Google AI Overviews Cut Clicks Even for Cited Brands

A citation is not a click

The pitch for optimizing content for AI answers usually rests on one assumption: if Google's AI Overview cites you, you've won. The latest click-through data complicates that. Being cited inside an AI Overview helps compared with being left out, but it still leaves publishers and brands well below the traffic they earned before the summaries appeared.

The clearest data point comes from a roundup of 2026 research. When no AI Overview appears, average click-through rate sits around 3.3%. When an Overview appears and cites the page, it falls to 2.1%. When the page isn't cited, it drops to 0.9%.3 Being cited is better than being skipped, but by that math a cited page still gives up roughly a third of its clicks compared with a results page that has no summary at all.

The 58% figure and what it measures

The most widely repeated number comes from Ahrefs. It reran its study on December 2025 data and found that an AI Overview now correlates with a 58% lower average click-through rate for the top-ranking page.2 The method used aggregated Google Search Console data to calculate desktop CTR by keyword. It then used the CTR trend for informational keywords without Overviews to estimate what CTR would have been had Overviews never launched. That estimate was about 3.7%, against an actual figure near 1.6%.2 Ahrefs summed it up bluntly: for every 100 clicks a top-ranking page could once earn, Google now keeps 58.2

One analysis of the Ahrefs study stresses that this is nearly double the 34.5% decline the same research found eight months earlier. That suggests the effect is accelerating rather than levelling off.1 It also describes the drop as a fall from roughly 7.3% to 1.6% for the number one result.1

Ahrefs adds a caveat that often gets lost. CTR has also fallen for keywords without AI Overviews. It places the Overviews in a long line of Google features, such as Local Packs and Top Stories, built to answer queries without sending users elsewhere.2

Why the statistics seem to disagree

Anyone following the debate has likely seen several different numbers, including 25%, 58% and 68%, cited as if they were interchangeable. They aren't.1 Some track referral traffic at particular publishers, while others track CTR at a single ranking position.1 Mixing them produces confusion more than insight.

Timing matters too. Seer Interactive tracked 53 brands, 5.47 million queries and 2.43 billion impressions from January 2025 to February 2026. It found CTR on AI Overview queries bottomed at 1.3% in December 2025, then recovered to 2.4% by February 2026.3 Ahrefs captured December 2025, which may be the trough of the decline. The Seer recovery is real but partial, and two months of data is too short to call a trend. Meanwhile, AI Overviews now appear on more than one in five Google searches.3

Who actually gets cited

Even if citation were enough, it is concentrated among a few sites. An index combining six studies and more than 680 million citations from August 2024 to April 2026 found that Reddit, Wikipedia, YouTube, LinkedIn and Forbes lead across AI engines. Together with ten other domains, they account for roughly 68% of citations across ChatGPT, Claude, Gemini, Perplexity and Google AI Overviews.4 Reddit appears in around 40% of citations, and Wikipedia makes up 26–48% of ChatGPT's top-ten sources.4

For a mid-sized publisher or brand, that means competing for a small slice of citations, and then earning fewer clicks even after winning one.

The AEO argument, and its limits

Marketers have responded with "answer engine optimization," the practice of shaping how a brand shows up inside AI-generated answers. One LinkedIn post captures the logic: AI pulls from reviews, Reddit threads, podcasts and comparison articles. Brands with a thin online presence get thin answers. Because much of the buying decision now happens before any click, visitors increasingly arrive only to confirm what an AI already told them.5 The post also points to HubSpot's AEO tool for measuring this.5

That reasoning fits the data, though not quite how its advocates frame it. If decisions are made inside the answer, citation works more like brand exposure than a traffic source. Its value has to be measured in awareness and downstream conversions rather than sessions.

The bottom line

Citation in an AI Overview should be treated as damage control, not a growth strategy. It recovers part of the clicks the summary takes away, but none of the reliable data shows it restoring pre-Overview traffic. Publishers that plan around a return to old CTR levels are likely to be disappointed. The more realistic approach is to treat AI visibility and site traffic as two separate goals, measured separately.