Growth Hacking Is a System, Not a Bag of Tricks

Ask ten marketers what growth hacking is and most will describe a trick: a referral bonus, a clever pop-up, a viral loop that worked for a famous startup. Copying tricks rarely works, because the trick was the output of a process, not the process itself. Done properly, growth hacking is a system for running experiments quickly and honestly, so your team learns what drives growth faster than your competitors do.

What the term really means

Sean Ellis coined the term "growth hacker" in 2010 to describe a marketer whose main focus is growth. Since then it has been stretched to mean almost anything, from automation tools to aggressive spam. Strip away the hype and three things remain:

• One growth metric that the whole team works on.
• A steady flow of small experiments across the entire customer journey, not just in ads.
• A habit of writing down what each experiment taught you, including the failures.

It is not a separate department and it does not need a large budget. It needs discipline and a weekly rhythm.

Step 1: Choose one metric that matters

Pick the metric that best reflects the value customers actually receive. For a food delivery service it might be repeat orders per month; for a B2B software product, accounts that use the key feature every week; for a beauty salon, booked appointments. Followers and impressions are rarely the right answer, because they can grow while the business does not.

Then break the metric into its drivers. The classic AARRR model (acquisition, activation, retention, referral, revenue) is a useful map. Each experiment should aim at one driver, so you always know which part of the funnel you are trying to improve.

Step 2: Build an idea backlog

Collect ideas from everyone who touches the customer: sales, support, product, delivery staff. Customer questions and complaints are usually the richest source. In my experience, the best growth ideas rarely come from brainstorming sessions; they come from a salesperson who hears the same objection every day.

Keep one shared sheet. Write each idea in one line and do not judge it yet. At this stage the goal is volume.

Step 3: Turn ideas into hypotheses

An idea becomes testable when you write it in this format: "If we [change], then [metric] will [move], because [reason]." For example: "If we show the expected delivery time on the product page, the add-to-cart rate will increase, because uncertainty about delivery makes people hesitate."

The "because" part matters most. It is what you are really testing, and it is what you will learn even if the result is negative. Before the test starts, also write down what result will count as success and how long the test will run. Deciding that afterwards is how teams fool themselves.

Step 4: Prioritise with ICE

You will always have more ideas than capacity. ICE scoring is a simple way to choose. Rate each idea from 1 to 10 on three questions, then take the average:

• Impact: if it works, how much will it move the metric?
• Confidence: how sure are we that it will work, based on data, customer feedback or past tests?
• Ease: how quickly and cheaply can we run it?

ICE is subjective, and that is fine. Consistency matters more than precision: score as a group, quickly, using the same definitions every time. Watch for one trap: if ease always wins, your backlog fills up with button colours and headline tweaks. Keep some capacity every month for one bigger bet with high potential impact.

Step 5: Run the test properly

• Change one thing at a time, so you know what caused the result.
• Agree on the sample size or duration before launch, and do not stop the test the moment it looks good.
• Do not let tests on the same audience overlap, or the results will mix.

Small markets have a specific challenge: limited traffic. An A/B test on a page with a few hundred visitors a week may need months to show a reliable difference. In that case, test bolder changes (bigger effects are easier to detect), run tests in channels where you have more volume, such as paid social or messages to your customer base, or start with qualitative checks. Five customer calls can kill a weak idea before you spend a week building it.

Step 6: Learn and document

Every finished test gets a short record: the hypothesis, what you did, the result, what you learned and the next step. A failed test with a clear lesson is worth more than a "win" nobody can explain. Over time this library becomes the team's real asset. New people read it instead of repeating old mistakes, and discussions are based on evidence rather than opinion.

Common mistakes to avoid

• Copying a famous tactic without the context that made it work.
• Running many tests at once on the same small audience.
• Declaring a winner after two days.
• Testing only ads and ignoring onboarding, pricing pages and retention.
• Growth at any cost: spammy referral schemes or fake urgency bring short-term numbers and long-term damage to trust.

A realistic weekly rhythm

1. Monday: a 45-minute growth meeting. Review last week's results, score new ideas, choose one to three tests.
2. Tuesday to Thursday: build and launch.
3. Friday: write up every test that has finished.
4. Once a month: review the learning library and check whether your funnel priorities have changed.

Start next week with one metric, ten ideas and an ICE sheet. Run two tests and write down what you learn. In three months you will have something more valuable than any trick: a team that knows, from evidence, what makes your customers act.