What ACoS measures, and why the lowest ACoS is the wrong goal
ACoS answers one question: for every dollar of revenue your ads produced, how many cents did the ads cost? A 25% ACoS means you spent 25¢ in advertising to collect $1 of advertised revenue. Amazon reports it on every Sponsored Products, Sponsored Brands and Sponsored Display campaign, and sellers on eBay, Walmart Connect and Etsy Ads meet the same idea under other names.
Because ACoS is a cost ratio, sellers instinctively chase a smaller number. That instinct is wrong, and it is the single most expensive mistake in marketplace advertising. Driving ACoS to 8% usually means you have bid down to exact-match branded keywords — the searches from shoppers who were already looking for you and would have bought anyway. The ads look brilliant and the business stops growing.
The useful comparison is not against other sellers or against last month. It is against your own break-even ACoS, which is simply your profit margin before advertising. Any ACoS below that number adds profit; any ACoS above it subtracts profit. A 42% ACoS on a product with a 55% margin is healthy and probably under-invested. A 22% ACoS on a product with an 18% margin is quietly destroying money on every order.
So this calculator returns four percentages rather than one: your ACoS is the fact, break-even is the ceiling, target is where you want to bid, and TACoS says whether the advertising is building the listing or renting orders from it.
The four formulas and how they fit together
ACoS = ad spend ÷ ad-attributed sales × 100. Both figures come from the same campaign report over the same date range. Ad-attributed sales are not your total sales — they only include orders Amazon credits to a click within the attribution window, which for Sponsored Products is 14 days from the click.
ROAS = 100 ÷ ACoS%. ROAS and ACoS are reciprocals of each other, so a 25% ACoS is a 4.0× ROAS and a 50% ACoS is a 2.0× ROAS. Neither contains more information than the other. Amazon's console leans on ACoS, Google and Meta lean on ROAS, and if you run both you should convert once and pick a single language for your reporting.
Break-even ACoS = pre-ad profit margin. This equality does the real work, and it is worth seeing why it holds. Suppose a product sells for P with a margin of m, so each sale contributes P·m in profit before advertising. If the ACoS is a, advertising costs P·a on that sale. Profit is P·m − P·a = P(m − a), which is zero exactly when a = m. The price cancels out, so the break-even ACoS depends on your margin percentage alone — not on how expensive the product is.
Target ACoS = break-even ACoS − desired net margin. Follow the same algebra: to leave a net margin of mnet after advertising you need a = m − mnet. On a 35% margin, a 10% net margin means bidding to a 25% ACoS.
TACoS = ad spend ÷ total sales × 100. The denominator changes to every order in the period, organic included. TACoS is not a metric Amazon reports — sellers coined it — but it captures something ACoS structurally cannot: whether your advertising is lifting the listing's organic rank. Falling TACoS with rising total sales is the signal that ads are compounding rather than substituting.
Worked example: a $28 kitchen gadget on a 35% pre-ad margin
Start by building the margin from the unit economics, because everything downstream depends on it.
- Price: $28.00 per unit.
- Amazon referral fee at 15%: 28.00 × 0.15 = $4.20.
- FBA fulfilment fee for this size tier: $5.00.
- Landed cost of goods including freight and duty: $9.00.
- Total pre-ad cost: 4.20 + 5.00 + 9.00 = $18.20.
- Pre-ad profit: 28.00 − 18.20 = $9.80 per unit.
- Pre-ad margin: 9.80 ÷ 28.00 = 0.35 = 35%. That is your break-even ACoS.
Now the campaign. Over 30 days you spent $840 and Amazon attributed 120 orders worth $3,360 to your ads. Business Reports show $8,400 of ordered product sales in total, so 300 units moved.
- ACoS: 840 ÷ 3,360 = 0.25 = 25.0%.
- ROAS: 100 ÷ 25 = 4.00×.
- Break-even ACoS: 35.0%. You are 10 points inside it.
- Target ACoS for a 10% net margin: 35 − 10 = 25.0%. You are exactly on target.
- TACoS: 840 ÷ 8,400 = 10.0%.
Check the per-unit story to prove it. On an advertised sale you collect $28.00, keep $9.80 before ads, and spend 25% × $28.00 = $7.00 on advertising, leaving $2.80 — which is 10% of $28.00, the margin you asked for.
Across the whole product line: gross profit is 8,400 × 0.35 = $2,940, minus $840 of ad spend leaves $2,100 of net profit, a 25% net margin. The line is more profitable than any single advertised sale because 180 of the 300 units — three-fifths of the volume — arrived organically and carried no ad cost at all.
How to read your ACoS against break-even
Put your three ACoS figures in order and the decision writes itself.
ACoS below target. You are clearing the margin you wanted with room to spare. Raise bids or widen match types until ACoS climbs toward target. Leaving a campaign at half its allowable ACoS is not prudence, it is forgone volume — and on Amazon, forgone volume also means forgone organic rank, because sales velocity feeds the ranking algorithm.
