Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

Marketing Finance › Troubleshooting

Finance Does Not Trust Marketing Numbers

Why finance rejects marketing's numbers: attributed revenue above real sales, spend that differs from the ledger, ROAS instead of profit and no proof of lift.
Troubleshootingintermediate8 min readUpdated

First, ask which number is rejected and trace it

"Finance doesn't trust our numbers" is usually about two or three specific figures, and asking which ones is the fastest part of the diagnosis. Sit with the finance business partner and go through the last report: which lines were questioned, which were ignored, and what finance did instead. Common answers are that marketing's spend did not match the ledger, that claimed revenue was larger than the company's, that the return figure ignored fees or margin, and that nobody could show the sales would not have happened anyway.

For each rejected figure, trace its source: which system it came from, which definition it used, which costs or revenues it included, and whether the same figure was calculated the same way last month. Kaushik's measurement model starts from the business objective and derives the measures from it; finance distrust is often a sign that marketing's measures were derived from what platforms report rather than from what the business is trying to achieve. The trace usually points to one or two of the causes below.

Diagnosis table

Diagnosis

SymptomCheckLikely causeFix
Finance points out that claimed revenue exceeds total salesSum of platform-credited revenue against recognised revenue for the periodAttributed revenue summed across platformsReport one business standard; show the gap to actual sales
Marketing's spend differs from the ledger every monthSource of the spend figure: platform dashboards or finance's cost centresSpend reported from platformsReport spend from the ledger after reconciliation
Finance calls the returns "revenue, not profit"Whether the headline return uses margin and all marketing costsROAS presented where ROI is neededLead with return on investment on contribution
Finance asks "would we have sold this anyway?" and gets no answerWhether any test has measured the lift caused by the largest budgetsNo incrementality evidenceRun agreed holdout tests on the biggest spend
The same measure shows different values in different reportsWritten definitions and their change historyDefinitions changing between reportsHold definitions in one dictionary; date every change
Finance cannot see marketing in its own monthly viewWhether marketing's results appear in the P&L format finance reviewsResults outside the P&L viewBuild a marketing P&L view that reconciles to the accounts

Fix, verify and prevent each cause

Credited revenue that adds up to more than the business sold

Each advertising platform records the conversions it can connect to its own ads; Google's conversion tracking documentation describes measuring what people do after interacting with an ad, and every platform does the same for its own ads under its own rules. A customer who touched two platforms is claimed by both, so the sum across channels exceeds the company's sales, and finance concludes the whole report is inflated.

Fix: stop adding platform figures together; use the reconciliation method in the platform versus analytics discrepancies guide to adopt one business standard for Attribution and show credited revenue beside recognised revenue with the gap stated.

Verify: the report's revenue total matches or is reconciled to finance's revenue for the same period.

Prevent: keep a standing line in every report for the difference between credited and actual revenue, so it is explained before anyone asks. For Google Ads, the tutorial on reconciling conversions with sales records builds that line from the order system or CRM, cause by cause, and turns it into a ratio for setting targets.

Spend taken from dashboards rather than the ledger

Platform dashboards show media delivered; the ledger shows what was booked to marketing's cost centres, including fees, production and accruals, in the month the service was delivered. Reports built on dashboards rarely match, and every mismatch costs credibility.

Fix: run the monthly reconciliation in the tracking spend against budget guide: match the tracker to the ledger line by line, correct coding errors, supply accruals on time and list the known differences. Report spend from the reconciled figure.

Verify: for a full quarter, the spend in marketing's report equals the ledger or differs by listed, explained items.

Prevent: agree month-end deadlines with finance and make one named person responsible for accruals and the reconciliation. That person can work from the marketing budget month-end close checklist, which covers spend capture, accruals, reconciliation and variance notes in order.

A revenue ratio offered where finance needs profit

Finance reads any return as a profit return. A report that leads with return on ad spend, revenue over media, invites the response that it says nothing about whether marketing made money.

Fix: lead the finance report with return on investment (ROI) on contribution margin with all marketing costs, and keep the channel ratios in a separate section, following the ROI vs ROAS comparison. Where the profit picture itself is weak despite strong ratios, the high ROAS but not profitable diagnosis covers the causes.

Verify: finance discusses the return figure on its merits rather than disputing what it means.

Prevent: define both returns in the shared dictionary and send each to the meeting that uses it.

No evidence that marketing caused the sales

Even a reconciled, margin-based report shows only that marketing was present before sales. Finance's sharper question is whether the sales would have happened without the spend, and attribution cannot answer it.

Fix: agree with finance one or two Incrementality tests on the largest budgets, designed as the incrementality testing guide describes, with the method, the control and the success threshold written down before the test starts.

