Reporting Attribution to a CFO

Why the standard marketing report doesn't land with finance

ROAS, new-to-brand percentage, attribution model comparisons — these are the vocabulary marketing teams use to talk to each other, and they're mostly the wrong vocabulary for a finance conversation. A CFO's actual question is closer to "if we cut this line, does revenue drop, and by how much, with what confidence." Reporting a 6.04x ROAS answers a different, narrower question — what got credited — and a sophisticated finance leader will often ask the incrementality question back immediately, because ROAS alone has never told them what they actually need to know for a budget decision.

The three numbers worth leading with

First, incremental cost per acquisition — not attributed CPA, the number derived from an actual holdout or geo-lift test, because that's the number that reflects what a real new customer costs, not what the platform's own reporting claims. Second, a stated confidence level attached to that number — whether it's from a properly powered test (high confidence), a directional signal like new-to-brand rate (medium confidence), or an unvalidated attributed figure (low confidence, and worth saying so plainly). Third, the answer to the cut-test question directly: based on the evidence available, what would happen to revenue if this specific spend stopped — framed as a range, not a false-precision point estimate.

A worked translation, from marketing numbers to finance numbers

Take a DSP prospecting line spending $12,000 a month, reporting a 4.8x attributed ROAS and a 38% new-to-brand rate. Translated for a CFO: a geo-lift test on this line, run over 6 weeks, showed a 22% incremental lift, implying roughly $2,640 of the $12,000 monthly spend is producing sales that would happen anyway (the non-incremental portion), and the remaining spend is producing genuinely incremental revenue at an incremental CPA of $14.20 per new customer — a number to compare directly against acquisition cost on other channels the CFO already tracks, like Meta or Google. That's a materially more finance-native way to present the same underlying data than "4.8x ROAS, 38% NTB," and it answers the actual question a budget conversation needs answered.

What to admit is uncertain, and why that builds trust rather than costing it

A CFO who has sat through enough marketing reviews has learned to distrust reports with no visible uncertainty, because real measurement always has some. State plainly which numbers are directly measured (a completed incrementality test), which are modelled or estimated (an MMM coefficient, a projected lift from a partial-duration test), and which are simply attributed and unvalidated. A report that says "we're confident in this number because we tested it, and less confident in this one because we haven't yet" reads as more credible, not less, than one presenting every figure with equal, unearned certainty.

The common mistake, including ours

The mistake is presenting a blended, attributed portfolio ROAS as the headline number in a finance-facing report, because it's the biggest, most impressive-sounding figure available, without translating it into the incremental, cost-per-acquisition terms a CFO actually needs. We've led a board-level update with a strong attributed ROAS figure before a genuinely sceptical CFO pushed back with exactly the incrementality question the report should have already answered — and we redid the presentation around incremental CPA and a stated confidence level, which landed the underlying good result far better than the original framing had.

The lesson generalised past that one meeting. A report built around attributed ROAS invites the sharpest possible question a finance leader can ask — would this revenue exist without the spend — and answers it badly. A report built around incremental CPA answers that exact question directly, which means it survives scrutiny from the most skeptical reader in the room rather than depending on nobody asking the obvious follow-up.

Formatting the report itself

Lead with the incremental CPA and confidence level in the first line, not buried after a page of channel-by-channel breakdown — a CFO reading quickly should get the answer to the cut-test question before anything else. Follow with the supporting evidence, including what was tested, when, and at what scale, so the headline number is traceable rather than asserted. Save the marketing-native metrics — ROAS, new-to-brand, CTR — for an appendix, available for anyone who wants the underlying detail but not required reading for the decision itself.

When the honest answer is genuinely bad news for finance

If an incrementality test shows a channel finance has been funding on the strength of a strong attributed ROAS is producing little real incremental lift, say so plainly and immediately, with the incremental CPA number attached, rather than softening the finding or delaying the report until a more favourable period. A CFO who discovers a channel's real incrementality was known and unreported for months loses more trust than one who hears a disappointing number promptly, framed with the confidence level attached and a clear recommendation for what to do next — reallocate, retest at a different scale, or hold pending a longer-window read.