Direct answer
There is no correct percentage. Gartner's 2026 CMO Spend Survey says 7.8% of revenue; Duke's CMO Survey says 9.0%. Both describe what large companies spend, not what yours needs. The budget is the output of four decisions — who you're for, why they'd pick you, what you sell, what you say — priced against how many buyers you must reach. Decide those first; the number falls out. Copy a percentage and you have a number without an argument.
Every autumn someone in the leadership team asks what the right percentage is. The question sounds prudent. It is asked in place of a harder one — what is the money for — and it gets asked because the harder one has no dashboard.
Most businesses set next year's marketing budget by looking at this year's and adjusting it. Some look up a benchmark. Almost none can tell you, line by line, which buyers the money is supposed to reach and why those buyers would choose them once reached. The budget exists. The argument underneath it doesn't.
What the benchmarks actually say
Gartner's 2026 CMO Spend Survey, run January to March 2026 among 401 marketing leaders — the vast majority at companies with more than $1 billion in revenue — found marketing budgets at 7.8% of company revenue, up from 7.7% in 2025. Fifty-six percent of those CMOs said they lack the budget to deliver their own 2026 strategy. Duke University's CMO Survey, 35th edition, March 2026, 308 marketing leaders at US companies, put marketing at 9.0% of revenues and 9.6% of overall budgets — the lowest share in several years. Same year, same job title, two surveys, and the answers are 15% apart.
Those numbers are not wrong. They are irrelevant. They describe the habits of companies with a thousand employees and a marketing department, averaged across industries that have nothing in common. A firm with 120 people, thirty-five new customers a year and an average contract of €180,000 does not share a single input with that average, except the word marketing.
A percentage is what you use when nobody has decided what the money is for.
Why a percentage is a reflex, not a decision
The appeal of a percentage is that it removes the need to answer anything. It scales with revenue instead of with the job. Revenue goes up, marketing gets more; revenue dips, marketing gets cut — Duke's survey found that when profits miss, executives cut marketing 45.4% of the time, more often than any other line. That is not a budgeting method. It is a thermostat.
And the cost of a copied percentage is not paid once. It is paid in installments: a year of spend aimed at "the market" rather than at the buyers who actually decide, a sales team that inherits leads nobody qualified, a board that asks what the money produced and gets a report about reach. The number looked responsible in October. Nobody can defend it in June.
Build the number from the four decisions
The budget that can be defended is built the other way round. It starts from four decisions that sit upstream of every euro: who this company is for, why that buyer would pick you over the alternative, what exactly you are selling them, and what you say to them. Call it the upstream audit. Until those four are answered in writing, there is nothing to price.
Once they are, the arithmetic is unglamorous. How many new customers does next year's plan need? What share of qualified conversations do you win? So how many conversations do you need, and how large is the pool of companies you have to be present in front of to get them? The budget is the cost of being present in front of that pool, often enough, with the message you decided on. It might come out at 3% of revenue. It might come out at 11%. Either is defensible, because either has an argument attached.
| Percentage method | Decision method | |
|---|---|---|
| Starting point | Last year's number, or a survey average | How many of which buyers you have to reach |
| What it can answer | "Are we in line with others?" | "What is this money for, and what should it return?" |
| What it can't | Whether the spend reaches anyone who matters | Nothing it wasn't asked — every line maps to a decision |
| In a slow quarter | Cut first, because nothing defends it | Protected, because cutting it has a named cost |
| Who can defend it to the board | Nobody, beyond "it's the benchmark" | Anyone who can read the four decisions |
From the book: an executive programme with thirty seats at €30,000 each had never sold out, year after year, on a budget that was perfectly respectable by any benchmark. The budget was not the variable. The decision about who the programme was really for, and what it actually sold, was. Once that changed, the first season together filled all thirty seats with a waiting list, on spend that did not materially move. The price today is above €40,000. The percentage was fine the whole time. The argument underneath it was missing.
What to do with next year's number
Start from the buyer count, not the revenue line. Write the four decisions down before the spreadsheet is opened — if the leadership team cannot agree on who the company is for, that disagreement is the budget problem, and no percentage resolves it. Then price the reach. Then strike every line that does not map to one of the four decisions; that is where the money quietly leaks, and it is the same leak I describe in why spending more on marketing won't fix a slow month. And once the number is set, protect it for the year. A budget built from decisions has a cost of cutting that can be stated in customers not reached. A percentage has no such defence, which is why it is the first thing to go.
If the budget is going to outside help, the same logic sets the fee, which is the argument in what a fair marketing retainer actually costs: price the job, not the revenue. The longer version of all of this — why the upstream decisions come first and what happens when they don't — is the whole of Reports, Not Revenue.
- Ask three people in the leadership team, separately, who the company is for. If the answers differ, stop budgeting and decide that first.
- Write down how many new customers next year needs, your win rate on qualified conversations, and the size of the pool you must be present in front of.
- Price the reach from those three numbers. Compare the result to 7.8% of revenue only out of curiosity.
- Strike any line that cannot be traced to one of the four decisions.
- Record the cost of cutting the budget in customers not reached, so the number survives the first slow quarter.
If you can't trace next year's marketing number back to a decision about who it's for and why they'd choose you, that is what a Commercial Immersion Diagnostic settles first: €2,900 fixed, no retainer, and if it doesn't surface something concrete and actionable, you don't pay for it.
A budget copied from a benchmark costs the same as one built from a decision. Only one of them can be defended in June.
Book an introduction call See what's included