ROAS is a revenue metric. Your business runs on margin.

Your ROAS target is 4x.

You have been hitting it. The platform confirms it. Your marketing team confirms it. The account is performing.

Here is the question that rarely gets asked: where did 4x come from?

If the answer is "that's what we were told" or "it's the industry benchmark" or "it's what we've always used," then you do not have a ROAS target. You have a number that feels like a target. The distinction matters because these two things produce identical-looking dashboards and very different P&Ls.

What ROAS actually measures

ROAS measures revenue attributed to ad spend divided by ad spend.

At ROAS 4x, every $1 you spend in advertising returns $4 in attributed revenue. That is all ROAS tells you. It says nothing about what that $4 of revenue costs to fulfill, what it returns after returns and refunds, how it moves relative to your overhead, or whether the customer acquired was worth acquiring at the economics required.

ROAS is a revenue ratio. It says nothing about margin, profit or the health of the business. It measures one thing: platform-attributed revenue relative to platform spend, and it measures that accurately. The trouble starts when it is treated as a proxy for something it does not measure.

The margin calculation ROAS ignores

The breakeven ROAS for any business is determined by one number: gross margin.

If your gross margin is 40%, you need every dollar in revenue to cover $0.60 in cost of goods before you have anything left over. For advertising spend to be profitable at the contribution level, your ROAS needs to exceed 1 / gross margin.

At 40% gross margin, that's 1 / 0.40 = 2.5x. At 40% margin, ROAS of 4x is profitable.

At 22% gross margin, that's 1 / 0.22 = 4.55x, about 4.5x. At 22% margin, ROAS of 4x is unprofitable on contribution. Every sale attributed to the ad is losing money before you pay for the ad team, the platform fees, the warehouse, or anything else.

This is not a marginal error. This is the direction of the business.

If you want the line for your own margin and return rate, it is worked out in one table in The ROAS your margin demands, in one table.

In the accounts I have audited, thin-margin e-commerce businesses were often running ROAS targets set against different economics: the industry benchmark, a standard set for another business, the number that happened to be in place when the account launched. The margin calculation was never run. The target was never validated. The account has been "performing" against a benchmark that may have been wrong from the start.

Why this gap persists

The ROAS target exists in the marketing dashboard. The gross margin exists in the P&L or the ERP. These two systems are managed by different people, updated on different cycles, and rarely put in the same conversation.

The marketing team knows ROAS. They may not know the current gross margin, and the number changes: product mix shifts, supplier costs move, promotional discounting compresses margin on specific SKUs while the blended average stays stable.

The finance function knows margin. They may not know ROAS by campaign type, and they are not evaluating the marketing account against that number directly.

The result is a gap that can persist for months or years without anyone identifying it. The campaigns are managed competently. The financials are tracked correctly. The connection between them, the calculation that would tell you whether the spend is commercially sound, never gets run.

What the correct calculation requires

To set a ROAS target that reflects the actual economics, you need four numbers:

Gross margin by product or category. Not blended average. The margin on the products being advertised, because ROAS and margin need to be matched at the same level of granularity. A 45% margin category and a 15% margin category sitting in the same campaign produce a blended ROAS that is accurate and useless for either one.

Target contribution margin from advertising. How much profit do you need the advertising to produce, net of ad spend, before overhead? This is a business decision, and no formula makes it for you. But once you have it, the required ROAS follows directly.

Return and refund rate. Platform attribution counts attributed revenue at the sale. Your business counts revenue after returns are processed. A 12% return rate on a 4x ROAS account is effectively a 3.5x ROAS on net revenue. The platform does not adjust for this.

Overhead allocation. Contribution margin covers cost of goods. Full profitability accounting includes platform fees, agency or team costs, and fixed overhead. How much of that overhead does advertising need to carry? This shapes whether the breakeven ROAS is the floor you're trying to beat or the ceiling you're trying to stay comfortably above.

With these four numbers, the ROAS target becomes an output of the economics. The number is defensible because it is derived from the business itself.

The practical consequence

A business running a ROAS target that is below its margin-adjusted breakeven is not underperforming. The dashboard says otherwise. The business looks healthy by every platform metric.

What is actually happening: the advertising is generating attributed revenue, the cost of goods on that revenue exceeds the revenue net of ad spend, and the P&L absorbs the difference. This can run invisibly for long periods because the P&L pressure gets attributed to rising costs, supplier issues, seasonality, or any number of other explanations that do not point directly at the marketing economics.

The fix requires connecting the ROAS target to the margin data: a calculation that takes about an hour with the right numbers and produces a target that the business can defend rather than one it inherited.

The harder version of the same problem

There is a more complex version of this problem, and it tends to surface as soon as the numbers are looked at properly.

The blended ROAS is not the only level at which the calculation can be misleading. In the accounts I have audited, the usual picture is a mix of campaign types with very different economics: branded search capturing existing demand at high ROAS, generic search at moderate ROAS, Performance Max at variable ROAS that was often propped up by branded attribution, prospecting campaigns at low ROAS that are justified on new customer acquisition grounds. Branded revenue is the weakest part of that picture: when eBay switched its brand search ads off in a large field experiment, almost all of that traffic arrived through the free listings instead (Blake, Nosko, & Tadelis, 2015).

When these are averaged, the blended number looks reasonable. When they are separated, the picture is often that one or two campaign types are profitable and the rest are below the margin breakeven.

Budget is being allocated to the blended performance, which means the unprofitable campaigns grow proportionally with the profitable ones.

This does not show up in a standard performance report. It requires building a margin-adjusted model by campaign type.

The ROAS on the dashboard is accurate. It is also incomplete. The complete version includes the margin adjustment, the return rate, the customer acquisition cost (CAC) at the campaign level, and the answer to whether what you're running can profitably sustain the scale you're planning.

References

Blake, T., Nosko, C., & Tadelis, S. (2015). Consumer heterogeneity and paid search effectiveness: A large-scale field experiment. Econometrica, 83(1), 155–174.