Google AI traffic drop: Here’s what publishers are doing about it

AI Overviews have caused publishers plenty of headaches this year. So how are they fighting back?

2025 has been a tough year for many publishers – and many are pointing the finger in one direction.

AI summaries are now widely reported to have caused significant drops in traffic across publisher sites as the year has gone on. Google’s AI Overviews are cited as the main culprit, with publishers who have traditionally relied on organic search feeling the greatest impact.

The Gemini-powered AI Mode that the search giant released earlier in the year has also been identified as a factor, although as a relatively new tool, there is less data to show what kind of impact this is having on its own.

What’s changing

This isn’t the only area where changes are reducing traffic to publisher sites.

Alongside traditional search engines – where AI summaries have joined paid advertising in nudging out organic results – a growing reliance on platforms such as ChatGPT and generative-AI search engines like Perplexity is coming at the expense of more conventional search behavior.

The integration of AI tools into other platforms, such as Meta AI appearing under news stories on Facebook and Grok within X, has also meant that users who weren’t necessarily seeking AI assistance to begin with are easily drawn into using these tools.

These join demographic shifts that are also redistributing traffic, such as younger audiences increasingly seeking information through video content published on social media platforms.

Publishers that depend on traffic to sustain their business models are understandably concerned.

For many, eyeballs aren’t only about ad impressions, but also key in promoting memberships and subscriptions, generating affiliate sales, and cross-promoting other titles in the same stable.

The need to be agile, of course, is hardly new; fail to keep up with changing user habits, technological developments, or regulatory shifts, and you won’t be publishing for long. But this particular change has nonetheless proven to be a brutal complication.

Beyond traffic losses, it has also sparked concerns about the accuracy of AI-generated summaries. Outdated information, hallucinations, and questionable source credibility have highlighted shortcomings in these systems. A fast and professional-sounding answer is not necessarily a reliable one.

More crucially, the sustainability of this ecosystem has been questioned. If it becomes unprofitable to produce the original reporting and trusted journalism that these systems rely on, what exactly will they be drawing from in the future?

All of this can leave publishers feeling despondent as we head into 2026. But the good news is that many have already begun adjusting their strategies to weather the impact.

Here’s what they’re doing – and what you can think about adopting if you’re seeing the same effects.

The scale of the issue

Before we examine some practical approaches, it’s worth looking at the scale of the problem.

Some figures make for sobering reading. At the start of the year, a study from Bain & Company indicated that around 60% of searches now yielded no clicks through to a publisher’s site.

Separate research published in July showed that AI Overviews had led to drops of 47.5% in click-through rate on desktop and 37.7% on mobile, and that sites previously ranking first for a query could see a drop of around 79% if their result was placed beneath an AI Overview.

That same month, Daily Mail publisher DMG Media claimed that AI Overviews had led to a drop in click-through rates of up to 89%.

Engaging audiences elsewhere

Search engines may be a valuable source of traffic, but many publishers have already diversified their efforts to gain audiences elsewhere for good reason.

UK publisher Future plc, for example, whose titles include TechRadar, Marie Claire, and Go.Compare, claims that only 27% of its sessions originate from Google search results. It has even developed a strategy dubbed Google Zero to address these challenges by engaging audiences directly through other channels.

The constant threat of algorithm changes and the rise in zero-click searches have persuaded many publishers to place more attention on social media platforms.

While these platforms are also vulnerable to volatility, they offer certain advantages, such as the increased likelihood of shareability within trusted networks.

This isn’t simply a way to counter a drop from search engines; it also helps build authority and trust with audiences who encounter your brand repeatedly across channels, rather than only fleetingly through a search result.

It’s also proving to be a valuable revenue stream for some. Reach plc, for example, is reportedly making five-figure sums per day from Meta’s Content Monetization program, which rewards publishers for engagement with images and videos posted on Facebook.

This follows reports earlier in the year that Meta’s algorithm changes have made Facebook a more dominant source of social referral traffic.

Many publishers are also using messaging platforms to engage existing audiences more directly, building communities through channels such as WhatsApp, Messenger, and Instagram broadcasts.

