Growth Modeling: Build the Number Before You Spend the Budget

Growth modeling is the process of mapping out how revenue, customers, or market share will change over time based on specific inputs: acquisition rates, conversion rates, retention, average order value, and the cost of each. Done well, it turns a growth target from a number someone wrote on a whiteboard into a set of levers you can actually pull.

Most businesses skip it. They set a revenue target, allocate a budget, and hope the channels deliver. When they miss, they blame the channels. The problem usually started much earlier, in the planning room, before a single pound or dollar was spent.

Growth modeling sits at the centre of go-to-market planning. If you are building or pressure-testing a growth strategy, the wider context around how modeling connects to channel selection, market prioritisation, and commercial planning is worth working through systematically. The Go-To-Market and Growth Strategy hub covers that ground in full.

Why Most Growth Targets Are Fiction

I have sat in a lot of planning sessions where someone has written a number on a slide and called it a growth target. Thirty percent revenue growth. Double the customer base. Top-five market position by end of year. The number is usually chosen for one of three reasons: it sounds ambitious enough to impress stakeholders, it matches what a competitor claimed to be doing, or it is what last year’s number plus an optimistic percentage looks like.

None of those are growth models. They are growth wishes.

A real growth model starts from the opposite end. Instead of announcing a destination and assuming the route will present itself, you build the route first and see whether the destination is reachable. You ask: given our current conversion rates, our expected traffic volumes, our average order value, and our churn rate, what growth is actually achievable? And then, if that number is not enough, you ask what would need to change and by how much.

This is not pessimism. It is the only honest way to plan. Forrester has written about intelligent growth modeling as a discipline distinct from revenue forecasting, and the distinction matters: a forecast predicts what will happen, while a model maps what would need to be true for a given outcome to occur. Those are very different exercises.

What Goes Into a Growth Model

A growth model is not a single spreadsheet. It is a structured set of assumptions about the inputs that drive revenue, connected in a way that lets you see what happens when any one of them changes.

The core inputs depend on your business model, but for most commercial organisations they include some version of the following:

• Addressable audience size. How many people or businesses could realistically buy from you in the period you are modeling? Not the total addressable market as typically inflated in pitch decks, but the serviceable, reachable segment you can actually get in front of.
• Awareness and reach. Of that addressable audience, what proportion will encounter your brand through paid, organic, earned, or owned channels? This is where most models are weakest, because it requires honest thinking about channel capacity and diminishing returns.
• Conversion rates by stage. What percentage of reached prospects become leads? What percentage of leads become customers? These rates are rarely flat across volume ranges. They tend to fall as you push into less-engaged audiences.
• Average order value or revenue per customer. What does a customer actually spend? Broken down by segment if the variance is significant.
• Retention and repeat purchase rate. For subscription or repeat-purchase businesses, this is often the most important input in the model. A small improvement in retention compounds faster than almost any acquisition lever.
• Cost per acquisition by channel. What does it cost to generate a customer through each channel, fully loaded? This includes media spend, agency fees, content costs, and a realistic attribution of overheads.

Connect these inputs and you have a model that tells you whether your target is achievable, which levers have the most impact, and where the biggest risks sit. Change one assumption and the whole picture shifts. That is exactly the point.

The Assumption That Breaks Most Models

Earlier in my career I overvalued lower-funnel performance. I watched cost-per-acquisition numbers and conversion rates with the kind of attention most people reserve for their bank balance. The channels looked efficient. The numbers looked good. Growth looked like it was working.

It took me longer than I would like to admit to recognise that a significant portion of what performance marketing was claiming credit for was going to happen anyway. People who were already looking for the product, already close to a decision, already familiar with the brand. The channel was capturing existing intent, not creating new demand. It was efficient, but it was not growth in any meaningful sense.

Think about a clothes shop. Someone who walks in and tries something on is far more likely to buy than someone who walks past the window. Performance marketing tends to optimise for the people already walking through the door. Growth requires getting more people to walk past the window and feel something. That is a different problem, and it requires different inputs in your model.

The assumption that breaks most growth models is the assumption that existing channels will scale linearly. They will not. As you push paid search beyond the high-intent queries, CPAs rise. As you exhaust your retargeting pool, frequency goes up and returns go down. As you scale social, you move from engaged audiences to cold ones. Every channel has a capacity curve, and most growth models ignore it entirely.

SEMrush has a useful breakdown of market penetration strategy that illustrates this well: the further you push into a market, the harder and more expensive each additional percentage point becomes. A growth model that does not account for this will always overpromise.

How to Build a Growth Model That Is Actually Useful

The goal is not to build a model that is right. No model is right. The goal is to build a model that is honest about its assumptions, structured enough to test them, and specific enough to be actionable.

Here is how I approach it.

Start with what you know, not what you hope

Anchor every input in historical data or comparable benchmarks. If your current conversion rate from lead to customer is 12%, do not model 20% unless you have a specific, funded plan to get there. If your average order value has been flat for three years, do not assume it will increase without a pricing or packaging change to justify it.

This sounds obvious. It is routinely ignored. I have reviewed growth models where every single input was set at the optimistic end of the plausible range, which meant the model was not a model at all. It was a best-case scenario presented as a plan.

Build three scenarios, not one

A single-scenario model is a confidence trick. It tells you what happens if everything goes to plan, which is not a useful planning tool because things rarely go entirely to plan.

Build a base case using current performance rates and realistic channel scaling. Build a downside case where one or two key inputs underperform by a meaningful margin. Build an upside case where specific improvements, with specific owners and timelines, deliver better-than-base results.

