What numbers does a CFO want to see from marketing?

Usually three: gross profit after ad spend, how long spend takes to turn into sales, and how sure you are about each channel. Shopify's profit reports give the first, GA4's days to key event the second, and a test or a stated method the third. Platform ROAS gives none of them.

Why doesn't my CFO trust platform ROAS?

Because each platform counts sales by its own rules and windows, so their claims overlap and can add up to more than you sold. The rules also change, as Meta's move to count link clicks only shows. A CFO trusts numbers that tie to the books and keep the same method.

What do I say when the CFO asks what happens if we cut a channel?

Say what you know and how you know it. If a test covered the channel, give its lift and confidence. If not, say it is untested, name the test that would settle it, and offer a cut in some regions as that test. A guess dressed as a forecast costs trust.

How do I defend my channel mix to my CFO?

Speak the CFO's language: gross profit, not platform ROAS. Show one table that ties to the revenue in your books. Say whether a test, a model or only a click report backs each channel, and bring one test plan for the biggest unproven line.

By Joris van Huët, Founder & CEOUpdated 2026-10-017 min read

Run the numbers for your store: the free multi-channel budget calculator.

Usually by speaking in margin, not in platform ROAS. Show one table that ties to the revenue in the books. Say whether a test, a model or only a click report backs each channel. Then bring a test plan for the biggest unproven line. Defend what you can show; offer to test the rest.

A CFO is not asking which channel is best. They want three answers: does the spend pay back in margin, how sure are you, and what happens if we cut it. Platform dashboards answer none of the three, which is why decks built on them lose the room.

What one store's data shows

One store's anonymised GA4 export, 1 January 2024 to 21 August 2026. It holds shares of revenue only: no ad spend, no order counts.

What the export shows	Value	Source cell
All channels in the touched view, added up	110.4% of revenue	Channels sheet, Touched column total
Direct, in last click, first click and touched views	57.7% of revenue	Channels sheet, Direct row
Journeys with 2 to 3 touches (12.5 days to buy)	12.2% of revenue	Journeys sheet, 2 to 3 touches row
Break-even ROAS at a 40% margin (1 / 0.40)	2.5x	Break-even sheet, 40% margin row

Start where a CFO starts: the total. Add up the touched view in that store's Channels sheet and you get 110.4% of revenue. Nothing is broken; a journey that touched two channels counts in both. But a column that sums to more than the whole looks like double counting to anyone who closes the books. Lead with revenue from the books, then show shares that add up to the whole.

Then comes the question you will get. Direct holds 57.7% of revenue in all three views of that store's Channels sheet. A CFO reads that as: most sales come free, so why pay for ads? The export can't answer that either way. Direct means GA4 saw no source for the visit, and a buyer who saw an ad and later typed your address can land there. Say what you don't know, then offer the test that settles it.

Timing is the third trap. Journeys with 2 to 3 touches took 12.5 days to buy in that store's Journeys sheet. Spend booked late in a month can turn into sales the next month, so a month-end ROAS punishes whatever ran last. Accountants match costs to the revenue they earn. Give ads the same courtesy.

The last row speaks the CFO's own language. In the Break-even sheet, a 40% margin means each euro of ads must bring back 2.5x in sales just to break even. It is arithmetic, not a result for that store, and each channel should be judged against it on the sales it caused.

What the export can't show is spend, so it holds no return for any channel. It shows where credit landed. Your CFO needs what each channel caused, and that takes your spend, your margin and at least one test.

Why does platform ROAS lose the argument?

Three platforms, three clocks. Meta counts a purchase up to 1 or 7 days after a link click, or 1 day after an impression. Google Ads counts a click-through conversion for 30 days unless you change the window. GA4 looks back 90 days for most key events by default. One sale can be claimed under all three rules at once.
The rules move. Meta now counts click-through only after a link click. It says some accounts may still use prior versions while the change rolls out. A chart that jumps between quarters may be measuring a rulebook, not your customers.
Revenue is not margin. Platform ROAS divides revenue by spend. A CFO divides gross profit by spend, and Shopify counts profit only for products with a cost recorded when they sold. For illustration, a 4x ROAS on a 20% margin returns 80 cents of gross profit per euro: a loss.
A claim is not a cause. Every platform number counts sales that followed an ad, including sales that were coming anyway. Only a test separates the two.

What will a CFO accept as evidence?

Grade each channel by the strongest evidence behind it, and say the grade out loud.

A test. A lift study or a regional holdout, with its confidence. Google advises testing before major budget decisions, or using studies to check what a media mix model suggests.
A model. A media mix model or a causal model of your paths, with the method named.
Attribution. GA4's credit for the channel under two models, shown as a range.
A claim. The platform's own number, labelled as the platform's.

Expect most of your mix to sit on the lower grades. Google says that, based on historical data, most advertisers run about 1–2 studies per year. So aim the next test at the line where being wrong costs most, and tell your CFO when it reads out.

One test also earns trust for the rest of the page. Meta lists checking whether your attribution model aligns with a controlled experiment among the questions its lift test answers. Even one tested channel shows how far a click report can drift from what a channel caused.

What to do this week

Check how much revenue has a margin. In Shopify, go to Analytics > Reports, filter the Category by Finances and open the Finance Summary. In the Gross profit breakdown card, compare Net sales with cost recorded against Net sales without cost recorded. Pass: nearly all net sales have a cost recorded. Fail: a big slice has none, so add Cost per item before the meeting.
Ask Meta's own model for a discount. In Ads Manager, click the Columns: Performance dropdown, scroll down to Compare attribution models and select Incremental. Pass: incremental results sit close to standard ones. Fail: there is a wide gap, so even Meta's model says much of the claim was coming anyway, and that line gets tested first.
Find the tests you already have. In Google Ads, go to Lift measurement within the Goals menu. Select the columns icon, then Modify columns, then Conversion Lift, and apply. Pass: a finished study gives you one tested line for the page. Fail: there is none, so book one for your biggest Google line before the next budget round.

Check the homework. Your GA4 Attribution paths export already holds the evidence. Causality Engine reads that one file and shows what each channel caused next to what last-click gave it, in 1 to 2 minutes, for €99 once (excluding VAT), refundable within 30 days. Check the homework

Sources, 1 October 2026: About attribution models and attribution settings (Meta Business Help Center); About conversion windows (Google Ads Help); Select attribution settings (Google Analytics Help); Profit reports (Shopify Help Center); Set up Conversion Lift based on users (Google Ads Help); About Conversion Lift (Meta Business Help Center); Finance reports (Shopify Help Center); How to view results for incremental attribution in Meta Ads Manager (Meta Business Help Center); Set up Conversion Lift based on geography (Google Ads Help).

Frequently asked questions

What numbers does a CFO want to see from marketing?
Usually three: gross profit after ad spend, how long spend takes to turn into sales, and how sure you are about each channel. Shopify's profit reports give the first, GA4's days to key event the second, and a test or a stated method the third. Platform ROAS gives none of them.
Why doesn't my CFO trust platform ROAS?
Because each platform counts sales by its own rules and windows, so their claims overlap and can add up to more than you sold. The rules also change, as Meta's move to count link clicks only shows. A CFO trusts numbers that tie to the books and keep the same method.
What do I say when the CFO asks what happens if we cut a channel?
Say what you know and how you know it. If a test covered the channel, give its lift and confidence. If not, say it is untested, name the test that would settle it, and offer a cut in some regions as that test. A guess dressed as a forecast costs trust.

Go deeper: Causal attribution, explained.

Sixty-second versions of these ideas: Causality Engine on YouTube Shorts.

Keep reading

Terms in this article

Browse the full glossary