Google Shopping Ads: fewer impressions, more money. Sound familiar?
The reverse crocodile — nothing to do with reptiles
There's a term — "crocodile effect": when the graph's jaws split apart, impressions climb, but clicks or revenue drop. Everyone's used to that one. But right now Google Shopping Ads is showing the reverse pattern — and almost nobody's ready for this version.
Impressions on product ads are shrinking. CTR is going up at the same time. Fewer ads get shown, but people click them more readily and more precisely. On the surface that sounds like good news. In practice, it's a signal that the rules of the game already changed — quietly.
What's actually happening
Strip away the noise and the logic is simple. Google has learned to tell "just browsing" queries apart from "wallet already open" queries. On low-intent, vague, research-y searches, it increasingly serves an AI Overview — a full answer right in the results. It saves Shopping ads for queries that are clearly commercial and ready to convert.
Net result: total product impressions drop — because AIO ate a chunk of the queries. But the impressions that remain are far sharper — because that's filtered, hot traffic. CTR isn't rising because your creative got better. It's rising because Google did the initial traffic filtering for you.
This isn't a bug or a temporary dip. It's the early stage of something much bigger: Shopping ads getting built directly into AI Overviews. Not next to them, not below them — inside them. Google is already signaling it's ready to embed product cards straight into the AI answer itself, not just in the block above organic results.
What to do if you run a product site
First — stop panicking about falling impressions as a number on its own. If that drop comes with rising CTR and conversion, that's not a loss, that's filtering working in your favor. Watch conversion and revenue per click, not raw reach. Impression volume is a vanity metric if it doesn't turn into money.
Second — start optimizing your product feed for offer quality, not "maximum impressions." Accurate attributes, honest prices, clear descriptions that parse well for both humans and algorithms. The clearer your product is to the machine, the better your odds of being that "filtered" impression inside AIO down the road.
Third — and this matters — don't build your business on one leg. If your entire store's traffic depends on Google Shopping, you're at the mercy of one company's decision about who gets shown inside its AI answer, and when. Channel diversification isn't a buzzword from a slide deck — it's literal insurance against Google reshuffling the format again tomorrow, and you finding out after the fact from a falling graph.
Fourth — start treating your product assets as raw material for AI-native formats. Descriptions, specs, reviews, structured data — all of it will eventually feed the generation of answers inside AIO. Whoever preps their feed for this early gets the spot in the new results once the integration goes mainstream.
Where this is all heading
Google isn't just bolting AI on top of search — it's gradually rebuilding the mechanics of who gets shown ads, and when. The rules used to be transparent: bid, relevance, ad quality. Now there's a new player in the equation — the AI Overview itself, deciding whether to answer the query in its own text or pass it along to a Shopping ad.
If you run product projects, look at this soberly: fewer impressions isn't a catastrophe if quality is rising. But dependence on a single traffic distribution channel is a risk that hasn't gone anywhere — it just changed shape. If you want to figure out how to rebuild your feed and strategy for this new distribution mechanic, we can dig into it during an audit — details on the services page.