When a high ROAS is still unprofitable

ROAS is one of the most misunderstood metrics in Amazon advertising. It measures how much ad-attributed revenue comes from one euro of ad budget. It does not measure how much of that is left after product costs, Amazon fees, logistics, returns and fixed costs. Exactly in this gap you get accounts that look good in the ad dashboard and still earn little or no profit at the end of the month.

The second value is the real yardstick. A ROAS of 6 is disastrously unprofitable at a 15% margin and comfortably profitable at a 45% margin. Without the break-even figure, a team evaluates a campaign in a vacuum.

Why "high ROAS" does not automatically mean "profitable"

Four patterns explain most cases in which a good ROAS hides a bad business.

1. The product margin does not carry the ad costs

A product with a low purchase price, high shipping costs or strong price competition often has a thin margin before advertising. Even a ROAS that looks good compared with the industry can then sit below break-even. The metric that counts is not ROAS compared with other campaigns, but ROAS compared with the product-specific break-even threshold.

2. Ads buy revenue that would have come organically anyway

Sponsored Brands and Sponsored Products campaigns on your own brand name often achieve a very high ROAS, which is to be expected because these buyers were searching for the brand anyway. The measured ROAS is real; the additional marginal revenue compared with the situation without an ad often is not. The account shows success, while the company pays for revenue that would have happened without the click.

A simple test: If the campaign is paused, total revenue barely drops while the ad costs disappear completely. That points to cannibalization rather than growth.

3. Returns, fees and costs after the purchase are not counted

Ad-attributed revenue in Amazon Ads refers to the order value at the time of purchase. Returns, refunds, goodwill gestures and chargebacks happen later and no longer show up in the ad report. For products with a high return rate, such as clothing, shoes or items that need technical explanation, a seemingly profitable ROAS can look much worse after returns.

4. Fixed costs and economies of scale are ignored

The break-even ROAS at product level takes variable costs into account. Rent, salaries, software, warehouse and other fixed costs are not included. An account can look marginally profitable at product level and still not generate enough contribution margin to cover the company's fixed costs. This level can only be judged with a full calculation, not with a single advertising metric.

The metrics that make the gap visible

• Break-even ROAS per product: calculated from selling price, purchase price, Amazon fees, shipping and packaging.
• TACoS: shows how much the overall business depends on advertising, independent of the ROAS of individual campaigns.
• New-to-brand vs. existing customer share: distinguishes additional from repeat revenue in brand campaigns.
• Return rate per ASIN: corrects gross revenue to a more realistic net value.
• Contribution margin after advertising: the actual steering metric, in euros rather than in percent or ratio.

Special case: evaluating brand campaigns correctly

Brand campaigns should not be switched off completely just because their marginal benefit is unclear. They defend the Buy Box against competitors on your own brand, support product launches and can promote cross-selling to other products of the brand. Rather than a blanket shutdown, it makes more sense to run a deliberately lower budget with a clear objective: defense instead of growth contribution, with correspondingly lower expectations for marginal revenue.

A practical approach for the next account review

1. Calculate the break-even ROAS for each main product, rather than setting only a blanket target value for the whole account.
2. Group campaigns by product and check the actual ROAS against the product-specific threshold, not against the account average.
3. Look at brand campaigns separately and check them against new-to-brand data for genuinely additional revenue.
4. Include the return rate per category, especially for new or little-tested products.
5. Report TACoS and contribution margin monthly alongside ROAS, so decisions do not rest on a single metric.

A worked example: ROAS 5 and still a loss

A ROAS of 5 sounds like a clear success: For every euro invested, five euros of revenue come back. Let us calculate what is actually left.

A shop spends €2,000 on advertising and generates €10,000 in revenue with it, a ROAS of 5. The products have a gross margin of 40%, so €4,000 in gross profit. After deducting the ad costs, €2,000 remain. Looks good.

Now add the items that do not show up in the ad account:

• Returns, 12% in the apparel space: €1,200 of revenue drops away, of which €480 is gross profit, and processing costs roughly €150 on top.
• Payment fees, 2%: €200.
• Shipping and packaging, not fully passed on: €600.
• Proportional logistics and warehouse costs: €350.

