Amazon ACoS: What a Good Target Is and How to Lower It
Amazon ACoS is the number most ad accounts get judged on, and it is also the number most accounts set once and never revisit. That combination is expensive in both directions.
A target left too loose keeps funding clicks that were never going to convert, and with Amazon seller fees compressing contribution margin, that gap comes straight out of profit. A target pulled too tight does something quieter and worse. It retires the terms that were holding organic rank, and you find out six weeks later when organic sales slide and nobody can point at what changed.
ACoS is a constraint, not a strategy. It tells the bidding system what you are willing to pay for an ad-attributed dollar. It does not know which SKU is launching, which one is defending margin, and which query is carrying rank for a whole subcategory. That judgment is yours, and it has to be made per job rather than per account.
This guide covers what ACoS measures and what it leaves out, what a good ACoS looks like by goal and by category, what Amazon will and will not tell you about your peers, and how to reduce ACoS on Amazon without cutting the spend that was doing the work.
Want your break-even and target in seconds? Use the free Amazon ACoS calculator.
Quick Answer
There is no single good Amazon ACoS. A workable target is your break-even ACoS, which equals your contribution margin before ad cost, minus the profit you intend to keep, then adjusted for the job that SKU is doing right now. Mature SKUs defending margin usually run tightest. Launches and rank plays run above break-even on purpose, for a defined window. To reduce ACoS, cut the queries that are wrong for the product, lower CPC on placements that convert poorly, and raise conversion rate before you touch bids. Blanket bid cuts lower ACoS on the report and remove ranking volume in the account, so put a floor under the terms you cannot afford to lose.
What ACoS measures, and what it leaves out
ACoS, or Advertising Cost of Sale, is ad spend divided by ad-attributed revenue, expressed as a percentage. Amazon reports it on Sponsored Products, Sponsored Brands, and Sponsored Display, and it is the efficiency metric the console defaults to.
The formula and a worked example
ACoS = (ad spend / ad-attributed revenue) x 100
Spend $60 on a campaign that returns $200 in attributed sales and you are at 30%. You paid $0.30 for every $1.00 the campaign earned.
The three costs ACoS does not see
ACoS counts one cost and one revenue line. It ignores everything else in the cost stack, which is why an account can hit its target and still lose money:
- Cost of goods, fulfillment, and referral fees. A 25% ACoS on a product with a 22% contribution margin is a loss.
- Organic sales. ACoS only sees ad-attributed revenue, so a campaign lifting organic volume looks worse than it is. TACoS is the metric that catches this.
- Inventory position. Efficient spend on a SKU that runs out in eleven days buys you a stockout and a rank reset.
If ACoS is the only number in your weekly review, add contribution margin and days of cover next to it. Our rundown of the Amazon PPC metrics worth tracking covers the rest of the set.
What a good ACoS is depends on the job the SKU is doing
The question "what is a good ACoS" has an answer, and it is not a number. It is a sequence: find the ceiling, subtract the profit you want, then adjust for the job.
Break-even ACoS is the ceiling, not the target
Break-even ACoS equals your contribution margin before advertising. At that point every ad dollar is exactly paid for by the gross profit on the sale it produced.
Take a $150 product with $40 in COGS, $20 in Amazon fees, and $30 in inbound and outbound shipping. Contribution is $60, so margin is 40% and break-even ACoS is 40%. Run above 40% and that unit loses money on an ad-attributed basis. Run below it and the difference is yours.
Break-even is a ceiling. It is not a goal, because a campaign sitting at break-even contributes nothing.
Target ACoS is break-even minus the profit you intend to keep
Decide the margin you want after ad cost, then subtract. On the same product, keeping 15 points of profit gives you a 25% target ACoS. Keeping 25 points gives you 15%.
Two practical notes. Break-even is per product, so grouping SKUs with similar margins into the same campaigns keeps the target meaningful instead of averaging two different economics into one number. And a target set on a blended catalog margin will systematically overspend on your thin-margin SKUs and underspend on your best ones.