ACoS between target and break-even. Profitable, but thinner than you planned. This is the right zone for a product you are deliberately pushing for rank, or for a launch you expect to subsidise for a quarter.
ACoS above break-even. Every advertised order loses money. That is defensible for a launch, for a product whose repeat-purchase rate makes the first order a loss leader, or for defending a branded search term against competitors' ads. It is not defensible as a steady state on a one-off purchase.
Published category benchmarks are close to useless here. Compare against your break-even, never against a category average.
Read TACoS on a trend rather than as a level. A TACoS of 10% that drifts down to 7% while total sales rise means your ads are earning organic placement that then sells without further spend. A TACoS that climbs while total sales stay flat means you are paying more each month for the same orders — usually rising competition on your main keywords, and a signal to check search-term reports for waste before raising budgets. If you want to trace the same logic on a per-customer basis instead of per-order, run the numbers through the cost per acquisition calculator.
Break-even ACoS, target ACoS and equivalent ROAS by pre-ad margin
| Pre-ad margin | Break-even ACoS | Break-even ROAS | Target ACoS (10% net) | Target ROAS |
|---|---|---|---|---|
| 15% | 15.0% | 6.67× | 5.0% | 20.00× |
| 20% | 20.0% | 5.00× | 10.0% | 10.00× |
| 25% | 25.0% | 4.00× | 15.0% | 6.67× |
| 30% | 30.0% | 3.33× | 20.0% | 5.00× |
| 35% | 35.0% | 2.86× | 25.0% | 4.00× |
| 40% | 40.0% | 2.50× | 30.0% | 3.33× |
| 50% | 50.0% | 2.00× | 40.0% | 2.50× |
| 60% | 60.0% | 1.67× | 50.0% | 2.00× |
ROAS figures are 100 ÷ ACoS. Notice how brutal thin margins are: at a 15% margin you need a 20× ROAS to net 10%, which almost no paid channel delivers.
Attribution makes ACoS a lagging number
Sponsored Products credits an order to a click for 14 days after that click. Pull today's report and the spend is complete but the sales are not — orders from the last fortnight's clicks are still arriving. A campaign that looks like a 60% ACoS this morning can settle at 35% two weeks later.
Never judge a campaign on fewer than 14 days of closed data, and never re-bid on yesterday's ACoS. TACoS is looser still: its denominator comes from Business Reports and is dated by order rather than by click, so treat small TACoS movements as noise.
Mistakes that make an ACoS number lie to you
- Using gross margin as the pre-ad margin. The break-even ACoS needs margin after every variable cost of the sale: cost of goods, referral fee, fulfilment or FBA fee, storage, returns allowance and inbound freight. Leave out the FBA fee and you will overstate your allowable ACoS by ten points or more.
- Comparing ACoS across ad types. Sponsored Products, Sponsored Brands and Sponsored Display do not all report the same way, and Amazon Attribution for off-Amazon traffic is different again. Compare like with like.
- Judging a launch by steady-state rules. During a launch you buy rank and reviews as well as orders. Set an explicit subsidy budget and an end date.
- Ignoring the branded-search problem. Ads on your own brand name report a beautiful ACoS while mostly re-buying orders you already had. Segment branded and non-branded campaigns before you draw any conclusion about efficiency.
- Forgetting returns. Amazon's sales figure is gross of returns in most reports. A 12% return rate makes a 30% ACoS behave like roughly 34%, because the spend stays and the revenue goes back.
- Treating TACoS as a target. TACoS has no break-even of its own — it is diagnostic. Set targets on ACoS, watch the trend on TACoS.
- Averaging ACoS across products. An account-level ACoS blends a 60%-margin hero with a 12%-margin accessory and hides which one is losing money. Calculate per ASIN, or at least per margin band.
ACoS, ROAS, TACoS and CAC: which one to report
These four metrics answer different questions, and mature sellers track all of them.
ACoS or ROAS is your bidding metric. It sits at the keyword level, updates fast, and compares directly to a break-even you control. Use whichever direction your team thinks in and convert the other; the ROAS calculator handles the conversion and adds the gross-profit version.
TACoS is your channel-health metric. Report it monthly per product line alongside total sales to see whether advertising is buying durable rank or renting temporary orders.
CAC — customer acquisition cost — is the business metric marketplace sellers most often skip. ACoS divides by revenue; CAC divides by new customers. For a consumable, an ACoS well above break-even can be rational if the second and third orders arrive organically. If you also sell direct, the customer lifetime value calculator and the LTV to CAC ratio calculator give you the framework.
Marketing ROI restates the same data as a percentage return a finance team recognises — see the marketing ROI calculator. Before any of it, confirm your margin by rebuilding the unit economics in the Amazon FBA profit calculator; an ACoS target built on a wrong margin is worse than no target.