Growth marketing for startups: a working method, not a bag of tricks

What is growth marketing?

Growth marketing is marketing that follows the customer past the sign-up. Classic marketing often stops at the lead or the sale. Growth marketing keeps asking: did they activate, did they come back, did they bring anyone, did they pay more over time? Every 'no' marks a place where growth leaks, which makes it a place to experiment.

Growth marketing

An approach that treats growth as a system measured across the whole customer journey, from first touch to renewal, and improves it through experiments: hypothesis, test, measurement, decision.

For a startup this matters because money and weeks are scarce. You cannot afford 'awareness' campaigns that teach you nothing. So growth work starts with narrow questions, such as whether a pricing page converts better with a demo button or a free trial, rather than with an annual plan.

Is growth marketing the same as growth hacking or performance marketing?

Not quite: growth hacking is its ancestor and performance marketing is one of its parts. Sean Ellis popularised the phrase in his July 2010 post Find a Growth Hacker for Your Startup, defining a growth hacker as someone whose true north is growth. The label later drifted towards hacks and shortcuts, and 'growth marketing' became the more common name for the same numbers-first mindset applied to the full lifecycle.

The table is our own summary, not an official taxonomy; companies draw these lines differently. The useful row is the last one. Each approach has a typical failure, and growth marketing exists precisely so that you do not stop at cost per acquisition.

When should a startup start growth marketing?

Start once at least some users stick around without being chased. Before that, pushing acquisition is pouring water into a leaking bucket: you pay to bring in people who leave. What you are looking for is product-market fit, even if only within one narrow customer group.

In the same essay, Graham calls recruiting users by hand the most common unscalable thing founders have to do at the start. Calls, one-to-one demos and hand-held onboarding are not growth marketing in the strict sense, yet they produce the hypotheses you will later test at scale.

How do you run growth experiments week by week?

Run a short, repeating loop with one north-star metric and a ranked backlog of hypotheses. This is the loop we suggest to the founders we work with; one turn takes one or two weeks.

How fast is fast enough? The best-known yardstick comes from Y Combinator.

In Startup = Growth Graham adds that 10% a week is exceptional and that 1% suggests you have not yet worked out what you are doing. Use it as an order of magnitude that pushes you to think in weeks, not as a target for a board slide: it describes selected companies, measured on a metric of their choosing.

Which metrics matter at each stage of the funnel?

Track one metric per stage of AARRR, the 'pirate metrics' Dave McClure set out in his Startup Metrics for Pirates talks. The point of the model is to show where people drop out between stages.

You do not need to polish all five on day one. Start with the weakest stage; in a young startup it is usually activation or retention rather than acquisition.

Which channels should you test first?

Test first where your customers already look for a solution or already gather, and cap it at two or three channels. One channel tested properly teaches more than five tested halfway.

• Paid search captures existing demand; size the budget around cost per customer, as our guide to Google Ads for startups explains.
• Paid social creates demand where none exists yet and lives on creative testing; see Meta ads for startups.
• SEO and content build a durable channel but take months; start with SEO for startups.
• Founder-led sales and communities are slow and manual, and often the fastest route to the first ten customers.
• The product itself: invites, sharing and free tiers are the territory of product-led growth.

Which of these fits depends on how you take the product to market: who buys, at what price, through which sales motion. If that is still open, settle your go-to-market strategy first.

How much should a startup spend on marketing?

Enough to learn something decisive per experiment, and more only where the cost per customer holds up. Published benchmarks rarely fit an early startup. Gartner's 2026 CMO Spend Survey puts average marketing budgets at 7.8% of company revenue, but it surveyed 401 marketing leaders in North America, the UK and Europe, most of them at companies with over $1 billion in revenue.

A percentage of revenue means little when revenue is close to zero. A better frame is the budget per test: what you need to reach a sample big enough to decide, multiplied by the number of tests you can run before your next funding milestone.