ROAS is the wrong number to run your ecommerce ads on

Most ecommerce founders can tell me their ROAS to one decimal place. Ask them the margin on the products those ads sold and the answer gets vague. That gap is where the money goes.

ROAS feels like the truth. Revenue divided by ad spend, one clean number, higher is better. It stops measuring anything useful the moment your products carry different margins, which is to say always.

Take two products, both running at a 4x ROAS. The first is a £50 candle that costs you £12 to make and ship. The second is a £50 supplement bundle that costs you £34. Same ads, same return on paper. The candle hands you real profit. The supplement is close to break-even before you have paid for the office, and if returns or discounts creep in it goes underwater. The dashboard shows two identical green numbers. Your bank account knows the difference.

ROAS also counts customers you already had

There is a second problem, and it is quieter. A lot of the revenue credited to paid was going to arrive anyway. Someone who typed your brand name into Google. A repeat buyer you retargeted three days after they had already decided to come back. The ad takes the credit, but the sale was not incremental. You paid to reach a customer who was walking through the door regardless.

Brand search and warm retargeting are the usual culprits. They post beautiful ROAS because they catch demand you already created, and they make a paid account look far healthier than it is. Strip them out and the real question appears. How much did the ads add that would not have happened otherwise?

A 4x ROAS on a 25% margin product is a way to lose money with a great-looking dashboard.

Report contribution instead

The number we put in front of clients is contribution: what is left after cost of goods and ad spend. It is the figure a finance director recognises, because it is the one that says whether the activity made or lost money. This is separate from the wasted spend already sitting in most accounts, which is its own audit.

Contribution is less flattering and far more honest. A campaign at 2.2x ROAS on a high-margin product can contribute more profit than one at 5x on a thin one. Once you are looking at contribution, you stop chasing the ratio and start making the decisions that change the P&L: money into the products and audiences that carry margin, and out of the ones that only looked good.

Where we would start

Get cost of goods into the reporting, per product, so ROAS stops being an abstraction. Separate prospecting from brand and retargeting, and judge each on what it adds rather than what it claims. Then set the target in contribution and let budget follow the profit.

Sometimes the honest read is that the highest-return campaign in the account is the one worth spending less on. That sits badly with a lot of agencies, because it means a smaller number on the report. It is also why we run paid inside one strategy with SEO, CRO and email rather than as a channel graded on its own vanity metric. If your ads look fine on ROAS and you are still not sure they are making money, that is worth a fit call.

Common questions

What is a good ROAS for ecommerce?

There is no single figure, because a good ROAS depends on your margin. A 2x can be profitable on a high-margin product and a 5x can lose money on a thin one. Work out your break-even ROAS from your margin first, then judge everything against that.

What is contribution margin?

What is left after you take cost of goods and ad spend off your revenue. It tells you whether the activity actually made money, which ROAS on its own cannot.

Is ROAS useless?

No. It is a fine efficiency check inside a single product or campaign. It just cannot compare across products with different margins, and it cannot tell you what was incremental.