The Growth Model

Budget in, growth curve out — and what that curve does to the value of the business. It plans the year, then becomes the monthly report that proves the return.

The play

Marketing is the largest line on the P&L nobody can defend in the language finance uses.

Every year the same meeting happens. Marketing asks for a budget and presents activity. Finance asks what it returns and gets a number about clicks. Both leave irritated, the budget gets held flat, and the argument repeats twelve months later with worse data.

This turns that meeting into an arithmetic conversation instead of a political one.

Why the pressure is worse this year than last

Finance is now watching directly

Top 3

Where CFOs now rank marketing-spend visibility among financial planning priorities, in Gartner's February 2026 CFO survey. Two years ago it barely made the top ten. Board audit committees are asking for marketing ROI alongside the financial statements.

Gartner CFO Survey, Feb 2026

And the money has stopped growing

3.3%

Marketing budget growth over the last twelve months, down from 5.8%. Budgets sit at roughly 7.8% of company revenue — effectively flat year on year, against rising expectations. Flat budgets are where measurement stops being optional.

The CMO Survey · Gartner CMO Spend 2026

The cost of not having an answer

4.2 yrs

Average CMO tenure — the shortest of any C-suite role — attributed largely to marketing leaders being unable to defend their programmes in the financial language their organisation uses. The person who buys this play has skin in it personally.

2026 CMO tenure analysis

Why it fails, mechanically

61%

Cite data integration as the biggest barrier to proving marketing ROI; 53.7% cite accurate data collection. Not a analytical failure — a plumbing failure. The numbers exist in four systems that do not reconcile.

Marketing intelligence research, 2026

The gap is not that marketing directors are bad at maths. It is that nobody has built the one model that connects spend to pipeline to revenue to enterprise value, agreed it with finance in advance, and then reported against it every month without renegotiating the definitions.

What it actually is

A single model, built on the client's own numbers, that answers three questions in one place.

What does the current spend return? Channel by channel, with the assumptions visible and the confidence stated honestly — including where the data cannot support a claim.
What happens if we change it? Move the budget up, down or sideways between channels and see the growth curve move. Scenarios, not a single forecast, because a single forecast is always wrong and always argued with.
What does that growth do to the value of the business? The question the owner is actually asking, and the one almost no marketing report answers.

The valuation link, done properly

For an owner heading toward an exit, growth is not an operational metric — it is a multiple. In the mid-market the growth premium runs at roughly half a turn of additional multiple for every five percentage points of EBITDA growth above the industry median; a firm growing at 20% where the median is 5% can command 1.5–2.0× above the median multiple. Professional-services businesses currently trade around 4–7× EBITDA, so that premium is not a rounding error — it is often the difference between two exit outcomes.

Two disciplines keep this honest rather than promotional. First, buyers in 2026 reward growth and margin together — growth bought by burning cash is discounted, so the model shows the margin consequence alongside the growth curve. Second, the range is stated, not a point estimate, and the sensitivity is shown. A marketing supplier presenting a single confident valuation number to a business owner deserves the scepticism it gets.

The growth premium

~0.5×

Additional EBITDA multiple per five percentage points of growth above the industry median, in the mid-market. Sustained outperformance compounds into 1.5–2.0× over the median.

Lower-mid-market valuation benchmarks, 2026

The base it applies to

4–7×

Typical professional-services EBITDA multiple range; 5–7× manufacturing, 5–9× healthcare services. Within each band the multiple moves on growth rate, recurring revenue, customer concentration and management depth — three of which marketing touches directly.

Industry multiple benchmarks, 2026

How it is delivered

1 · Reconcile

Get to one set of numbers

Spend, pipeline, conversion and revenue pulled from the systems they actually live in and reconciled to what finance already reports. This is the 61% problem, and it is where most of the effort goes.

If marketing's revenue number and finance's revenue number do not match, nothing built on top of them will survive the first board meeting.

2 · Agree the definitions

With finance, in writing, first

What counts as a lead, a qualified opportunity, an attributed win, a marketing-influenced deal, and over what window. Signed off by the FD before any modelling happens.

Almost every failed marketing ROI exercise fails here, months later, in an argument about definitions nobody had at the start.

3 · Model

Budget in, curve out

Channel response, lags, saturation, and the cost of acquisition at each level of spend. Scenarios at three budget levels, with the assumptions exposed rather than buried, so the board can argue with the inputs instead of the conclusion.

Then the valuation layer on top: growth curve → EBITDA path → multiple range.

4 · Report

The same model, monthly

The planning tool becomes the reporting tool. Predicted versus actual, every month, in the same shape — which is what makes it credible, because it is falsifiable and it does not get quietly reshaped when the numbers disappoint.

This is where the AI earns its place: the assembly and reconciliation run automatically, so the monthly report costs hours rather than a fortnight of an analyst.

5 · Recalibrate

The model learns

Quarterly, the actuals are fed back and the response curves are refitted. Year two is materially better than year one, which is the argument for the retained relationship rather than a one-off deck.

A model nobody recalibrates is a spreadsheet with a date on it.

Where it sits

Inside the planning engagement

It is the quantitative half of the Strategy & Planning Model — the part that turns a set of recommendations into budget scenarios a board can choose between.

Then it survives the engagement as the standing report, which is how the planning work earns the retained execution.

What it does not claim

It is not attribution. Nobody can cleanly attribute a B2B deal with a nine-month cycle and eleven touchpoints to one channel, and anyone selling that is selling a story. This models contribution and response at portfolio level, which is both defensible and enough to make budget decisions with.
It is not causal proof. Observational data with confounders. Where a claim needs proving, the model says what test would prove it — a holdout, a geo split, a spend step-change — rather than asserting it.
It is only as good as the plumbing. If the CRM stages are fiction, the model inherits the fiction. The reconcile step surfaces that early and honestly, which occasionally means the first finding is "your pipeline data is not usable yet".
The valuation number is a range. Multiples move with the market, the buyer and the quality of the business. We show the sensitivity and say so out loud.

Who buys it

Who: the owner or MD heading toward an exit, or the marketing director whose budget conversation is about to get harder. Occasionally the FD, and when it is the FD it closes fastest.
Status: a v1 exists and has been run. Needs templating and a standard data-intake before it is sold as a repeatable product.
Sells with CompanyOS: the reconciliation work in step 1 is the first slice of the spine. If a client buys both, they pay for that once.
How it is judged: whether the FD signs the definitions, whether the monthly predicted-versus-actual holds within its stated range, and whether next year's budget conversation takes one meeting instead of three.