Most teams talk about “growth” as if it were a single number to push up. It is not. Growth is the product of several smaller behaviors — sign-ups, activation, frequency of use, retention — and each of those behaviors responds to different levers. The growth equation breaks the abstract goal of “grow the business” into the specific variables a team can actually move.
This post covers three connected ideas that turn growth from a slogan into a system: the growth equation, which decomposes growth into component variables; the North Star Metric (NSM), which focuses the team on a leading indicator of user value; and the growth experiment framework, which converts hypotheses into weekly learning. Together, they form the operating layer of growth hacking — the layer that sits between product-market fit and channel-specific tactics like acquisition or activation.
What is the Growth Equation? Breaking Growth into Component Levers
A growth equation describes how a business actually grows, expressed as the product of a few component variables. Instead of staring at one top-line number, the team can see which lever — sign-ups, activation rate, value per user, retention — is moving and which is stuck.
The framework was popularized by Andy Johns, a growth leader at Facebook, Twitter, and Wealthfront. The core idea is that every growth conversation should start from a clear equation — without one, the team cannot tell whether a tactic is treating the cause or the symptom.
A growth equation does not have to end in revenue. For many products, focusing on a leading indicator of value is more actionable than chasing revenue directly. A messaging app that obsesses over revenue per user will miss the upstream variable — messages sent — that actually drives monetization. The equation forces you to name that upstream variable.
Two things change once a team adopts a growth equation:
- Debates become specific. “We need more growth” becomes “activation rate dropped last quarter — what happened?”
- Roadmaps map to levers. Each initiative ties back to one variable in the equation, which makes prioritization easier.
Growth Equation Examples by Product Type
The shape of the equation depends on the product. Below are six common product types and the variables that tend to matter.
Messaging app:
(New sign-ups) × (Activation rate) × (Messages per user) × (Retention rate)
Subscription meditation app:
(Subscribers) × (Monthly price) × (Average subscription length)
Content platform:
(Visitors) × (Articles read per visit) × (Time on article) × (Return rate)
Online marketplace:
(Category expansion) × (Inventory per category) × (Product page traffic) × (Purchase conversion) × (Average order value) × (Repurchase rate)
B2B collaboration tool:
(Team sign-ups) × (Activation rate) × (Collaboration events per team) × (Weekly active rate)
E-commerce store:
(Visitors) × (Conversion rate) × (Average order value) × (Purchase frequency)
Notice that each equation optimizes a different target. Some focus directly on revenue, others on the user behaviors that create value. Behavioral variables — messages sent, articles read, collaboration events — are often leading indicators of future revenue and retention. They move first, and the financial number follows.
Choosing the Right Variables: Directness and Actionability
A useful growth equation balances two properties.
- Directness — the variable should sit close to the value the user receives. A user who sends more messages is getting more out of a messaging app; a user who completes more workouts is getting more out of a fitness app.
- Actionability — the team should be able to move the variable through product, marketing, or onboarding decisions.
“Messages per user” passes both tests. It is direct (more messages means the product is doing its job), and it is actionable (onboarding flows, notifications, and feature design can all influence it). A variable like “user satisfaction in surveys” might be direct but is harder to move week by week. A variable like “ad impressions” might be easy to move but says little about whether users are getting value.
Selecting Your North Star Metric: A Leading Indicator of User Value
Once the growth equation is in place, the next question is which component best signals that users are receiving the product’s core value. That component becomes the North Star Metric — the single number that, if it goes up, means everything else is on track.
A few well-known North Star Metric examples:
- Airbnb: nights booked (not revenue, not listings)
- Slack: messages sent by teams (not workspaces created)
- Medium: total reading time (not page views)
- Spotify: time spent listening (not sign-ups)
Each of these is a leading indicator, not a financial outcome. When Airbnb’s nights booked rises, revenue follows. When Slack’s messages-per-team rises, retention and paid conversion improve. The NSM is chosen precisely because it moves first.