Targets by job, not by account
The same target applied to a launch SKU and a mature hero SKU under-serves both. Set it against what the SKU is being asked to do this quarter.
| The job | What you are buying | How the target behaves | The signal it is working |
|---|---|---|---|
| Launching a new ASIN | Keyword rank and conversion data | Loosest of the five, deliberately above break-even, for a window you set in advance | Organic rank climbing while TACoS flattens |
| Scaling a proven ASIN | Incremental volume at a known margin | Just under break-even, held steady rather than driven down | Sales growing while ACoS stays flat, not while it falls |
| Defending margin on a mature ASIN | Efficiency on demand you already own | Tightest, set from margin minus target profit | Profit per unit holds while ad sales dip only slightly |
| Clearing slow inventory | Units out before storage costs compound | Judged against the cost of holding the unit, so a high ACoS can be the cheaper option | Days of cover falling faster than margin |
| Defending your own branded terms | Keeping competitors off traffic that was already yours | Reported ACoS flatters this work, since much of the revenue would have arrived anyway | Competitor share of your branded search falling |
What Amazon will and will not tell you about your category
Category ACoS averages get quoted constantly and almost none of them are sourced. Here is what Amazon itself publishes.
The category benchmark report for Sponsored Brands gives you real peer comparisons on impressions, CTR, ROAS, and ACoS, expressed as the median, the top-performing quartile, and the bottom quartile within your shopping category. That is the one place Amazon hands you a category ACoS figure you can trust, and it is Sponsored Brands only.
The newer benchmarks reporting that went generally available worldwide in May 2026 covers Sponsored Products, Sponsored Brands, Sponsored Display, Sponsored TV, and Amazon DSP, but its metric set is CTR, CPC, CPM, and the new-to-brand family. ACoS and ROAS are not in it. So for Sponsored Products, the format carrying most accounts' spend, Amazon does not publish a category ACoS benchmark at all.
The practical consequence: derive your target from margin, use the Sponsored Brands report to sanity-check whether your efficiency is normal for the category, and treat any blog-quoted "electronics average ACoS" as decoration. Our Amazon ads benchmarks by category and ad type post covers the wider metric set with the same caution.
Why lowering ACoS often costs more than it saves
Most ACoS reduction requests arrive as a single instruction: get it under X. Executed literally, that means cutting bids across the account until the number complies. It works, and it is usually the most expensive way to hit the target.
The starve line
There is a point in every account where cutting bids stops removing waste and starts removing the sales that hold organic rank. Call it the starve line. Above it, every dollar you cut was buying a click that was never going to convert. Below it, you are cutting the volume that keeps your ASIN ranking on the terms that matter, and the loss shows up in organic sales one to two months later, in a different report, with no obvious cause.
The starve line is not the same as break-even. Break-even is arithmetic on one unit. The starve line is a property of the account: it sits wherever your ranking terms stop getting enough attributed velocity to defend their position.
You find it by segmenting, not by watching the account-level number. Split your spend into three buckets and look at each separately:
- Waste. Queries that are wrong for the product, placements converting well below account average, match types pulling irrelevant traffic. Cut freely.
- Marginal. Converting, but above your target. Reduce CPC here, do not pause.
- Load-bearing. The handful of terms driving both ad and organic velocity for your top ASINs. Put a floor under these and leave them alone during any efficiency push.
That third bucket is usually small, often under 20 terms in a mid-sized account, and it is where a blanket cut does its damage.
The three inputs ACoS is made of
ACoS is arithmetic on CPC, conversion rate, and average order value. Falling CPC lowers it, rising conversion rate lowers it, and a higher average order value lowers it. Nothing else moves it directly, and CTR only matters because it feeds volume and relevance rather than because it appears in the formula.
That ordering matters, because conversion rate is the lever that lowers ACoS without lowering volume. Bid cuts trade one for the other. Listing and price work does not.
How to reduce ACoS on Amazon without starving winners
Six moves, in the order that does the least collateral damage.
1. Segment before you cut
Run the search term report and sort by spend, not by ACoS. High-ACoS terms with trivial spend are noise. The terms worth acting on are the ones where spend is real and intent is wrong. Add those as negatives, exact where the query is specific and phrase where a pattern repeats.
