A North Star metric is the single measurement that best captures the value customers get from your product, and that your team can move through its daily work. Revenue is not that metric. Revenue shows you were paid. A North Star shows that customers received value first. For a B2B SaaS company, the fastest way to find yours is to work backward from the moment a new user first gets that value, the "aha moment," and then confirm it with data.
This guide covers the definition, the tests a good metric must pass, real examples, how to map the aha moment, and the traps that make a North Star useless. It is written for product leaders and founders who run a free trial or freemium product and want more users to become paying accounts.
TL;DR
- A North Star metric measures customer value delivered, ideally as a leading indicator. Revenue is a lagging scoreboard, not a North Star.
- Good ones are simple, tied to value, and influenced by the team's daily decisions. Most companies need supporting input metrics and guardrails too.
- The aha moment is the early user action that predicts retention. Define it with correlation analysis, not with a milestone like "onboarding finished."
- For AI products, track what users do with the output (accepted, edited, rejected), not only accuracy or latency.
- Pair every North Star with counter-metrics, or the team will game it.
Table of Contents
- What a North Star metric is
- What makes a good one
- North Star metric examples
- The aha moment: where the North Star starts
- How to map and discover your aha moment
- North Star metrics for B2B SaaS
- North Star metrics for AI products
- Guardrails: how to stop the metric from backfiring
- Our perspective
- Where we can help
- FAQ
- Sources
What a North Star metric is
The idea is simple: one number everyone in the company can repeat, that tells you whether customers are getting what they came for. Metabase's glossary describes the best ones as measuring value delivery to the customer before the business captures the money, which lets product teams influence the number in their day-to-day work.
That is why the usual SaaS suggestions of MRR, churn rate or lifetime value are weak choices. They matter, but they sit downstream. By the time they move, the product decisions that caused the move are months old.
A useful North Star also has a unit that reads like customer behavior. "Projects shared with a teammate each week" describes something a person did. "Revenue" describes something a finance system recorded.
North Star metric vs. KPI
People often ask how a North Star differs from a KPI. A KPI is any metric you track against a target: support response time, win rate, gross margin. You can have dozens. The North Star is the one KPI you place at the top, the one that decides which of the others deserve attention when they conflict. All North Stars are KPIs. Very few KPIs deserve to be the North Star.
What makes a good one
The engineering framework published by Vatsal Shah sets out three criteria. The first is that the metric must capture the core utility users get, because if users do not find value they will not retain. In practice we turn these ideas into a short checklist you can run in a workshop:
- It reflects value received, not value captured. Time in app can rise because the product is confusing. A completed task rarely does.
- It leads revenue. If it moves, retention and expansion should follow a few weeks or months later. You can test that against your own cohorts.
- The team can move it. Designers, engineers and support should each be able to name something they could ship this quarter that would push it.
- It is understandable in one sentence. If a new hire cannot repeat it after one meeting, it will not guide decisions.
- It has a clear time window. "Weekly" or "within 14 days of signup" beats an undefined total.
If a candidate fails the third test, it is probably a business outcome. Keep it on the executive dashboard, but do not ask a product squad to own it.
North Star metric examples
The table below lists commonly cited examples. These are widely repeated by commentators, not official statements from the companies, so treat them as illustrations of the logic rather than confirmed internal targets. The pairs come from a roundup of company examples.
| Company | Commonly cited metric | Why the logic works |
|---|---|---|
| Airbnb | Nights booked | Captures both sides of the marketplace getting value |
| Spotify | Time spent listening | Reflects the core product experience |
| Uber | Weekly active riders | Frequency of the core transaction |
| Zoom | Meeting minutes | Usage of the core job |
These are consumer examples, and they flatter the concept because the value is easy to see. B2B SaaS is harder. The buyer is not the user, usage is spread across a team, and value often shows up as a result (a report sent, a deal closed) rather than as time spent.
Here are B2B-shaped candidates, written as sentences, so you can see the pattern:
- Collaboration tool: teams with three or more members who completed a shared project in their first 14 days.
- Analytics product: accounts that viewed a saved dashboard weekly.
- Workflow automation: automations that ran successfully at least once a week.
- Fintech onboarding: customers who completed their first reconciled transaction.
None of these is universal. Each has to be tested against your retention data.
Netflix: the example people get wrong
People often name a single North Star for Netflix. The company's own words are more careful. In its Q1 2024 earnings communications, Netflix's co-CEOs described a set of North Stars rather than one, covering engagement, revenue and profit, linked through member satisfaction. Engagement is the leading indicator: members who are satisfied watch more, and so are more likely to stay and to bring in others. The lesson for you is not "pick hours viewed." It is that a mature company can hold several outcomes, as long as one of them leads the others.