The gap between your downside and upside tells you something important: it tells you how much variance you are carrying in the plan, and whether your business can absorb the downside without serious damage.

Identify the two or three inputs that move the model most

In every growth model there are inputs that matter enormously and inputs that are almost irrelevant to the outcome. Sensitivity analysis, running the model with each input varied while holding the others constant, tells you which is which.

When I was running an agency and we were working through growth planning for a retail client, we discovered that a 5% improvement in their retention rate had more impact on three-year revenue than doubling their acquisition budget. That single insight changed where we recommended they invest. Without the model, we would have defaulted to the obvious answer: spend more on acquisition.

BCG’s work on go-to-market strategy and commercial planning makes a similar point: the inputs that drive the most value are not always the ones that get the most attention in planning conversations.

Connect the model to channel investment decisions

A growth model that sits in a spreadsheet and never connects to budget allocation is a planning exercise, not a planning tool. The model should directly inform how you split budget across channels, what targets you set for each, and what early indicators you will monitor to know whether the model is tracking.

If the model says you need to reach 400,000 new people in a quarter to hit your acquisition target, and your current organic reach is 80,000, that gap needs to be filled somehow. The model forces the conversation about where that reach comes from, at what cost, and whether the economics still work when you factor it in.

The Difference Between a Growth Model and a Revenue Forecast

These two things get conflated constantly, and conflating them causes real problems.

A revenue forecast is a prediction: based on current trajectory, here is what we expect to happen. It is useful for financial planning, but it is largely passive. It tells you where you are going if nothing changes.

A growth model is active. It says: here is where we want to go, here are the specific inputs that would need to change to get there, and here is what we will do differently to change them. It is a decision-making tool, not a reporting tool.

The confusion matters because organisations often produce revenue forecasts and call them growth plans. They are not. A forecast that extrapolates from current performance does not tell you how to grow. It tells you what happens if you carry on as you are.

For businesses thinking about how growth modeling connects to pipeline and revenue planning, Vidyard’s research on untapped pipeline potential for GTM teams is worth reading as context for how much revenue opportunity typically sits unaddressed in existing go-to-market structures.

Where Growth Models Go Wrong in Practice

I have seen growth models fail in predictable ways. Not because the mathematics were wrong, but because the assumptions feeding them were either dishonest or untested.

Optimistic conversion rates. Teams benchmark against their best month rather than their average. The model looks achievable. The reality underperforms consistently and nobody can explain why.

Ignored channel saturation. The model assumes you can spend three times as much on paid search and get three times as many customers. You cannot. At some point you run out of high-intent queries and CPAs climb sharply. The model needs to reflect this.

Missing the new audience problem. Growth that relies entirely on converting existing demand is not really growth. It is market capture. Sustainable growth requires reaching people who do not yet know they need you. That is harder to model because you do not have conversion data for audiences you have never reached, but ignoring it produces a model that is structurally incapable of delivering the growth it promises.

No ownership of the inputs. A model is only as good as the people responsible for hitting its component parts. If the model assumes a 15% improvement in email conversion but nobody owns that improvement with a specific plan and a deadline, it is a number in a spreadsheet, not a commitment.

SEMrush’s overview of growth strategy examples illustrates how the most effective growth initiatives tend to be specific and testable rather than broad and aspirational. That specificity is what separates a model from a wish.

Growth Modeling in a New Market or Category

Modeling gets harder when you are entering a market where you have no historical data. New geography, new product category, new customer segment. The temptation is to use analogies from existing markets and assume similar performance. Sometimes that works. Often it does not.

BCG’s analysis of go-to-market launch planning in high-stakes environments makes a point that applies well beyond biopharma: in new market entry, the sequencing of your go-to-market moves matters as much as the inputs themselves. A model that gets the inputs right but sequences the activity wrong will still underperform.

In new market situations, I tend to run a tighter, shorter-horizon model with explicit review points built in. Rather than modeling 12 months with high confidence, model 90 days with honest uncertainty ranges, execute against it, and then recalibrate with real data. The model becomes more accurate over time as you replace assumptions with actuals.

The first time I had to build a growth model for a market I had no data on, it was uncomfortable. I was handed a whiteboard and expected to produce something credible for a client who had far more domain knowledge than I did. What I learned then, and what has held true across every market since, is that a model built on transparent assumptions with honest uncertainty ranges earns more trust than a confident model built on shaky ones. Clients and stakeholders can engage with uncertainty. What they cannot work with is a plan that falls apart the moment it meets reality.

Making the Model Work Over Time

A growth model is not a document you produce in January and revisit in December to explain why things went differently. It is a live tool that should be updated as actuals come in and assumptions are tested against reality.

Set a cadence for reviewing the model against performance data. Monthly is usually right for most businesses. Identify which inputs are tracking ahead of assumption and which are behind. When an input underperforms, the question is not just what happened but what that means for the rest of the model and whether the overall target is still achievable through other levers.

This kind of ongoing model management requires a degree of commercial discipline that not all marketing teams are set up for. Forrester’s work on agile scaling in go-to-market functions is relevant here: the teams that manage growth most effectively tend to combine structured planning with the flexibility to reforecast and reallocate as conditions change.

The model should also evolve as the business evolves. A growth model built for a business at £5m revenue will not serve a business at £20m. The inputs change, the channel mix changes, the competitive environment changes. Treating the model as a fixed document rather than a living one is how organisations end up making decisions based on assumptions that stopped being true two years ago.

Growth modeling is one piece of a broader commercial planning process. If you want to see how it connects to market selection, channel strategy, and go-to-market sequencing, the Go-To-Market and Growth Strategy hub pulls those threads together in one place.

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