Growth Modelling: Build a Model That Reflects Reality

Growth modelling is the process of mapping out how a business expects to grow, which levers drive that growth, what it costs to pull them, and what the realistic return looks like over time. A good growth model does not just project revenue. It forces you to articulate your assumptions, stress-test them against what you actually know, and make decisions based on something more durable than optimism.

Most businesses have a spreadsheet they call a growth model. Fewer have one that genuinely guides decisions. The difference is usually not technical sophistication. It is honesty about what the numbers are actually saying.

Why Most Growth Models Fail Before They Are Used

I have sat in a lot of planning sessions where a growth model was presented with confidence and then quietly abandoned within a quarter. The model looked credible. The assumptions were plausible. But it was built to win approval, not to guide decisions. That is the most common failure mode I see, and it happens at every level of marketing maturity.

The problem is usually not the maths. It is the inputs. A model built on conversion rates from your best-ever month, cost-per-acquisition figures from a channel that is already saturating, or audience size estimates that assume you can reach everyone in your total addressable market is not a model. It is a story told in spreadsheet format.

When I was running iProspect and we were growing from around 20 people to over 100, one of the things that separated the periods of real growth from the periods of just being busy was whether we had a clear picture of what was actually driving revenue. Not what we thought was driving it. Not what the attribution reports said. What was genuinely moving the business forward. Building an honest model of that, even a rough one, was more valuable than any sophisticated forecasting tool.

If you are building or rebuilding your growth strategy from first principles, the broader thinking behind go-to-market and growth strategy is worth reading alongside this. Growth modelling does not exist in isolation. It is the quantitative expression of a set of strategic choices you have already made, or need to make.

What Should a Growth Model Actually Contain?

A growth model should map the relationship between inputs and outputs across your business. That sounds obvious. In practice, most models only capture part of it.

At minimum, a functional growth model needs to cover four things. First, your demand sources: where new customers come from, how they find you, and what triggers their consideration. Second, your conversion architecture: what happens between first contact and first purchase, and where people drop out. Third, your economics: what it costs to acquire a customer, what that customer is worth over time, and how those two numbers relate to each other at different volumes. Fourth, your constraints: the things that will limit growth regardless of how much you spend, whether that is market size, operational capacity, or brand awareness in a given segment.

The fourth one is where most models fall down. It is uncomfortable to model your ceiling. But if you do not, you will keep investing in channels past the point where they can deliver, and you will misread the results when growth slows.

Forrester’s work on intelligent growth modelling makes a similar point: growth strategy needs to account for where demand actually comes from, not just where it converts. That distinction matters more than most planning processes acknowledge.

The Attribution Problem That Distorts Every Growth Model

Earlier in my career, I overvalued lower-funnel performance channels. I thought we were generating demand. In many cases, we were capturing it. The distinction sounds academic until you try to scale a channel that has already reached everyone who was already going to buy, and you watch your cost-per-acquisition climb while your conversion rate falls. Then it feels very practical.

The issue is that performance channels, particularly paid search and retargeting, are very good at being present at the moment of conversion. That makes them look, in most attribution models, like they caused the conversion. Sometimes they did. Often, the customer had already decided and was just looking for a way to complete the transaction. The channel got credit for closing a door that was already open.

Think about how a clothes shop works. Someone who tries something on is far more likely to buy than someone browsing the rail. But the fitting room did not create the desire. The brand, the window display, the recommendation from a friend, the memory of seeing it somewhere, all of that happened before they picked up the garment. If you only measured fitting room usage, you would conclude that the fitting room is your most important growth lever. You would be wrong, and your model would reflect that wrongness at scale.

This is why market penetration thinking matters in a growth model. If you are only optimising for conversion among people who already know you, you are not modelling growth. You are modelling efficiency within a fixed pool. Those are different problems with different solutions.

A growth model that does not account for the cost and mechanics of reaching genuinely new audiences will consistently underestimate what growth actually requires. And it will consistently overestimate what performance channels can deliver once you have exhausted existing intent.

