ACoS vs Profit: Why Our Best ACoS Period Made the Least Money
TL;DR
Three periods of our own Amazon ads: the one with the lower ACoS made a third less profit per day. How to judge an ad by profit, with the formulas and real numbers.
✨ Written with AI assistance from our own settlement and advertising data, and reviewed by Avanta Global EOOD before publication. Learn more
Short answer: ACoS tells you what your ads cost against the sales Amazon credits to them. It does not tell you whether you made money. On our own listing, two periods at the same price had ACoS of 25.8% and 30.8%. The one with the worse ACoS made A$32 a day; the one with the better ACoS made A$22.
To judge an ad you need two more numbers: your break-even ACoS, which comes from the price and fees, and the profit left after all ad spend, which also counts the sales ads did not get credit for.
Three periods from one listing
We sell one product on Amazon Australia, a bubble mailer bundle. Between August and September 2026 its price moved from A$34.99, to a deal at A$29.74, and back. We split the period by the price buyers actually paid, and joined three sources: the settlement report for price and fees, the advertising report for spend and ad sales, and our product cost of A$7.40 per unit.
| A: before the deal | B: deal | C: after the deal | |
|---|---|---|---|
| Dates | 12–29 Aug | 30 Aug – 10 Sep | 11–28 Sep |
| Units sold | 98 | 144 | 94 |
| Price actually paid (per unit) | A$33.88 | A$29.70 | A$34.67 |
| Contribution before ads (per unit) | A$11.22 | A$7.58 | A$11.91 |
| Break-even ACoS | 33.1% | 25.5% | 34.4% |
| ACoS | 25.8% | 23.5% | 30.8% |
| TACoS (ad spend ÷ all sales) | 21.4% | 8.7% | 16.5% |
| Profit after all ad spend, per day | A$21.6 | A$59.8 | A$32.3 |
Contribution = price paid − referral fee − FBA fee − product cost. It excludes storage, refunds and the monthly subscription, which do not depend on the ad.
What ACoS got wrong
A vs C is the clean comparison: nearly the same price, nearly the same number of days. By ACoS, A is five points better. By profit, C made half as much again per day.
The gap is in what ACoS leaves out. ACoS divides ad spend by ad-attributed sales only. In period A we spent A$7.26 on ads for every unit we sold in total; in C, A$5.74. More of C's sales came without an ad click, and ACoS cannot see those sales. TACoS can: 21.4% in A against 16.5% in C.
B shows the other blind spot. The deal cut the price by 15%, and the contribution per unit fell by a third, from about A$11.20 to A$7.58. That dropped the break-even ACoS from 33% to 25.5%. The ACoS we paid barely moved, so the ads went from seven points of headroom to two. The chart looked fine, but the margin for error had nearly gone. The deal still earned more per day, because it sold 12 units a day instead of 5, most of them without an ad click.
One large order helps period B: a single buyer took 25 units at the deal price. Without that order, B still made about A$48 a day, the best of the three.
The three numbers to use instead
1. Break-even ACoS: the ACoS at which an ad-attributed sale makes exactly zero.
Break-even ACoS = contribution before ads ÷ price paid
It belongs to the price, not to the account. Change the price, run a deal, or absorb a fee increase, and the bar moves the same day. Our referral fee rose from about 5.6% to 13% in July (what that did to our numbers), and our break-even ACoS dropped by seven points without a single ad setting changing.
2. Headroom: break-even ACoS minus actual ACoS. It shows how far a campaign can drift before each ad sale loses money. Seven points in A, two in B, three and a half in C.
3. Profit after all ad spend: the number that decides.
Profit = units sold × contribution per unit − total ad spend
This counts every unit, including the organic ones ads may have helped rank, and charges every dollar of ad spend. If ACoS and this number disagree, this number wins.
What we got wrong first
In our own notes on 31 August we recorded that the deal had pushed our ads "4.3 points past break-even". Those figures were taken while the deal was still running, and before the settlement had caught up with the orders. With the full window, the deal ran two points inside break-even and was our most profitable stretch of the three.
Two rules came out of that. Judge a period after its orders have settled, not while it is running. And split by the price buyers paid, not by the calendar: our deal price shows up in the settlement from 30 August to 10 September, not on the dates we had written down for it.
How to check this on your own account
- From the settlement report, take units, item price after promotions, referral fee and FBA fee per SKU (our line-by-line settlement guide shows which lines).
- Subtract your product cost per unit to get contribution before ads.
- From the advertising console, take spend and ad sales for the same dates and SKU.
- Compute break-even ACoS, ACoS, TACoS and profit per day for each period, and compare the last one.
The break-even ACoS calculator and the profit and TACoS calculator do the arithmetic in your browser. AgentXray runs the same calculation continuously from your settlement and ads data, and prices every bid or budget draft against the break-even of the SKU it touches.
FAQ
What is a good ACoS on Amazon? There is no universal good ACoS. A good ACoS is one below your break-even ACoS for that SKU at its current price. For us that was 33% at full price and 25.5% during the deal.
Is TACoS better than ACoS? TACoS shows the whole business: ad spend against all sales. ACoS shows the ads alone. Use ACoS against break-even for each campaign, and TACoS with profit per day to judge the listing.
Can a campaign with a high ACoS be profitable? Yes, if it stays under break-even. It can also pay off above break-even while it builds rank that later brings organic sales. That is a deliberate bet, and profit per day over the following weeks is how you tell whether it paid.
About this article
Figures come from Avanta Global EOOD's own Amazon.com.au data for one SKU, 12 August – 28 September 2026: settlement line items for price and fees, Sponsored Products reports for spend and ad sales, and a product cost of A$7.40 per unit. Orders are assigned to a period by their first settlement date. Measured on 30 September 2026. Currency is Australian dollars. The periods are consecutive, not a controlled experiment; demand also changed between them.
About this article
This article was researched and drafted with AI assistance. Before publication, it passed automated editorial review against Avanta Global EOOD's published editorial standards (factual accuracy, source attribution, voice & readability). Our editorial standards page documents exactly what we check. We continuously monitor published content for accuracy and update articles when new information emerges. Learn more about our editorial process and the team behind AgentXray.