ROAS vs Profit: Why a High ROAS Can Still Lose Money

ROAS is the most widely quoted number in ecommerce advertising and one of the least informative about whether the advertising is working. It is not a bad metric. It is a metric that answers a narrow question — how much attributed revenue came back per dollar spent — and gets used to answer a much wider one.

This guide covers what ROAS measures, the four things it structurally cannot see, and how to judge a campaign on the money it actually leaves behind.

What ROAS measures

Both sides of that fraction deserve scrutiny. The denominator is reliable: you know what you spent. The numerator is an estimate produced by the advertising platform, using its own attribution window and its own rules about which conversions to claim.

That has a consequence people underrate: attributed revenue is not additive across platforms. If two platforms each claim a $60 order because the customer saw both, and you sum their reported revenue, you have $120 of revenue from a $60 order. Your bank account disagrees.

MER is a different question

MER uses your actual revenue and your actual total spend, so it has no attribution problem and no double counting. What it cannot do is tell you which campaign caused anything. ROAS is per-campaign and estimated; MER is business-wide and real. Use both, and never label one as the other.

The four things ROAS cannot see

1. What the goods cost

ROAS compares revenue to ad spend with no reference to product cost. A campaign at 3x ROAS on a product with a 40% gross margin and the same campaign on a product with a 70% gross margin are not comparable events, and the ratio is identical in both cases.

2. What came back

Attribution is recorded at purchase. Returns happen afterward, and generally do not flow back into the reported ROAS. A campaign selling a product with a 25% return rate is reporting revenue that partly does not exist, and the cost of shipping those units out and back is real.

3. What it costs to run the business

Fulfillment, payment processing, support, software and salaries are invisible to ROAS. They are not invisible to your bank balance.

4. Whether the revenue was incremental

A campaign that retargets people already on their way to buying will report excellent ROAS. Whether it caused any additional purchases is a different question, and the platform reporting the ROAS is not a neutral party to it.

The number that does see them: contribution after ads

This is gross profit, not revenue, minus the advertising that produced it. It answers the question ROAS is usually being asked to answer: after paying for the goods and paying for the advertising, is there money left?

It is a product-level contribution figure, not company net profit — it does not carry fulfillment, salaries, software or any other operating expense. A campaign can produce healthy contribution after ads and the business can still lose money, if overhead exceeds the total contribution of everything it sells. Treating contribution after ads as profit is how businesses conclude they can afford overhead they cannot. See contribution margin for where it sits relative to gross profit and net profit.

Break-even ROAS depends entirely on your margin

There is no universal "good ROAS", and quoting one without a margin attached is meaningless. What does exist is a break-even point you can calculate for your own products.

At a 50% gross margin, a 2x ROAS means the gross profit exactly equals the ad spend: contribution after ads is zero, and everything else the business costs is unfunded. At a 25% gross margin you need 4x just to reach that same zero.

Note what the table implies: a 3x ROAS is comfortably profitable at a 70% margin and loses money at a 25% margin. The same number, opposite conclusions. This is why "what's a good ROAS?" has no answer and "what's my break-even ROAS on this product?" has a precise one.

Two campaigns, same store

Both campaigns spend real money in the same month. Illustrative figures.

Campaign A has the ROAS you would put in a report. It produced $483 of contribution on $10,000 of spend — before a single dollar of fulfillment, processing or overhead. Campaign B, at two-thirds of A's ROAS, produced fourteen times the contribution.

If the store scaled on ROAS it would put more money into A. If it scaled on contribution it would put more into B. Only one of those decisions makes the business more money, and the metric on the dashboard decides which one gets made.

What actually differed

Nothing about the advertising. A sells a low-margin product with a high return rate; B sells a high-margin product that customers keep. The campaigns are being judged on a metric that cannot see either fact, which means the metric is measuring the product and reporting it as advertising performance.

How to use ROAS well

ROAS is not useless — it is fast, available in real time, and comparable within a product at a fixed margin. Reasonable practice:

• Compare ROAS only across campaigns selling similar-margin products. Across different margins it is comparing different questions.
• Calculate your break-even ROAS per product and treat it as the floor, not a target.
• Use ROAS for fast in-flight signals — a campaign collapsing from 4x to 1.5x overnight is worth knowing about immediately. Use contribution for decisions about where budget goes.
• Use MER as the sanity check. If campaign ROAS is rising while MER is flat, the platforms are claiming more credit, not producing more sales.
• Watch returns by campaign, not just overall. A campaign can have a return rate well above the store average and its ROAS will never say so.

How ORVX reports this

ORVX keeps attributed revenue and realized revenue structurally separate. Attributed ROAS is reported as an advertising metric; MER is reported against real business revenue; and neither is allowed into the profit calculation, because platform-reported revenue is not money the business received.

Ads Intelligence reports contribution after ads alongside ROAS, so the comparison in the table above is the default view rather than something you assemble by hand. Where ad spend cannot be linked to specific products, it is shown as unlinked rather than spread across products — see product profitability for why that distinction matters more than it sounds.

Frequently asked questions