ROAS can rise while the business earns less

A stronger revenue-to-ad-spend ratio can conceal weaker margins, heavier discounting or a less valuable customer mix; reconcile the advertising number with contribution before celebrating.

The paid-media report is green. Return on ad spend has risen from four to five. The finance report is less cheerful: the campaign leaves less money after serving the orders.

There is no mathematical contradiction. Revenue divided by advertising spend is not the same calculation as contribution after advertising. A campaign can improve the first while weakening the second.

The CMO needs to know what the numerator contains and which costs sit outside the dashboard. Otherwise optimisation can reward the appearance of efficiency while the business loses economic quality.

Read the value definition first

Google Ads documentation explains that conversion values can represent different kinds of business value. A ROAS label is therefore incomplete without knowing which values the account receives.

If the value is order revenue, the ratio says how much reported revenue is associated with each unit of ad spend. If it is a different value, the interpretation changes. Neither should be silently compared with a report using another definition.

Ask whether the figure includes tax, shipping, discounts and later adjustments. Ask how cancellations and returns are handled. These are basic commercial questions, not a challenge to the media team’s competence.

Also record the attribution approach and window. A change in credited revenue can arise from reporting rules as well as customer behaviour.

Work through a simple comparison

Consider two hypothetical campaigns, each with €10,000 of advertising spend. Campaign A records €40,000 revenue and has a 40% contribution margin before advertising. Campaign B records €50,000 revenue and has a 25% contribution margin before advertising.

A has revenue ROAS of four and contributes €16,000 before ads, leaving €6,000 after ads. B has revenue ROAS of five and contributes €12,500 before ads, leaving €2,500 after ads.

The higher-ROAS campaign leaves €3,500 less. In this illustration, contribution margin already includes the relevant variable costs of delivering those revenues; fixed overhead is excluded from both comparisons.

These invented figures demonstrate the mechanism. They do not imply that one ROAS threshold works across products, customers or businesses.

Find what changed underneath

Inspect product mix first. The campaign may have shifted toward high-revenue products with weaker margins. A promotional bundle may raise order value while adding expensive components or fulfilment work.

Then inspect discounts and returns. An offer can create attractive initial order values that fall after refunds, or encourage customers to buy products with a high probability of being returned.

Check customer mix too. Existing customers, new customers and low-fit bargain seekers can behave differently after the first transaction. A blended total can conceal a shift that matters to the business.

Finally, examine service capacity. Additional demand may require overtime, expedited shipping or external support. A campaign forecast based on normal delivery costs may become inaccurate as volume grows.

Reconcile reports rather than forcing agreement

Commerce and advertising systems can use different dates and definitions. An order may be credited to one reporting period while its return appears later. That makes a short-period comparison difficult without a cohort view.

Shopify’s profit-report documentation illustrates another practical dependency: profit reporting needs the relevant cost data. Missing or outdated costs can produce a misleading view even when the sales totals are correct.

Ask finance and analytics to reconcile one representative cohort from order through refund and fulfilment. Document the differences rather than changing definitions until the dashboards happen to match.

The aim is an agreed account of the economics, with each platform report used for the job it can actually perform.

Give optimisation a commercially useful signal

If margin varies substantially, the team may need more differentiated values or targets. The right implementation depends on data quality, platform capability and the business’s objectives.

Do not feed an elaborate estimate into automation merely because it looks more sophisticated than revenue. A noisy or stale margin model can direct spending badly too.

Start with a reliable distinction the business can maintain, such as clearly different product groups or customer outcomes. Test the change and observe both volume and contribution, because improving the ratio can reduce total profitable demand if the target becomes too restrictive.

A high ratio on a tiny volume is not automatically the best commercial result.

Put the board discussion on two axes

Report efficiency and total contribution together. Show how much value remains after the relevant costs and how that changes when spending increases or decreases.

Include cash timing where it matters. A profitable campaign can still consume cash before customer payments arrive, particularly when fulfilment or rewards are paid early.

Set a review trigger for material margin, return-rate or mix changes. The target that was sensible last quarter may no longer fit the same headline revenue ratio.

ROAS remains a useful operating metric when its definition is clear. The mistake is asking it to stand in for the entire business model. A CMO should be able to explain both why the ratio moved and whether the company is better off as a result.