A good North Star Metric meets four conditions:
- It is a leading indicator of business outcomes, not a lagging financial metric.
- It reflects a behavior that shows users are getting value, not a vanity count.
- It can be measured in real time, not just quarterly.
- It is movable through product and growth work, not fixed by external factors.
Put together: use the growth equation to decompose the business into component variables, identify which variable best signals user value, and then orient the team around that leading indicator instead of around lagging revenue numbers. The team still tracks revenue, of course. But the daily work points at the upstream variable that creates revenue downstream.
The Growth Experiment Framework: Building a System for Rapid Learning
The companies that grow fastest do not have one brilliant insight. They run more experiments per week than their competitors do per quarter. What looks like an overnight win is usually the cumulative result of dozens or hundreds of small tests.
Before any of that works, the data infrastructure has to be in place. Page views from a generic analytics tool are not enough. The team needs event-based tracking that captures four kinds of data:
- User actions — clicks, sign-ups, purchases, feature use
- User attributes — demographics, acquisition channel, device, plan tier
- Timestamps — when each action happened and the time between actions
- Context — which page the user was on, what they did before, the full path through the journey
Tools like Mixpanel and Amplitude can help, but the harder work is deciding what to track. Three questions are worth asking before instrumenting anything:
- Which events actually matter for understanding user behavior?
- Which attributes will the team use to segment users?
- Can the team trace the full journey from first touch to conversion?
Once that foundation exists, the experiment cycle has four steps — analyze, ideate, prioritize, test — followed by a weekly rhythm that keeps the cycle moving.
Step 1 — Analyze: Finding Bottlenecks in Funnels, Cohorts, and Segments
Every experiment starts with a clear read of the current state. The team looks at user behavior from three angles.
Behavior patterns:
- How are users moving through the product now?
- Where do they spend the most time?
- Which features drive the most engagement?
User characteristics:
- Who are the most valuable users?
- What do they have in common?
- How does behavior differ across segments?
Friction points:
- Where in the funnel do users drop off?
- Why are they dropping?
- What patterns appear in customer support tickets?
The goal is not exhaustive analysis. The goal is to surface specific, actionable insights — outliers, unexpected patterns, obvious opportunities — that can power an experiment.
Step 2 — Ideate: Generating Hypothesis-Driven Experiment Ideas
Once the analysis surfaces a problem, the team generates ideas for how to solve it. This step works best as a cross-functional exercise. Engineers, designers, marketers, and data analysts each bring a different angle, and the combination produces ideas no one person would have reached alone.
The most important rule at this stage: do not evaluate yet. Quantity and creativity matter here. Anyone should be able to propose an idea without fear of being judged.
Each idea should include the following fields:
- Title — a clear, descriptive name
- Target — who is affected (all visitors, new users only, a specific segment)
- Change — exactly what will be modified
- Location — where in the product or customer journey this happens
- Hypothesis — why this is expected to work, with data or research support
- Success metric — how impact will be measured
- Success threshold — what counts as “success”
Example:
Title: Simplify mobile onboarding
Target: First-time mobile app users
Change: Reduce the sign-up form from seven fields to three (email, password, name)
Location: Initial sign-up screen
Hypothesis: Analytics show 45% of mobile users drop out mid-sign-up, and average completion time is 3.2 minutes on mobile versus 1.1 minutes on desktop. User interviews also surfaced complaints about typing on mobile. Reducing friction should improve completion.
Metric: Sign-up completion rate, time to complete
Threshold: 15% improvement in mobile sign-up completion
A well-formed idea answers the basic questions of any growth experiment before the team commits engineering time to it.
Step 3 — Prioritize: ICE Scoring for Test Selection
Once a backlog of ideas exists, the team needs a way to decide which to test first. ICE scoring is the simplest framework that works for this.