2. Reduce CPC where intent is wrong, not where ACoS is high
A high ACoS on a converting term with strong intent usually means your bid is above what the placement is worth, not that the term is bad. Lower the bid and keep the term. A high ACoS on a term with no purchase intent means the term should not be there at all. Those are different fixes and the console reports them identically.
3. Use placement adjustments instead of an account-wide bid cut
Amazon lets you set placement bid adjustments of up to 900% for Sponsored Products across top of search, rest of search, and product pages, on both automatic and manual targeting and on fixed or dynamic bidding. Most efficiency problems are placement problems: top of search converts well and costs a lot, product pages often do the opposite. Pull the adjustment down on the weak placement and you cut cost without touching the keyword's base bid. Our Amazon ad placements breakdown goes into which placements suit which objective.
4. Fix conversion rate before you fix bids
Every point of conversion rate is a proportional cut in ACoS at the same CPC. Images, A+ content, review velocity, price relative to the buy box competitor, and whether the top three bullets answer the objection that stops the purchase. This is slower than a bid change and it does not cost you volume, which is why it goes first on any SKU you intend to keep.
5. Change one lever at a time and wait out the attribution window
Sponsored Products reports on a 7-day attribution window by default and Sponsored Brands on 14 days, per Amazon's ad campaign attribution documentation. Attributed sales are recorded against the click date and can take up to 12 hours to appear, and Amazon recalculates attribution at intervals after the conversion event, so a figure you read on day one is provisional.
The operational rule: judge a bid change on a rolling 7 or 14-day window matching the ad type, not on yesterday. Reading three days of Sponsored Products data as final is how accounts end up cutting a term that was working.
6. Judge the outcome on TACoS and profit, not ACoS alone
If ACoS falls 8 points and TACoS is flat, you removed waste. If ACoS falls 8 points and TACoS rises, you cut ad-attributed sales that were supporting organic volume, and the ad account is now taking credit for an account-level loss. Check both. RoAS is the same information inverted and is worth having alongside if your stakeholders prefer it.
For a deeper tactical pass on the reduction levers themselves, our guide to reducing ACoS on Amazon drills into the campaign-level mechanics.
Where a weekly manual pass runs out
Everything above is doable by hand. On one product line with thirty campaigns, a careful operator with a search term report and a spreadsheet will beat most automation, because the judgment about which query is wrong for the product is the hard part and humans are good at it.
The break happens at volume and at frequency. Four hundred campaigns across a catalog where each SKU has a different job means four hundred separate target decisions, and what happens instead is that one target gets applied to all of them because that is the only thing a person can maintain. Then the efficiency push arrives, bids come down across the board, and the load-bearing terms come down with everything else because nobody had time to mark them.
Console-native tooling closes part of the gap. Rule-based bidding lets you set a ROAS guardrail and have Amazon adjust bids toward it, and dynamic bids down-only reduce bids on impressions less likely to convert. Both help. Neither knows which twenty terms hold your rank, because you never told it, and there is no field for it.
| What the work requires | Careful weekly manual pass | Console rules and dynamic bidding | Platform running on connected data |
|---|---|---|---|
| Spot a query that is wrong for the product | Reliable, this is where human judgment wins | Reads performance, not intent | Classifies intent semantically, using your definition of wrong |
| Hold a different target per job across 400 campaigns | Not maintainable at that volume | One rule shape applied broadly | Optimization logic selected per campaign |
| Keep a floor under a ranking term during a cut | Possible, and easy to forget under time pressure | Down-only limits damage, no explicit per-term floor | Bidding zones with an operator-set floor and ceiling |
| Explain why a bid changed three weeks ago | Depends entirely on the notes someone kept | Rule history, limited detail | Change log per campaign |
| Decide what the target should be | Yours to make, and it should stay that way | Not addressed | Not addressed, the target is still your call |
| Test a change before it spends | A judgment call, checked after the fact | No preview | Simulator run before deployment |
Where Qinetix ads automation fits
Qinetix is the part of Trellis that places and adjusts your bids once you have decided what each campaign is for, the execution layer sitting under the strategy. For ACoS work, four mechanisms carry the weight: optimization logic you select per campaign rather than per account, bidding zones with a floor and a ceiling so a load-bearing term cannot be cut below the level that keeps it ranking, a learning phase and down-only defaults that stop a new configuration from moving fast in the wrong direction, and a change log plus a simulator so you can test a target change and later explain what it did. Pricing runs as a parallel mechanism with shared visibility rather than as part of the same bid decision, which matters when a margin change is what moved your break-even in the first place.