The aha moment: where the North Star starts
The aha moment is the point at which a new user first experiences the value of your product. It is the early action, or short sequence of actions, after which they are far more likely to stay. If the North Star is the long-run measure of value, the aha moment is its first-week version, and it is the part of the funnel you can influence most during a free trial.
The best-known example is Facebook. Chamath Palihapitiya described getting to seven friends in ten days as the company's focus. Analyst Benn Stancil argues in a piece for Mode that the figure worked less as precise science and more as a memorable rallying point for the whole company. That is the correct way to hold your own number: precise enough to test, simple enough to repeat.
Why defining it wrongly is expensive
Growth research summarized by Prodara, citing a16z and 2026 industry playbooks, says that misdefining the activation event is the highest-leverage error a product-led growth company can make. It also reports that defining and instrumenting it well can improve trial-to-paid conversion by 20 to 30 percent within two quarters, without price or feature changes. Treat that as a claim from a secondary source, not a promise. Your result depends on how wrong the current definition is.
The common failure has a name: activation theater. Spike AI describes teams that optimize for milestone completion, such as "onboarding finished," instead of actual value. A user can click through a checklist and never come back. If the milestone does not statistically predict retention, you are optimizing a metric that produces activity and not revenue.
How to map and discover your aha moment
This is the process we use as a starting point. It works with a spreadsheet and a product analytics tool, and it works better when someone also walks through the product as a new user.
- List candidate actions. Collect every action a user can take in the first 7 to 14 days: invited a teammate, connected a data source, created a first report, ran a first workflow.
- Split users into retained and churned. Use a cohort that has had enough time to show the difference, for example accounts still active at day 30 versus those that are not.
- Compare early behavior. For each candidate action, check how many retained and churned users did it within the first days, and how often.
- Look for the action and threshold that separates the groups. You are hunting for something like "created and shared a report within three days," not a single click.
- Test causality, not only correlation. Users who do hard things may simply be more motivated. Change the product to make that action easier or earlier, and see whether retention and trial conversion move for the new cohort.
- Write it as a sentence and instrument it. One event definition, one time window, one dashboard.
- Check the path to it. Count how many steps, fields and decisions stand between signup and that action. Every one that does not serve it is a candidate for removal.
Step 7 is where qualitative work pays off. Analytics tells you where people stop. It does not tell you why. Signing up as a real user and going through setup shows you the setup that comes before any value, the prompt that arrives at the wrong moment, and the pricing screen that raises doubts.
A benchmark to sanity-check retention
Amplitude's 2025 Product Benchmark analysis of more than 2,600 companies found that getting 7 percent of a cohort back on day seven puts a product in the top 25 percent for activation performance. In enterprise products, the median reaches only 2.1 percent day-seven retention, against 12.4 percent for the top 10 percent. If your day-seven number is well under those, your activation event is probably either wrong or too hard to reach. Compare against your own segment, not the headline figure.
North Star metrics for B2B SaaS
One mistake we see in B2B SaaS is trying to push acquisition, retention and monetization at the same time. Datapile's PLG playbook argues for a sequence: establish the user flow, prove the free product delivers value using cohort retention, scale quality acquisition, and only then focus on conversion to paid. In that view, the North Star changes as the company matures, and monetization is a packaging challenge once retention of core value is proven.
For a company with a free trial, that suggests a practical rule:
- Before you have retention: your North Star is an activation or first-value metric. Example: accounts reaching first value within seven days.
- Once cohorts hold: move toward a repeated-value metric. Example: weekly accounts completing the core job.
- When growth is steady: add expansion and conversion layers, and keep the value metric on top.
If you are choosing between tuning pricing pages and fixing first-run experience, the sequence says to fix first value first. A better pricing page cannot convert users who never saw why they would pay.
North Star metrics for AI products
AI features add a trap: the technical metrics look great while users quietly stop trusting the output. A write-up on AI product metrics separates system performance metrics, such as token consumption, latency and accuracy, from product performance metrics, such as task completion rate, edit rate and return rate. High-performing teams, it says, instrument what users do with the output: accepted, edited or rejected. When redoing the work is faster than accepting the AI's answer, trust is low, however accurate the model is on paper.
The same source calls task completion rate, the share of started tasks that reach a successful end state, the closest thing to a North Star for AI products. The Institute of Product Management makes a related point: accuracy is not a North Star, and neither is engagement by itself. The right one captures the user's benefit when the AI works and the cost to them when it does not.
A workable AI North Star therefore looks like "tasks completed with output accepted or lightly edited, per active account per week," paired with a guardrail on rejection and retry rates.