- Impact — how much can this move the core metric? (1–10)
- Confidence — how sure are we that it will work? (1–10)
- Ease — how easy is it to implement? (1–10)
Teams differ on whether to multiply the three scores or average them, but the important part is applying a consistent rule across all ideas. Start with the highest-scoring ideas.
One caveat: an ICE score is a starting point, not an absolute ranking. A lower-scoring experiment can still be the right next move if it aligns with long-term strategy or unblocks a follow-up test. Use the score as a guide, not a verdict.
Step 4 — Test: Running Clean Experiments
Once priorities are set, the team runs the test. This is the step where discipline matters most, because sloppy execution invalidates whatever was learned.
Before the test:
- Confirm tracking is in place so results can be measured.
- Decide sample size and test duration in advance.
- Document exactly what is being tested and why.
- Notify other teams to avoid conflicts (a large feature launch landing mid-test, for example).
During the test:
- Watch for technical issues.
- Watch for unexpected user behavior.
- Resist the urge to peek at results too early. Statistical significance takes time to accumulate.
After the test:
- Analyze the results carefully.
- Share what was learned with the rest of the organization.
- Decide whether to ship, iterate, or kill the change.
Every experiment should produce a short report with the following:
- Experiment details — what was tested, on whom, how
- Test type — feature change, message change, pricing change
- Results — impact on the core metric and statistical significance
- Duration — start and end dates
- Hypothesis vs. outcome — what was expected versus what happened
- Possible confounders — external events that may have affected results
- Conclusion and next steps — what was learned and what to do next
Share the report widely. A failed experiment that produces a real learning is as valuable as a successful one — sometimes more so, because failures tend to correct mistaken assumptions that would have led to bigger investments later.
Weekly Growth Cadence: Meetings, Backlog, and Experiment Velocity
Experiments lose momentum without a regular rhythm. A weekly cadence keeps the cycle moving.
A standard weekly growth meeting fits in one hour:
- 15 minutes — review core metrics and focus areas
- 10 minutes — discuss results from the past week’s tests
- 15 minutes — discuss what was learned from completed experiments in depth
- 15 minutes — decide which tests to run this week
- 5 minutes — check that the idea pipeline is healthy
Tuesday tends to work well as the day. The team has Monday to prepare and review results, and the rest of the week to execute.
One rule matters more than the schedule itself: the meeting is for discussion and decisions, not for generating ideas. Ideas should be added to a shared backlog between meetings, where teammates can read and react before the meeting starts. A meeting that turns into a brainstorm wastes everyone’s hour and produces shallower decisions.
Speed matters most in growth hacking. The teams that learn fastest win. The point is not to run bigger experiments but to run more of them. Start with a sustainable pace — two or three experiments per week — and increase the rate as the process settles in.
Conclusion
The growth equation breaks the abstract goal of “grow” into specific variables the team can actually move. The North Star Metric focuses everyone on a single leading indicator of user value. The growth experiment framework — analyze, ideate, prioritize, test, repeat weekly — converts hypotheses into learning at a pace that compounds.
These three pieces sit between product-market fit and the channel-specific work that comes next. Once a team can name its equation, its North Star, and its weekly experiment cadence, it has the operating system it needs to grow with intent rather than by accident.
The next post in this series covers the AARRR funnel and the first stage of that funnel — user acquisition.
Growth Hacking Series
(1) What is Growth Hacking? Definition and Why Product-Market Fit Comes First
(2) Growth Equation and Experiment Framework: How to Decompose Growth into Levers
(3) User Acquisition Strategy: From AARRR Funnel to Channel Optimization
(4) User Activation Strategy: From Onboarding to the First Aha Moment
(5) User Retention Strategy: Cohort Analysis and the 3 Stages of Retention
(6) Monetization Strategy: From Pinch Points to Price Optimization
(7) Sustainable Growth Hacking Techniques: 6 Principles and a Growth Readiness Checklist