ROAS

A company spends money on advertising and hopes that this will lead to more sales. ROAS is the number that measures exactly that. It answers a simple question: How much revenue comes back if I spend one euro on advertising? The abbreviation stands for the English term “Return on Ad Spend,” roughly translated as “return from advertising expenditure.” The calculation behind it is a division: revenue divided by advertising spend. Anyone who pays 1,000 euros for ads and thereby earns 4,000 euros has a ROAS of 4.

What a ROAS of 4 really means

ROAS is so popular because the number is immediately understandable. A value above 1 means: more comes back than was put in. A value below 1 means the advertising is burning money. Marketing departments can use it to directly compare individual campaigns and shift budget to wherever the number is highest.

However, there is a typical misconception. ROAS measures revenue, not profit. If a retailer sells a product for 100 euros but bought it themselves for 80 euros, only 20 euros of margin remain. A ROAS of 2 then looks good, but in reality means a losing deal. That’s why experienced companies calculate what is known as break-even ROAS: the value from which advertising even becomes worthwhile in the first place.

Closely related is the metric ROI, the Return on Investment. It includes all costs and shows true profit. ROAS is more narrowly defined and only looks at advertising expenditure. In return, ROAS can be calculated faster and on a daily basis, which is why it is preferred in day-to-day operations.

How advertising platforms optimize toward a target value

For ROAS to be calculated at all, a sale must be attributed to an ad. This is called attribution. If someone clicks on an ad and shortly afterward makes a purchase in the online shop, the shop reports this revenue back to the advertising platform. The metric emerges from many such reports.

Modern systems like Google Ads or Meta go a step further. The advertiser only specifies a target, for instance a ROAS of 5. A machine learning model, that is, a program that has learned patterns from past data, then estimates for each individual user how likely a purchase is and how high the resulting revenue might be. Based on this estimate, the system automatically determines how much it bids for the ad placement.

The whole process runs as an auction and takes milliseconds. Whoever promises the most gets the ad slot. The time delay is important here: some people only buy days after the click. The ROAS of a campaign therefore often rises further after the fact.

ROAS in quarterly results and in practice

In the news, ROAS mainly comes up around major platform corporations. When Meta or Amazon present their quarterly results, they like to argue with the rising ROAS of their advertising clients. The message behind this is: our advertising works, so give us more budget. Data protection debates are also linked to this, since stricter rules make attribution harder and push down the measurable values.

Anyone running their own small shop sees ROAS directly in the advertising account as a simple number or percentage figure. Typical target values range between 3 and 10, depending on the industry. For expensive products with high margins, a low value is sufficient, whereas for groceries with tight margins, significantly more is needed.

One final point is often overlooked. ROAS rewards advertising to people who would have bought anyway. Anyone who shows an ad to loyal repeat customers achieves fantastic values but doesn’t gain any new customers. That’s why larger companies additionally look at metrics that measure real growth.