Guardrails: how to stop the metric from backfiring
Once a metric becomes a target, people start optimizing the number instead of the thing it stood for. That is Goodhart's Law, and Shah's framework applies it directly: without paired counter-metrics, teams can push "tasks completed" up by letting error retries balloon, or reach for dark patterns that inflate vanity numbers while eroding trust.
The same piece describes the North Star as the top tier of a system. The star sets the focus, and HEART (Happiness, Engagement, Adoption, Retention, Task Success) plus guardrails such as error rates and system costs act as sensors. A PM-focused breakdown lays the system out in layers, starting with the North Star and followed by the input metrics that teams can act on.
A compact version you can put on one page:
| Layer | Question | Example for a trial product |
|---|---|---|
| North Star | What core value are we scaling? | Accounts completing the core job weekly |
| Input metrics | What can teams change? | Time to first value, invites sent, setup completion |
| Guardrails | What must not get worse? | Error rate, support tickets per account, refund rate |
Our perspective
We approach the North Star from the free trial inward. Most people who start a trial never pay, and the reason is rarely the price. It is usually that setup asked for effort before the product showed any value, that a prompt interrupted at the wrong time, or that pricing was unclear when the user was ready to decide. That is why we do not start with a dashboard. We sign up like a real user, go through the trial step by step, and mark where a real person would stall.
The usual advice, "pick one metric and rally behind it," skips the harder question: is the activity you are counting actually value? We have seen teams count a completed onboarding checklist as success when the checklist was the very thing delaying the first useful result. In our work on product setup at Supliful, changes to how the product was set up supported $1.1M in GMV growth. The lesson we took was that the shortest honest path to first value is a growth lever, not a polish item.
Three opinions we hold firmly:
- Start with the aha moment, then the North Star. If you cannot say what a new user does in their first week that predicts they will stay, a company-level metric will be decoration.
- Prove it, then say it. Check that the early action predicts retention, then change the experience and see whether the outcome follows.
- For AI, measure trust in the output. Whether people accept, edit or discard what the AI produces tells you more than model accuracy.
With more than ten years across SaaS, fintech, gaming and regulated industries, we have learned that regulated and complex products raise the stakes. You cannot simplify a workflow by deleting a step that compliance requires. You find the steps that do not serve the user's first value and cut those.
— Edijs
Where we can help
If you already track a North Star but trial users still do not convert, the gap is usually between the metric and the experience. Our AI + UX Discovery engagement is built for that: we go through your trial as a user, map the drop-off points, and show what to fix first and why, with research behind each recommendation. If the findings point at AI features or workflows that need prototyping, see AI Agents & Workflows, or browse all services. Our work suits product teams that are actively investing in conversion and workflow improvement.
FAQ
What is the difference between a KPI and a North Star metric?
A KPI is any metric tracked against a target, and a company can have many. A North Star is the one metric placed above the rest because it best reflects customer value and predicts long-term success. When two KPIs conflict, the North Star decides which wins.
What is Netflix's North Star metric?
Netflix does not name a single one. In its Q1 2024 communications, the co-CEOs described several North Stars, namely engagement, revenue and profit, linked by member satisfaction, with engagement as the leading indicator. Claims that its North Star is simply hours viewed are commentators' shorthand.
How do I define the aha moment for my SaaS product?
List early user actions, compare how retained and churned users behaved in the first days, and look for the action and time window that separates them. Then change the product to make that action easier and check whether retention improves. If the action does not predict retention, it is a milestone, not an aha moment.
Can a company have more than one North Star metric?
Most teams do best with one at a time, and the metric can change as the company matures, for example from first-value activation to repeated value to expansion. Large companies sometimes hold several outcomes, as Netflix describes. Even then, one leading indicator should anchor the rest.
What is a good North Star metric for an AI product?
Task completion rate, the share of started tasks that reach a successful end, is a strong starting point. Pair it with signals about what users do with the output: accepted, edited or rejected. Accuracy and latency are useful system metrics but do not show whether users trust the result.
Sources
- Metabase: North Star metric glossary
- Designing the North Star Metric: The Engineering Framework
- North Star Metric Examples (Teknicks)
- Netflix Q1 2024 earnings call transcript (The Motley Fool)
- Facebook's "Aha" Moment Was Simpler Than You Think (Mode)
- Product-led growth metrics (Prodara)
- B2B SaaS Product Metrics (Spike AI)
- The 7% Retention Rule Explained (Amplitude)
- The PLG Playbook for B2B Startups (Datapile)
- AI Product Metrics Nobody Uses
- AI Product North Star Metrics (Institute of Product Management)
- North Star Is Not Enough: HEART, Guardrails, and Counter-Metrics