North Star Metric Examples, and How to Pick One for Your Startup

Common north star metric examples include daily active users at Duolingo (56.5 million in Q1 2026), Premium Subscribers and monthly active users at Spotify, Nights and Seats Booked at Airbnb, and trips at Uber. Each counts a unit of value customers get. Revenue metrics such as annual recurring revenue (ARR) were the focus of about 50% of companies in a Future survey, but revenue fails the customer value test on its own.

What is a north star metric?

A north star metric is a single number that tracks the value customers get from your product and signals whether the business will last. Mixpanel defines it as "the one measurement that’s most predictive of a company’s long-term success" (Mixpanel). For a short version to keep on hand, see our glossary entry.

Revenue was the most common north star category, at about 50% of companies, and growth efficiency and user experience were the least common, at about 10% each.

Share of companies focused on each north star category% of companies

1. Revenue50 % of companies
2. Customer growth35 % of companies
3. Consumption growth30 % of companies
4. Engagement growth30 % of companies
5. Growth efficiency10 % of companies
6. User experience10 % of companies

Share of companies focused on each north star category

Item	Value, % of companies
Revenue	50 % of companies
Customer growth	35 % of companies
Consumption growth	30 % of companies
Engagement growth	30 % of companies
Growth efficiency	10 % of companies
User experience	10 % of companies

The three tests

Mixpanel sets three conditions a candidate must meet: it must lead to revenue, reflect customer value, and measure progress. Run each candidate through all three before you adopt it.

Here is a worked example. Mixpanel describes an ecommerce company whose north star is the number of new customers purchasing each week.

• Leads to revenue: each purchase is money in.
• Reflects customer value: a completed order means the buyer found what they came for.
• Measures progress: a weekly count shows you which way you are moving within a month.

Daily visits would fail the first test, and revenue alone would fail the second. In our reading, a metric that passes only two of the three tests is an input metric, not the north star.

Where the term came from

Sean Ellis coined the term. He meant it to reduce administration, simplify meetings, and align teams around the singular goal of growth.

Keep it separate from the One Metric That Matters (OMTM). Mixpanel draws the line this way: the north star is a long-term guide, while the OMTM covers specific teams and projects for two to six months.

What companies choose

Category	Example metrics	Share of companies
Revenue	annual recurring revenue (ARR), gross merchandise value (GMV)	~50%
Customer growth	paid users, market share	~35%
Consumption growth	messages sent, nights booked	~30%
Engagement growth	monthly active users (MAU), daily active users (DAU)	~30%
Growth efficiency	lifetime value to customer acquisition cost (LTV/CAC), margins	~10%

Source: Future.

The shares add up to well over 100%. We read that as many companies tracking more than one category at a time, which fits a north star built from a few sub-metrics.

Revenue is the most common category, but it fails the customer value test on its own. In our reading, a startup does better to pick the usage or customer measure that comes before revenue and to watch revenue alongside it.

Which metrics do Duolingo, Spotify and Airbnb put first?

Public companies reveal their north star choices in the metrics they put at the top of their results. Below is what three consumer businesses led with, with figures exactly as reported. The last two columns are our reading, not the companies' statements.

Company	Headline metric in its release	Latest figure	Value it measures (our reading)	Likely input metrics (our reading)
Duolingo	Daily active users (DAUs)	56.5 million, Q1 2026	A learner coming back to study that day	New users, next-day retention, streaks kept, reactivated users
Spotify	Premium Subscribers and monthly active users (MAUs)	293 million subscribers, 761 million MAUs, Q1 2026	Listening people return to, and value it enough to pay	Free-to-Premium conversion, reactivations, days listened per month, churn
Airbnb	Nights and Seats Booked, with Gross Booking Value (GBV)	Nights and Seats up 10%, GBV up 16%, Q4 2025	Trips guests actually book	Search-to-booking conversion, active listings, first-time bookers, repeat bookings

Duolingo: one number, defined tightly

Duolingo states that it primarily evaluates engagement using DAUs, with MAUs as a supplementary metric. In Q1 2026, DAUs grew 21% year over year to 56.5 million, and the letter repeated a medium-term goal of 100 million DAUs in 2028. That leaves a gap of 43.5 million daily users. Every team's input metric has to add up to closing it.

