Blog/ACoS and Returns: how Merch sellers read ad results after refunds

ACoS and Returns: how Merch sellers read ad results after refunds

P
PPC Optimizer Pro Team
•November 16, 2026•9 min read
ACoS and Returns: how Merch sellers read ad results after refunds

ACoS and returns live in two different reports, and that is the whole problem. Your ad console counts an order within days of the click and keeps showing it as an ad sale. If the shirt comes back, the return lands weeks later in your Merch sales report, where it takes the royalty away.

So an ACoS that looks safely under break-even can still lose money once returns come in. The fix is simple: lower your break-even ACoS by each design's return rate before you compare.

For a royalty seller this matters more than for a brand. A brand with a 40% margin can absorb a few returns. A Merch shirt paying a few dollars of royalty has very little room, and one return out of six orders can turn a profitable campaign into a losing one.

This guide covers why the two reports disagree, the break-even ACoS after returns at several return rates, a worked example, and a weekly check you can do in a spreadsheet.

Why your ad console and your royalty disagree

Amazon's own ACoS guide defines ACoS as ad spend divided by ad revenue, times 100. Ad revenue is the price the buyer paid for orders that followed a click on your ad.

Three things separate that number from what you actually keep.

The ad console counts sales, not royalty. A $19.99 shirt counts as $19.99 of ad revenue. Your royalty on it might be a few dollars. That is why a Merch seller's break-even ACoS is so low; Break Even ACoS for Merch by Amazon has the table by price. The order is counted fast, the return comes slowly. The console credits an order that follows a click inside its attribution window; the column names in your report (for example "7 Day Total Sales") tell you which window it uses. A buyer who returns a shirt usually does it after it arrives, tries it on and sends it back, which is weeks after the click. The return shows up in your Merch sales report. That report lists returned units per design, and the royalty you keep is net of them. Amazon's public ACoS guide says nothing about removing returned orders from ad revenue, so don't count on the console to correct itself. Judge profit from your royalty, and use ACoS only as a quick signal.

The 60 days of Merch returns data on this blog is one account's real numbers: 149 of 1,645 units returned, 9.1%, with most designs never returned and a few returned at 20% or more.

Break-even ACoS after returns

Without returns, break-even ACoS is royalty divided by price. With returns, only the units that stay sold pay you, so:

Break-even ACoS after returns = royalty x (1 - return rate) / price

Here it is for two worked examples. The royalties are examples, not Merch's rates; use what your own design pays at its price.

Return rate$16.99 shirt, $2.02 royalty$19.99 shirt, $4.55 royalty
0%11.9%22.8%
5%11.3%21.6%
9.1% (the data above)10.8%20.7%
15%10.1%19.3%
25%8.9%17.1%

The drop looks small in percentage points, but it is a large share of a thin margin. On the $19.99 shirt, a 9.1% return rate takes 2.1 points off a 22.8% ceiling. If you were happy running at 21%, you are now under water.

The break-even ACoS calculator does the base number for any price and royalty; multiply its royalty input by (1 - return rate) to get the version after returns.

A worked example: one design, two readings

Say one $19.99 design with a $4.55 royalty ran for 30 days at a $0.20 bid. These are made-up numbers to show the method.

What the ad console showsAfter returns come in
Clicks120120
Spend$24.00$24.00
Ad orders66 (1 returned later)
Ad sales$119.94$119.94 (unchanged in the console)
ACoS20.0%20.0%
Royalty earned$27.30 (6 x $4.55)$22.75 (5 x $4.55)
Profit after ad spend+$3.30-$1.25
Break-even ACoS22.8%19.0%

In the console the campaign looks fine: 20% is under the 22.8% ceiling. Three weeks later one shirt comes back, and the same 20% ACoS is now above a 19.0% ceiling. Nothing changed in the ad report. The campaign was losing $1.25, not earning $3.30.

The bid tells the same story. A bid ceiling is royalty times conversion rate. At a 5% conversion rate that is $4.55 x 5% = $0.23 before returns, and $3.79 x 5% = $0.19 after this design's returns. The $0.20 bid sat between the two.

