North Star Metric: How to Pick the Right One (Checklist + Examples)
Most teams pick the north star that is easiest to move. Here is how to choose one that measures customer value, break it into input metrics your teams control, and check it against real company examples.
Share
A north star metric is the single number that best captures the value customers get from your product and, because of that, predicts long-term revenue. Airbnb's is nights booked; Slack's is paid teams. Choose yours by asking which customer behaviour would most grow the business if it rose today, test it against a short checklist, then break it into 3-5 input metrics your teams can move.
Key takeaways
The term is credited to growth marketer Sean Ellis (Mixpanel, 2018).
In Lenny Rachitsky's survey of employees at more than 40 growth-stage companies, about 50% focused on revenue, 35% on customer growth, 30% on consumption, 30% on engagement, and 10% each on growth efficiency and user experience (a16z Future, 2021).
Amplitude recommends 3-5 input metrics under the north star, using breadth, depth, frequency and efficiency as a starting heuristic (Amplitude, 2024).
Daily active users and registered users make poor north stars, according to the same Amplitude guide, because they do not show realised value.
A metric is defined by three things: the unit of value, a quality bar, and a frequency (Reforge, 2022). Most bad north stars skip the quality bar.
What is a north star metric?
It is the one product metric the whole company agrees best reflects customers getting value. The idea is that if this number grows for the right reasons, revenue follows.
Amplitude's North Star Framework, set out in the North Star Playbook co-authored by John Cutler (Amplitude), says a good north star has three traits (Amplitude, 2024):
It reflects customer value, ideally the "aha moment" when the product does its job.
It represents the product strategy, so someone reading it understands what you are betting on.
It is a leading indicator. Monthly revenue and average revenue per user describe the past. The further upstream of revenue the metric sits, the earlier you can act.
Amplitude groups products into three games. Your north star should fit yours:
Game
Question
Typical north star shape
Attention
How much time will customers spend in the product?
Time or sessions with meaningful engagement
Transaction
How many transactions do users complete?
Orders, bookings, rides, purchases
Productivity
How efficiently can someone get their work done?
Completed jobs, active teams doing core work
Most B2B SaaS tools play the productivity game, which is why "hours spent in the app" is usually the wrong north star for them. Less time can mean the tool works better.
North star metric vs KPI
A KPI is any number you track to judge performance; a company might have dozens. The north star is one KPI promoted above the rest because it links customer value to future revenue. Input metrics sit below it, and health metrics (uptime, support load, margin) run alongside it as guardrails.
What are some real north star metric examples?
These come from Lenny Rachitsky's 2021 survey of current and former employees at over 40 growth-stage companies (a16z Future, 2021). They are what employees reported at the time, not official statements, and companies change them.
Company
Reported north star
Airbnb
Nights booked
Uber
Rides taken
Twilio
Messages sent
Slack
Number of paid teams
Dropbox
Teams using Dropbox Business
Airtable
Weekly paid seats
Asana
Weekly active paid users
Shopify
Active merchants
Miro
Number of collaborative boards
Amplitude
Weekly learning users
Netflix
Median view hours per month
Calm
LTV/CAC
Robinhood, Superhuman
Net Promoter Score
Two patterns stand out. Marketplaces pick a completed transaction. Team-based B2B products count paying teams, paid seats or active paid users, not every login.
Note too that public examples disagree. Lenny's list has Netflix on view hours; Amplitude's guide suggests Netflix more likely maximises subscribers; Reforge describes Netflix retention in hours consumed per week. Treat any company's north star as a snapshot, not a rule to copy.
Why is your north star probably wrong?
Most north stars are picked in one meeting and never tested. Five failure patterns cover most of them.
1. It is revenue. Revenue was the most common focus in Lenny's survey, at roughly half of companies, so this is not a fringe mistake. But revenue is what the customer pays, not the value they get, and it lags. Lenny advises against revenue as the only north star when it is spiky, when the company is too early, or when it does not inspire the team. Keep revenue as the business outcome and pick the customer behaviour that leads it.
