2026-11-04 · Étienne Laurent
Every Amazon seller tracks ACoS. Almost none of them can answer a harder question: what would sales look like if the ad spend stopped tomorrow? That gap between "campaign ACoS looks fine" and "I actually know what my advertising is buying me" is where most European accounts leave money on the table — not in bid strategy, in measurement. With BFCM three weeks out, now is the window to fix how you measure, before the surge makes every number noisier.
Campaign-level ACoS tells you whether a single campaign's ad spend was profitable against that campaign's attributed sales. It says nothing about halo effect, nothing about organic lift, and nothing about whether the campaign cannibalised a sale you would have gotten anyway. Most sellers manage their entire advertising strategy from this one number, which is like steering a car by watching only the speedometer on one wheel.
TACoS (Total Advertising Cost of Sale — ad spend divided by total revenue, not just attributed revenue) forces a different question: is overall revenue growing faster than ad spend? Pull a rolling 8-week view per marketplace and watch the ratio. A campaign with a mediocre ACoS that is clearly pulling your organic rank up — visible as falling TACoS over consecutive weeks even as ad spend holds steady — is working. A campaign with excellent ACoS but flat TACoS is probably just capturing demand that existed already. This is the level where most of our Avidsen and Blaupunkt reviews already operate, and it's the minimum bar before BFCM.
Sponsored Display and DSP rarely get credited for what they actually do: push traffic to detail pages that convert on other ASINs in the catalogue, or convert days later through a direct search instead of an ad click. Standard attribution reporting misses this almost entirely — it's built around last-click, seven-day windows. To see the halo, compare total catalogue sell-through velocity in weeks with active Sponsored Display/DSP spend against matched weeks without it, isolating for promotions and seasonality. Brands running multi-SKU catalogues (gate motors plus accessories, projectors plus mounts) see the clearest signal here, because the halo lands on adjacent SKUs that a single-ASIN attribution report will never surface.
The top of the model, and the one almost no mid-sized European seller runs: a deliberate holdout. Pause a specific campaign type (commonly Sponsored Display remarketing, since it's the easiest to isolate) in one marketplace for two weeks while keeping it live in a comparable marketplace, then compare the revenue delta against the historical baseline for both. This answers the actual question — not "what sales does the ad console attribute to this campaign" but "what sales would not have happened without it." It is the only level that tells you whether a channel is additive or redundant with organic demand you already had.
A seller running FR, DE, IT, ES and UK separately usually builds five versions of Level 1 reporting and stops. The attribution maturity model gets harder to skip at scale, not easier — because halo and incrementality effects differ by market. German Sponsored Display remarketing often shows a stronger halo than French, driven by longer consideration cycles on higher-ticket items; Spanish and Italian markets show more same-session conversion and less delayed halo. A consolidated cross-market view at Level 2 is necessary just to stop making the five-market budget decision on five different scales of noise.
Don't try to reach Level 4 overnight. Move one level up before BFCM distorts every number for six weeks:
ACoS will keep telling you whether this week's campaigns paid for themselves. It was never going to tell you whether advertising is actually growing the business. Those are different questions, and only one of them survives contact with BFCM-season noise.
This article was drafted by Étienne Laurent (Head of Strategy & AI, Miumx) based on Miumx's internal Amazon operations data.
AI involvement:
Final content was reviewed, edited, and verified for accuracy by Étienne Laurent, Head of Strategy & AI, Miumx.
EU AI Act Article 50 compliance · Reviewed 2026-11-04