These offer high open rates and immediate reach, with features like content links and polls providing a more conversational alternative to traditional page posts.

For some users, this is also a more convenient way to engage with these brands over other channels, such as newsletters or first-party apps.

Going beyond what AI summaries can deliver

An AI summary might extract relevant text from an article and present many users with a satisfactory answer. But it’s worth remembering that text isn’t the only thing on your site that people value.

This is where the benefit of making your content rich and diverse becomes clear.

Strong, engaging images – particularly those that tell a story and invite interaction – alongside informative video content and interactive features are the value-adds that encourage people to visit your site and build a relationship with your brand, rather than feel sated by a quick AI summary.

You can also offer readers the best of both worlds. Many publishers now include a short summary at the top of an article, followed by the full piece for those who want more detail.

This won’t be the right approach for every type of content, but if readers come to expect this format, it increases the likelihood of them clicking through.

Understanding what kind of content is likely to trigger AI Overviews

Some content lends itself well to AI-overview-style answers, such as definitions, informational queries, and straightforward questions. Other types do not – and therein lie opportunities for publishers.

Recent research from Ahrefs into 146 million search results found that searches less likely to trigger AI Overviews include those related to shopping or product comparison, real estate, local searches, time-sensitive content, and sports. While the data isn’t absolute, it’s a strong directional indicator.

Many publishers already specialize in the latter two, although much of the traffic to these stories comes from direct visits rather than organic search. Content adjacent to this – explainers or analysis pieces, for example – could therefore be a good way to capture additional search traffic.

Many of these same publishers also work heavily in shopping and product comparison, so it may be worth doubling down on those efforts, particularly as this carries the added benefit of affiliate revenue.

That said, the exact form this content should take is evolving. Traditional reviews and evergreen “best-of” buying guides haven’t escaped AI Overviews unscathed, so experimenting with more tangential buying guides and product-led features could be worthwhile.

Other content less likely to be surfaced in AI Overviews includes interviews, experience-based stories, and opinion pieces.

Remembering SEO principles when optimizing for AI Overviews

Some people want their answers to be picked up by the AI-powered components of search engines and LLMs – and this is a perfectly sensible goal if you’re looking to get noticed.

There is, however, a danger of this Generative Engine Optimization (GEO) creating a distraction from maintaining good SEO, which should still be the cornerstone of your site and content strategy.

It used to be that people wrote with Google in mind. Best practice then shifted to writing naturally for humans. Now, with AI-generated snippets playing a larger role in search results, the winning approach is to combine the two – writing for humans in a way that AI systems can easily parse and understand.

Exploring opportunities in image search

Text-based content has clearly been impacted by AI Overviews, and text remains the dominant way people search. But this isn’t the only way people look for information.

Image search has become a more popular method for certain types of discovery, particularly when shopping for products. It’s now estimated that Google Images processes over a billion requests per day.

Whether this works for you depends in part on whether your site is image-led and whether those visuals draw people in.

Publishers with original, high-quality images that stand out in image results – rather than the generic product photography found everywhere else – are likely to reap the greatest benefits.

And since engaging visuals can directly influence metrics such as dwell time, creating or sourcing strong images is well worth the effort.

Final thoughts

Staying agile, diversifying traffic and revenue sources, and engaging audiences in ways that AI Overviews cannot will help publishers remain in the strongest possible position as AI tools evolve and create new challenges.

But practical steps like these are only part of what’s likely to shape future success. Regulatory changes are also likely to play a significant role.

The UK-based Professional Publishers Association (PPA) recently submitted recommendations to the Competition and Markets Authority (CMA) that aim to create a fairer value exchange between publishers and search engines, giving publishers greater visibility and control.

Suggestions include clearer transparency over acquisition sources in GA4, more prominent links to publisher sites within AI Overviews, and requiring Google to disclose the data it scrapes for those overviews – remedies that would help address several of the structural issues publishers now face.

These are, of course, only recommendations. But with growing concern over the sustainability of the current system, and those concerns only getting louder, the likelihood of such remedies being adopted is increasing.