More than $800 of advertising spend on a supplements account was going to one SKU that was not converting. Our team found it in the weekly report, the spend stopped, and after that listing was improved the returns on that SKU fell 12%.
The $800 is not really the point. The point is that a single SKU can burn money for weeks inside an account where the campaign numbers look ordinary, because campaign numbers describe campaigns. They do not tell you whether the business is growing.
That is the job TACOS does, and it is the reason we ask for it before almost anything else when a seller says their advertising is not working. It is also a number Amazon does not put on any dashboard, so somebody has to build it, which is why so few accounts have it.
What ACOS is actually measuring
ACOS is advertising cost of sales: ad spend divided by the revenue Amazon attributes to those ads. Spend $200, get $1,000 in attributed sales, and ACOS is 20%.
It answers its own question well: is this campaign buying revenue efficiently. It answers the question most founders are actually asking badly: is the money working.
Two things get in the way. The first is that attributed sales are attributed on Amazon's terms. Sponsored Products attributes click-based sales in a 7 day window for sellers, while Sponsored Brands and Sponsored Display both use 14 days. Reported by Intentwise and Seller Labs, both third party, and they agree. So a weekly ACOS comparison is partly comparing different amounts of attribution, and the ad types are not measured on the same clock.
The second is more serious. ACOS can improve while the business shrinks. Cut spend on a campaign that was feeding rank, and the ACOS falls immediately because the expensive discovery clicks are gone, while the organic sales those clicks were building fall a few weeks later, quietly, in a number nobody was watching.
What TACOS measures instead
TACOS is total advertising cost of sales: advertising spend divided by total sales, organic and ad-attributed together, expressed as a percentage. The formula is published identically by Helium 10 and Jungle Scout, both third party, 2024 and 2023, and they agree.
Take a month with $4,000 of ad spend, $12,000 of ad-attributed sales and $40,000 of total sales. ACOS is 33%, which most people would call uncomfortable. TACOS is 10%, which says that a tenth of everything the brand sold that month went back into advertising and the other 90% arrived without being paid for twice.
Same account, same month, two very different conversations. One is about a campaign. The other is about a business.
- Is this campaign buying revenue efficiently
- Denominator is ad-attributed sales only
- Moves with bids, placements and match types
- Improves when you switch spend off
A campaign diagnostic
- Is the business less dependent on ads than it was
- Denominator is every sale, organic included
- Moves with rank, conversion and repeat purchase
- Gets worse when you switch spend off badly
A business diagnostic
Where the two numbers live
Nothing in the advertising console shows TACOS, because the console only knows about advertising. Total sales sit in Seller Central, so the number has to be assembled by hand or by a report somebody built.
The spend comes from the advertising console. The path is Measurement and Reporting, then Sponsored ads reports, then a campaign report for the period. The total sales come from Reports, then Business Reports, then the sales and traffic view for the same period, which is also where ordered product sales, sessions and unit session percentage live.
Monthly is the right grain for most accounts. Weekly TACOS bounces around for reasons that have nothing to do with performance, particularly on catalogues with a lumpy order pattern or a subscription tail, and a number that jumps every Monday gets ignored by everyone within a month.
How to tell what your TACOS is telling you
TACOS on its own means very little. Read against the direction of total sales it means almost everything, because together they describe four situations that call for four different decisions.
| Total sales | TACOS | What it usually means |
|---|---|---|
| Rising | Falling | Ads are building organic rank and the business is compounding. This is the pattern to want. |
| Rising | Rising | Growth is being bought. Acceptable during a launch, expensive as a permanent state. |
| Flat | Rising | Spend is replacing organic sales rather than adding to them. Usually a rank or conversion problem. |
| Falling | Falling | Spend was cut and organic has not held. The most commonly misread of the four, because the ACOS looks better than ever. |
Five checks that turn that table into a decision.
- Plot twelve months, not three. A single month is noise. A year shows whether the line is genuinely bending, and it also shows you your own seasonality instead of somebody else's.
- Read TACOS and total sales on the same chart. Separately they are two numbers. Together they are a diagnosis, and it takes about ten minutes to build once.
- Do it per ASIN as well as per account. Account-level TACOS hides exactly the problem you are looking for, which is the one SKU quietly consuming the budget while the rest of the catalogue carries the average.
- Check it against contribution margin, not gross revenue. The same TACOS figure can be comfortable on a high margin product and unsustainable on a thin one, which is why it reads as good news and bad news inside the same account.
- Compare TACOS in the months you spent heavily against the two or three months after. If organic did not pick up when the spend came off, the spend was buying sales rather than building position, and that is the single most useful thing this metric will ever tell you.
