The growth marketing framework: the experiment process step by step

When you hire an agency for growth marketing, you are not buying a one-off campaign. You are buying a way of working. A growth marketing framework is a repeatable process that lets you systematically discover what drives growth and what does not, instead of betting on one big idea. TL;DR: you collect ideas, prioritise them objectively, test them in short cycles and build on what works. In this article we make that process concrete, so you know exactly what you are buying and what a collaboration looks like week after week.

Plenty of buyers rightly ask us: where does my budget go? The honest answer is that it does not go into one spectacular launch, but into an engine that gets a little smarter every month. That is less exciting to pitch, but it is what makes growth predictable. If you want the broader picture first, read our explanation of what growth marketing is and how it connects SEO, CRO, content, paid and lead generation into a single growth engine.

Why a process, and not a one-off tactic

It is tempting to see growth as a bag of tricks: a new ad here, a button you change there. But one-off tactics have a problem. You never know for sure why something did or did not work, so you cannot repeat it or scale it either. A process solves that. By treating every idea the same way, you build knowledge that sticks, even when the market, the season or the team changes.

The second argument is focus. An organisation always has more ideas than capacity. Without a process, the idea belonging to the person with the highest job title or the loudest argument wins. With a process, the idea with the best expected impact wins. That difference decides whether your budget goes towards growth or towards opinions.

Before we unpack the steps one by one, here is the process at a glance. The five steps are not a straight line but a cycle: each round feeds the next, so your growth engine gets sharper sprint after sprint.

Step 1: collect ideas in a central backlog

It all starts with a backlog: one shared list of every idea that could accelerate your growth. Those ideas come from everywhere. From data on where visitors drop off, from conversations with sales, from customer interviews, from what a competitor is doing and from the gut feeling of someone who knows your audience well. Gut feeling is allowed, as long as the idea then goes through the same process as all the others.

A good backlog is not a vague dumping ground, but a structured list. Every idea gets at least a short description, the place in the funnel where the effect should land and the assumption underneath it. That way ideas do not evaporate and the same proposal does not resurface every month. The backlog is the memory of your growth process.

Step 2: prioritise with ICE scoring

This is where it gets concrete, because this is where most teams get stuck. You have fifty ideas and capacity for five. Which five? For that we use ICE scoring, a simple model that rates every idea on three dimensions:

• Impact: how big is the expected effect on your goal if this experiment succeeds?
• Confidence: how sure are you, based on data or experience, that the effect will actually happen?
• Ease: how easily and quickly can you execute it, in time and resources?

You give every idea a score per dimension, for example from 1 to 10, and combine those into a single number. Ideas with a high score go to the top of the list. The beauty of ICE is not its mathematical precision, because there is none. The beauty is that it shifts the discussion. Instead of debating whether an idea is good, you debate why you are giving it a high confidence score when you have no data. That conversation makes assumptions visible, and visible assumptions are testable assumptions.

ICE keeps the roadmap honest. An idea with enormous expected impact but low confidence and high effort sinks on its own, while a small, certain, quick win rises to the top. That way you avoid burning a quarter on one gamble while ten fast learnings sit on the shelf.

Step 3: sharpen the hypothesis

An experiment without a hypothesis is just a change with a hopeful look on its face. So for every idea you pick up, you write a sharp hypothesis in a fixed format: if we do X, then we expect Y to change, because Z.

That last clause, the “because”, is the most important one. It forces you to explain which assumption about your audience you are actually testing. Are you testing a new form layout because you think visitors drop off at too many fields? Then you learn not only whether the new layout wins, but also whether your assumption about friction holds up. That is knowledge you can reuse in ten other places.

Every hypothesis comes with one primary success metric, agreed up front. Not three, not five. One. And that metric is tied to something that counts: leads, pipeline or revenue, not the number of clicks or likes. Anyone who picks the metric that happened to turn green afterwards proves nothing. How to stay clear of that trap is covered in our explanation of vanity metrics in marketing.

Step 4: test in sprints

With a prioritised backlog and sharp hypotheses you move to execution, and you do that in sprints: fixed cycles of two weeks, for example, in which a defined set of experiments goes live. The rhythm is deliberate. A fixed cadence makes sure growth does not depend on inspiration or coincidence, but on a calendar that keeps turning.

Within a sprint you build, you launch and you let the experiment run long enough to get a reliable signal. That patience is crucial. Drawing conclusions too early from too little data is one of the most expensive mistakes in growth, because you then build on noise.

A sprint never just ends. It ends with a review in which you hold each experiment’s outcome up against the hypothesis. Did it win? Did it lose? And above all: what does that teach you about your audience?

Step 5: learn and build on it

This is the step that separates an agency that runs campaigns from an agency that builds growth. Every result, win or loss, you feed back into the system. A winning experiment becomes a fixed part of your approach and often raises a follow-up question: can we make this bigger, or does it work elsewhere in the funnel too? A lost experiment is not a failure, but the removal of an assumption that did not hold, plus a new idea for the backlog.

That is how a flywheel emerges. Every sprint adds knowledge, that knowledge raises your confidence scores on new ideas, and better scores lead to better priorities. Over time the process becomes not only faster but also sharper, because you keep getting better at knowing what works in your market. That cumulative effect is exactly what a single one-off campaign can never deliver.

Important to be honest about: not every sprint delivers a breakthrough. Sometimes the harvest of two weeks is mainly knowing what does not work. That feels slow, but it is the opposite of waste. You pay once for a lesson learned and reap the rewards for years afterwards.

How this process compares to standalone services

A fair question from buyers is how this approach differs from hiring an SEO specialist or an advertising agency. The difference is orchestration. A specialist optimises within their channel. The growth process decides across channels: maybe your biggest lever this month is not more traffic, but better conversion of the traffic you already have. The backlog knows no channel borders, only expected impact on growth. That is why the same sprint can contain an SEO experiment, a page test and a new lead gen offer, chosen on the basis of what promises the most growth.

Want to know how this approach compares to a more channel-focused way of working? Read growth marketing versus demand generation. Also dig into the difference between B2B and B2C growth marketing, because the approach and the channels diverge sharply. And if you would rather not build this whole process yourself but have it run by a team that does it daily, that is exactly what you bring in a growth marketing agency for.

What to look for as a buyer

When you assess a proposal, ask about the process, not about the promise. A strong agency can show you how it manages the backlog, how it prioritises and how it feeds learning back in. Watch for these signals:

• Transparent backlog: can you see at any moment which ideas are on the list and why they have that priority?
• Metrics agreed up front: is the success metric for each experiment set beforehand, or only chosen afterwards?
• Honest reporting on losses: does the agency also report the experiments that did not work, along with what was learned from them?
• Link to revenue: does it steer on pipeline and revenue, or on numbers that look good but deliver nothing?

An agency that scores green on all four is not selling you luck but a system. And a system is what makes growth predictable.

Ready to build your own growth engine?

A growth marketing framework is not a secret recipe, but discipline: collect ideas, prioritise honestly, test sharply and keep learning. The gain is not in one experiment, but in the sum of hundreds of them, sprint after sprint. Want to talk through what this process looks like concretely for your growth? Get in touch and we will walk through your biggest growth opportunities together.