The Maze: A better click-through rate can hide a smaller opportunity to win customers. In a Google Shopping analysis covering 175 billion impressions, median ad impressions fall while the share turning into clicks rises. The pattern spans January 2025 to roughly August 2026, across AI Overviews' European expansion. It is tempting to call that better targeting. It could also mean fewer low-probability impressions survive. The difference matters: retailers buy access to customers, not attractive percentages on a dashboard.
The ratio improves as the available volume contracts. Median impressions move from roughly 1.9 million to 1.4 million, while median click-through rate rises from about 1.2% to 1.6%. These are approximate levels, read from the published series rather than a downloadable data table. Impressions stay relatively steady through spring 2025 before declining; click-through rate improves more clearly from autumn. The sample is large, but its country coverage and the unit behind each median are undisclosed. This is an observed sample trend, not a census of Google's advertising market.
A rising rate does not establish rising clicks. Click-through rate divides clicks by impressions. Remove impressions that rarely earn a click and the ratio can improve without attracting another shopper. The remaining audience may be more inclined to click, but that does not tell a retailer whether its total customer pipeline expanded. Nor can we recover the answer by multiplying these two medians: they need not describe the same account, and their product is not the median number of clicks. Actual click totals are needed before celebrating traffic growth. Orders, conversion rates and profit require their own evidence too.
The AI explanation remains a testable hypothesis. Mike Ryan of Smarter Ecommerce, an advertising software and services business, proposes that Google may favor AI Overviews for queries with a lower predicted advertising click rate. That could leave Shopping ads with a smaller, more clickable query pool. Google's March expansion reached nine additional European countries, giving the timing a plausible anchor. It does not prove the mechanism. Query composition, advertiser budgets or sample changes could also matter. We would need query-level exposure and a credible comparison group to isolate the effect of AI answers.
Google's economics and a retailer's traffic can diverge. During the July 2025 earnings call, management described AI Overviews as monetizing at roughly the same rate. That historical statement is not a promise of stable Shopping-ad reach or better retailer margins. A platform can preserve monetization while an individual advertiser sees fewer opportunities; auction prices and the kinds of queries shown can also change. For operators, the useful follow-up is a consistent-account comparison of impressions, clicks, cost per click, conversion rate and contribution profit. Segment by query intent before crediting creative quality for a better ratio.
Why it matters: A higher click rate is useful only alongside the volume and value of the clicks. Before shifting budget, establish whether the same campaigns are reaching fewer searches, buying more expensive visitors or producing more profitable orders. The AI-selection theory deserves investigation, but retailers do not need to settle Google's motives to improve their reporting. Put impressions and absolute outcomes beside CTR. A denominator can shrink faster than a business grows.
