Neither ACOS nor TACOS knows what the product costs you

ACOS is ad spend divided by the sales Amazon attributes to your ads. TACOS is the same spend divided by all the product's sales, organic included. ACOS grades the advertising; TACOS shows how much of each sales dollar advertising takes. Neither shows profit: subtract TACOS from the product's margin before advertising, and what is left is its margin after advertising.

Both ratios divide by sales. Profit is what is left after the costs, and neither ratio has seen them.

Same spend, two denominators

Amazon defines ACOS as ad spend divided by ad revenue. TACOS keeps the numerator and widens the denominator to every sale the product made, which Amazon's own guidance describes as showing whether total sales, organic and paid, are growing in relation to ad spend. Because ad sales are part of total sales, the two are tied by one line of arithmetic:

A product that spent $1,000 on ads, got $4,000 of ad-attributed sales and sold $10,000 in total has an ACOS of 25% and a TACOS of 10%. Ads produced 40% of its sales, and 25% × 40% is 10%. Same dollars of spend, read against two different totals.

Why the two move apart

TACOS moves when ACOS moves, and also when the share of sales coming from ads moves. That second part is why a falling TACOS can be good news or bad news. A baseline month for one product, then two different ways it can change:

Month 2 is the story TACOS is usually quoted for: the ads cost the same, organic sales rose, and advertising takes a smaller share of the product's revenue. Month 3 has the best ACOS and the best TACOS of the three, and the product sold $2,000 less than in month 1. Both ratios improved because spend fell by half while ad sales fell by 37.5% and total sales by 20%. Whether month 3 made more money than month 1 depends on the product's margin, and neither ratio contains it.

What the product's costs leave for advertising

Amazon's own test for ACOS is that it needs to be lower than your profit margin to keep a profit. The margin that matters is the one before advertising: what is left of a sales dollar after landed cost of goods, Amazon fees, storage and refunds. It is also the product's break-even ACOS, and the break-even ACOS calculator works it out from price, cost and fees. TACOS connects to it without any approximation, because both are shares of the same total sales:

Here are two products with the same ad results. Same spend, same ad sales, same total sales, so the same 25% ACOS and the same 10% TACOS:

Spend, ad sales, total sales, ACOS and TACOS are identical for the two. One product keeps $2,000 and the other pays $200 to be sold: 30% minus 10% is 20%, and 8% minus 10% is −2%. By Amazon's test, Product B's ads lose money on their own as well. A 25% ACOS is above an 8% margin: $4,000 of ad sales kept $320 of margin and cost $1,000 of ads. Indirect expenses come off after all of this to reach net profit.

So the TACOS ceiling that says anything about profit is the product's own margin before advertising, less the profit you want to keep. A benchmark borrowed from another account was set by that account's costs.

What happens when you raise the budget

Raising the budget is the decision both ratios get used to justify, and it is where Product B gets expensive. Suppose more budget buys more ad sales at the same 25% ACOS and the costs per sales dollar stay where they are. Every extra $100 of ad sales then costs $25 of ad spend:

• Product A keeps $30 of margin on that $100, less $25 of ads: $5 more profit.
• Product B keeps $8, less $25 of ads: $17 more loss.

Ads that look efficient on Product B send more volume through unit economics that lose money. Its ACOS will look the same the whole time.

Which number answers which question

At the target level, the same ratio-versus-money problem has its own article: why an ACOS threshold cannot tell $8 from $800.

Where this still needs a person

The check is only as good as the costs that go into it. Amazon does not supply cost of goods, and if it is missing, profit is overstated by the whole cost of the product. Fees that have not been reported yet make the most recent days look better than they will. Vendlen's Dashboard lays out sales, advertising, every Amazon fee on its own line and cost of goods down to gross and net profit, so the costs behind the margin come from the account's own fee lines and the cost of goods you entered.

Product-level ad spend is not always measured. Vendlen's per-product figures include Sponsored Brands spend split across products by daily product sales, which makes that part of a per-product TACOS an estimate.

And the arithmetic does not see what the ads did to organic sales. A product that loses money this month may be buying rank that pays later, or it may simply be losing money. Which one it is depends on the product, its stage and what you are trying to do with it. That call is yours.

What to do with this

1. Pick one product and one complete month. Take its ad spend, ad sales and total sales for exactly those dates, and work out ACOS, TACOS and the share of sales that came from ads.
2. Work out its margin before advertising: total sales minus landed cost of goods, Amazon fees, storage and refunds, divided by total sales. Count inbound freight once, inside cost of goods or as its own line; freight Amazon carried inbound is already among its fees. The break-even ACOS calculator does the per-unit version.
3. Subtract TACOS. Below zero means the product lost money that month before any overheads, whatever the ACOS says.
4. Before raising a budget, compare ACOS with that margin. If ACOS is above it, each extra ad sale costs more in ads than it keeps in margin, so profit falls.
5. When TACOS falls, look at total sales over the same months before you call it progress.

← All articles