Use ACOS or ROAS to judge advertising efficiency, TACoS to track advertising spend against total sales, and contribution to decide whether the economics work. None of these ratios alone tells you whether advertising created additional demand or whether your business earned a net profit.
For the same spend and attributed sales, ACOS and ROAS express exactly the same relationship. TACoS changes the denominator, so it answers a different question. The useful choice is not which number to trust universally, but which number fits the decision—and whether its underlying data is comparable.
1. Three definitions in plain terms
Amazon Ads’ ACOS guide, checked on 2026-10-22, defines ACOS as advertising spend divided by advertising revenue, expressed as a percentage. It defines ROAS as advertising revenue divided by advertising spend.
Throughout this article, advertising revenue means ad-attributed sales, not profit and not proof of sales that would otherwise never have happened.
Metric
Formula
What it answers
When it misleads
ACOS
Ad spend ÷ ad-attributed sales × 100
What percentage of attributed sales went to advertising?
When you ignore product margins or compare incompatible reports
ROAS
Ad-attributed sales ÷ ad spend
How many dollars of attributed sales were reported per advertising dollar?
When revenue return is mistaken for profit or incremental sales
TACoS
Ad spend ÷ total sales × 100
What percentage of total sales went to advertising?
When unrelated sales conceal weak campaigns or sales scope changes
For TACoS, total sales includes ad-attributed and other sales within the defined business scope. Do not add advertising sales to total sales: that would double-count revenue.
ACOS and ROAS are reciprocals when they use identical inputs. An ACOS of 25%, written as 0.25, corresponds to ROAS of 1 ÷ 0.25 = 4. A 4x ROAS means $4 in attributed sales per $1 spent—not $4 in profit.
TACoS is used here as a seller management calculation. Do not assume every Amazon report offers a field with that name.
Zero denominators need special handling. With positive spend and no attributed sales, ROAS is zero, while ACOS has no finite value. With no spend, ROAS is undefined. Mark undefined ratios clearly rather than turning them into apparently excellent results.
2. One hypothetical example, calculated three ways
Consider an illustrative US Amazon business selling one product. These figures are assumptions, not observed seller results or current Amazon fee rates.
For a completed reporting period, assume:
Advertising spend: $3,000.
Ad-attributed sales: $10,000.
Total sales: $20,000.
Selling price: $40 per unit.
Units sold: 500, including 250 attributed to advertising.
No discounts, refunds, or taxes in the example.
Sales scope and timing are aligned for this simplified calculation.
The three calculations are:
ACOS: $3,000 ÷ $10,000 × 100 = 30%.
ROAS: $10,000 ÷ $3,000 = 3.33x, rounded.
TACoS: $3,000 ÷ $20,000 × 100 = 15%.
All three numbers are correct. Advertising consumed 30% of attributed sales and 15% of total sales. Each advertising dollar corresponded to about $3.33 in attributed sales.
The distinction matters when someone says advertising costs are too high. Too high compared with attributed revenue, total revenue, or the contribution available to pay for advertising? Those are different tests.
Under this deliberately aligned example, $20,000 − $10,000 = $10,000 of sales sits outside the attributed-sales figure. That does not establish that advertising had no influence on those purchases. In real reports, do not even label this subtraction as organic sales until you have reconciled timing, product coverage, and attribution definitions.
3. Set targets from contribution, not a benchmark
There is no universally good ACOS. A 30% ACOS can be workable for one product and loss-making for another because the money available before advertising differs.
Continue the hypothetical example with these variable costs per $40 unit:
Landed product cost: $12.
Selling fees: $6.
Fulfillment costs: $5.
Other variable costs, excluding advertising: $1.
These illustrative costs total $24 per unit. Each cost appears once; fulfillment is not also included in the landed product cost.
Pre-advertising contribution per unit is $40 − $24 = $16. The pre-advertising contribution margin is $16 ÷ $40 × 100 = 40%.
With a consistent product mix and revenue basis:
Contribution break-even ACOS is 40%.
Corresponding break-even ROAS is 1 ÷ 0.40 = 2.5x.
To retain a 10% contribution margin after advertising on attributed sales, target ACOS would be 40% − 10% = 30%.
At the example’s 30% ACOS, contribution associated with attributed sales is $10,000 × 40% − $3,000 = $1,000. This is a simplified allocation, not a measurement of incremental advertising profit.
Across the whole business, contribution after advertising is $20,000 × 40% − $3,000 = $5,000, or 25% of total sales. That is contribution, not net income: fixed overhead and other excluded expenses still need funding.
For your own model, use actual product costs and verified fees. The Amazon FBA calculator can support a preliminary cost estimate, but reconcile estimates with your actual charges before setting targets.
Discounts must reduce revenue once. Refunds also remove revenue; do not subtract the same refund again as an expense. Record separate return-related costs separately. Recoverable returned inventory is an asset, not new revenue.
If advertised products have different margins, use their relevant sales mix rather than applying the catalog’s highest margin to every campaign.
4. Choose the metric for the decision
The campaign–account–business framework below is an editorial decision aid, not Amazon’s official reporting taxonomy.
For a bid or targeting decision, start with ACOS or ROAS. Choose whichever your team finds easier to interpret. Reviewing both does not provide independent confirmation when their inputs are identical.
Compare the ratio with the relevant product margin, then examine spend, sales, relevance, and the quality of the evidence. There is no universal click count after which every target should be paused. A costly product, a low-cost accessory, and a new campaign do not share one sensible cutoff.
For budget allocation, examine campaign economics alongside absolute contribution. A small campaign with excellent ROAS may have limited expansion opportunities. A larger campaign with a weaker ratio may still generate more contribution. Neither observation proves what the next dollar will produce.
