Ads in AI Overviews: What Happens to Shopping Ad CTR Now
Shopping ad click-through rate is going up for a lot of advertisers right now. That sounds like good news and mostly it is not. Impressions are falling faster than clicks, which pushes the percentage up while the actual number of visits stays flat or drops. If your Google Ads report shows a healthier CTR this month, this is the first thing to check.
- What actually changed
- The “reverse crocodile” pattern
- Why CTR rises when impressions fall
- The reporting trap
- How to check your own account
- What to do about it
- Why search ads behave differently
- Frequently asked questions
What actually changed
AI Overviews have expanded onto commercial queries, including the kind of product searches that used to return a clean row of Shopping ads. Visibility Labs studied 20.9 million shopping keywords and found AI Overviews appearing on roughly 14% of them, up more than fivefold from late 2024.
Where an Overview appears, the layout above the fold changes. The panel takes the top of the screen, and the ad units that used to sit there move down or, on some queries, do not serve at all. Search Engine Land reported the effect as AI Overviews reducing Shopping ad impressions while concentrating clicks among higher-intent searches.
That second half of the sentence is the part worth slowing down on, because it is what makes the metrics look strange.
The “reverse crocodile” pattern
Search Engine Roundtable gave the shape a name: a reverse crocodile. In the original crocodile chart, impressions rise while clicks stay flat, so the two lines open like a pair of jaws. In the reverse version, impressions steadily decline while median CTR steadily increases, so the jaws close from the other direction.
Relevant Audience reported Shopping ad CTR up around 20% while impressions fell over the same period. The reading offered there is straightforward: the ads are being shown less often on lower-intent queries where AI Overviews now appear, so CTR rises without extra clicks arriving.
The distinction that matters: a higher CTR because more people are clicking is good. A higher CTR because fewer people are being shown the ad is not. The percentage cannot tell you which one happened. Only the raw click count can.
Why CTR rises when impressions fall
This is a composition effect, not a performance improvement.
Your Shopping impressions were always a mix. Some came from people ready to buy a specific product. Others came from vague, early-stage or informational queries where the ad appeared but almost nobody clicked. Those low-intent impressions dragged the average CTR down, which is normal and was never a problem in itself.
AI Overviews now absorb a large share of exactly that low-intent traffic. The queries where somebody was asking rather than buying are the ones getting answered on the results page. So those impressions disappear from your account.
What remains is a smaller, higher-intent pool. The same buyers click at the same rate they always did, but the denominator shrank, so the percentage rises. Nothing about your campaign got better. The measurement floor moved.
The reporting trap
This is the situation that produces a green dashboard and a quiet phone, and it is worth naming because it is easy to walk into.
A monthly report shows CTR up 20% and cost per click roughly flat. Read on its own, that looks like a win, and it will be presented as one. Meanwhile total clicks are down, sessions are down, and the number of orders has slipped in a way that gets attributed to seasonality or to a competitor.
The mistake is treating CTR as an outcome. It is a ratio, and a ratio can improve because the numerator rose or because the denominator fell. Only one of those puts money in the business.
| What you see | What it probably means | Good or bad |
|---|---|---|
| CTR up, impressions up, clicks up | Genuine improvement | Good |
| CTR up, impressions down, clicks flat | Low-intent impressions removed | Neutral |
| CTR up, impressions down, clicks down | Losing real volume | Bad |
| CTR down, impressions up, clicks up | Broader reach, still working | Usually fine |
| CTR down, impressions flat, clicks down | Relevance or position problem | Bad |
How to check your own account
Twenty minutes in Google Ads settles it. Do not rely on anybody else’s percentage, including the ones quoted above.
- Compare year on year, not month on month. Shopping is seasonal enough that a month-over-month comparison will mislead you.
- Put impressions, clicks, CTR and conversions in the same view. The story only appears when they sit side by side.
- Look for the signature: impressions down, CTR up, clicks flat or down. If clicks held steady while impressions fell, you lost waste. If clicks fell too, you lost demand.
- Segment by search term where you can. The queries that disappeared tell you whether you lost buyers or browsers.
- Check impression share and where it was lost. Budget-limited and rank-limited are different problems; an Overview taking the slot is a third thing again, and it will not always be labelled clearly.
- Finish on revenue per product group. That is the number that cannot be flattered by a shrinking denominator.
What to do about it
Most of the useful responses are unglamorous.
- Stop reporting CTR without impressions beside it. On its own it is now actively misleading. Any report that shows one without the other should be rebuilt.
- Judge campaigns on clicks and revenue. Absolute numbers do not move because of a layout change; ratios do.
- Do not cut budget because impressions fell. If clicks and orders held, you are paying for the same demand with less waste in front of it. That is fine.
- Do look at the feed. When fewer queries reach you, the ones that do matter more. Titles, product types, images, price competitiveness and stock availability decide whether you appear on those remaining high-intent searches.
- Rebase your targets. A target CPA or ROAS set against last year’s traffic mix is now set against a different mix. Leaving it untouched quietly changes how aggressively the system bids.
- Expect CPCs to firm up. Fewer, better queries means more advertisers competing for the same impressions. A rising cost per click here is not necessarily a failure, but it should be planned for rather than discovered.
None of that needs a new tool. It needs somebody looking at the account every week and acting on the right number, which is the whole job of ongoing PPC management rather than a one-off audit.
Why search ads behave differently
Worth separating the two, because the same headline gets applied to both and the mechanics are not identical.
For standard search ads, the widely reported effect runs the other way: on queries where an AI Overview appears, click-through rate tends to fall, sometimes sharply. Neil Patel’s write-up puts the reduction above 50% for affected queries and argues that impression share becomes the metric to watch.
Shopping behaves differently because of what it is. A product listing with a price and an image answers a buying question in a way a generated paragraph cannot. So the Overview absorbs the research query and leaves the purchase query, which is why the remaining Shopping impressions convert at a higher rate rather than a lower one.
The practical consequence is that a single “AI Overviews are hurting ads” narrative is too blunt to act on. It depends on the campaign type and the query mix. The wider picture for organic sits in the companion piece on AI Overviews and SEO, and the ongoing management side is on the Google Ads page.
For the management side in the UK, the PPC agency Leicester page covers how campaigns are built and reported, and if you are weighing ads against organic, the UK SEO cost guide has a section on SEO versus Google Ads for a smaller budget.
To see what a change in Shopping click-through rate or conversion rate does to the money, run the account through the free PPC ROI calculator.
Frequently asked questions
Are ads actually shown inside AI Overviews?
Google has been expanding ad placements in and around AI Overviews and AI Mode. From an advertiser’s point of view the practical change is where your ad sits on the screen and how often it serves, rather than a new campaign type to opt into.
My Shopping CTR went up. Is that good?
Only if clicks went up too. Check impressions in the same view. If impressions fell and clicks are flat, the rise is arithmetic rather than performance.
Should I pause campaigns where impressions dropped?
Not on impressions alone. If clicks and revenue held, you are simply appearing on fewer low-intent searches. Pause on lost revenue, not on lost impressions.
Does this affect every product category?
No. AI Overviews appear on roughly 14% of shopping keywords studied, and the distribution is uneven. Categories with more research-style queries are affected more than categories where people search by exact product or model number.
What is the single number to watch?
Revenue per product group, or cost per acquisition if you sell leads rather than products. Both are immune to the denominator problem that makes CTR unreliable at the moment.