Definition
Return on ad spend (ROAS) is the revenue an ad platform attributes to a campaign, ad group or account, divided by the amount spent on it over the same window. A ROAS of 5 means the platform credited five dollars of orders to every dollar spent. Because each platform measures its own attributed revenue, ROAS is a channel-level view rather than a statement about the whole store.
Formula
ROAS = Attributed revenue ÷ Ad spend
- Attributed revenue
- Order value the platform credits to its ads under its own attribution settings, for the period being measured
- Ad spend
- Media cost billed by that platform for the same period, before agency fees or tooling
One Shopping campaign over four weeks
- Ad spend on the campaign
- $12,000
- Conversion value reported by the platform
- $54,000
- ROAS as the platform reports it
- 54,000 ÷ 12,000 = 4.5
- Orders from the same campaign confirmed in the store
- $46,000
- ROAS on store-confirmed orders
- 46,000 ÷ 12,000 = 3.8
Illustrative figures. The gap between 4.5 and 3.8 is normal: the platform counts view-through and assisted orders that the store's own data may credit elsewhere. Neither number is wrong, but they answer different questions, and a bid strategy set against the first will behave differently from one set against the second.
Why it matters
ROAS matters because it is the number the ad platforms optimize toward and the number most bid strategies accept as a target. Google Ads and Meta both let a store set a ROAS goal, and their algorithms then bid more where the predicted return is high and less where it is low. That makes ROAS the lever a store actually pulls day to day. It is also the fastest signal that something has changed: a product going out of stock, a competitor undercutting on price, a feed error dropping the best sellers. The catch is that ROAS only reports what that platform saw, so it is a steering metric for a channel, not a verdict on whether the business made money.
Where it goes wrong
- Treating ROAS as profit: a ROAS of 3 on a product with a 30 percent gross margin loses money once cost of goods, shipping and fees are counted, which is why break-even ROAS has to be worked out first
- Adding ROAS across platforms: Google and Meta can both claim the same order, so summing their attributed revenue produces a figure the store never received
- Chasing a higher ROAS by cutting spend: the last dollars in a campaign usually return less than the first, so trimming budget lifts the ratio while shrinking total profit
- Comparing ROAS across campaign types: a branded search campaign will almost always show a higher ROAS than prospecting, because it captures people who were already coming, not because it is a better use of money
- Changing the attribution window and reading the jump as improvement: a longer window credits more orders to the same spend and the ratio rises without any change in results
Questions about ROAS
Is ROAS the same as return on investment?
No. ROAS divides attributed revenue by media spend only. Return on investment subtracts all costs, including product cost, fulfilment, fees and the people running the campaigns, and divides the remaining profit by the total investment. A campaign can show a strong ROAS and a negative ROI at the same time.
Why does my ROAS change when I switch attribution models?
Because the numerator changes. Data-driven, last-click and first-click models split credit for the same orders differently, so the revenue assigned to a campaign moves even though the orders and the spend did not. Pick one model, note the date you chose it, and compare periods only under the same model.
Should ROAS targets be the same for every campaign?
Usually not. Campaigns that reach new customers can justify a lower ROAS if the product has a strong repeat rate, while remarketing and branded campaigns should carry a higher target because much of their revenue would arrive anyway. Set targets from each campaign's role and margin, not from one account-wide number.