How to Structure Your Assumptions Without Lying to Yourself

Every growth model is built on assumptions. The question is whether you are explicit about them or whether you have buried them inside numbers that look like facts.

The most useful thing I ever did in a planning process was separate what we knew from what we believed from what we hoped. It sounds simple. It is surprisingly hard to do honestly, particularly when you are presenting to a board or a client who wants certainty.

What we knew: our current conversion rates, our average order value, our retention rates from the last 12 months, our cost-per-click in channels we had been running for over a year. These were facts, albeit facts with some variance.

What we believed: that a new channel we were testing would perform similarly to our existing channels once it matured. That our retention rate would improve if we invested in a CRM programme. That expanding into a new segment would not cannibalise existing revenue. These were informed hypotheses, not facts. Treating them as facts in a model is where the trouble starts.

What we hoped: that the market would grow, that a competitor would stumble, that a PR moment would drive organic demand. These should not be in your model at all, or if they are, they should be clearly labelled as upside scenarios, not base case assumptions.

The discipline of separating these three categories is what makes a model useful. It tells you where your risk is concentrated, which assumptions need to be validated before you commit capital, and what you should be measuring to know whether your model is holding.

Scenario Planning: The Part Most Marketers Skip

A growth model with a single set of projections is not a model. It is a forecast. The difference is that a model lets you ask “what happens if this assumption is wrong?” and get a meaningful answer.

I judged the Effie Awards for several years, which gave me an unusual vantage point on how the industry thinks about effectiveness. One thing that struck me consistently was how rarely the submitted cases included any honest account of what did not work, what was assumed and turned out to be wrong, or how the strategy adapted when reality diverged from the plan. The winning cases were compelling. They were also, almost without exception, told backwards. The strategy was presented as if it were inevitable, which made it look smarter than it was and less useful as a learning for anyone trying to replicate it.

Scenario planning is the antidote to that kind of backwards storytelling. You build three versions of your growth model: a base case using your most defensible assumptions, a downside case where two or three of your key assumptions are wrong, and an upside case where conditions are more favourable than expected. Then you look at what decisions would be different in each scenario.

If the answer is “none, we would do the same things regardless,” your scenarios are not meaningfully different. If the answer is “we would make materially different investment decisions in the downside case,” then you have identified the assumptions that matter most and the triggers you should be watching.

BCG’s work on go-to-market strategy and pricing highlights a related point: the assumptions that feel most fixed, things like pricing, channel mix, and customer segment, are often the ones most worth stress-testing, because they have the largest downstream impact on the model when they shift.

The Metrics That Belong in a Growth Model

Not every metric that appears in your marketing dashboard belongs in your growth model. A growth model should contain the variables that have a direct, causal relationship with revenue or profit. Everything else is operational data that helps you manage the inputs, not the model itself.

The metrics I consistently find most useful in a growth model are these. Customer acquisition cost, calculated honestly across all channels including brand spend, not just the channels that get last-click credit. Customer lifetime value, segmented by acquisition channel and customer type, because a customer acquired through word of mouth often behaves very differently from one acquired through paid search. Payback period: how long it takes to recover the cost of acquiring a customer, which tells you how sensitive your model is to changes in retention. Market penetration rate: what share of your total addressable market you have reached, which tells you how much headroom you have before you start hitting diminishing returns in your existing strategy.

The relationship between these four metrics tells you almost everything you need to know about whether your current growth strategy is sustainable. If your CAC is rising, your LTV is flat, your payback period is extending, and your penetration rate is high, you are not in a growth model. You are in a harvest model, and your strategy should reflect that.

Vidyard’s analysis of why go-to-market feels harder than it used to points to something real here: the channels that used to deliver predictable acquisition economics have become more competitive and more expensive, which means the metrics that anchored older growth models need to be revisited more frequently than most teams do.

Connecting Your Growth Model to Channel Investment Decisions

The point of building a growth model is to make better decisions about where to invest. That sounds obvious, but in practice, many organisations build a model and then make channel investment decisions through a completely separate process, usually based on what worked last year or what the channel teams are lobbying for.