2. It counts people, not value. Sign-ups, registered users and raw daily actives rise when you buy ads or loosen a definition. Amplitude names DAU and registered users as poor north stars.
3. It has no quality bar. Reforge's G2 example: all site visitors is a weak unit, while visitors who write reviews is the one that drives growth. "Teams that created a report" beats "teams that logged in".
4. It moves too slowly to steer by. Annual retention is real value, but it takes a year to read. Gibson Biddle's fix is a proxy metric: a faster stand-in that predicts the slow one. His Netflix example was the share of new members who gave at least 50 ratings early in their membership (Ask Gib, 2025).
5. Nobody trusts the number. If finance, product and sales compute it three different ways, the north star becomes a debate rather than a guide. Write the definition down to the event and the time window.
How do you choose a north star metric? A 6-step method
Budget two working sessions of about 90 minutes, a week apart, so you can pull data in between.
Write your strategy and value moment first. What job does the customer hire the product for, and when do they first get it done? If your strategy is unclear, fix that first with a one-page product strategy canvas.
Ask Lenny's lead question. Which metric, if it went up today, would most speed up the business? List 5-8 candidates.
Name your game. Attention, transaction or productivity. Strike candidates that do not fit it.
Define each survivor precisely. Unit of value, quality bar, frequency. "Weekly accounts with 2+ members who completed a report" is a definition; "engagement" is not.
Check it against retention data. Pull 6-12 months of cohorts. Do accounts that hit the candidate behaviour retain better than those that don't? If you cannot see a gap, the metric is not capturing value.
Score it with the checklist below, then commit for at least two quarters. Changing the north star every quarter teaches the company to ignore it.
A north star selection checklist you can copy
Copy this and score each candidate yes or no. A good north star gets at least 8 of 10.
NORTH STAR CANDIDATE: ______________________________
Definition (unit + quality bar + time window): ______
[ ] 1. Value: it rises only when customers get the job done
[ ] 2. Strategy: someone reading it can guess what we are betting on
[ ] 3. Leading: it moves weeks or months before revenue does
[ ] 4. Retention link: accounts that hit it retain better (checked in cohort data)
[ ] 5. Not a headcount: it is not sign-ups, registered users or raw DAU
[ ] 6. Quality bar: it excludes low-value activity
[ ] 7. Game fit: it matches our game (attention / transaction / productivity)
[ ] 8. Decomposable: it breaks into 3-5 inputs that teams can move
[ ] 9. Trustworthy: one written definition, one data source, no manual stitching
[ ] 10. Hard to game: a growth hack cannot spike it overnight without real value
Score: __ / 10 Decision: adopt / refine / reject
Review date: __________ (at least quarterly)
How do you build an input metric tree?
A north star you cannot act on is a poster. Input metrics are the 3-5 levers that most directly drive it and that a team can influence through the product. Amplitude's heuristic is to look at four dimensions:
Breadth: how many users or accounts get value
Depth: how much value each one gets
Frequency: how often they come back for it
Efficiency: how quickly or easily they get it
Worked example: a B2B reporting tool (illustrative)
Say a 12-person team sells a reporting tool to finance teams. They pick a north star of weekly active reporting teams: accounts where 2 or more people opened or shared a report in the last 7 days. Today the number is 1,150. The target is 1,500 in two quarters.
North star: Weekly active reporting teams (1,150 → 1,500)
│
├── Breadth: new teams activated in their first 14 days
│ baseline 31% of trials → target 40%
│ owner: onboarding squad
│
├── Depth: reports shared outside the creator, per active team per week
│ baseline 2.1 → target 3.0
│ owner: collaboration squad
│
├── Frequency: active teams returning 3+ of the last 4 weeks
│ baseline 58% → target 65%
│ owner: core reporting squad
│
└── Efficiency: median minutes from data import to first shared report
baseline 46 min → target 20 min
owner: data import squad
Each squad now owns one number that rolls up to the one that matters. Every roadmap item should name the input it moves. An outcome-based roadmap is a natural home for these inputs, with each Now item tied to one branch of the tree.