What a healthy TACOS looks like, and why nobody can hand you the number
We get asked for a target figure constantly, and a single number would be a guess dressed up as a benchmark.
Published guidance says the same thing. Perpetua states plainly that TACOS varies widely with marketplace, advertising goals, ad types, seasonality, competition and product price, and Helium 10 declines to give a good TACOS figure at all, pointing instead at a decreasing or flat trend over time. Both third party, 2024, and they agree with each other. Anyone quoting you one number across every category is quoting a number they made up.
What does hold up is the shape of the curve across a product's life.
At launch, TACOS is at its highest and it should be, because you are paying for position you do not have yet. That spend is buying ranking data and velocity, and reading it as waste in month one is how launches get strangled.
In the growth phase, TACOS should start bending downward while total sales keep climbing, which is the moment the advertising begins to compound instead of just to buy. On a mature product it should settle at whatever level your margin can carry indefinitely, with the spend doing defence and incremental reach rather than heavy lifting.
The only benchmark worth anything is your own trailing twelve months against your own contribution margin. Your margin sets the ceiling. Your trend tells you which direction you are travelling. Everything else is somebody else's category.
What to check before you change a single bid
A rising TACOS is a symptom, and the reflex is to attack it at the bid level, which is usually the wrong end of the problem.
- Check the page can convert before you spend more on sending people to it. Amazon publishes its own retail readiness checklist and it asks five questions: is there enough inventory, are you winning the Featured Offer, are there enough positive reviews, are the images inspiring, is the description clear. It also names thresholds, a star rating of 3.5 or higher and at least 15 customer reviews. Source: Amazon Advertising, primary.
- Confirm you hold the Featured Offer. Sponsored Products ads only serve when your offer is the Featured Offer, so losing it stops delivery and distorts every number you are about to compare.
- Check inventory and days of cover. Spend against a SKU that is about to run out buys rank you will hand straight back, and the TACOS damage lands in the following month.
- Look at unit session percentage before ACOS. If traffic is steady and conversion has slipped, the page changed rather than the auction, and no bid adjustment will fix a page.
- Account for the seasonality in your own category rather than in general, because a TACOS that rises every Q4 in the same shape every year is a calendar, not a problem.
The order matters. When our team audits an account with a climbing TACOS, the cause is more often a page that stopped converting than a campaign that got worse, and the cheapest fix is rarely in the ads console.
Reading TACOS alongside organic share
TACOS tells you how dependent the business is on advertising. It does not tell you why. For that, two other reports do most of the work.
The first is Search Query Performance in Brand Analytics, which lives under Brands, then Brand Analytics, and needs Brand Registry. It maps the funnel per search query and gives your share at each stage: impression share, click share, cart add share and purchase share, measured against the whole marketplace rather than your own account. Reported by Amalytix and Kapoq, third party, and they agree.
The gaps between those shares are the diagnosis. Impression share far below purchase share means you win when you are seen and are simply not seen enough, which is a ranking job. Impression share far above click share means the search result is not persuading anyone, which is a main image, title and price job.
The second is the search term impression share report for Sponsored Products, which Amazon describes as showing how your account-wide impression share for each search term compares to other advertisers. Amazon's own example: an impression share of 20% for a search term on a given date means you won 20% of all Sponsored Products impressions for that term that day. Source: Amazon Advertising, primary. The report is in the Report Center in the advertising console.
Put those next to TACOS and the picture completes itself. Falling TACOS with rising organic purchase share on your head terms is real growth. Falling TACOS with flat organic share usually means you just spent less, and the bill for that arrives later.
What each ad type is actually for
TACOS improves when advertising builds something. Which ad type builds what is worth being clear about, because blending them muddies the reporting and the strategy together.
- Sponsored Products is the foundation and where most budget belongs. It appears in search results and on product pages, drives the most direct conversions and produces the cleanest search term data, which is what everything else is built from.
- Sponsored Brands carries the headline, the brand and the benefit message. It is where brand defence lives, and where a category message can be made to a shopper who has not decided on a product yet.
- Sponsored Display is retargeting and placement on other people's pages. It picks up sessions that did not convert and defends your own detail pages from competitors doing the same thing to you.
- Amazon DSP sits above all of it, buying audiences programmatically on and off Amazon. It is judged on audience behaviour and incrementality rather than raw ACOS, which is why dropping it into a TACOS conversation without that framing produces an argument. Our Amazon DSP page covers where it fits.
Structuring campaigns so the numbers mean something
Most TACOS problems we see are structure problems wearing a bid costume. When one campaign is asked to rank a new keyword and defend a brand term at the same time, the algorithm gets two contradictory targets and averages toward mediocre, and the reporting cannot separate the two either.