Verify: the test produces a lift estimate with its uncertainty, and finance accepts it as the basis for that channel's budget.

Prevent: schedule a test on the largest or fastest-growing channel at least yearly, and use the results to calibrate how much of the attributed figure to trust.

Definitions that move between reports

If "revenue", "customer" or "acquisition cost" means one thing in March and another in June, or one thing in the marketing deck and another in the board pack, finance cannot compare anything and stops trying.

Fix: write each measure once, with its formula, source, filters and owner, in a shared dictionary as the metrics layer and definitions guide sets out, agree it with finance, and calculate reports from it rather than in each spreadsheet.

Verify: the same measure for the same period returns the same value in every report that uses it.

Prevent: change definitions only by agreement, with the change dated and noted in the next report.

Results that never appear in finance's own view

Finance runs the business on the profit and loss account. If marketing's results live only in dashboards of clicks, leads and credited revenue, finance has to translate them, and translation breeds doubt.

Fix: build a marketing view in profit and loss form, from attributed net revenue through contribution to marketing cost, tied back to the ledger, as the marketing P&L guide describes, and prepare answers to the questions finance asks about it.

Verify: finance uses the view in its own monthly review without re-working it.

Prevent: keep the view's basis stable, including the people-cost convention, so trends mean something year to year.

A prevention routine for credibility

Trust is built monthly and lost in a single meeting, so the routine is about consistency, and a wrong figure that has already been sent is handled openly, as the tutorial on correcting a number after a report has gone out describes. After each month-end, reconcile spend to the ledger and send finance the reconciliation with the report, not after it is challenged. In the same report, show credited revenue beside recognised revenue, with the gap and its explanation. Lead with the return on contribution with all marketing costs, and put channel ratios in their own section. Each quarter, review the metrics dictionary with the finance business partner and record any change with its date; agree which tests will run in the coming quarter and report the results of the last. Each year, rebuild the marketing P&L view with finance and confirm the conventions it uses. Throughout, answer questions from finance in finance's terms: which line, how much, timing or permanent, and what it means for the forecast. The routine holds up best inside a settled relationship, and the guide to working with a finance business partner sets out the written split of ownership, the monthly review with its decisions log and the habit of raising risks early.

Illustrative example — not a real result

A software company earns back finance's attention

A business software company's finance team had stopped reading the monthly marketing report, which claimed £1.8 million of influenced revenue and a return on ad spend of 7. The finance partner identified three objections: media spend in the report was £142,000 while the ledger showed £171,000 after agency fees and accruals; influenced revenue summed from four platforms exceeded new-business bookings of £1.1 million; and nothing showed what paid search added. Marketing rebuilt the report from the ledger, replaced influenced revenue with bookings by source on one first-touch standard, and reported return on contribution with all costs, which came to 1.9. A six-week regional holdout on paid search estimated that 62% of its credited bookings were incremental. Finance adopted the report as the basis for the next budget round.

If finance asks whether a model or an experiment should settle the question, the comparison of marketing mix modelling and incrementality testing shows how tests supply the causal evidence a model is then anchored to.

Common mistakes

Defending the dashboard instead of reconciling it. Summing platform revenue. Reporting media spend as total marketing cost. Leading with a revenue ratio. Claiming causation without a test. Changing a definition and not saying so. Sending the reconciliation only after finance finds the gap.

Frequently asked questions

What is the quickest way to rebuild finance's trust in marketing reports?

Reconcile one closed month with finance, line by line, and show the result. Match marketing's spend to the ledger, show platform-credited revenue beside the revenue the business actually recognised, and explain every difference. A report that admits its gaps and shows how they were measured earns more trust than one with better-looking numbers, and the same reconciliation, repeated monthly, becomes the basis finance stops questioning.

Finance wants proof that marketing caused the sales. Is attribution enough?

No. Attribution shows which marketing touches were present before a sale; it cannot show which sales would have happened without the marketing. That question needs an incrementality test, such as holding out a region or an audience and comparing results. One or two well-designed tests on the largest budgets, with the method agreed with finance beforehand, usually settle more arguments than any attribution model.

Should marketing report to finance in finance's format or its own?

In finance's format for anything finance will use to make decisions: cost by cost centre and line matching the ledger, return on investment on contribution margin, and the same definitions every month. Marketing's channel metrics still matter, but they belong in an appendix or in channel reviews. A bridge that shows how the channel figures become the finance figures lets each audience see both without confusing them.

Terms used on this page

Where this page is used

Sources

  1. Digital Marketing and Measurement Model — Avinash Kaushik
  2. About conversion tracking — Google Ads Help

Related