Separate campaigns with different objectives. Capturing existing branded demand and exploring unfamiliar search terms are not interchangeable tests. Strong reported efficiency does not prove that every attributed sale was additional.
For account-level spending, use TACoS with total sales and contribution. TACoS shows how much of the business’s sales base advertising consumes. It cannot identify which target needs a lower bid.
For a product launch, define a bounded learning budget. Specify what you want to learn, how much you can afford to spend, and what evidence would justify continuing. A launch label does not make losses harmless or guarantee later organic sales.
If traffic arrives but purchases disappoint, inspect offer clarity and product-page accuracy before assuming bids are the only problem. A listing audit can help structure that review without promising a conversion improvement.
5. Reconcile reports before changing bids
Two ACOS figures can differ because their inputs differ. Before choosing one, write down the numerator and denominator behind each report.
Use this reconciliation checklist:
Marketplace and currency: Are both reports measuring the same marketplace in the same currency?
Date range and time basis: Do they cover matching dates, and what event determines when sales are recorded?
Attribution: What window, interaction types, and sales eligibility does each report document?
Product coverage: Are you comparing the same advertised products and purchased products?
Advertising coverage: Does spend include the same campaigns and advertising products?
Revenue treatment: How are discounts, cancellations, refunds, and taxes handled?
Data status: When was each report downloaded, and is either period still subject to adjustment?
Verify these details in the documentation for the specific report and in your account. Do not assume a particular attribution window applies everywhere or impose an unsupported waiting period before reviewing results.
Keep reported advertising metrics intact. If finance uses a different revenue definition, create a separate reconciled management view and label it rather than silently replacing one report’s denominator.
Also avoid averaging campaign percentages. Account ACOS is total included spend divided by total included attributed sales—not the simple average of individual campaign ACOS values. The same principle applies when combining ROAS or TACoS across periods or products.
Finally, avoid mismatched scope. Dividing all account advertising spend by one product’s sales does not produce a meaningful product-level TACoS unless that spend genuinely belongs to that product.
6. Read trends without claiming organic lift
A falling TACoS can accompany healthy growth, but it does not establish why sales changed. Seasonality, price changes, promotions, stock availability, product mix, and advertising decisions can all complicate the comparison.
Consider a second hypothetical period using the same 40% pre-advertising contribution margin:
Advertising spend falls from $3,000 to $2,000.
Attributed sales fall from $10,000 to $8,000.
Total sales fall from $20,000 to $12,000.
ACOS improves to $2,000 ÷ $8,000 × 100 = 25%, and ROAS improves to 4x. But TACoS rises to $2,000 ÷ $12,000 × 100 = 16.67%, rounded.
Contribution after advertising falls to $12,000 × 40% − $2,000 = $2,800, down from $5,000. Advertising looks more efficient while the business produces less contribution. This arithmetic does not prove that reducing advertising caused the decline.
Additional search data can help investigate. Amazon’s Brand Analytics overview, checked on 2026-10-22, describes Search Catalog Performance and Search Query Performance dashboards for examining shopping and search activity. It states that access requires a Professional selling account and the Brand Representative role for a brand enrolled in Amazon Brand Registry.
Where available, use that evidence to investigate changes in discovery and purchasing. It does not turn a TACoS trend into proof of advertising-driven organic growth.
7. A reporting layout you can keep
Keep the headline report small enough to support a decision, with definitions available underneath it. The following is a practical reporting template, not an Amazon-required format.
Reporting field
What to record
Why it belongs
Scope
Marketplace, currency, products, campaigns
Prevents incompatible comparisons
Period and status
Start date, end date, download date, provisional or reviewed
Makes timing differences visible
Advertising inputs
Spend and attributed sales
Allows ACOS and ROAS to be recalculated
Business inputs
Total sales and units sold
Gives TACoS its denominator and adds volume context
Efficiency ratios
ACOS, ROAS, TACoS
Separates advertising efficiency from business spending intensity
Economics
Pre-ad contribution margin and contribution after ads
Connects ratios to available money
Context and action
Price, promotions, availability, decision, owner
Explains changes and records the next step
Use a consistent review cadence appropriate to your sales volume. Compare similarly defined periods, and flag incomplete data rather than presenting it as final.
Record amounts beside ratios. A lower ACOS on much less attributed revenue may not support the same decision as a lower ACOS on stable or growing revenue.
Write actions as testable decisions: reduce exposure to irrelevant traffic, investigate an offer change, or maintain a bounded test while gathering clearer evidence. Include the reason and a review condition. Avoid instructions such as “improve ROAS” without identifying what will change.
The final line should state whether contribution increased or decreased, what remains uncertain, and which decision the evidence supports.
8. Frequently asked questions
How is ACOS different from ROAS?
ACOS divides advertising spend by ad-attributed sales and expresses the result as a percentage. ROAS reverses the division. With identical inputs, 20% ACOS equals 5x ROAS. They are two presentations of the same efficiency relationship, not separate evidence of success.
What is TACoS?
TACoS is advertising spend divided by total sales, multiplied by 100. Use matching marketplace, product, and period scopes. It measures advertising spending relative to the whole sales base, but does not measure profit, organic ranking, or the incremental effect of advertising.
What is a good ACOS?
A useful target depends on pre-advertising contribution margin and the campaign’s objective. If that margin is 35% and you want to retain 10% contribution on the same sales basis, an illustrative target is 25% ACOS. Fixed overhead remains separate; there is no universal benchmark.
Why does my ACOS look different in two reports?
Check spend coverage, attributed-sales definitions, attribution settings, product scope, dates, currency, and adjustment status. Recalculate both figures from their underlying amounts. If the definitions differ, label the reports separately rather than treating one number as automatically wrong.