Separating campaigns by objective is what makes the metric readable.
- Launch and rank campaigns exist to push a target keyword onto page one. A high ACOS here is the price of the position, not a failure, and judging these on ACOS is how brands accidentally cancel their own launches.
- Scale campaigns hold proven converters and push volume at an efficiency your margin can carry.
- Brand defence campaigns protect your own name cheaply, and they belong on their own, because branded terms convert so well that mixing them into a general campaign flatters the average and hides everything underneath.
- Conquesting campaigns target competitor products, and they only make sense where you have a real advantage a shopper can see, such as rating, price or a feature difference.
- Harvesting is the routine that moves converting search terms out of discovery campaigns and into exact match, and adds the terms that never convert as negatives. Weekly is the cadence that keeps it honest.
The discovery layer underneath is the familiar three: an automatic campaign to surface terms you would not have guessed, broad or phrase for controlled expansion, exact for the proven ones. Terms flow up as they prove themselves and get negated when they do not. That flow turns paid demand into organic rank, and organic rank is the only thing that makes TACOS fall.
Keeping TACOS honest month after month
The supplements account matters here mostly for the cadence. That SKU was caught in a weekly review rather than a quarterly one, and the gap between those two habits is the difference between $800 and considerably more.
- Record spend, ad sales and total sales monthly in one sheet and let it accumulate. Twelve rows is a trend line. One row is an opinion.
- Track TACOS per ASIN for your top ten SKUs. The account average is exactly where a bad SKU hides, and ten rows is not a reporting project.
- Review search terms weekly. Wasted spend compounds quietly, and the terms that burn budget are rarely the ones anybody suspected.
- Set target ACOS from margin, per SKU. A single account-wide target treats a 45% margin product and an 18% margin product as the same business.
- Write down what changed and when. A bid change, a price change, a new main image, a stockout. Six weeks later, when TACOS moves, that log is the only thing that will tell you which one did it.
What we would do first if this were our account
Build the twelve month chart before touching anything. Spend, total sales and TACOS on one line each. It takes an afternoon and it usually settles an argument that has been running for months about whether the advertising is working.
Then split it per ASIN for the top ten SKUs, because the account average is where the problem hides and the outlier is usually visible within minutes.
Then, and only then, look at the campaigns, starting with retail readiness on any ASIN carrying serious spend. If the page cannot convert, no bid on earth fixes it, and that is the work our listing optimization side handles. If the structure is blending objectives so nothing can be read, that is where our high ACOS work starts.
We cannot promise a TACOS figure, or a date by which it will fall, and we would be careful with anyone who offers one, because the number depends on your margin, your category and how much rank you currently hold.
What we can tell you before you spend anything is which SKUs are carrying the account, which are being carried, and whether the spend is buying position or just buying sales. That is what the free, no-obligation audit looks at, and if the advertising is already doing its job, we will say so and tell you where the real constraint is instead.
Related guides
Common questions about TACOS
What is a good TACOS percentage?
There is no single figure, and the sources that publish guidance on this say so themselves. TACOS varies with category, price, margin, competition, season and how established the product is. The useful version of the question is whether your own TACOS is falling while your total sales rise, and whether the level you are at is one your contribution margin can carry.
Why did my sales drop when my ACOS improved?
Usually spend came off a campaign that was feeding organic rank. ACOS improves immediately because the expensive discovery clicks disappear, and the organic sales those clicks were supporting fade a few weeks later. TACOS catches this because the denominator includes the organic sales that quietly went missing.
Should I calculate TACOS weekly or monthly?
Monthly for the trend, because weekly figures swing for reasons unrelated to performance and a noisy number stops getting read. Weekly still has a job, but it is search term review and waste control rather than trend analysis.
Does a rising TACOS always mean something is wrong?
No. During a launch it is expected, because you are buying a position you do not hold yet. It becomes a problem when it persists after the launch window and total sales are not rising with it, which is the point at which spend is replacing organic sales instead of building them.
Where do I find total sales for the calculation?
In Seller Central under Reports, then Business Reports, in the sales and traffic view. Ad spend comes from the advertising console under Measurement and Reporting. The two have to be pulled for the same date range, and mismatched ranges are the most common reason a TACOS figure looks wrong.
Does TACOS include sales outside Amazon?
Not as it is normally calculated. Standard TACOS uses Amazon total sales, so external traffic sent to Amazon listings shows up as organic sales in the denominator and pulls TACOS down with no change to advertising. Worth knowing before reading a sudden improvement as an advertising result.