Google on AI Overview Traffic Loss

Google’s Head of Search, Liz Reid, has continued to defend the impact of AI Overviews on publisher traffic, arguing that the feature is not removing meaningful visits but instead reducing what she describes as “bounce clicks”. Her comments were made during an appearance on Bloomberg’s Odd Lots podcast, where she revisited points she has been making publicly since AI Overviews were introduced more widely.

The idea of “bounce clicks” refers to users who click on a search result, briefly view a page, and then quickly return to Google. According to Reid, these are users who were mainly seeking a quick fact or short answer rather than in-depth information. With AI Overviews now providing summaries directly within the search results page, she suggests that some of these users may no longer need to visit external websites at all.

In her explanation, Reid emphasised that this shift does not necessarily mean a loss of engaged traffic. She argued that users who want more detailed information will still click through to publisher websites in the same way as before. In other words, the traffic that remains is, in her view, more meaningful and intentional.

She also acknowledged that there may be a reduction in ad clicks for certain types of queries. However, she suggested that this is balanced out by an increase in overall search activity. From Google’s perspective, the introduction of AI Overviews may be changing user behaviour rather than reducing total engagement with search.

This is not a new narrative from Google. In an earlier blog post published in August, Reid stated that overall organic click volume from Google Search to websites had remained relatively stable year-on-year. She also introduced the idea of “quality clicks”, which she defined as visits where users do not immediately return to search results.

Later, in an interview with the Wall Street Journal, she again used similar language around “bounced clicks” and suggested that advertising revenue had remained broadly stable following the rollout of AI Overviews. Across these appearances, Google has maintained a consistent position that the nature of clicks is changing rather than overall traffic collapsing.

However, one of the key criticisms of this argument is the lack of publicly available data to support it. In both the blog post and subsequent interviews, Google has not provided detailed figures, charts, or breakdowns showing how different types of clicks are measured or classified. This has made it difficult for external observers to independently verify the claims being made.

During the Bloomberg interview, Reid noted that Google does track how often users return to Search, describing it as an important internal signal. However, no specific data or quantitative evidence was shared to illustrate how this behaviour has changed since the introduction of AI Overviews.

This absence of detailed transparency has led to a reliance on independent research to understand the wider impact on traffic and engagement. Several external studies suggest that the overall effect on publishers may be more significant than Google’s framing implies.

For example, data published by Chartbeat as part of the Reuters Institute’s Journalism and Technology Trends and Predictions 2026 report found that global publisher traffic from Google Search declined by roughly one-third. In the same analysis, referrals from Google Discover were also reported to have fallen by around 21% year-on-year across thousands of publisher websites.

Other research points to similar trends in click behaviour. Analysis from Seer Interactive found that click-through rates for queries featuring AI Overviews dropped sharply, falling from 1.76% in 2024 to 0.61% in 2025. This represents a decline of around 61%, particularly affecting informational search queries that traditionally rely on organic clicks for traffic.

Further evidence from the Pew Research Centre, which examined 68,000 search queries, found that users clicked on a result just 8% of the time when AI Overviews were present, compared with 15% when they were not. This suggests that the presence of AI-generated summaries may significantly alter user engagement patterns.

In addition, Digital Content Next, a trade association whose members include major publishers such as The New York Times, Condé Nast, and Vox, reported a median 10% year-on-year decline in Google search referrals across its members between May and June 2025. The organisation described its findings as reflective of real-world traffic changes experienced by publishers rather than estimates or models.

Taken together, these independent datasets point towards a broader reduction in referral traffic and click-through rates. However, they do not directly isolate the specific category of “bounce clicks” described by Google. This creates a gap between how Google explains the data and how external organisations measure it.

That gap is central to the ongoing debate. Google’s position is that AI Overviews are primarily reducing low-value, short-duration visits that were unlikely to generate meaningful engagement. In contrast, independent studies tend to measure overall traffic and behaviour without distinguishing between different levels of user intent.

As a result, both sides are effectively looking at the same phenomenon through different lenses. Google is focusing on the quality of remaining clicks, while publishers and researchers are focusing on total traffic and engagement levels.

Another important factor is that Google has not released granular datasets that would allow third parties to test the distinction between “bounce clicks” and more engaged visits. Without this data, external analysis remains based on aggregate traffic trends rather than segmented behaviour.

This means that while Google’s explanation is internally consistent, it cannot yet be independently verified. As things stand, the “bounce clicks” theory is a narrative supported by internal interpretation rather than publicly available evidence.

Looking ahead, this issue is likely to remain a key point of discussion as AI Overviews become more widely integrated into search results. Publishers are already closely monitoring referral traffic, while advertisers and SEO professionals are trying to understand how user behaviour is evolving in response to AI-generated answers.

For Google, the focus remains on engagement within Search itself, including how often users return and continue searching. For publishers, however, the primary concern is visibility and traffic from Google as a referral source.

Until more detailed data is shared, the true impact of AI Overviews on different types of clicks will remain open to interpretation. For now, “bounce clicks” should be seen as Google’s explanation for changing behaviour, rather than a fully evidenced conclusion that has been independently verified.