High ROAS, Failing Business: What the Ratio Leaves Out

"The campaign is a success, ROAS is excellent." I hear this sentence in most marketing reviews, and I have learned to treat it as the start of the conversation, not the end. ROAS is a useful number for comparing campaigns inside one ad account. It is a poor number for deciding whether the business is healthy, because almost everything that decides profit happens after the moment ROAS is recorded.

What does ROAS actually measure?

ROAS is the conversion value an ad platform records, divided by what you spent on ads. Google's own documentation writes it as conversion value ÷ ad spend × 100%, so 5 dollars of reported sales for 1 dollar of spend is a 500% ROAS.

The important word is "reported." Google describes conversion values as values you report through conversion tracking. The platform does not know your revenue. It knows what your tag told it at the moment of the order or the form submission. If that value is the basket total at checkout, ROAS is built on orders placed, before anyone has canceled, returned or failed to pay.

What does ROAS leave out?

It leaves out every cost and every event that sits between an order and money in the bank. In practice that means six things:

• Margin. Two products with the same price and very different cost of goods produce the same ROAS and very different profit.
• Canceled and refused orders. In markets where cash on delivery is common, an order is a promise until the courier collects.
• Returns and refunds, including the shipping you paid in both directions.
• Uncollected payments: installments, invoices on credit, bank transfers that never arrive.
• Lead quality. For a service or training business, the "conversion value" is often an estimate assigned to a form. A form is not a sale.
• Sales follow-up. A lead answered the next day is worth less than the same lead answered in minutes, and no ad report shows that difference.

Discounts add a quieter problem. A deep offer can raise ROAS and lower profit at the same time, because the ratio rewards revenue.

How can a campaign with a ROAS of 5 lose money?

It loses money when the costs and losses after the order are larger than the gap between revenue and ad spend. Here is a hypothetical example with round numbers, for an online store that spends 10,000 dollars on ads in a month.

The dashboard says ROAS 5. The business kept 4,000 dollars to cover salaries, rent, software and the owner's income. Move cancellations up a little or margin down a little and the same "excellent" campaign is paying customers to take products away.

What should I read beside ROAS?

Read three numbers beside it: revenue actually collected, contribution margin after advertising, and a blended efficiency ratio.

Collected revenue is what reached the account after cancellations, returns and failed payments. It comes from your store, your CRM or your accountant, never from the ad platform.

Contribution margin after advertising is collected revenue minus the variable costs of delivering those orders (cost of goods, shipping, payment fees) minus ad spend. It is the amount left to pay for the fixed costs of the business. If this number is small or negative, no ROAS figure rescues the month.

The blended ratio, which many practitioners call MER (marketing efficiency ratio), is total revenue divided by total marketing spend for the same period. It is not a platform metric. Each platform can claim the same sale, but the bank statement counts it once. I calculate it on collected revenue.

Can I make the platform's ROAS more honest?

Yes, partly, by sending the platform what happened after the order. Google Ads supports conversion adjustments: you can retract a conversion entirely or restate its value, and its documentation uses partial returns as the example. Google Ads also supports offline conversion imports, which let you measure what happened after an ad led to a click or a call, such as a sale closed by phone or in your office.

This matters for more than reporting. Bidding strategies optimize toward the values they receive. If they are fed orders that later get refused, they learn to find more people who place orders and refuse them.

At Argan Package in Saudi Arabia, I built a campaign for slow-moving products from sales and warehouse data instead of from the ad account's own numbers, and according to my campaign records, sales of those products rose 400% within 7 days. The business data chose the campaign, and the ad platform only delivered it.

What should I do this week?

Set up one weekly check that takes less than an hour, on the same day each week.

Take the last full week that is old enough for most deliveries and returns to have settled. Write the platform's reported revenue and ROAS on one line. Under it, write collected revenue for the same orders, then the cancellation rate and the return rate. Subtract variable costs and ad spend to reach contribution after advertising. Divide total collected revenue by total marketing spend for the blended ratio.

Then look at the gap between the first line and the collected line. If it is stable from week to week, ROAS remains a usable daily signal. If it moves, find out which campaign, product or payment method moved it before you raise any budget. And change the sentence in the next meeting: report what was collected and what was left after costs, and mention ROAS last.