Publishers have become accustomed to Google’s familiar playbook of PR — promising more traffic and higher-quality clicks, to calm down backlash, whenever it launches major search changes. But the numbers often tell a very different story. Google AI Overviews is no exception.
Google, in recent weeks, has tried to calm publisher fears about the impact of AI Overviews. Liz Reid, vice president and head of Google Search wrote an article. In an extensive blog post it is stated that third-party attempts to measure the impact on referral traffic of AI overviews “inaccurately suggests dramatic declines in aggregate” traffic.
But Publishers have seen this film before. New publisher data confirms the familiar pattern: search traffic referral is decreasing year over year.
Digital Content Next, which includes the New York Times and Conde Nast as members, surveyed 19 of its member companies between May and July to determine what was happening with their Google search referral traffic. Google AI Overviews are indeed harming publisher traffic.
The organic search referral traffic from Google has declined broadly. The majority of DCN members — both news and entertainment sites — have experienced traffic losses between 1% and 25 % from Google search. Twelve of the respondents were news brands and seven were not.
In May and June 2025 the median Google Search Referral was down every week for eight weeks, with losses exceeding gains by two to one. The downward slope was unbroken and constant for the seven brands that did not belong to the news category.
According to the results, over the eight-week period, the median YoY decrease in referred traffic was -10% for all brands, -7% news brands and -14% non-news brands.
Jason Kint of DCN stressed that these losses were a direct result of Google AI Overviews as many publishers claimed. He added that the latest data provides a “ground-truth” of what is actually happening and cuts through Google’s vague statements about “quality clicks” made in its most recent post. “I think that all publishers are ignoring Google’s post. Kint added that this “probably helps ground” the Google post.
These findings follow a recent Pew survey found that AI summaries make users less likely click on links.
Meanwhile, U.K. trade group Professional Publishers Association It was announced that it had The Competition and Markets Authority (CMA), in response to the CMA’s request for evidence and recommendation on how Google’s AI-based search features are harming publishers, submitted both evidence and recommendations.
According to the PPA, Google controls 93% of the U.K. market for search and that AI Overviews as well as the U.K. launch on July 28 of AI Mode are changing user behavior away from source websites and towards zero-click results. A spokesperson said that AIOs are also being used in Google Discover which reduces the number of clicks to citations.
According to its evidence, it has collected from members including Conde Nast Magazines, Future Magazines, Immediate Media, and Hearst Magazines, many brands report 10 to 25% year-over-year CTR decreases, even when rankings are stable. The PPA cited a lifestyle publisher who shared data about a popular search term “how to get away from [insect]”. It found that even though the CTR was still 5.1 percent, and impressions were relatively stable, it had dropped sharply to 0.6 percent in the past year.
According to the PPA, an automotive content publisher also experienced similar drops in traffic despite having a 7% boost in search visibility. The publisher saw a 25% drop in traffic for articles that ranked first in organic searches. According to the evidence, over the same time period, CTR decreased from 2.75 percent down to 1.71 percent.
Hope comes in the form of Department of Justice remedy
There’s some hope in the form of an pending remedies order by Judge Amit Mehta, in the Department of Justice Google Search antitrust lawsuit. Kint said that among the many proposals being considered, one is of particular importance to publishers. It involves the separation of Google’s AI crawler from the search crawler. Publishers are currently unable to block Google’s AI-crawler without completely disappearing from the search index. Unless a DOJ directive is issued forcing Google to make a change, they’re stuck in a holding pattern.
Kint stressed that the judge could issue an order requiring immediate action, rather than waiting on an appeal. The DoJ will need to prove irreparable damage and that they are likely to win on appeal. He said that the harm was occurring in real time. A ruling could force Google into providing the choice and separating the services immediately, which would be an important move.
European Publishers are also calling on EU regulators to act quickly on the same issue, with Google AI Overviews. They claim that AI Overviews has a detrimental impact on independent publisher’s sites, and the inability to opt-out of Google’s AI Crawler without dropping off the search rankings is causing irreparable damage. Google maintains that the claims about search traffic are “highly inaccurate.” A Google spokesperson did not immediately respond to a request for comment.
Prevention strategies
Publishers are well-versed in adapting to Google changes. They have learned to navigate algorithm changes, product updates, new features, and find ways to maintain traffic and revenues.
Many argue that staying up to date with breaking news and maintaining quality journalism is the key to performing on AI-driven platforms. Five publishing executives told Digiday that SEO practices are still important, and they don’t fall for the hype surrounding GEO. They won’t be changing their techniques. Others are more proactive in adapting their SEO strategy.
Jess Sholtz agrees with the core SEO strategy not changing. She was a marketing consultant for Ringer Media in Switzerland, where she led AI strategies. “[AI summary platforms like] ChatGPT is not building up an index, it’s RAG-ing existing indexes,” said she. “If you’re in the [search engine] Index you can’t get into the [AI summary] Citation and you fail at this first point.” Instead of trying to memorize every piece of information in its model the AI retrieves relevant information from an external source (or “index”) and then generates a response based on this retrieved information.
Sholtz insists that all publishers should invest in their brand to ensure they are ahead of any future traffic drops. However, smaller titles and less well-known publications must not neglect this investment. She said that in the past, ranking high on Google searches provided a certain level of credibility, even if a user did not know the publisher. This automatic trust has been eroded by AI-driven interfaces such as ChatGPT, Perplexity and Google’s AI Overviews. Sholz said that without clear brand signals even quality content struggles with clicks and engagement. The focus shifts from pure editorial strength towards a combination of brand recognition, content, and user trust.
She said that on these new AI surfaces the top citation does not have the same ‘oh, it’s ranking, so I trust it – it’s fine’ factor. “And branding is the first pillar which is coming in more because someone won’t blindly click on that top citation.”
