All 29 modules are open from the start — nothing here is
locked, and nothing costs anything. Sign in so your progress, titles and
credentials stay with you, on every device you use.
Free forever, with your Google account. No password, no payment.
"Growth" gets stretched to mean signups, downloads, revenue, anything pointing up and to the right. But durable growth is a system: bring people in, get them to real value, and keep them long enough that each new group adds up instead of leaking away. This guide decodes what growth actually is and where it comes from.
Ready?
1
Growth Is More Than Acquisition
Nothing was broken about the marketing. The problem was that acquisition had never been the constraint, and pouring money into the top of a leaky bucket simply makes the leak more expensive.
This is the first idea worth internalising about growth: it is not one thing, and “we need more users” is a sentence that refuses to say which thing. Growth is a chain of separate modules, each with a different failure mode and a different fix, and adding effort to a healthy module does nothing if a later one is broken.
A restaurant makes it concrete. People walk past and notice it. Some come in and order. Some come back next month. Some spend more when they do. Some tell a friend. Those are five different businesses problems — a great location does not fix bad food, and great food does not fix an empty street.
The practical move is diagnostic rather than motivational. Before spending anything on growth, work out which module is actually constrained. The constrained stage is almost never the one currently receiving the most attention, because attention tends to follow whichever stage is easiest to buy.
The gap between the first bar and the second is the whole diagnosis: 72 of every 100 arrivals are lost before anything else gets a chance. Three times the top bar still hands 28 people to the second one — the money never reaches the stage that is actually holding the number down.
Everyday example, the leaky bucket
Picture growth as filling a bucket. Acquisition is water pouring in the top; churn is
water leaking out holes in the sides. If the holes are big enough, pouring faster barely
raises the level, you just spend more water to stay in the same place. Patch the holes
(retention) and the same inflow finally starts to fill the bucket.
Filling a leaky bucket faster
A company tripled its paid acquisition budget and watched total users stay almost flat. Every new cohort was leaving at roughly the rate the new spend brought them in. Acquisition was never the constraint — the bucket was — and money spent on the top of the funnel had simply made the leak more expensive.
Quick check
A team triples its ad spend and monthly signups double, yet total monthly active users stay flat. What's most likely happening?
2
The AARRR Funnel
The standard way to break growth into stages is a five-part funnel, usually written AARRR and pronounced, inevitably, like a pirate. The stages are Acquisition, Activation, Retention, Referral and Revenue.
Acquisition is how people find you. Activation is whether they reach first real value — the moment the product does the thing it promised. Retention is whether they come back. Referral is whether they bring someone else. Revenue is whether they pay.
The order matters, because each stage is fed by the one above it. Doubling acquisition while activation is broken simply doubles the number of people who bounce off. That is not growth; it is a larger disappointment, bought at full price.
So the useful work is diagnostic before it is creative. Measure each module separately, find the one where the drop is both large and plausibly recoverable, and make that the quarter’s growth problem. It is a far narrower and more answerable question than “how do we grow?”
A
Acquisition
How people first find and arrive at your product.
A
Activation
Whether they reach first real value — the moment the product does the thing it promised.
R
Retention
Whether they come back after that first experience.
R
Referral
Whether they bring someone else in.
R
Revenue
Whether, and how well, the product makes money from them.
Every stage is fed by the one above it, so the funnel can only be as wide as its narrowest point. The keylined stage is the one that decides whether any of the spending above it compounds — without it the other four just convert budget into churn, faster.
Everyday example
A restaurant has the same five modules: people walk past and notice it, come in and order, come back next month, spend more when they do, and tell a friend. Every one of those is a different problem with a different fix, and "we need more customers" is a sentence that refuses to say which.
Airbnb and the photographs
Airbnb's early growth was held back at activation rather than acquisition: listings with poor amateur photographs did not get booked. The founders hired professional photographers to shoot listings, and bookings rose sharply. It is the standard example of finding the constrained stage rather than adding traffic to a funnel that could not convert it.
Quick check
A user downloads the app (they arrived fine) but never finishes setup and never reaches the first value moment. Which AARRR stage is failing?
3
Retention Is the Engine
Of the five stages, one decides whether the other four were worth paying for. It is retention, and the reason is arithmetic rather than sentiment.
If customers leave at the rate you acquire them, your user count is flat no matter how much you spend — and every customer has to be bought again. If they stay, each acquired customer keeps paying, keeps referring, and the money spent to get them is spent once. Retention is what converts acquisition spending from a rental into a purchase.
