Growth hacking, and what it means once the hype is gone

Growth hacking was a real idea before it became a job title. What survives is a method, and the method is worth keeping even though the word has been worn out.

The method is this. Find the one thing currently limiting growth, run cheap tests against that one thing, and keep only what compounds.

The part that made it different

Traditional marketing buys attention and measures the result. The original growth work changed the product so that using it produced distribution.

The examples people quote are the ones where the mechanism was built in. Referral credit that both sides receive. A file-sharing link that necessarily reaches someone who does not have an account yet. A free tier that is useful enough to be recommended.

None of that is a campaign. It is a product decision with a distribution consequence, which is why it needed someone who could work across both.

The constraint is the whole job

At any time one thing limits growth. Acquisition, activation, retention or revenue. Working on the other three produces measurable improvements that do not move the business.

A product where people sign up and never return has a retention constraint. Doubling the traffic into it doubles the number of people who leave. That is the most expensive mistake available and it is made constantly, because acquisition work is easier to commission and easier to report.

Find the constraint first. Everything else is second.

Testing cheaply

The cheap part matters as much as the testing part. A test that takes a quarter to build has already cost more than the information is worth.

Run the version that takes a day. A landing page for a feature that does not exist yet, to see whether anyone clicks. A manual process behind a form, to see whether anyone wants the automated version. A price change on one segment.

Most tests fail. That is the design. A programme where everything succeeds is testing things that were already known.

What compounds and what does not

A discount produces a spike and trains people to wait for discounts.

A referral mechanism produces users who bring users. A content asset keeps earning after it is published. A product improvement that raises retention raises the value of every acquisition channel at once.

Prefer the compounding one even when it is slower, because the spike has to be repurchased every time.

What it is not

A list of tricks. The articles promising fifty growth hacks describe other companies’ answers to other companies’ constraints. Copy an answer without the constraint and you spend a quarter on a referral programme for a product nobody retains.

It is also not an excuse to skip the fundamentals. A product people do not want cannot be hacked into growth.

Where it applies

Best in products with self-service sign-up, short feedback loops and usage data. Worst in long enterprise sales cycles, where the loop is quarters long and the sample is small.

In the second case the method still applies, and the tests are qualitative and the cadence is slower.

Starting

Write down your funnel with real numbers at each stage. Find the stage with the worst drop. List ten cheap things that might change it. Run them in order of cost.

That is the whole practice, and it does not need the name.

If the constraint turns out to be demand, that is growth work. If it turns out to be the product, that is engineering.