What ROAS is and how to calculate it

What ROAS is

ROAS, return on ad spend, is a division: the revenue you attribute to the ads, divided by what those ads cost you. If a campaign cost 1,000 and the account attributes 4,000 of revenue to that campaign, the ROAS is 4. Sometimes it is written 4:1 or 400 percent. It is the same sum.

The formula is the easy part. People also ask what a good ROAS is, and whether it should be high or low. This page answers all three without publishing a Growth Digital benchmark.

The formula, with an example that is not a case study

ROAS = attributed revenue / ad cost. You do not need a real client to understand it. Imagine a shop that spends 200 in a week and a correctly installed tag attributes 800 of orders to those clicks. 800 / 200 = 4. The dashboard will say 4, or 400 percent, depending on the column. Check that both columns talk about the same period and the same campaign. Mixing the month of the cost with the week of the revenue inflates or sinks the number without anything changing in the shop.

The example breaks if the 800 is not from those ads. If the order would have come in anyway through the brand, or if it was counted twice because Analytics and the ads platform are not aligned, the ROAS describes the measurement, not the business. How that measurement is sent is in GA4 and Tag Manager.

Why a "good ROAS" is not a number from this studio

A 4 is often cited in sector guides as a comfortable range for ecommerce. It is someone else's citation, not our result and not a promise. It helps you understand the conversation. It does not help you set your budget.

The sum that is yours: if the product leaves a 20 percent margin after making and shipping it, a ROAS of 4 leaves little once you subtract the ad, and it can leave nothing if there are returns. If the margin is 70 percent, a ROAS of 2 can be healthy. The same number, two businesses, two answers.

• Look at the margin of the product or the service, not the ROAS in a social thread.

• Subtract returns and cancellations if the platform does not.

• Do not compare a brand campaign, where people were already looking for you, with a prospecting campaign.

• Do not declare victory in three days. A long sales cycle shows the click today and the revenue next month.

High ROAS, low ROAS, and volume

People ask whether ROAS should be high or low. Higher, if everything else stays the same, is better than lower. Everything else almost never stays the same. Raising the return by turning off the campaigns that bring new customers leaves a pretty number and a smaller business. Lowering the target so you can spend more only makes sense if the margin can hold each extra unit.

That is why ROAS is not the only column. Next to it you need spend, the number of conversions, and the average value. A return of 12 on 40 of spend is not a strategy. It is a small sample.

Where the calculation breaks

1. Revenue never reaches the account. The ad is there, the form is there, and nobody has sent the value. ROAS comes out at 0 and the conclusion is false. Measurement first, judgment second.

2. Revenue is counted twice. The platform and Analytics attribute the same order. Adding them duplicates it.

3. Brand and generic are mixed. The campaign for your name usually has a very high return because those people were going to find you anyway. Using it as proof that "the ads work" hides the campaign that actually brings people who did not know you.

4. A lead is not revenue. In a service business, the form is not invoicing. If you assign an invented value to each lead, the ROAS is a hypothesis. Say so, or wait for the accepted quote.

That fourth point is the one that most resembles an agency case study, and we are not going to decorate it with a percentage. If you sell a service, useful ROAS appears when the CRM says which lead became an invoice, not when the form says thank you.

What to do with the number on Monday

Export the cost and the attributed revenue for the last four weeks, campaign by campaign. Do the division by hand once, so you understand the column in the dashboard. Separate brand and non-brand. Compare the result with the margin, not with a 4 you remember.

If the number cannot be calculated because there is no revenue in the account, the next step is not to change the bid. It is to understand Tag Manager and send the missing event. If the number can be calculated and it does not fit the margin, the next step is the offer or the page, not an agency that promises a return. The channel, when it is time to work on it, is paid media.