ACoS vs TACoS vs ROAS on Amazon: Formulas, Benchmarks and Which One to Track (2026)

Three metrics describe the same ad spend from three angles, and sellers argue about which one matters. ACoS tells you whether the ads pay for themselves on the orders they touch. ROAS is the same number inverted, and is what Amazon's own dashboards and budget rules lean on. TACoS tells you whether the ads are moving the whole business or just replacing sales you would have got anyway. This guide gives the formulas, a worked rupee example, where each metric misleads, how Flipkart's ROI maps onto Amazon's ACoS, and what 2026 benchmarks say, with their sources. For the mechanics of campaigns themselves, start with our Amazon ads guide for Indian sellers.

Formulas are Amazon's, from its advertising library. Benchmarks are from third-party 2026 reports and are attributed; Amazon publishes no averages, and its ACoS guide says "there isn't a definitive number for a good Amazon ACOS". Checked 25 September 2026.

What are the formulas?

ACoS and ROAS are exact inverses: ROAS = 100 ÷ ACoS when ACoS is a percentage. A 25% ACoS is a 4x ROAS; 20% is 5x; 50% is 2x; 33.3% is 3x. Anyone who says ROAS is "better" than ACoS is talking about which direction feels natural, not about information.

Worked example: a ₹1,299 product over one month

Assumptions: a personal-care brand, selling price ₹1,299, product cost ₹380, Amazon fees (referral, closing, shipping and GST on fees) ₹310, so contribution before ads is ₹609, or 46.9% of price. That is the break-even ACoS. In the month: ₹42,000 ad spend, 2,800 clicks at a ₹15 average CPC, 196 ad-attributed orders (7% conversion), and 310 organic orders.

Now the same account three months later. Spend has risen to ₹70,000, ad orders to 300 (ACoS 18.0%), and organic orders to 520. Total sales are ₹10,65,180 and TACoS is 6.6%. ACoS got slightly worse; TACoS is flat; the business is 62% bigger. That is a healthy account, and a seller watching only ACoS would have been tempted to cut spend.

When does ACoS mislead?

• It ignores organic. A campaign on your own brand name can show 5% ACoS while cannibalising orders that would have arrived free. A launch campaign at 50% ACoS may be building the rank that makes next quarter's organic sales.
• It ignores margin. A 30% ACoS on a 25% margin product loses money; on a 50% margin product it is fine. Compare each ASIN's ACoS to its own break-even, worked out from our Amazon seller fees guide.
• It lags. Sponsored Products attributes orders for up to 14 days after the click, so this week's ACoS always looks worse than it will in a fortnight. Read it on a trailing 30-day window.
• It counts the wrong revenue. Ad revenue in the console is gross of returns and cancellations. In apparel, where returns can run high, true ACoS on net sales is materially worse than the console shows.
• It flatters halo sales. An ad for one ASIN that produces an order for another ASIN in your catalogue is counted as ad revenue, which is fair, but it means a single ASIN's ACoS is not that ASIN's economics.

When does ROAS mislead?

ROAS is ACoS in different clothes, so it shares every weakness above. It adds one of its own: it is the metric Amazon uses in performance-based budget rules ("increase budget by 20% when ROAS reaches 5", in Amazon's example), and a ROAS rule fires on gross attributed revenue including halo and pre-return sales. A rule that raises budget on a 5x ROAS can raise it on a campaign that is unprofitable after returns and margin. Set ROAS thresholds from break-even (100 ÷ break-even ACoS), not from round numbers. Our guide to budget rules and bidding covers this.

When does TACoS mislead?

• It hides a bad campaign inside good organic. A brand with strong organic sales can run a loss-making campaign for a year with TACoS at 4%. Read it with ACoS, not instead of it.
• It moves with seasonality. TACoS falls in a festival month because organic sales jump; it does not mean ads got better.
• It needs clean total sales. Use ordered product sales from the business report for the same date range as ad spend. Mixing a 14-day attributed ad figure with a calendar-month sales figure produces nonsense.

How does Flipkart's ROI compare?

Flipkart Ads does not report ACoS. Its headline metric is ROI, defined as revenue from ads divided by ad spend, which is the same thing as Amazon's ROAS. So a Flipkart ROI of 4 is an Amazon ROAS of 4 and an ACoS of 25%; ROI = 1 ÷ ACoS, and ACoS = 1 ÷ ROI. A brand running both marketplaces should convert to one metric so the two ad accounts can be compared against one margin. We use ACoS on both because it reads directly against margin. See our Flipkart ads guide for how PLA and PCA campaigns report.

What are the 2026 benchmarks, and how much should you trust them?

Amazon publishes none. Third-party 2026 reports give a consistent shape. Autron's benchmark report, compiled from industry data and its own managed accounts (sample size not disclosed, marketplace not specified), puts the cross-category median ACoS at 34%, with most accounts between 25% and 40%; category medians run from 23% for food and grocery, through 30% for electronics and 31% for beauty, to 57% for clothing, which it attributes to return rates. For TACoS, agency commentary generally treats under 10% as healthy for an established brand and 15% to 25% as normal during launch; that is practitioner consensus, not a measured dataset. Indian CPCs are lower than US ones, and Indian conversion rates in many categories are also lower, so the ACoS outcome is similar; but no published India-specific ACoS study exists as of September 2026 that we would cite.

Which metric should you track, and how often?

• Weekly: ACoS by campaign and search term, on a trailing 30-day window, against each ASIN's break-even. This drives negatives and bid changes; our guide to reducing ACoS is the routine.
• Monthly: TACoS for the account and for each product line, alongside organic rank on the top ten keywords. This decides whether to raise or cut total spend.
• Quarterly: profit after ads per ASIN. This decides which products to keep advertising at all.
• ROAS: only where the console forces it on you, such as budget rules; convert it in your head.

If you would rather someone else built the spreadsheet, our marketplace ads management reporting shows ACoS and TACoS per ASIN against break-even every month.

FAQ

Is ACoS or ROAS better? They carry identical information; ROAS = 100 ÷ ACoS. Use ACoS because it compares directly with margin: a 30% ACoS against a 40% margin is an instant read. A 3.3x ROAS needs a second step.

What is a good TACoS on Amazon? Amazon publishes no figure. Practitioner consensus treats under 10% as healthy for an established product and 15% to 25% as acceptable during a launch. The direction matters more than the level: flat or falling TACoS while sales grow is the goal.

Why is my ACoS in the console different from my own calculation? Attribution windows, halo sales on other ASINs, and gross versus net of returns. The console uses attributed gross sales within the window; your accounts use net shipped sales by calendar month.

Can TACoS be lower than ACoS? Always, unless you have zero organic sales. Total sales are at least as large as ad sales, so the same spend divided by a larger number gives a smaller percentage.

How do I compare Flipkart and Amazon ad performance? Convert Flipkart ROI to ACoS (ACoS = 1 ÷ ROI) and compare each against that marketplace's margin after fees, which differ because the fee structures differ.

Does Amazon show TACoS anywhere? Not as a standard column in the ads console as of September 2026. Calculate it from ad spend and total ordered product sales for the same period.

Reading the metrics the right way

ACoS for the campaign, TACoS for the business, break-even for the target. If you want those three tracked per ASIN, with the negatives and bid changes that follow from them, see our marketplace ads management service.