The definition is the part to copy. Duolingo counts unique users who engage with our Duolingo App or the learning section of our website each calendar day, and reports a period by taking the average of the DAUs for each day in it.

Use the same four parts to write your own definition this week:

1. Who counts: unique users, accounts or workspaces.
2. What action: the step that delivers value, not a login.
3. What window: day, week or month, matched to natural use.
4. How you aggregate: an average across the period, or a count at period end.

Spotify: reach plus payment

Spotify's results list opened with Premium Subscribers grew 9% Y/Y to 293 million, followed by Monthly Active Users (MAUs) climbed 12% Y/Y to 761 million. One number tracks audience, the other tracks willingness to pay. In our reading, a freemium startup can run the same pair and watch the ratio between them against free-to-paid conversion benchmarks.

Airbnb: count the transaction

Airbnb reported that Nights and Seats Booked rose 10 percent, while Gross Booking Value grew 16 percent year-over-year. Both measure completed bookings rather than visits or searches. We think the volume metric makes the cleaner north star for a marketplace, because GBV also moves with price even when the number of trips does not.

Which metrics do Uber, Coinbase and Slack put first?

In all three cases the definition decides what counts more than the headline figure does.

Company	Metric it reports first	How the company defines it	Reported figure	What the definition leaves out (our reading)
Uber	Trips	Growth split into two drivers: Monthly Active Platform Consumers (MAPCs) and monthly trips per MAPC	More than 40 million trips every day	Trip value: a short ride and a long one count the same
Coinbase	Total Trading Volume	Spot plus derivatives trades in U.S. dollar value	Up 156% year over year in 2025	Fees earned per dollar traded
Slack (2019 filing)	Daily active users	Created or consumed content in a 24-hour period, free or paid	Exceeded 10 million	Whether the user pays

What each definition counts

Uber. Fourth-quarter 2025 trips grew 22% year over year to 3.8 billion, driven by MAPC growth of 18% and trips per MAPC growth of 3%. That split tells a team which lever moved. Uber pairs trips with money: $193 billion in Gross Bookings for the year.

Coinbase. Spot volume includes half of the value of trades routed off the platform. That rule changes the total. Over the same year, transaction revenue grew 2%. In our reading, that gap shows a volume metric can climb while fee income barely moves.

Slack. An organization became a Paid Customer only once it had three or more paid users. On that rule Slack reported more than 88,000 Paid Customers among more than 600,000 organizations with three or more users. That works out to roughly 15% (88,000 ÷ 600,000). A one-seat team paying alone never entered the paid count. For your own ratio, compare against our free-to-paid conversion benchmarks.

A definition card to fill in this week

Write your north star on one card before you build the dashboard:

1. Unit: the event you count (a trip, a trade, a content action).
2. Counts if: the condition that qualifies it, such as "created or consumed content."
3. Excluded: what you drop on purpose, such as accounts under a seat threshold.
4. Window: a day, a week or a quarter.
5. Drivers: two numbers that multiply to the total, the way Uber splits users and frequency.
6. Money check: the revenue figure you watch beside it so volume cannot drift from income.

What are input metrics, and how do they differ from KPIs and OKRs?

An input metric is a lever a team can move directly this week, one that rolls up into the north star. The north star tells you where the company is going. The inputs tell each team what to push on to get there.

Inputs are the translation layer. Each one names a customer behavior, a product area that shapes it, and a line back to the business outcome.

Mixpanel describes the structure as a metric tree, a logical hierarchy of your growth model that maps how input metrics (at varying levels of granularity) feed into focus metrics, which in turn impact your North Star. The practical payoff is speed. Shifts in engagement or churn often happen at the input metric level, well before they show up in your North Star.

A metric tree you can copy

Keep the list short: a few inputs, each with one owner. We group them by four questions.

Level	What it asks	Example for a team collaboration product (our illustration)
North star	Is customer value growing?	Weekly active teams
Breadth input	How many accounts get value?	New accounts reaching activation in week one
Depth input	How much of the product do they use?	Active members per team
Frequency input	How often do they come back?	Days active per team per week
Efficiency input	How fast do they reach value?	Hours from sign-up to first shared project

To fill it in, write the north star at the top, then ask of each candidate input: can one team move it inside a quarter, and does it plausibly move the north star if it rises. Drop anything that fails either test. Our growth model template has a blank version.