The way to see it is the retention curve: plot the share of a sign-up group still active over time. It always falls at first. The question is where it stops.
A curve that flattens at 30% says a real group found lasting value, and that flat portion is a foundation the business compounds on. A curve that keeps sliding toward zero is a leaky bucket — growth is being rented from the acquisition budget, and it ends the day the budget does.
For the first months the dashed curve is the better-looking one, and a dashboard read in that window would say so. It has no floor — it is renting its users from the acquisition budget — while the solid one gives up more early and keeps 22% indefinitely. You are not looking for a high number; you are looking for where it stops falling.
Everyday example
A gym that signs up two hundred people in January and keeps eleven of them in March does not have a marketing problem. Retention is the stage that decides whether every other stage was worth paying for.
The shape of the retention curve
Plot the % of a signup group still active over time. If the curve keeps sliding toward
zero, every user eventually leaves, you're renting users, not keeping them. If it
flattens at some level (say, 25% still active at 6 months and holding),
you have a stable base each new group adds to. That flattening is one of the strongest
signals of product-market fit.
A flat retention curve means growth compounds. A curve heading to zero means it can't.
A floor worth more than a spike
Two products launched the same month. One retained 45% at week one and 4% by month three; the other retained 28% at week one and flattened at 22%. The second looked worse on launch day and was worth several times more a year later, because the flat part is the part you stop paying for.
Quick check
Product A retains 25% of each signup group at 6 months and the curve flattens there. Product B's retention curve keeps sliding toward 0%. Which has the healthier growth foundation?
4
Earned vs. Bought Growth
Growth comes in two kinds, and confusing them is how companies end up with impressive numbers and a business that stops the moment spending does.
Bought growth is growth you pay for directly — advertising, paid placement, outbound sales. It is predictable, it can be turned up with money, and it stops when the invoice stops. It is a billboard.
Earned growth comes from the product itself — word of mouth, referrals, content that ranks, network effects. It is slower to start, harder to control, and it keeps running after you stop spending.
Neither is virtuous. Bought growth is genuinely useful, especially early, when you need customers in order to learn anything at all. The danger is only in mistaking one for the other on a chart, because they behave completely differently under pressure — and pressure always arrives eventually.
The distinguishing feature of earned growth is that the output feeds back into the input: the thing the product produces is what brings the next person in. Bought growth has no such loop, which is why its cost per customer tends to rise over time while earned growth’s tends to fall.
The question worth asking of any growth number: what happens to this if we stop paying next month? The answer tells you which kind you have, and most teams have never asked it out loud.
Owned
Earned growth
Self-sustaining: happy, retained users refer others and come back on their own.
Rented
Bought growth
Works only while you pay. Fine as fuel, dangerous as the whole engine, and only if each customer is worth more than they cost to acquire.
There is no entry point drawn on this ring, because there does not need to be one after the first turn. Bought growth is the same four boxes with the last arrow missing — every customer has to be paid for separately, which is why its cost per customer climbs while this one’s falls.
Everyday example
Rented growth is a billboard: it works while you pay for it and stops the day you do not. Earned growth is word of mouth: slower to start, and it keeps running after the invoice stops.
Dropbox and the referral
Dropbox's referral programme gave both the inviter and the invitee extra storage, and is widely reported to have driven a very large share of its early sign-ups. What made it earned rather than bought is that the reward was the product itself — the more the loop ran, the more it cost Dropbox in storage and the less it cost in cash, which is the opposite of an advertising budget.
Quick check
A startup's growth comes entirely from paid ads; the instant it pauses spending, growth stops cold. What's the concern?
Drill what you learned
Scenario 1
easy
A dashboard proudly shows "cumulative registered users" crossing 5 million, and leadership calls it proof of strong growth.
What should you check before agreeing?
Scenario 2
easy
Signups are up 40% this quarter, but 6-month retention dropped from 30% to 18% over the same period.
How should you read this together?
Scenario 3
easy
Analysis shows most new users sign up but never complete the one setup step that unlocks the product's core value.
Which AARRR stage needs the most attention?
Scenario 4
easy
A PM argues: "Retention is a nice-to-have; our real job is filling the top of the funnel."
What's the strongest counter?
Scenario 5
easy
Two cohorts: January's retention curve flattens at 22%; June's keeps sliding toward zero.
What does this comparison tell you?
Scenario 6
medium
A startup grows 20% month over month, entirely from paid ads, with CAC creeping up and retention low.
What's the risk in this growth?
Scenario 7
medium
A team wants to boost referral (the second R) but users churn within a week of signing up.
