Ask most Amazon sellers how their ads are doing and you’ll get one number back: ACOS. It’s the headline metric in Amazon’s own console and in most agency reports. It’s also a poor target to manage a business by.
What ACOS actually measures
ACOS is advertising cost of sales: ad spend divided by the sales those ads produced. Spend $300 to make $1,000 of ad sales, and your ACOS is 30%.
That tells you the cost of getting an ad sale. It doesn’t tell you whether the sale was worth having. For that you need to know what’s left after the product cost and Amazon’s fees, and ACOS knows nothing about either.
Same ACOS, opposite results
Take two products, both running at a 30% ACOS:
| Product A | Product B | |
|---|---|---|
| Price | $40 | $20 |
| Profit before ads | $20 (50%) | $5 (25%) |
| Ad cost per sale at 30% ACOS | $12 | $6 |
| Profit per ad sale | $8 | −$1 |
Product A makes $8 on each ad sale. Product B loses $1. An account-wide target of 30% treats them as equally good. In reality, you should probably be spending more on A and less (or differently) on B.
The break-even ACOS
The fix starts with one simple number per product: its break-even ACOS. That’s the profit margin before advertising. Product A breaks even at 50% ACOS. Product B breaks even at 25%.
Once you know the break-even for each product, your ACOS report starts to mean something. You can see at a glance which products make money from ads and how much room each one has.
Lower isn’t always better
The trap with ACOS is that lower always looks better. But the lowest ACOS usually comes from bidding only on your cheapest, most certain traffic: branded searches and a few long-tail terms. You keep a lovely ACOS and miss most of the sales you could have made at a profit.
As you bid on more traffic, ACOS rises. Profit often rises with it, up to a point. Past that point, the extra sales cost more than they earn. The goal is to find that point for each product, not to push ACOS as low as it will go.
We show this with real numbers in the worked example on our approach page, where a month at 28% ACOS makes more money than a month at 20%.
What to target instead
Target profit. In practice that means:
- Work out profit per unit before ads for each product: price, less product cost, less Amazon fees.
- Set a target ACOS per product from its margin and its job: launch, grow or harvest.
- Judge the result by profit after ad spend, not by the ACOS alone.
- Watch organic sales as well, because ads can support organic rank and that profit counts too.
ACOS still has a place. It’s a good gauge of how efficiently a campaign is running against its own target. It just shouldn’t be the target.
If you’d like to see what your ads really earn, ask for a free profit audit.