A 4.2x ROAS campaign can lose money on every order. Here is the arithmetic Meta does not run for you.
Meta reports 4.2x ROAS on a campaign. You spent $7,200 and Meta attributes $30,240 in revenue. By every number on the dashboard, that campaign is working, so you scale it.
Run the same campaign through your actual cost structure and it produced a contribution margin of negative $2,043. Every additional dollar you put into it makes the business slightly worse.
Both statements are true. ROAS is not a profitability metric and was never meant to be one. It is a revenue-to-spend ratio, and revenue is the number furthest from your bank account.
ROAS divides attributed revenue by ad spend. That is the whole formula. It contains no information about:
A brand with 30 percent COGS and a 5 percent return rate can be genuinely profitable at 2.5x ROAS. A brand with 49 percent COGS and a 41 percent return rate can be losing money at 4.2x. The ratio alone tells you nothing without the cost structure behind it, and the cost structure varies by SKU.
Crestline Co. ran $7,200 of Meta spend against the CL-7732 Trail Boot over thirty days. Here is what Meta saw, and what actually happened.
| What Meta reports | Amount |
|---|---|
| Ad spend | $7,200 |
| Attributed revenue, 168 orders at $180 | $30,240 |
| ROAS | 4.2x |
Now the same 168 orders, carried all the way through.
| Line | Amount | Why |
|---|---|---|
| Gross revenue | $30,240 | 168 orders at $180 |
| Less refunds | ($12,420) | 41% return rate, 69 units back |
| Net revenue | $17,820 | 99 orders actually kept |
| Less COGS on kept units | ($8,712) | 99 at $88 landed cost |
| Less outbound shipping | ($1,512) | All 168 orders shipped at $9 |
| Less return shipping | ($1,242) | 69 at $18 round trip |
| Less restocking labor | ($290) | 69 at $4.20 per unit |
| Less payment processing | ($907) | ~3% charged on gross, not net |
| Contribution before ad spend | $5,157 | |
| Less ad spend | ($7,200) | The 4.2x campaign |
| Contribution margin | ($2,043) | About $12.16 lost per order |
Crestline Co. is a simulated DTC brand used for demonstration. These figures are illustrative and are not a client result. Our interactive demo cites roughly $12.40 per order on marginally different COGS assumptions.
Notice which line does the damage. It is not the ad spend. Contribution before ad spend was $5,157 on $30,240 of gross revenue, a 17 percent margin, and the 41 percent return rate is why. Two thirds of the gross revenue never stayed.
Notice also that returns are charged three times: you refund the sale, you eat the return shipping, and you pay someone to put it back on the shelf. Only the first of those appears anywhere near a marketing dashboard.
A 41 percent return rate on one SKU when the brand average is 22.4 percent is not a marketing signal. It is a product signal that happens to be showing up in your ad account.
For CL-7732, 74 percent of return reasons said some version of "runs small." The fix was a sizing guide update, which had been flagged to the product team eight weeks earlier and was still in a backlog. In the meantime the ad account kept spending, because the ad account could see a 4.2x ROAS and nothing else.
This is the shape of the problem worth internalizing. The support tickets knew. The returns data knew. The product backlog knew. The ad account, which controlled the spending, knew none of it, and none of those four systems talk to each other.
Pausing the campaign is the smaller half of the fix. It stops the bleeding at roughly $7,200 a month. Fixing the sizing guide is what lets the SKU sell profitably again. Doing only the first turns a product problem into a permanently dead SKU.
Compute contribution margin per SKU, then per campaign. The formula that matters:
Net revenue after returns, minus COGS on units kept, minus all shipping, minus restocking, minus payment fees, minus ad spend.
From there you can derive your real break-even ROAS, which is specific to each SKU rather than a single company-wide target. For CL-7732 at a 41 percent return rate, break-even sits near 5.9x, which no boot campaign is going to hit. At the brand's normal 22.4 percent return rate, the same SKU breaks even around 3.4x and the 4.2x campaign would have been genuinely profitable.
That is the entire story: the campaign was not badly targeted, it was selling a product with a defect nobody had connected to the spend.
The reason this persists is organizational, not analytical. Ads, returns, support and product each hold one piece, and no single team can see the whole shape. Postlytix reads all of them at once, calculates true contribution margin per SKU, and surfaces the ones where spend is outrunning margin, along with the root cause and a pause plan you approve before anything runs.
This is one of eight places the numbers come apart. Why your e-commerce revenue numbers never match explains why fragmentation happens across the whole stack, and links to the rest.