What Is ACOS? The Honest Math Behind Amazon Ad Profitability

Every "what is ACOS" post tells you ACOS is the percentage of sales revenue spent on ads, then tells you lower is better. Both are true. Neither is useful. The honest answer is that ACOS is just a step in the math toward real margin, and the operators who chase low ACOS usually do it by capping spend on winning campaigns. BTB Audits' core managed spend is on Meta and Google. The Amazon perspective in this post comes from auditing accounts across DTC supplements, baby care, electronics, and home goods over the past 6 years. The math is straightforward enough that any operator can run it on their own account in 10 minutes.

What ACOS actually is

ACOS stands for Advertising Cost of Sale. It is Amazon's primary advertising metric. The formula is simple.

ACOS = (Ad Spend / Sales Revenue from Ads) × 100

A campaign that spent $300 on ads and generated $1,000 in sales has a 30% ACOS. The inverse of ACOS is ROAS (return on ad spend). Both numbers describe the same relationship. The choice of which to display is conventional, not analytical.

Why Amazon uses ACOS instead of ROAS. ACOS is structured as a cost ratio (lower is better) which fits the merchant mental model of "how much does each sale cost me in ads." ROAS is a return ratio (higher is better) which fits the advertiser mental model. Both numbers carry the same information.

ACOS calculations are happening at scale across Amazon. Amazon's own Q4 2023 earnings report put the company's advertising services revenue at $46.9 billion for the full year 2023, up 24% year over year. That number is roughly the size of Google Ads' first decade. Every sponsored product campaign feeding that number produces an ACOS, and most of them are being optimised against the wrong target.

ACOS vs TACOS, and why TACOS is more honest

ACOS measures ad spend against ad-driven sales only. TACOS (Total Advertising Cost of Sale) measures ad spend against total sales revenue, which includes both ad-driven sales and organic sales.

TACOS = (Ad Spend / Total Sales Revenue) × 100

Why TACOS is more honest for most decisions:

• ACOS can look great (low) while the brand is bleeding money to ads that are eating searches the brand would have won organically. Branded keyword campaigns are the most common offender. They report a low ACOS because the customer was searching the brand name, but those sales would have happened without the ad spend.
• TACOS catches the cannibalisation that ACOS hides. If your ad spend rises but TACOS stays flat or rises, the ads are not adding incremental revenue. They are taxing the same sales twice.
• A healthy DTC brand on Amazon typically runs ACOS in the 25% to 40% range at the campaign level but TACOS in the 8% to 15% range at the account level. That gap means the ads are amplifying organic ranking, not replacing it.

The one-sentence rule: optimise for TACOS at the account level, use ACOS at the campaign level. The campaign-level ACOS tells you which keywords are working. The account-level TACOS tells you whether the whole programme is paying for itself.

What "good ACOS" actually means by category and stage

This is where the SERP fails most readers. Generic "20-30% ACOS is good" benchmarks are misleading because ACOS is a function of three things: gross margin, stage of business, and category competitiveness.

The math floor. The maximum ACOS you can afford before ad spend exceeds gross profit equals your gross margin percentage. At 60% gross margin, every dollar of sale generates 60 cents of gross profit, so an ACOS up to 60% breaks even on contribution. Anything below that target leaves margin to cover overheads and profit.

The stage modifier shifts the practical target inside that ceiling.

• Launch stage. Brands often accept ACOS above their margin floor (60% to 80%) for the first 8 to 12 weeks to drive ranking. The ad spend is paying for organic rank, not for the sale itself.
• Optimisation stage. Once the listing ranks, the target drops to 30% to 40% on most categories, with TACOS pulled in tight.
• Cost-cut stage. A brand pulling back on spend pushes ACOS down further by killing the long tail and concentrating on top-converting keywords.

The category modifier is real. Competitive high-LTV categories (supplements, beauty, electronics) tolerate higher ACOS because the customer comes back. Low-LTV one-time-purchase categories need tighter ACOS because the first transaction has to carry the cost. The same logic applies on Meta and Google: the gross-margin math behind a good ROAS is the same shape on every platform, just expressed in different units.

Why "lower ACOS" is the wrong goal

The most common misconception in Amazon advertising. ACOS is a ratio. You can lower a ratio either by lowering the numerator (ad spend) or by raising the denominator (revenue). Operators who chase low ACOS usually do it by capping spend on winning campaigns, which leaves absolute dollars on the table.

The right question is not "what is my ACOS?" It is "at the current ACOS, am I generating positive contribution margin after all costs?" If yes, the next question is "can I scale spend on this campaign without ACOS climbing past my margin ceiling?"

A worked example. A skincare brand at $80 AOV (average order value) with 60% gross margin spends $2,000 a month on a campaign at 20% ACOS, generating $10,000 in sales. Gross profit on those sales: $6,000. Ad spend: $2,000. Contribution after ads: $4,000.

The same brand scales that campaign to $8,000 a month. ACOS rises to 35%. Revenue rises to about $23,000. Gross profit on those sales: $13,800. Ad spend: $8,000. Contribution after ads: $5,800.

The "worse" 35% ACOS earns $1,800 more in monthly profit than the "better" 20% ACOS. Every operator who would cap that campaign at 20% ACOS to "improve the metric" is leaving $1,800 a month on the table. This is the same scaling-discipline principle from the Meta audit method, where chasing a higher reported ROAS on a capped budget hides a smaller business behind a better-looking number.

The reader's job: run the math (1 minus 1 divided by gross margin, then multiplied by 100) on their own product. That number is the ceiling. Anything below it is contribution-positive. The only question after that is how much volume the listing can absorb before ACOS hits the ceiling. Stop trying to drive ACOS below half the ceiling. That is where the leak is.