How to defend a budget without making a promise you can't keep
For the marketing lead who sits across from a CFO every quarter. The structure of the budget ask, the three questions a CFO always asks, and how to answer them without promising a return.
You know the moment. The CFO looks up and asks what this amount is going to deliver. Your reflex is to name a number, because a number ends the meeting and "we'll see" doesn't. That number saves you today and sends you the bill next quarter.
The problem is not that you lack an answer. It's that you give the wrong answer to the wrong question. What the CFO is really asking is: why put this money here instead of somewhere else. You answer with a prediction you cannot keep, because a prediction reassures faster than a line of reasoning does. That works, once.
What's inside
- The structure of the budget request, written out and ready to copy, with the parts in the right order
- What you can state instead of promising a return
- The three questions a CFO will ask, and how to answer them without promising anything
Where should we send the guide?
Why a promise weakens your position
Every promised return creates a contract you never signed but still have to honor. The moment you say a number, a ROAS, a revenue figure, a payback period, that number becomes the only thing that matters in the next conversation. Not the assumptions underneath it, not what changed along the way.
Miss the number, and next quarter you're not making a budget ask, you're answering for a failure: explaining what went wrong instead of discussing what you learned. Hit the number, and you're expected to hit it again, in a field where the variables outside your control shift every quarter. You win today's meeting and lose the next four.
A board thinks in certainties, and an ad budget doesn't provide that certainty, never did and never will. Your answer has to bridge that gap, not hide it.
What you can say instead
Three things replace the promise, and they stay true no matter how the quarter turns out.
What you're going to learn: not "this grows revenue" but "we're finding out whether this audience responds to this message on this channel, at what cost per interaction." That's a promise you can keep, because it's about a process you control, not an outcome the market decides.
What decision the budget unlocks: a budget is the price of an answer, not a goal in itself. "With this amount we can decide in eight weeks whether to expand this channel, adjust it, or shut it down." That shifts the question from "what will this return" to "what decision does this buy us".
What uncertainty you name: whether the audience is large enough, whether the message lands, whether the channel still works at scale the way it does at small volume. Unnamed uncertainty comes back later as an excuse. Named uncertainty is an agreement.
Asking to learn versus asking to scale: never in one ask
"Give me budget to learn" is a request for an experiment with an open outcome: it might return something, it might return nothing, and either outcome is a success, because you get the answer you paid for. "Give me budget to scale" is a request to do more of something that has already shown a pattern. Different question, different evidence, different yardstick.
Ask for budget to learn while hinting at scaling later, and you're implicitly promising a result while formally asking for an experiment, and the CFO will hold you to that promise later. Ask for budget to scale while admitting you're still learning, and you undercut your own ask. Separate the two on your agenda where you can, and always in your language. One conversation, one ask.
Making uncertainty explicit without sounding weak
A range instead of a point estimate. Not "this costs 20,000 euros" but "between 15,000 and 25,000, depending on how fast we find the signal we're looking for." That sounds controlled, not uncertain, because it shows you've mapped the limits of your own not-knowing instead of hiding them.
Assumptions on the table, named: "we're assuming this audience exists in sufficient numbers on this channel, we'll test that in the first two weeks." A named assumption is one you can disprove, and that's exactly what a pilot is for.
And a stop rule: what signal, or the absence of it, means you stop, adjust, or push on. A board that knows you operate your own brake trusts you sooner with the accelerator.
When you come back, and with which number
Agree right away when you'll come back and with what number. Not "we'll keep you posted" but "in eight weeks we'll come back with the number of qualifying interactions at what cost, and with a conclusion on whether we continue." You're not saying upfront what the number should be, only that you'll report it and what you'll do with it. That gives the CFO a fixed date instead of a standing question, and forces you to decide now which number drives the decision.
The ask for an increase
Never ask for an increase out of a good feeling or pressure to grow. Three things need to be in place before you ask: the signal from the previous period, repeatable and tied to the number you agreed to report; the limit you're now hitting, why the current budget became a ceiling, such as an audience running out; and a new range with the same honesty as the first ask. Without these three, you're asking to learn again, dressed up as scaling, and the CFO will sense it.
Fictional example: at Veldbrug, a made-up software company, the marketing lead came back with the agreed pilot number, the finding that the audience had run out, and a new range for broader reach. No result promised, just a repeatable signal and a clear limit.
The instrument: the structure of the budget ask
Copy this and fill it in before you walk into the meeting. The order is fixed: what you know comes before what you ask.
1. What we learned last period. The number you agreed to report, and whether the assumption you tested held up.
2. This ask is: to learn or to scale. Pick one, and write down why.
3. The amount, as a range. A minimum and a maximum, never a single number without margin.
4. The assumptions behind this amount. Three to five, named, with how you'll test each.
5. What we'll measure, and when we come back. One metric, one date. No backup metric in case the first disappoints.
6. The stop rule. The signal, or its absence, that means we stop or adjust.
7. What this isn't. One sentence: no return promised, no timeline to profit, just an answer to the question from point two.
The three questions a CFO always asks
"What if it doesn't work?" Answer with the stop rule from point six, not with reassurance. "Then we stop at the agreed moment, and we know this channel isn't it. That itself is an answer we don't have yet."
"When will we know?" Answer with the date and metric from point five, never "soon". A vague timeline sounds cautious but reads as evasion.
"Why this amount and not half?" Answer with the assumptions from point four: what half the amount can't test, such as an audience too small to produce a signal. Derive the amount from what you need to measure, not from what was available last year.
How to read the response
Questions about the assumptions, the stop rule, or the date are normal, a sign the conversation is working: the CFO is testing, and you have the answers ready because you wrote them out beforehand.
Alarm looks different. If "what will this return" keeps coming back after you've explained three times what you're going to learn, the CFO doesn't want a budget ask, they want a promise, and this instrument won't resolve that. If there's no room to discuss a stop rule, "we don't stop, we have to make this work," the budget is already locked in emotionally or politically before the conversation started. Catch that early: it's no longer a budget conversation at that point.
Closing
This instrument guarantees no approval and no return. What it does guarantee is that you're asking an honest question you can keep, no matter how the quarter turns out. If the question is about the document that lays out a full plan, our guide "From rough marketing plan to a decision document a board signs" covers that structure. This guide is about the moment you sit across from the person who decides.
Take your next budget ask and run it through the seven points above before you walk into the meeting. The point where you get stuck is the point you need to write out first.