A growth model should be the thing that resolves those conversations. If your model says that reaching new audiences is your primary growth constraint, the investment decision should follow from that. If your model says that retention is the biggest lever, the investment should follow from that. The model should be doing work, not just existing.

One of the most useful exercises I have run with leadership teams is to take the existing channel budget allocation and ask: if we built our budget from scratch using only what the growth model tells us, what would it look like? The gap between that answer and the current allocation is usually revealing. It shows you where investment is driven by habit, by internal politics, or by the availability of measurement rather than by the actual growth logic of the business.

This connects directly to the broader challenge of scaling go-to-market operations. Forrester’s thinking on agile scaling makes the point that the processes and investment structures that work at one stage of growth often become the thing that constrains growth at the next stage. A growth model helps you see that transition coming rather than discovering it after the fact.

There is also a useful distinction to draw between growth hacking tactics and structural growth modelling. Growth hacking tends to focus on short-term conversion optimisation and channel experimentation. That has its place. But it is not a substitute for understanding the structural economics of how your business grows. A model gives you the context to know which experiments are worth running and which are solving for the wrong problem.

When to Rebuild Your Growth Model Rather Than Update It

A growth model is not a permanent document. It is a representation of how you believe your business works at a given point in time. When that belief becomes sufficiently disconnected from reality, updating the model is not enough. You need to rebuild it from different assumptions.

There are a few signals that tell you a rebuild is needed rather than a refresh. Your actual results have diverged from your model projections for three or more consecutive quarters and you cannot explain why. A significant structural change has occurred in your market, a new competitor, a channel that has fundamentally changed its algorithm or pricing, a shift in customer behaviour. Your business has entered a new stage of growth where the economics are materially different from the previous stage. Or you have acquired new data, through a proper incrementality test, a media mix model, or a customer research programme, that challenges a core assumption in your existing model.

I have seen businesses hold onto an outdated growth model for years because rebuilding it would require acknowledging that the strategy it justified was not working. That is an expensive form of denial. A model that does not reflect your current reality is not giving you information. It is giving you permission to keep doing what you are doing, which is a different thing entirely.

BCG’s work on evolving go-to-market strategy in changing markets captures something important here: the businesses that maintain growth through market transitions are usually the ones that update their strategic assumptions before the market forces them to, not after.

Making the Model Useful for People Who Did Not Build It

A growth model that only makes sense to the person who built it is not a strategic asset. It is a personal document. For a model to actually influence decisions across a business, it needs to be legible to the people who will use it, which usually means finance, commercial leadership, and the channel teams whose budgets it is meant to inform.

The best growth models I have worked with share a few characteristics. They are transparent about their assumptions, which means the assumptions are visible and labelled, not buried inside formulas. They are connected to operational metrics that people are already tracking, so there is a clear line between what the model predicts and what the business is measuring. And they have a clear owner: someone who is responsible for keeping the model current and for flagging when reality is diverging from the projections.

That last point matters more than most organisations acknowledge. A growth model without an owner becomes stale very quickly. The market moves, the business changes, the channel economics shift, and the model just sits there reflecting a reality that no longer exists. At that point, it stops being useful and starts being misleading, which is worse than having no model at all.

Early in my career, I was handed a whiteboard pen at a Guinness brainstorm when the founder had to leave for another meeting. My internal reaction was something close to panic. But the discipline of having to articulate a point of view under pressure, to make a case that held together in front of people who knew the business better than I did, taught me something I have carried ever since. The value of being explicit. Vague thinking survives in private. It does not survive when you have to write it on a whiteboard and defend it. A growth model is the same. The act of making your assumptions explicit and visible is what makes them useful, and what makes them honest.

Growth modelling sits at the centre of how good marketing strategy gets made. If you want to think through the broader strategic context, the go-to-market and growth strategy hub covers the decisions that a growth model should be informing, from market selection to channel architecture to how you measure what is actually working.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions