Most marketing budgets are approved on a spreadsheet that shows, line by line, what will be spent and says almost nothing about what will be proven. The agency retainer rolls forward. Paid search goes up by whatever the platform says it can absorb. The trade show stays in because it was in last year. Then, around month nine, the finance director asks what the money did, and nobody can answer in a way both sides accept.
For businesses with a December year-end, that conversation starts again in the next few weeks. It is worth having it differently this time.
The real argument is about proof
The usual story is that marketing and finance want different things. Marketing wants long-term brand. Finance wants this quarter's return. Recent research suggests that story is mostly wrong.
Bain & Company and Google surveyed almost 1,400 senior marketing and finance executives this year. Both groups named revenue impact and return on marketing investment as the measures that matter most. Their payback expectations were close too: about 70% of each group expect performance spend to pay back within months or quarters, and about 40% of each expect brand investment to take a year or more. Where they differ is on the evidence. Only 41% of marketers said they had the data, tools and measurement to tie their work to business outcomes.
McKinsey found the same gap from a different angle. In its State of Marketing Europe 2026 survey, which included UK marketing leaders, 72% of CMOs planned to raise budgets relative to sales. Only 3% could show a return on more than half of their spend.
Budgets are going up and proof isn't keeping pace. In the UK, the IPA's Q2 2026 Bellwether Report showed companies revising marketing budgets upwards for a second quarter, while their view of their own financial prospects turned negative. When confidence drops, spend that cannot show its working is the first to be cut. It rarely gets cut in the right order.
Two ways to get it wrong
Most leadership teams have fallen into one of these at some point.
The first is demanding a return on everything. If every pound has to show attributed revenue within the quarter, money drifts towards the channels that report well: branded search, retargeting, offers to existing customers. Those channels are good at catching demand that already exists, and some of the sales they claim would have come anyway. The cost shows up twelve to eighteen months later as a thinner pipeline, and by then nobody connects it to the budget meeting that caused it.
The second is waving brand spend through because it "can't be measured". Brand effects are slower and harder to isolate. They are not invisible. A budget line with no success measure, no review date and no owner is a habit, whatever it is called.
The fix for both is the same. Before the money is committed, agree what kind of proof each line can reasonably produce, and by when.
Step one: sort the budget into four pots
Put every line of next year's marketing budget into one of four pots. They exist because they need different evidence and different patience.
Demand capture
Spend that converts people who are already looking for what you sell: paid search, marketplace listings, retargeting, sales promotions and most email to existing contacts. Expect payback within weeks or a quarter. Judge it on incremental revenue, meaning the sales that would not have happened without it, rather than the revenue the platform attributes to itself.
Demand creation
Spend that makes more people consider you in future: brand advertising, events, PR, sponsorship, thought leadership and content aimed at people who aren't buying yet. Expect twelve months or more. Judge it first on leading indicators agreed in advance, then on commercial results over time.
The operating base
The fixed cost of having marketing at all: salaries, agency retainers, software, the website and data. Judge it on what it produces for the money and whether it is fully used. Retainers and software licences are where unused capacity tends to hide.
Tests
New channels, new markets, new messages. Cap this pot. We would suggest somewhere between 5% and 10% of variable spend for most SMEs, though the right figure depends on how settled your growth model is. Every test gets a question it is answering, a budget, an end date and a rule for stopping or scaling it.
Expect two things to surface the first time you do this. The operating base is often larger than anyone assumed. And some of what has been called "brand" turns out to be demand capture with a logo on it, or the other way round. Both are worth knowing before anyone argues about the total.
Step two: put each line through six questions
For every line above an agreed threshold, the person who owns it should answer six questions in writing before the budget is signed off. For a business spending £1m a year on marketing, a threshold of around £25,000 keeps the exercise to a manageable number of lines.
- Which pot is it in, and does finance agree? Disagreement here is useful. It usually means the line is doing two jobs, or neither.
- What will count as success? The measure and the target, written down now and fixed before launch. Bain found that leading companies lock their metrics before campaigns begin. That one habit prevents most of the arguments that come later.
- When should it pay back? Weeks, a quarter, a year or longer. Agree it between marketing and finance, then leave it alone unless the evidence changes.
- What evidence will we accept? Name the level on the proof ladder below and how you will reach it.
- What would we see if it wasn't working? If nobody can describe the failure signal, nobody will spot it in time.
- What would happen if we cut it by a fifth? This is the question most budgets skip. Lines are usually defended on their average return, but the decision is about the next pound. The last 20% of spend in a channel often earns far less than the first 80%.
A line that cannot get through all six is not necessarily wrong. It just isn't ready to be funded at full value. Fund it as a test, with a smaller budget and a date, until it can.
Step three: agree the evidence standard
Not all proof is equal, and much of the friction between marketing and finance comes from treating it as if it were. Naming the level of evidence each line should reach settles most disputes before they start.
A workable standard for most businesses:
- Demand capture lines should reach level 3 within the year. If you have never paused branded search to see how much of that traffic arrives anyway, start there.
- Demand creation lines should have level 1 leading indicators agreed before launch, plus a level 3 test wherever one is practical, such as running a campaign in one region and not another.
- Operating base lines should show what they produced: campaigns delivered, content published, leads handled, hours used against hours paid for.
- Nothing material should stay at level 0.
Leading indicators for demand creation
Brand spend is hardest to defend in B2B and professional services, where sales cycles are long and the buyer rarely clicks an advert. These indicators tend to move before revenue does, and most can be tracked without new software.
- Share of search, meaning searches for your brand name as a share of searches for you and your named competitors combined.
- Direct and branded website traffic, tracked as a trend.
- The share of new enquiries that arrive already knowing who you are. Ask every new prospect how they heard of you and record the answer.
- Your win rate when a buyer is comparing you with competitors.
- Price realisation: the average discount given and how often you are asked for one. Our 10-deal margin review shows how to measure this from a small sample.
- The time a qualified opportunity takes to close.
Pick the two or three that fit your business, set a baseline this quarter and agree what movement would count as progress. Then leave them alone for long enough to mean something.
Step four: change the meeting
In many businesses the marketing review is a slide deck of activity, presented to people who have already decided what they think. Replace it with a quarterly session of 90 minutes, with the CEO, the finance lead and the marketing lead in the room and one shared budget ledger on the table. Keep the agenda the same every quarter.
- Pot totals against plan (10 minutes). Where the money has gone so far, by pot.
- Lines that have hit a trigger (30 minutes). Decisions only: stop, scale or hold.
- Test results, including the failures (20 minutes). Bain found that marketers with strong finance relationships share failures alongside successes. Putting failures on the agenda makes that normal.
- Leading indicators (15 minutes). Trend only. No reopening of agreed payback windows unless the evidence has changed.
- Where the next pound goes (15 minutes). Any money released by stopped lines is reallocated here, deliberately, rather than drifting back into last year's channels.
The last item matters most. Budgets set once a year tend to be defended once a year. A standing reallocation slot, run on evidence, is how spend moves towards what works without waiting for a crisis. The Reallocation Ledger covers the same principle for larger strategic moves.
If you don't have a CMO
Many SMEs and growth companies don't. Marketing is a manager and one or two agencies, and finance is a finance director who is also looking after HR and IT. The test still works in a lighter form: run the four pots and the six questions once a year at budget time, then hold the quarterly review.
Two adjustments help. Ask your agencies to answer the six questions for every line they run on your behalf; a good agency will welcome it, and one that resists is telling you something. And make the CEO the referee. The CEO's job is to hold both sides to what was written down in the autumn, and to stop the payback window being quietly moved when results are inconvenient.
Signs your budget process needs this
- The largest line is justified mainly by the fact that it was there last year.
- Marketing reports platform-attributed revenue and finance quietly discounts it.
- Nobody can say what was cut the last time budgets tightened, or why those lines were chosen.
- Brand spend has no named indicators and no review date.
- Tests have no end date.
- Agency retainers renew without anyone reviewing what they produced.
Two or more of these usually means the budget is being set on habit and cut on instinct.
Where this leaves next year's budget
Running the test may well lead to a larger marketing budget. Once the capture spend that claims credit for sales that would have happened anyway has been trimmed, the money can go into demand creation with an agreed way of judging it, and finance has a reason to trust the case. Either way, next autumn's conversation starts from evidence instead of from last year's number.
Allington Advisors works with founders, CEOs and leadership teams on marketing strategy and growth planning, and we sit comfortably on both sides of the marketing and finance table. If you would like an independent view of next year's plan before it is signed off, we run a short marketing budget review with the CEO, finance lead and marketing lead together.