Where KPIs and OKRs fit

A key performance indicator (KPI) is any measure you track to judge health: churn, payback, support tickets, uptime. Your north star and inputs are KPIs, but most KPIs are neither.

An objective and key result (OKR) is a time-boxed goal. The objective states the intent and the key results state the measurable change. In our reading, the cleanest OKRs target an input metric, because a team can own an input in a way it cannot own the north star. For the tree above, an OKR might read: objective, get new teams to value faster; key result, cut hours from sign-up to first shared project from [baseline] to [target] by quarter end. Then run the work as a series of tests, as in our guide to running a growth experiment.

Attention, transaction and productivity products need different north stars

Before you shortlist candidate metrics, decide which game your product plays. Amplitude's engagement playbook frames retention around the "three games" of customer engagement. We use three labels: attention, transaction and productivity. The game sets what one unit of delivered value looks like, and the north star metric should count that unit.

Game	Unit of value	North star shape	Input metrics to watch (our reading)	Companies from earlier sections
Attention	Time or days spent	Active users per day, week or month	New user activation, day-7 return rate, sessions per active user	Duolingo, Spotify
Productivity	Work done efficiently	Teams or seats doing core work each week	Workspace activation, invites per new workspace, core actions per team	Slack

Attention: match the cadence to real habit

The choice inside this game is the time window. Facebook and Snap target daily active users (DAU) because social media is a daily habit for most. Pinterest uses weekly active users (WAU), since it doesn't expect its users to need the product daily. Spotify's podcast business hones in on monthly active users (MAU), likely because listening is sporadic.

Transaction: count completed units, not money

Marketplaces that take a cut focus on the volume of transactions: nights booked, rides taken, orders placed. Flat-fee platforms such as Twilio and Plaid spotlight activity: messages sent and bank accounts linked.

Productivity: engagement early, paying teams later

Freemium team products start with usage. Coda uses DAU14 and HubSpot uses WAU, while Airtable tracks Weekly Paid Seats and Asana tracks Weekly Active Paid Users. The Future survey found that the more mature and sales-driven the product is, the more companies focus on customers over engagement.

Worked example: placing a clinic scheduling tool

Take a scheduling product where patients book visits and front-desk staff manage calendars. It touches two games, so ask one question: which unit, if it doubled next quarter, would revenue follow?

• If clinics pay a flat monthly fee per seat, it plays the productivity game. North star: clinics where three or more staff schedule in a given week. Inputs: staff invited per new clinic and weekly scheduling actions per seat.

Write your answer down this week, then pick two candidates from your column of the table and test each against your pricing model.

How to choose a north star metric in one week

It takes five working days and one shared spreadsheet. Each candidate faces five tests. The first three come from Mixpanel's definition, which says a metric must lead to revenue, reflect customer value and measure progress. We add two of our own: it moves within weeks, and at least one team can change it this quarter.

Day 1: write the value statement

Write one sentence in this shape: [customer] gets [outcome] when they [action]. Keep rewriting until sales, product and support all sign it. If they cannot agree on the sentence, they will not agree on the metric.

Day 2: list five candidates

Turn the action in your sentence into five countable events. Use Mixpanel's examples as models, such as "Number of weekly customers completing their first order" and "Number of trial accounts with over 3 users in their first week." Each candidate needs a count, a time window and a qualifying threshold.

Day 3: pull 12 months of cohort data

Group users by signup month into cohorts. For each candidate, compare the cohorts that scored high on it with those that scored low. Check whether the high group retained and paid at higher rates. Our guide to reading retention curves shows what a healthy curve looks like.

Day 4: score each candidate

Score each test 0, 1 or 2, where 2 is a clear pass. The rows below are a worked example for a team scheduling tool, using invented scores.

Candidate	Customer value	Leads to revenue	Measures progress	Moves within weeks	Team can change it	Total (our scoring)
Weekly meetings booked	2	1	2	2	2	9
Workspaces with 3+ active bookers	2	2	2	1	1	8
Monthly signups	0	1	1	2	2	6
Paid seats	1	2	1	0	1	5

Paste the table into a sheet and add columns for your own tests if you need them. Any 0 on customer value rules a candidate out, whatever its total. Signups score well on speed, but they say nothing about whether customers got value.

