Growth MarketingExperiments across the lifecycle

What makes it different#

Traditional marketing teams often own acquisition and hand over at the point of sale. Growth marketing looks at the whole lifecycle and asks where the biggest constraint on growth sits. Sometimes it is acquisition. Often it is activation or retention: customers arrive but do not get value, or do not come back.

The second difference is method. Growth marketing runs as a series of experiments with hypotheses, measures and decisions. It borrows from product management and data analysis as much as from advertising. That makes it powerful in businesses with fast feedback loops, such as apps, subscriptions and e-commerce, and harder to apply where cycles are long.

The term is sometimes used loosely, or as a synonym for “growth hacking”. The useful core is not tricks or shortcuts. It is disciplined learning about which levers actually move revenue.

The lifecycle view#

The value of this structure is diagnostic. If many people sign up but few activate, buying more traffic only multiplies the leak. If customers activate but churn quickly, the issue may be product fit or onboarding. Growth marketing directs effort to the stage with the biggest gap. See funnel analysis and cohort analysis.

The experiment loop#

A good hypothesis states the change, the audience, the expected effect and the reasoning. Writing the reasoning matters, because a failed test with a clear hypothesis teaches you something about customers; a failed test without one teaches nothing.

Prioritisation frameworks that score ideas on impact, confidence and ease are common. Their value is less in the precise scores than in forcing a conversation about why one idea should go before another. See A/B testing for test design.

The metrics that matter#

Growth teams usually anchor on a north star metric: a single measure that reflects customers getting value and the business growing because of it. Beneath it sit stage metrics such as activation rate and retention by cohort, and economic metrics such as acquisition cost and lifetime value.

This simple model ignores discounting, returns and the time value of money, so treat it as a first check. For a fuller method see CAC and LTV.

Growth loops versus funnels

A funnel is linear: you pour people in at the top and some come out at the bottom. A growth loop is circular: the output of one cycle becomes input for the next. A customer who invites colleagues, creates content others find through search, or generates reviews that attract new buyers is part of a loop.

Loops matter because they compound. A business whose growth depends entirely on paid acquisition must keep paying for every new customer. A business with working loops gets some growth from its existing customers. Look for loops you can strengthen: referral, user-generated content, integrations, word of mouth.

Retention: the quiet growth lever#

Acquisition gets the attention, but retention often determines whether growth compounds or stalls. A business that keeps customers longer earns more from each one, can afford to pay more to acquire the next, and benefits from more referrals. A business that loses customers quickly has to run faster every month just to stand still.

Cohort analysis makes this visible. Group customers by the month they joined and track how many remain active or buy again over time. If retention curves flatten at a healthy level, the product is delivering lasting value. If they fall steadily towards zero, acquisition spend is filling a bucket with no bottom.

• Improve onboarding so more customers reach value early.
• Use lifecycle messaging triggered by behaviour, not by a calendar. See lifecycle marketing.
• Find the habit moments when customers naturally return, and support them.
• Listen to churned customers to learn what failed.

Running an experiment programme#

A growth programme is a rhythm, not a one-off project. Many teams hold a weekly session to review results, decide what ships next and update the experiment log. The log is the most valuable asset the programme creates: a record of what was tried, why, what happened and what it taught.

Set a sensible pace. Running more tests than you have traffic to support produces inconclusive results. Before each test, estimate the sample size needed to detect the change you care about, and avoid stopping tests early because an early result looks exciting. When traffic is limited, test bigger changes that are more likely to show a clear effect.

Who does growth marketing#

Growth work cuts across marketing, product, engineering and data. Some companies form a dedicated growth team; others give a growth lead authority to run experiments across functions. Either way, growth marketing stalls when every test needs a long approval process or depends on another team’s backlog.

• Analyst skills to find constraints and read results.
• Marketing skills for acquisition channels, messaging and lifecycle communication.
• Product skills for onboarding, activation and in-product prompts.
• Engineering access to build and ship tests quickly.
• Design and copy to create variants worth testing.

Growth marketing outside software#

The discipline grew up in apps and online services, but its method travels. An e-commerce brand can test onboarding for first-time buyers, post-purchase messages, bundles and repeat-purchase prompts. A services business can test enquiry forms, response speed, proposal formats and referral asks. A retailer can test loyalty mechanics and local offers.

The constraint is usually feedback speed and volume. Where results take months to appear, use leading indicators that correlate with the outcome and accept that fewer, larger experiments will be possible. The habit of writing hypotheses and recording what was learned is valuable even when formal testing is not.

Where growth marketing goes wrong#

• Running many small cosmetic tests while the real constraint goes untouched.
• Declaring winners on too little data. Use a sample-size estimate before testing.
• Optimising a stage metric that harms another, such as aggressive discounts that lift conversion but damage retention.
• Ignoring brand: short-term experiments can erode trust if they rely on manipulation.
• Letting learnings vanish. Keep an experiment log the whole company can read.

More traffic into a leaking bucket is not growth; it is a more expensive leak.

Published by Fabulous.Media, a network of specialist marketing agencies. Updated 9 October 2026. Platform features change often; check current official documentation before acting on platform-specific detail.