What is ROAS, and why is it not enough on its own?

ROAS is the most-quoted number in performance marketing. On its own it gets campaigns switched off, budgets raised and agencies replaced. But ROAS is not a measure of profitability — it is a ratio, and depending on what you put in the numerator and which window you look through, the same account can report a 2 or a 6 in the same month.

Below: how it is calculated, why it never matches the business’s own numbers, how margin turns a “good” ROAS into a loss, and what belongs beside it.

What is ROAS?

ROAS (Return on Ad Spend) tells you how many times over your ad spend came back as revenue:

ROAS = Revenue from advertising ÷ Ad spend

A hypothetical example: ₺20,000 of spend produces ₺80,000 in revenue, so ROAS is 4. Some dashboards print that as 4x, others as 400%. Same thing.

Both sides of the formula are more arguable than they look. Does “revenue” include VAT and shipping, and are refunds deducted? Does “spend” cover only media, or also agency fees, creative production and tooling? Comparing ROAS before those definitions are settled is like comparing distances without a unit.

ROAS and CPA do not measure the same thing

CPA is a cost; ROAS is a ratio. Both can fall at once — if average order value drops, you buy more orders more cheaply while revenue stays flat. They are read side by side, never swapped.

Platform ROAS is not the same as real ROAS

The ROAS in your Meta dashboard is calculated from the revenue Meta attributes to itself, and Google Ads does the same with its own model. Add two dashboards’ revenue together and you will usually clear the company’s actual turnover — because the same sale can be claimed twice.

The number that reflects the business is blended ROAS:

Blended ROAS = Total revenue ÷ Total ad spend

“Total revenue” is what accounting sees; “total spend” is every channel’s spend. This number appears in no dashboard and usually sits well below the platform figures.

Why the two numbers never reconcile

• Attribution windows differ between platforms.
• View-through conversions are included in some reports and not others.
• Cross-device matching and modelling fill in gaps with estimates.
• The dashboard cannot see refunds, cancellations or uncollected orders.
• If VAT and shipping are baked into the value parameter, revenue is inflated.
• Where measurement is incomplete, some real sales are never reported at all.

Do not force the two numbers to match. Track the gap as a ratio: while platform ROAS holds a stable multiple of blended ROAS, in-platform comparisons still mean something. When that multiple drifts, check the measurement setup before the campaigns — where the loss comes from is covered in our Pixel and Conversions API article.

How attribution windows inflate ROAS

An attribution window decides how many days after an ad a conversion can still be credited to it. A seven-day click window counts more conversions than a one-day window: the ads have not changed and the sales have not changed, only the counting rule got wider. So when comparing two periods, two accounts or two agency reports, ask first whether the window is the same.

View-through conversions are a separate argument. Someone sees an ad, does not click, arrives the next day and buys — and the sale is credited to the ad. Sometimes that is fair. Sometimes they were going to buy anyway.

The real question is incrementality

Retargeting shows the highest ROAS in almost every account, because it harvests demand that already exists. An ad shown to someone who abandoned a full cart can post a spectacular ROAS, and some of those people would have come back anyway.

Proving that a campaign generates additional revenue takes a controlled test: a holdout that keeps part of the audience out of the ads, a geographic split, or a platform conversion lift study. Such tests usually make ROAS look worse. They also make it true.

Without margin, ROAS means nothing

ROAS is calculated on revenue, not profit, so the same ratio produces opposite outcomes in two businesses. A hypothetical example: two stores both take a ₺1,000 order and both report ROAS 4 — ₺250 of advertising per order.

• Store A. Product cost ₺750, gross margin ₺250. Ads take ₺250 and leave nothing — before shipping, returns, payment fees and packaging. The store is losing money.
• Store B. Product cost ₺400, gross margin ₺600. Ads still take ₺250, leaving ₺350 of contribution per order.

One threshold captures the difference:

Break-even ROAS = 1 ÷ gross margin rate

At a 25% gross margin, break-even ROAS is 4; at 60%, it is 1.67. The same “ROAS 4” is break-even in one business and healthy profit in the other — which is why industry-average ROAS targets are useless. Your target is a derivative of your margin structure, not your sector.

To make it honest, use contribution margin: what remains after product cost, shipping, packaging, payment fees, returns and any marketplace commission. Set break-even from that figure and you close campaigns against a threshold instead of a feeling.

New customers, or the ones you already had?

An account’s total ROAS hides the customer mix inside it. Ads served to people already searching for the brand, and to existing customers, produce high ROAS — so as those line items take more of the spend, total ROAS climbs even while the business stands still.

Growth comes from new customers, and they are always more expensive. That needs its own number: ad spend divided by genuinely new customers, read against the contribution margin of a first order. A model that cannot break even there survives only if people buy again — which is why lifetime value, payback and CAC sit at the centre of how we approach growth marketing.

Expect ROAS to fall as spend rises, because you are reaching colder audiences. The question is not “why did ROAS drop” but “did total contribution margin rise despite the drop?”

The metrics that belong next to ROAS

• Contribution margin. What an order leaves after every variable cost.
• New customer acquisition cost. The price of growth, not blended CPA.
• LTV and LTV:CAC. Whether a first-order loss is rational or fatal.
• Payback period. How many months acquisition cost takes to come back.
• Blended ROAS / MER. All channels combined — the reality check.
• Average order value and conversion rate. Did ROAS move because of the ads, or because of the site?
• Return and cancellation rate. Invisible in the dashboard, quietly eating the margin.

How to actually make the call

1. Calculate break-even ROAS from contribution margin.
2. Set target ROAS by adding the profit you want on top, not by copying a sector average.
3. Put platform and blended ROAS side by side monthly; track the gap as a metric.
4. Report new customer acquisition cost separately.
5. Never kill a campaign on one week’s ROAS. At low volume, ROAS is noise.

All of it assumes the measurement underneath is correct: conversions that go missing or get double-counted make any margin model meaningless.

Is your ROAS actually producing profit?

We can review your ad account, measurement setup and margin structure together, then send you in writing where your break-even ROAS sits and which line item the budget is leaking into. See how we work on the Meta side on our Meta Ads service page, or get in touch for a free audit — non-binding, and it does not require full access to your account.

Frequently asked questions

What counts as a good ROAS?

No single number works for everyone. The threshold is 1 divided by your contribution margin rate: at 25%, break-even is 4; at 60%, 1.67 is already profitable. Target ROAS is the profit you add on top, and in a growth push it can sit deliberately lower.

Why doesn’t Meta’s ROAS match my accounting?

It counts only the sales it attributes to itself, uses its own window, may include view-through conversions, and never sees refunds. If VAT and shipping sit inside the value you send, the gap widens. The two figures are not expected to match; the ratio between them should stay stable.

Can profit go up while ROAS goes down?

Yes, and in a scaling account that is normal. As you open up to colder audiences, ROAS declines while total orders and total contribution margin can both rise. The criterion is the contribution margin left above break-even.

Why does ROAS swing so wildly on small budgets?

With few conversions, one high-value order can double a weekly ROAS and one refund can halve it. At low volume, compare longer periods, or wait until enough conversions accumulate to make the comparison meaningful.