The honest limit, and it is the important one: a target ACoS is a constraint, not a strategy. Set the wrong number and the system will execute it faithfully and efficiently, all the way into a bad outcome. Qinetix will hold a floor you defined and log the change you made. It will not tell you that 18% was the wrong target for a SKU three weeks from a stockout. That call stays with the operator, and any vendor telling you otherwise is selling you something.
Set the target per job, then defend the starve line
Amazon ACoS is worth managing carefully and worth ignoring at the right moments. The version of this work that holds up over a year looks like this: calculate break-even per SKU from real contribution margin, subtract the profit you intend to keep, adjust for the job that SKU is doing this quarter, identify the small set of load-bearing terms and put a floor under them, then reduce spend from the waste bucket first and the marginal bucket second. Read the result on a rolling window that matches the ad type, and check TACoS before you call it a win.
Start with the arithmetic. Our free Amazon ACoS calculator gives you break-even and target from your cost inputs in a few seconds. If the harder problem is holding a different target per campaign across a catalog without losing the terms that carry your rank, book a walkthrough and we will go through your account structure with you.
Frequently Asked Questions
A good ACoS is your break-even ACoS minus the profit margin you want to keep, adjusted for what the SKU is doing. A mature product defending margin might target 15% to 20%, while a launch SKU buying rank may run well above break-even for a defined window. There is no universal number, and any single figure quoted as the industry average is ignoring margin structure, which is the only input that decides it.
Break-even ACoS equals your contribution margin before advertising. Take the selling price, subtract cost of goods, Amazon referral and fulfillment fees, and inbound and outbound shipping, then divide the remainder by the selling price. A $150 product with $90 of total cost has a 40% margin and a 40% break-even ACoS.
ACoS measures ad spend against ad-attributed revenue only. TACoS measures ad spend against total revenue including organic sales, which is what shows you whether advertising is building durable demand or just buying transactions. Watch them together: ACoS falling while TACoS rises usually means you cut spend that was supporting organic volume.
Partly. Amazon's category benchmark report for Sponsored Brands gives median, top-quartile, and bottom-quartile ACoS and ROAS within your shopping category, which is a genuine peer comparison. Amazon's broader benchmarks reporting, generally available worldwide since May 2026, covers Sponsored Products and other formats but reports CTR, CPC, CPM, and new-to-brand metrics rather than ACoS. For Sponsored Products specifically, there is no published category ACoS benchmark, so derive your target from margin instead.
Segment your spend into waste, marginal, and load-bearing before you cut anything. Cut the waste bucket freely, lower CPC rather than pausing in the marginal bucket, and put a bid floor under the small set of terms driving both ad and organic velocity. Placement bid adjustments and conversion rate work reduce cost without reducing the volume that holds rank, so use those before an account-wide bid cut.
That is the normal shape of a launch or a rank-building push, where you are paying above break-even to buy keyword position and conversion data. Check whether organic sales and keyword rank are climbing while TACoS stays flat or falls. If both are true the strategy is working, and you should have set an end date for the window in advance so it does not run indefinitely.
Attributed sales are recorded against the click date and can take up to 12 hours to appear in reports, and Amazon recalculates attribution at intervals after the conversion event, so early figures are provisional. Sponsored Products uses a 7-day attribution window by default and Sponsored Brands 14 days. Judge changes on a rolling window matching the ad type rather than on a single day.
On any SKU where the job is profit, yes, since break-even ACoS is your margin and running above it loses money per unit. On a launch, an inventory clearance, or a branded-defense campaign, no, because the return is rank, storage cost avoided, or share held rather than contribution on that order. Decide which job the SKU has first, then the answer follows.
eCommerce News You'll Actually Use