Day 5: map the inputs and name owners

Break the winner into three to five inputs, the levers covered earlier in this piece. Give each input one named owner and one target for the quarter. A single owner keeps the metric from becoming a number that every team reports and nobody moves.

Owners can set shorter goals under the north star. Mixpanel describes The One Metric That Matters as a goal for specific teams and projects over a fixed period (two to six months). Log each owner's first tests in a growth experiment backlog, and rank them with an ICE score.

Worked example for a B2B invoicing startup

This company is hypothetical, and every number below is ours, built to show the arithmetic.

Then compare the accounts that stayed with the ones that left. A north star that tracks retention should look very different across the two groups.

A customer pays for the plan whether or not they use it, right up to the day they cancel, so revenue reports churn after it happens. Signups measure intent. Invoices paid on time measure the moment the customer gets value: cash lands in their account.

So the team picks invoices paid on time per week as its north star.

Break the north star into inputs

Invoices paid on time per week = activated accounts × invoices sent per account per week × on-time payment rate

Reminder adoption sits underneath the on-time rate, because a customer who sends a reminder on day one gets paid sooner.

The weakest link in this plan is the assumption that reminders raise the on-time rate. Test it before you build the quarter around it. Turn reminders on by default for half of new accounts, leave the other half as is, and compare on-time rates after four weeks. Our growth experiment guide covers the setup.

Review the north star and its inputs every Monday. If the north star stalls, check the inputs first.

Three mistakes: revenue as the north star, a vanity count, a metric no team can move

Most north star mistakes fall into three patterns. Each one feels safe on the day you pick it, which is why it tends to stick.

Revenue too early

Revenue is the obvious candidate. In a Future survey of employees at over 40 growth-stage companies, revenue metrics such as ARR (annual recurring revenue) and GMV (gross merchandise value) were "the focus of ~50% of companies." The same author warned that "a focus on revenue as a NSM can lead companies astray."

Dropbox shows where revenue fits. Its annual report names Total ARR "as the key indicator of the trajectory of our business performance." It calculates the figure from users with active paid licenses multiplied by their annualized subscription price. Even so, Total ARR decreased from 2024 to 2025.

A startup with forty customers cannot. Revenue records value customers received weeks or months earlier. We think revenue lags too far behind the product to guide this week's work.

A vanity count

Signups and MAUs (monthly active users) look like engagement but carry no meaning until you define the activity. Future separates consumption from simple visits: "creating a video, for example, rather than merely visiting the site." Write the action into the metric. Accounts that sent an invoice in the last seven days measures value. Accounts created measures acquisition spend.

Counting rules matter as much as the action. Dropbox, for example, counts one person "as multiple paying users if the person had more than one active license." Write your rule down before the first dashboard goes live. Our activation glossary entry covers how to define the qualifying action.

A metric no team can move

If no team can name an input it owns, the north star becomes a report nobody acts on. Mixpanel describes sub-metrics that are "team-specific and actionable, so individual contributors can draw a clear connection between their daily duties and the North Star."

The opposite risk is holding one metric too long. Future notes that teams risk "short-term thinking, missing new opportunities, and sacrificing the user experience."

Mistake	Symptom	Fix (our reading)
Revenue as north star	Weekly reviews explain past deals	Track a usage action that predicts the next renewal
Vanity count	Number rises while retention stays flat	Add the activity and time window to the definition
Unmovable metric	No team owns an input	Map one input metric to each team

When to revisit, and when to run several

Mixpanel says companies should "reevaluate their North Star metrics to make sure they still point the right direction, and amend them when they prove flawed."

Complex models get more than one. "Companies with complex business models can have multiple North Stars," Mixpanel notes. A marketplace might hold one for buyers and one for sellers. Give each its own owner and input tree.

What this page does not cover

This page does not cover how to instrument events or build dashboards in an analytics tool, how to set numeric targets for each input, or how to report metrics to a board or investors. Company figures are shown as each company reported them, and nothing here is financial or investment advice.

Related on Race to Scale: Growth loops explained, with the four loop types and a template to draw your own; Growth model template for startups, with inputs, a cohort layer cake, a loop row and a sensitivity table; How to read a retention curve: flattening, cohorts and the three mistakes.