ASIN-level ACOS is the wrong metric
A deodorant brand I audited has an ASIN called Limitless. The reporting from the previous agency showed Limitless running at 100.8% ACOS. By the standard interpretation, the campaign was unprofitable. Cut it.
When I looked at the actual data, 52.5% of all sales attributed to the Limitless campaign were sales of different ASINs. Customers were clicking on the Limitless ad, then buying the body spray, the 3-step bundle, or the wipes. The campaign was producing meaningful revenue. None of it was showing up against Limitless directly.
Another ASIN in the same account, the Body Spray, had an even more extreme version. 76.3% of all sales from its campaigns went to other ASINs. By ASIN-level ACOS, the Body Spray campaign looked terrible. In reality, it was the catalog entry point. Customers discovered the brand through Body Spray ads, then bought higher-margin products.
The right way to evaluate either campaign isn’t ASIN-level ACOS. It’s catalog-level ACOS. The right way most agencies report is the wrong way.
Why this matters more than it sounds
The halo effect (sales of one ASIN driven by ads for a different ASIN) is real, measurable, and almost always present in multi-product accounts.
When an agency reports ASIN-level ACOS without surfacing the halo, three things happen.
First, profitable campaigns look unprofitable. Agencies recommend cutting them. The agency looks proactive. The brand loses revenue.
Second, the wrong ASINs get the credit. Top-line “best performers” in agency reporting are usually the ASINs receiving halo sales, not the ones driving them. The brand allocates more budget to the recipients. The drivers get starved.
Third, the brand can never see the actual unit economics of an ad campaign. They see fractions of the picture, sliced ASIN by ASIN, none of which adds up to reality.
Apparel example. The audit data showed 93 to 96% of ad-driven purchases landing on a different ASIN than the one being advertised. Almost all the sales came from the 3-pack bundle, which received almost no direct ad spend. The single-pack ASINs being advertised looked individually unprofitable. The bundle looked organically successful. Neither was true. The ads were driving the bundle sales. The reporting just didn’t connect the dots.
What to do about it
Three things.
First, ask your agency for catalog-level ACOS, not just ASIN-level. They might not be able to produce it cleanly. That’s the first signal.
Second, ask them to surface the halo share for each major ad-supported ASIN. If a campaign is generating 50% of its sales on other ASINs, you need to know.
Third, when evaluating whether to cut a campaign, use the catalog math, not the ASIN math. Most brands have one or two ASINs serving as catalog entry points (lower-AOV products that introduce the brand). Those are usually the ones whose individual ACOS looks worst and whose actual contribution looks fine once the halo is counted.
The deodorant brand I started this essay with would have cut the Limitless campaign on the old reporting. The catalog math said keep it. Same campaign. Same data. Two different conclusions depending on which view your agency builds.
Most agencies build the ASIN view because it’s easier. The catalog view is the one that pays the bills.
Want to ask your agency these questions?
I put together a one-page diagnostic with five sharp questions you can send to your current agency. If they can’t answer in 24 hours with real numbers, you have a diagnostic gap.