What should come first?
Scenario 8
medium
Leadership celebrates that revenue doubled, but it came from doubling discount-driven signups who mostly cancel after the first month.
What's the concern?
Scenario 9
medium
A PM says the "aha moment" for their note-taking app is "a user creates their third note."
Why is identifying this moment useful for growth?
Scenario 10
medium
A growth team optimizes a signup-page button color for weeks and gets a 2% signup lift, while activation and retention go untouched.
What's the strategic critique?
Scenario 11
hard
A subscription app has healthy retention but almost no new users are arriving.
Which part of the growth system is the bottleneck here?
Scenario 12
hard
A founder equates "growth" strictly with "month-over-month new user count."
How would you broaden this definition usefully?
Scenario 13
hard
A team considers pouring its entire budget into a viral referral campaign for a product with a weak, confusing first-time experience.
What's the flaw?
Scenario 14
hard
Growth has been flat for two quarters. Acquisition and activation both look strong in the data.
Where should you look next?
Scenario 15
hard
A PM proposes judging the growth team purely on "new signups this month."
What behavior might this incentive accidentally encourage?
Drilled it. Now apply it to a real situation.
Put it to work
From lesson 1
Dropbox stopped buying users and started earning them
Dropbox
Early Dropbox tried paid search and found the economics impossible: the
cost of acquiring a customer through AdWords ran far above what a $99
product could justify. Drew Houston has said publicly that they were
paying hundreds of dollars to sell a $99 subscription.
The replacement was the double-sided referral — free storage for the
person inviting and the person joining — which is widely reported to have
driven signups up dramatically, on the order of 60% of signups coming
through referrals at its height.
Why did the referral programme work where paid acquisition could not?
What actually happened
The distinction generalises. Paid acquisition is a funnel — money in,
users out, and it stops the day the money does. Referral was a loop, where
users produced users, and the same spend compounded rather than repeating.
Growth is not a synonym for acquisition. Acquisition is one of five things
that move it, and usually the most expensive.
From lesson 2
Fixing the wrong stage of the funnel
A grocery-delivery app
Growth has stalled. The team's plan is a bigger performance-marketing
budget. Before signing it off you pull the AARRR numbers.
Last quarter, per 1,000 app installs
Stage
Count
Rate
Installs
1,000
—
Registered
620
62%
Placed first order
180
29% of registered
Ordered again within 30 days
38
21% of first orders
Still ordering at 90 days
14
8% of first orders
Where should the money go?
What actually happened
They spent the quarter on the second order instead: a reminder tied to the
household's actual reorder cycle, and a discount on order two rather than
order one. Repeat-within-30-days went from 21% to 34%, which was worth
more than doubling the ad budget would have been, at a fraction of the cost.
Find the steepest drop before deciding where to spend. The top of the
funnel is the most expensive place to fix a problem that lives further down.
From lesson 3
The leaky bucket, with numbers
A SaaS design tool
Two companies, identical acquisition — 1,000 new customers a month, and no
growth in that rate. The only difference is monthly churn.
Steady-state customer base at 1,000 new customers per month
Monthly churn
Average lifetime
Customers at steady state
10%
10 months
10,000
7%
14 months
14,300
5%
20 months
20,000
3%
33 months
33,300
2%
50 months
50,000
What does this table establish about retention?
Select all that apply — there are 3 to find.
What actually happened
The same arithmetic explains why retention work is usually undersold
internally: its effect arrives slowly, as a rising steady state rather than
a spike, so it never produces the chart that an acquisition campaign does.
Retention is the engine because it is a multiplier, not an addend. Fixing
it improves the return on every user you have ever acquired.
From lesson 4
Two channels, same CAC, different futures
A note-taking app
Two experiments ran for a quarter, each with a £40,000 budget and each
producing about 8,000 new users — an identical £5 cost per acquisition.
The growth lead wants to double down on the better one.
Two channels, one quarter
Paid social
Public-page sharing
Users acquired
8,100
7,900
Cost per acquisition
£4.94
£5.06
Users acquired in the month after spend stopped
40
2,100
Average invites sent per new user
0.1
1.4
Same users, same cost. What separates them?
What actually happened
Both were kept, for different jobs: paid social to fill specific gaps on
demand, sharing as the thing to invest in structurally. Eighteen months
later sharing accounted for the majority of new users, at a marginal cost
close to zero.
Judging a channel on cost per acquisition alone treats bought and earned
growth as the same thing. Ask what each user produces after they arrive.