One caution: 1 return in 6 orders is a 17% rate, but 6 orders is far too few to know a design's real return rate. That is the next section.

Which return rate to use for each design

A per-design return rate is the right idea, because returns are uneven: in the 60-day data, 387 of 474 selling designs had no return at all. But a rate measured on a handful of units is mostly noise.

A practical rule:

  • 20 or more units sold in 60 days: use the design's own return rate.
  • Fewer than 20 units: use your account's 60-day return rate. One return in 5 units is 20%, and that says more about luck than about the design.
  • A design you know has a sizing or color problem: use its own rate even on fewer units, and fix the listing first.

Use a 60-day window, not 14 days. Returns arrive weeks after the sale, so a 14-day report counts returns from older sales against only two weeks of new ones. In the returns data above, the 14-day rate read 11.3% while the 60-day rate was 9.1%.

Seasonal gift designs deserve a second look in January. Gifts are bought weeks before they are worn, so their returns can arrive well after the holiday ads have stopped. Judge a Christmas design's return rate once January's returns are in, not in mid-December.

A weekly check in a spreadsheet

You don't need software for this; a spreadsheet and 20 minutes a week will do.

  • Download your Merch sales report for the last 60 days. Note units sold and units returned for each design.
  • Work out each design's return rate, using the 20-unit rule above to choose between its own rate and the account rate.
  • Compute the royalty after returns: royalty x (1 - return rate).
  • Compute break-even ACoS after returns: that royalty divided by the price.
  • Download the ad report (search term, targeting or a bulk file) for the same designs and compare each one's ACoS with its own ceiling, not one account-wide target.
  • Set the bid ceiling from the royalty after returns times the conversion rate. Amazon PPC Bid Strategy for Merch shows how to pick a cautious conversion rate when the design has few orders; the Merch max bid calculator does the sum.
  • Don't cut on returns you haven't seen yet. A campaign that just sold well will look better this week than it will in a month. Re-check it after the returns window, before a big bid raise.
  • If a design's return rate is well above your account's, look at the listing before the ads: size chart, color in the mockup, and the reviews often explain it. Ads only bring more buyers to a listing that sends shirts back.

    How PPC Optimizer Pro handles this

    The Optimizer Lab in PPC Optimizer Pro reads your 60-, 30- and 14-day bulk files together with your Merch sales report. It works out every design's royalty with its returns subtracted, and judges each bid, placement and pause against that, not against list-price ACoS.

    Every suggested change shows the rule that fired and the numbers behind it. You can untick or edit any row, then download one Amazon-ready file to upload yourself. There is a 7-day free trial, and the free tools work without an account.

    FAQ

    Does Amazon remove returned orders from ACoS?

    Amazon's public ACoS guide only defines ACoS as ad spend divided by ad revenue and says nothing about returns. Don't rely on the ad console to correct for them. Measure returns from your Merch sales report and lower your break-even ACoS to match.

    What is a normal return rate for Merch shirts?

    There is no official benchmark. One US Merch account on this blog measured 9.1% across 1,645 units over 60 days, with most designs never returned and a few above 20%. Use your own 60-day numbers, not someone else's average.

    Should I stop advertising a design with a high return rate?

    If its break-even ACoS after returns is below anything the ads can reach, yes. Keep it listed: organic sales cost nothing in ad spend. Fix the listing first, then test ads again once the return rate comes down.

    Does TACoS solve the returns problem?

    No. TACoS divides ad spend by total sales, so it shows how much the ads cost against the whole business, but it still uses sales, not royalty after returns. Amazon TACoS explains when it is useful.

    How long should I wait before judging a campaign's profit?

    Long enough for most returns to arrive: a few weeks after the orders. Use 60-day windows for return rates and profit, and the 14-day view only for recent movement in clicks and bids.

    Keep reading

    P

    Written by

    PPC Optimizer Pro Team

    The PPC Optimizer Pro Team consists of Amazon sellers and developers who built this tool after years of managing Sponsored Products campaigns manually. We share data-driven strategies to help sellers reduce wasted ad spend and improve ACOS.