or renting attention.
- Service businesses (clinics, agencies, anything with a multi-session or multi-stage delivery): full-course completion rate - the percentage of clients who finish what they started, not just the percentage who book the first session. A high intake number with a low completion rate means your sales team is good and your delivery isn't.
- Subscription / SaaS: net revenue retention (NRR). This single number already accounts for churn, downgrades, and expansion revenue in one figure - it tells you whether your existing base is growing or quietly shrinking regardless of how aggressively you're acquiring new customers.
Notice none of these are 'engagement' or 'brand awareness.' Those feel important and they're almost impossible to causally tie to revenue without a very long chain of assumptions. If you can't draw the causal line from your metric to money in under three steps, it's probably not your North Star.
The failure that actually kills companies
It's rarely that a company picks the wrong single number. It's that they pick five. Marketing rallies around leads generated, product rallies around feature adoption, sales rallies around deals closed, and nobody in the building can answer the question 'is the company winning right now?' in under thirty seconds. Everyone's technically right about their own number and the company has no shared language at all.
A real North Star Metric should be the one number a junior hire in any department can recite after their first week, and understand why it's the one that matters. If that's not true at your company today, that's the actual gap - not the metric itself, but the absence of one everyone agrees on.
Choosing it isn't a branding exercise or a dashboard decision. It requires someone to actually sit down, map out the causal chain of the business - what happens first, what that leads to, what that eventually becomes - and have the discipline to throw out every metric that feels important but doesn't pass the three tests. That work is uncomfortable, it takes longer than people expect, and it's exactly the kind of thing most internal teams never get around to doing because the daily fires never stop.
Frequently asked questions
What is a North Star Metric and how is it different from a KPI?
A North Star Metric is the single number that best predicts long-term business health before revenue moves, while a KPI is any operational metric a team tracks to manage its own work. You can have dozens of KPIs across departments, but only one North Star - it's the number that ties all those KPIs back to a shared definition of 'the company is winning.'
How do I choose a north star metric for my business?
Start by listing every metric currently treated as important, then run each one through three tests: does it causally predict revenue or retention (causal test), can every department influence it through their own work (company-wide test), and does it move before revenue does, not after (leading test). Whatever survives all three, and maps most directly to your specific business model's causal chain, is your North Star.
Should I hire a consultant or agency to help define our North Star Metric?
It depends on whether your team has the time and distance to map your own causal chain honestly - most internal teams don't, because they're too close to the daily numbers to question them. A good outside operator should ask for at least 90 days of cohort-level data before suggesting anything, be willing to say your current metric is fine, and show you the reasoning, not just hand you a number. If someone proposes your North Star in a single one-hour call, that's a red flag, not a shortcut.
How often should a company review or change its North Star Metric?
Review it every quarter, but expect to actually change it rarely - maybe once every 12-24 months, usually when the business model itself shifts, such as moving from one-time sales to subscriptions. Changing your North Star too often is itself a warning sign, because it usually means the company hasn't actually understood its own causal chain yet.