Return on Ad Spend (ROAS)

How ROAS is calculated

The formula is revenue attributed to the campaign divided by the cost of the campaign. Spend $2,000 on ads, attribute $8,000 of orders to them, and the campaign ran at 4× ROAS. It is usually written as a multiple rather than a percentage, and it is deliberately a ratio, so it stays comparable across campaigns of very different sizes.

Two versions of the number exist and get confused constantly. Channel ROAS uses the revenue one ad platform claims and the spend on that platform. Blended ROAS uses total revenue divided by total ad spend across everything. Channel ROAS is what you optimise a campaign with; blended ROAS is the one that reconciles with your bank, and it is always the lower of the two once more than one platform is running.

The window matters as much as the formula. Every ad platform counts a conversion within an attribution window it chooses — often days after the click — so the same week can show three different ROAS figures depending on when you look and which window is set. Pick one window, write it down, and compare like with like; a ROAS that improved because someone widened the window has not improved.

What counts as a good ROAS

There is no universal benchmark, and any article quoting one is quoting somebody else's margin. Your break-even multiple is one divided by your gross margin: at a 25% margin you need 4× just to stand still, while a software product at 90% margin breaks even near 1.1×. Everything below is where the number misleads people.

• Margin sets the floor, so the same number is a success and a failure. A 3× ROAS is a loss on a 30% margin product and excellent on software. Compute your own floor before you copy anyone's target.
• Platform-reported ROAS is not company ROAS. Each platform claims conversions it had a hand in, so one order can appear in two dashboards. Reconcile against blended spend and total revenue before believing a channel is carrying the business.
• First-order ROAS ignores everything after the first order. A subscription or a repeat-purchase store can accept a losing first sale if the second is profitable, which is a decision about customer lifetime value, not about the ad.
• A rising ROAS often means you shrank. Efficiency climbs when you cut spend down to your warmest audience, so a campaign can look better every week while producing fewer customers each week.

ROAS, CAC, and the number underneath both

ROAS and customer acquisition cost answer different questions, and small teams get further with both. ROAS asks how much revenue an advertising pound returned; CAC asks what one new customer cost, and it can include the things ROAS ignores — agency fees, creative production, the discount you offered to close the sale. A campaign can post a respectable ROAS and still produce customers who cost more than they will ever be worth, which is why the two are read together rather than one instead of the other.

A first-party note from running our own paid search at AgentCeres — the AI growth team at agentceres.com. The thing that broke for us was never the ratio, it was the event underneath it: our signup conversion fired correctly while the paid-plan conversion silently did not, because the analytics call was made in a shape the library ignored without raising an error. Nothing looked wrong on the page, in the console, or in the campaign. Every efficiency figure computed in that window was arithmetically perfect and factually false. Before trusting any ROAS number, buy from yourself at full price and confirm the conversion arrives with the correct value attached — a metric is only as honest as the event that feeds it.