Review the tree at least once a quarter, as Mixpanel also recommends (Mixpanel). When an input rises and the north star does not move, the tree is wrong, and that is useful to learn.
Should an early-stage startup have a north star?
Usually not a polished one. Lenny's guidance for teams before product-market fit is to ask whether you are building something people want and to watch cohort retention (a16z Future, 2021). A north star built on 30 accounts is mostly noise.
A practical middle ground for a 2-5 person team: track one value behaviour weekly (say, "accounts that completed the core job this week") and treat it as provisional. Promote it to a real north star once you have retention cohorts that show it predicts staying. Our product-market fit survey guide covers the other signal to watch at this stage.
Common mistakes and when a north star doesn't work
One number for a multi-sided business. Marketplaces often need a pair: one for supply, one for demand. Forcing one can hide a collapsing side.
A north star with no guardrails. Pushing engagement can raise support tickets or churn among new users. Track 2-3 health metrics next to it.
Copying another company's metric. The examples above reflect each company's game and strategy at one point in time.
Ownership by one team. If only growth owns it, other teams treat it as someone else's problem. Each input needs a named owner.
Treating it as the strategy. Lenny argues the north star is your strategy expressed as a number. That only works if the strategy came first.
When it does not fit: very early products (see above), agencies and services businesses where value is delivered off-product, and portfolios of unrelated products, which need one north star per product. For product-led companies, our guide to product-led growth metrics covers the activation and conversion metrics that sit beside it.
Where Specky fits
A north star only helps if you can see why it moved. Specky reads customer calls, feedback, Slack, Jira and analytics into a Product Graph, so when an input metric dips you can see the customer signals behind it, and it checks after launch whether a bet actually moved the number it targeted. Faster loops between metric, evidence and decision are what we call learning velocity, and keeping a shared north star across several squads is part of the product operating model for scaling teams.
Frequently asked questions
What is a good north star for a SaaS product?
For most B2B SaaS products, a good north star counts teams or accounts that completed the product's core job in a recent period, such as "weekly active teams that shared a report". It beats logins because it captures value, and it beats revenue because it moves earlier. Team-based tools in Lenny Rachitsky's survey often used paid or active teams.
Who invented the north star metric?
The term is credited to Sean Ellis, the growth marketer and author, according to Mixpanel. Amplitude later turned the idea into the North Star Framework and the North Star Playbook, co-authored by John Cutler, which added the three games, input metrics and the breadth, depth, frequency and efficiency heuristic.
Is revenue a good north star?
Usually not on its own. Revenue measures what customers pay, not the value they get, and it lags behind product changes by months. It was still the most common focus in Lenny Rachitsky's 2021 survey, at roughly half of companies. A better pattern is revenue as the business outcome and a customer value behaviour as the north star that leads it.
How many input metrics should a north star have?
Three to five. Amplitude recommends that range, and it maps neatly to the four dimensions of breadth, depth, frequency and efficiency. Fewer than three usually means one team owns everything; more than five spreads attention and makes it hard to explain which lever moved the north star.
How often should you change your north star?
Rarely. Commit for at least two quarters, and review the input tree every quarter. Change the north star itself only when the strategy changes or cohort data shows it no longer predicts retention. Changing it often teaches teams that the number does not really matter, which defeats the point of having one.
How is a north star different from OKRs?
The north star is a long-lived measure of customer value; OKRs are quarterly goals. In a healthy setup, most product key results are targets for the north star's input metrics, such as raising first-14-day activation from 31% to 40%. The north star stays stable while the key results change each quarter.
GDPR-aligned. EU workspace data is stored in Frankfurt, Germany. GDPR and EU data Built in Vienna, Austria. A product of influence solutions fm GmbH (FN 675374y). Imprint 14-day trial. Card required, and nothing is charged until day 14. 30-day refund after subscribing. See pricing