AI Overviews Cut Top-Ranked Clicks 58%, Ahrefs Study of 300,000 Keywords Finds

Ahrefs analyzed 300,000 keywords and found the top-ranking page earns 58% fewer clicks when an AI Overview appears above it. A separate randomized study of Chrome users measured a 39.8% drop in outbound clicks — the first causal evidence of AI answers suppressing traffic, not merely a correlation. Those two figures anchor a new guide from Mansi Rana, founder of EZ Rankings, published on TechBullion on the state of brand visibility in AI answer engines.

How large is the AI answer surface?

The scale figures come from the platforms themselves. Google stated at I/O in May 2026 that AI Overviews now reach more than 2.5 billion people per month. AI Mode crossed 1 billion monthly users within a year of launch, also per Google's shared figures. ChatGPT reported 900 million weekly users in February 2026.

The consequence, Rana argues, is that a brand can hold the number one organic position and still be absent from the answer a buyer reads. She frames the guide around that gap, which she calls the most frequent conversation with marketing leaders in 2026.

What does a citation actually mean?

The guide distinguishes three outcomes: being used as a source, being named, and being recommended. A Semrush and Kevin Indig analysis found 61.7% of observed citations were "ghost citations" — the AI drew on a page but never mentioned the brand behind it. Rana's position: the named recommendation is the real prize, because it shapes the buyer's shortlist.

Which signals drive AI visibility?

The correlation data points away from classic link-building. In Ahrefs' study of 75,000 brands:

• Branded web mentions correlated with AI Overview visibility at 0.664.
• Backlinks correlated at only 0.218.
• Mentions in YouTube titles, transcripts and descriptions showed the strongest single relationship, at 0.737.
• A separate Moz analysis found 88% of AI Mode citations do not appear in organic results at all.

Rana's reading is restrained: rankings still matter as a retrieval foundation, but reputation across sources a brand does not own now decides who gets named. Takeaway: independent consistency beats link counts.

What is the 4E Answer Readiness Model?

EZ Rankings assesses brands across four layers, claiming a weakness in any one caps the other three.

• Entity clarity. One consistent brand description across the site, Google Business Profile, LinkedIn, Crunchbase, Wikidata and directories. Organization and Product schema should match that description word for word.
• Extractability. Growth Memo's analysis of 1.2 million ChatGPT citations found most come from the first third of a page. Lead sections with direct, quotable answers, question-style headings, tables and numbered steps.
• Evidence. Original data, pricing ranges, benchmarks and named methodologies give an engine a reason to cite a page rather than a generic summary.
• Echo. Earned media, podcasts, YouTube appearances and community discussion. Rana calls this the most neglected layer and the one the correlation data rewards most.

Does citation pay beyond the click?

Seer Interactive's 2026 data showed brands cited in an AI Overview earned a 15.74% paid click-through rate on the same queries, against 11.19% when not cited. Rana recommends tracking AI visibility as its own metric set: AI share of voice, citation rate, accuracy and sentiment, and downstream lift in branded search, direct traffic and assisted conversions.

What timeline is realistic?

Rana warns against promises of AI visibility in weeks. Retrieval-led surfaces such as Perplexity and AI Overviews can reflect improved entity data and fresh content in roughly two to three months. Changing how a model describes a category from memory requires sustained third-party coverage and typically takes six months or more.

Her caveats are worth noting. No major engine has confirmed llms.txt as a ranking or citation factor. Mass-produced AI-generated pages rarely earn citations. Click data alone understates AI visibility, and inaccurate AI descriptions of a brand demand fixing at the source.

The skeptic's note: most figures here are tool studies and vendor correlations, not confirmed engine statements, and Rana's firm sells the audits the guide recommends. Still, the causal 39.8% click drop and the 0.737 YouTube mention correlation are the numbers to watch. Monitor whether Google or OpenAI ever confirms citation factors directly — until then, treat every visibility model as informed speculation.