175 Billion Impressions Show Shopping Ads Being Served Less and Clicked More. Your Dealership's Rising Click Rate Deserves a Second Look.

Somebody just measured a pattern that will show up in your dealership's Google Ads account as good news.

Mike Ryan, head of ecommerce at Smarter Ecommerce, published an analysis on LinkedIn this week covering roughly 175 billion impressions across thousands of Shopping and Performance Max campaigns from hundreds of advertiser accounts. Between mid-2025 and mid-2026, median Shopping ad impressions fell from roughly 1.85 million to 1.4 million per account. Over the same window, median click-through rate climbed from 1.20% to nearly 1.55%. Barry Schwartz covered it at Search Engine Roundtable and Anu Adegbola covered it independently at Search Engine Land on the same day.

The part that matters is what the clicks did. According to Search Engine Land, across that dataset "clicks were often relatively flat or slightly down, while impressions had fallen more noticeably." The click rate rose because the number underneath it got smaller.

Ryan framed the oddity himself on LinkedIn: "Here's an odd one. Ever since AIOs launched, publishers have been gaining impressions and losing clicks: the dreaded SEO 'crocodile effect.' For Shopping ads, it's the opposite. But why?"

Start with the boundary, because your vendor will not

This dataset is ecommerce. Shopping and Performance Max advertisers selling products. There is not a single vehicle listing ad in the figures above, and nobody has run this analysis against automotive inventory. Anyone who forwards you this study as proof of what happened to your store's vehicle ads is overselling it.

Here is why it still lands on your desk. Google's own Merchant Center documentation says vehicle ads require you to enable the program in Merchant Center, upload vehicle data sources, link a Google Ads account, and then create "Performance Max or Standard Shopping campaigns" in Google Ads. Those are the same two campaign types Ryan measured, running in the same auction, against the same AI Overviews, in a program Google confirms is live in the US. Different advertisers, identical machinery.

So treat the numbers as a description of the machinery your vehicle ads run on. Do not treat them as a measurement of your store.

Every dealer already knows this arithmetic

Think about a direct mail drop. You used to mail 50,000 households and get 500 responses, a 1% response rate. This year you cut the list to 30,000 and targeted only past customers and in-market equity owners. You got 480 responses. Your response rate jumped to 1.6% and the number on the report looks fantastic.

You sold the same number of cars. You just stopped mailing the people who were never going to respond.

That is the entire finding. A rate is a fraction, and there are two ways to make a fraction bigger. Shrinking the bottom half is the boring one, and it is the one that has been happening in Shopping for a year.

We wrote about the mirror image of this two days ago. Google's new Search Generative AI performance report gives your dealership impressions and withholds clicks on the organic side, where sites have been gaining impressions and losing clicks. Organic is being shown more and clicked less. Paid Shopping is being shown less and clicked more. Both of those trends make the headline metric move in the flattering direction while the thing you actually bank stays flat.

Why it might be happening, in Ryan's words

Ryan's theory, quoted directly: "Google checks the predicted CTR of a given query and serves AIOs preferentially for lower-probability queries in order to preserve revenue. Thus CTR climbs partly due to a shrinking denominator (fewer impressions) and partly due to selection/pruning (higher click propensity)."

In plain terms, Google looks at a search, estimates how likely somebody is to click an ad on it, and hands the low-probability searches to an AI Overview while keeping the high-probability ones for ads. Ryan calls the broader idea "experience switching," and he told Search Engine Land that anecdotally he rarely sees Shopping ads and AI Overviews together on the same results page. Usually it is one or the other.

He also thinks it is temporary. He describes the current state as a "bridging strategy" and expects Google to shift "from more 'either/or' experiences toward more 'both/and' experiences" as it launches AI-native ad formats and scales the presence of Shopping ads inside AI Overviews. We already covered the leading edge of that, when Google confirmed exact match keywords can serve inside AI Mode.

Now the honest part. Ryan says plainly that the data does not prove AI Overviews caused any of this. Search Engine Land put it directly: "For now, it's a hypothesis." The timing could be coincidental, the effect could run through some other mechanism, or something else in the ad auction between 2025 and 2026 could be responsible. A third write-up of the analysis makes the same point, that this establishes correlation and Google has confirmed nothing.

The correlation is worth a lot less than the arithmetic. You do not need to believe Ryan's causal story to act on this, because the measurement stands on its own: impressions fell, clicks did not follow them up, and the rate rose as a result. That happens whether the cause is AI Overviews or something else entirely.

What this means for your dealership

Your CTR trend line is no longer a performance signal by itself. If your vehicle ads or Performance Max click rate is up year over year and your team is reporting that as an optimization win, ask what impressions did over the same period. If impressions fell faster than clicks, nobody optimized anything.

Impression volume becomes the number you watch. Search Engine Land's guidance for advertisers is to "look beyond CTR and pay closer attention to impression volume, clicks and overall traffic when judging performance." That is a reporting change, and it is one most dealer ad reports have not made.

A quiet impression decline looks like an efficiency gain. Fewer impressions with flat clicks and a steady budget produces a higher CTR and often a stable cost per click. Every headline metric in the monthly deck stays green while your total reach into the market shrinks. This is the failure mode where a store keeps spending the same money and slowly stops being seen by anyone who has not already decided to buy.

Bid and budget decisions made off CTR will point the wrong way. Raising budget because CTR improved assumes there is more of that good traffic to buy. If Google pruned the query set, the extra money chases inventory that is not there and lands in worse placements.

Do not confuse this with your organic AI visibility problem. They are separate surfaces with opposite symptoms, and we have covered how Google's two shopping surfaces already disagree with each other. Diagnose paid and organic separately or you will fix the wrong one.

What to do about it this week

1. Pull a 24-month view of your vehicle ads and Performance Max campaigns. Chart impressions, clicks and CTR on the same axis. If the CTR line rises while the impression line falls and the click line is flat, you have the pattern in your own account.
2. Add impressions and total clicks to your standing dealer report. If your vendor's monthly summary leads with CTR and cost per click and omits impression volume, that report can no longer tell you whether you are losing reach.
3. Check Search impression share and lost impression share (rank and budget). This separates "Google showed my ads less" from "I got outbid," which CTR alone cannot do.
4. Segment by query intent before you conclude anything. Model and trim searches behave differently from broad research searches. If the losses concentrate in the upper-funnel queries, that is consistent with what Ryan described, and those are the shoppers you were reaching earliest.
5. Audit your vehicle feed anyway. Missing or invalid required attributes shrink your eligible impressions for reasons that have nothing to do with AI Overviews, and it is the cheapest thing on this list to rule out.
6. Stop rewarding a rate. If anyone at your store or your agency is compensated or evaluated on click-through rate, this year handed them a raise for nothing.

The part worth watching

If Ryan is right that this is a bridge, the pattern reverses on its own once Google puts Shopping ads inside AI Overviews at scale. Impressions come back, the click rate falls toward its old level, and a lot of dealers will read that as their advertising getting worse in a month when nothing about their advertising changed.

That is the real cost of a metric that moves for structural reasons. It sends you the wrong signal on the way up and the wrong signal again on the way down. The stores that come through this well will be the ones already reading impressions, clicks and rate together, so that when the denominator moves again they can see it happen instead of reacting to it.

If you want a second set of eyes on what your paid search numbers have actually been doing for the last two years, book a time with us and we will go through the account with you.