Google Shopping Impressions Are Dropping — But CTR Is Rising

A weird pattern in Google Shopping data

Search Engine Roundtable flagged something worth paying attention to if you run paid or organic e-commerce visibility: Google Shopping ads are showing what’s been called a “reverse crocodile” effect. Normally, when impressions and CTR are plotted together and one line spikes while the other drops, it looks like an open crocodile mouth — usually a bad sign, meaning you’re getting seen more but clicked less. What’s happening now is the opposite. Impressions on Shopping ads are falling, but click-through rate is climbing. Fewer eyeballs, more engaged clicks.

That’s not noise. It’s a signal about where Google is routing query volume and why.

Why this is probably happening

The working theory, and it’s a sound one, is that AI Overviews are absorbing a chunk of the lower-intent, exploratory product queries — “best running shoes for flat feet,” “what’s a good budget espresso machine” — the kind of searches where someone isn’t ready to click through to a retailer yet. Google appears to be resolving those with a synthesized answer instead of a page of Shopping tiles.

What’s left in the traditional Shopping ad slot is a more qualified pool: people who already have some intent and are closer to comparing prices or clicking to buy. Fewer people see the ads overall, but the ones who do are more likely to act. Impressions shrink, CTR grows. That’s the mechanism behind the reverse crocodile.

This lines up with the broader pattern of AI Overviews suppressing traffic on informational and comparison-style queries while leaving transactional intent largely intact — something we’ve seen across other verticals, not just Shopping.

Why this matters beyond the metric itself

The interesting part isn’t the CTR bump — it’s what it implies about where this is heading. The theory behind the reverse crocodile effect suggests this could be a bridging strategy: as Google leans on AIOs for lower-probability queries to preserve revenue, the natural next step is scaling the presence of Shopping ads directly inside AIOs. If low-certainty queries are being routed to AIOs and Google moves toward embedding product results inside those same AIOs, the logical endpoint is that Shopping ads stop living exclusively in their own SERP module and start becoming a native component of the AI-generated answer itself.

For e-commerce teams, that reshuffles the whole model of how product visibility gets earned. Right now, feed optimization is built around ranking for keyword-matched queries in a fairly stable ad auction. If Shopping placements migrate into AIOs, the criteria for showing up shift toward whatever signals the AI answer engine uses to decide which products are worth surfacing — likely a mix of structured product data, review depth, pricing clarity, and consistency across the feed rather than pure bid strategy.

What to actually do about it

A few concrete moves make sense now, before this becomes the default:

• Don’t panic over impression drops in isolation. Look at the impression-to-CTR-to-conversion chain together. A shrinking top of funnel with rising conversion quality can be a net positive if your margins support it.
• Shift optimization targets from volume to click quality and conversion rate on the traffic you do get. Impressions were never the goal; revenue per impression is.
• Avoid over-indexing on Shopping ads as a single channel. If Google is actively reshaping how much of that inventory gets exposed to top-of-funnel queries, treat organic product visibility, comparison content, and other paid channels as real hedges, not backups.
• Keep an eye on how Shopping ads inside AI Overviews evolve — pilot markets, product categories, and placement formats. The feed hygiene and structured data work that helps today will likely be the same foundation that determines inclusion in AI-native product surfaces tomorrow.
• Get product titles, attributes, and descriptions in shape for machine consumption, not just keyword matching. Clear specs, accurate pricing, and complete structured markup are what let an AI system confidently include a product in a generated answer.

None of this requires abandoning Shopping ads. It requires treating the current dip as a preview of a bigger structural shift, and building feed and content practices that hold up whether the product surface is a SERP module or a paragraph generated by an AI system. If you want a clearer read on how AI Overviews are reshaping your specific query set, our GEO audit is a good place to start.