That turns €2,000 into roughly €220 of contribution margin, before fixed costs, before staff, before taxes. The campaign is not profitable; at best it is neutral. And this with a metric that appears in the dashboard as a clear success.

The decisive point: None of these items is hidden or unusual. They are simply not in the ad account, and that is why they do not appear in the evaluation of the campaign.

Measuring incrementality without a data science team

The hardest question in advertising is not how much revenue it generates, but how much revenue arises additionally, meaning would not have come without the ads. Ad platforms answer this question systematically too optimistically, because they credit every attributed sale to themselves.

This does not need an elaborate test design. Three pragmatic methods are enough for reliable indications:

1. The switch-off test. Pause a clearly delimited campaign, typically brand search terms, for two to three weeks and watch the total revenue of the affected product range, not the advertised revenue. If total revenue stays largely stable, the campaign was mostly not incremental. A period without special effects such as promotions or seasonal peaks is important.

2. The regional comparison. Where the platform allows it, switch off a campaign in one region and keep it running in a comparable one. The difference in revenue development between the two regions is a much better estimate of the actual advertising contribution than any attribution figure.

3. The look at the overall ratio. Compare total ad spend with total revenue, which on Amazon means TACoS. If this ratio rises over months while ROAS stays constant, the ads are increasingly buying revenue that would have happened anyway.

None of these methods delivers an exact value. All three deliver something more important: an order of magnitude for how far the platform figures deviate from economic reality. In our experience, the incremental share for brand search terms is much lower than the account reports, while for generic search terms it is often higher.

What a report that shows the gap should look like

Most ad reports start at ad costs and end at advertised revenue. A report that supports decisions goes further in both directions: one level up to total revenue and one level down to contribution margin.

In practice that means a table per product or product group with seven columns: total revenue, advertised revenue, organic share, ad costs, ROAS or ACoS, contribution margin before advertising and contribution margin after advertising. The last column, the contribution margin after advertising, is the only one that allows a statement about success, and it is missing from almost all standard reports.

Two additions raise the usefulness considerably. First, a comparison with the previous month and the same month last year, because absolute values say little without a trend. Second, a marker on the rows where the contribution margin after advertising is negative. These are the cases that need a decision right away, regardless of how good the ROAS looks.

The effort is manageable: the advertising data comes from the account, the cost structure from the calculation. Bringing the two together cleanly takes a few days and permanently changes the discussion about ad budgets, because from then on it is held in terms of profit instead of efficiency metrics.

How to change the conversation in your team

The hardest change is rarely a technical one. It means stopping talk about ROAS in meetings and talking about contribution margin instead, especially when targets, bonuses or agency contracts are tied to the old metric.

Check the incentives first. If an agency is measured on ROAS or ACoS, it will optimize that metric, which is rational and predictable, even at the expense of growth. To change that, you have to change the yardstick, not the expectation. A target range for contribution margin after advertising is just as verifiable as a target, but it steers toward the right thing.

Introduce the new number in parallel. Show both values side by side for two to three months. This builds trust in the new metric and shows the cases where the two diverge. Those cases are the instructive ones.

Define the number once, binding for everyone. "Contribution margin after advertising" must be documented: which costs are included, which are not, and which data cut-off is used? Without this, several versions of the same metric appear within a few weeks, and the discussion goes back to numbers instead of decisions.

Conclusion: ROAS is an advertising indicator, not a profit indicator

A high ROAS is a useful signal for advertising efficiency. But it does not answer whether a company actually earns more with this campaign. That answer only comes from comparing against the product-specific break-even threshold, adjusted for cannibalization, returns and fixed costs.

Once you set up this calculation cleanly, you can shift budgets to where advertising actually generates additional contribution margin, instead of to where the ROAS looks best.

Further sources

• Amazon Ads: Calculating profitability and break-even ROAS
• Amazon Ads: What ROAS is and how it is calculated
• Amazon Ads: Understanding new-to-brand metrics