iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
The claim that Google’s AI Search Features Are Killing Traffic to Publishers is too broad: Google’s AI answers reduce clicks for many affected searches and can damage publisher traffic, but the evidence does not show a universal web-wide collapse. Pew found lower click rates when summaries appeared; Google says total organic click volume remains relatively stable year over year.
The contradiction is real because the major studies measure different things. Google reports aggregate organic clicks across the web, while Pew measures user behavior and Chartbeat measures traffic within its publisher network. Smaller publishers, news sites, particular query types, and Google Search referrals can suffer substantial losses even if Google’s overall click total remains comparatively stable.
Key takeaways
- Google AI summaries reduce clicks on many searches that display them, but the evidence does not prove a universal collapse in traffic across the entire web.
- According to Pew Research Center in 2025, a browsing panel of 900 U.S. adults produced 68,879 unique Google searches, including 12,593 searches with an AI-generated summary; users clicked result links less often when a summary appeared.
- According to Chartbeat’s 2026 publisher-network analysis, average weekly pageviews fell about 6% from 7.64 billion in 2024 to 7.19 billion in 2025, while Google Search pageviews fell 34% between December 2024 and December 2025.
- Chartbeat reported larger Google Search declines for smaller sites in its network: 22% for large publishers, 47% for medium publishers, and 60% for small publishers.
- Google says total organic click volume to websites has remained relatively stable year over year, showing that the dispute partly concerns aggregate web traffic versus specific publishers, topics, channels, and time periods.
- Publishers should measure Google Search, Discover, direct, social, email, referral, engagement, subscription, and revenue changes separately rather than treating every Google impression as a conventional referral opportunity.
What do users actually do after seeing a Google AI summary?
Users are less likely to click a conventional result when Google displays an AI summary, and clicks to the sources cited inside the summary are especially uncommon. That finding comes from observed browsing behavior, making it more useful than a simple ranking comparison, although it does not by itself prove that every publisher traffic decline was caused by AI Overviews.
According to Pew Research Center’s 2025 browsing study, researchers tracked 900 U.S. adults during March 1–31, 2025, reconstructed the Google results pages associated with their activity, and analyzed 68,879 unique searches. An AI-generated summary appeared for 12,593 of those searches.
#1 Best Overall
Pew found that users clicked a result link less often when an AI summary appeared. Pew also found that roughly two-thirds of searches, both with and without an AI summary, ended with users browsing elsewhere on Google or leaving without clicking a result link. AI summaries therefore appear to intensify an existing zero-click tendency rather than explain every instance of falling search traffic.
A later research preprint offers a sharper measure of source referrals. According to the authors of an August 5, 2026 preprint, clicks to sources cited in AI Overviews occurred on about 1% of AI-Overview visits, and AI Overviews were associated with fewer clicks and more browsing sessions that ended without a click. The 2026 click-behavior preprint is research evidence, not a settled consensus finding, so the approximately 1% result should not be treated as a universal click-through rate for every country, query, device, or Google AI product.
The practical distinction is important: a publisher can be visible in an AI answer without receiving the pageview that historically followed a search impression. Citation visibility can support awareness or branded demand, but citation visibility is not the same thing as a referral session, an ad impression, an affiliate interaction, an email signup, or a subscription conversion.
Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchPC Slower Than It Used to Be?
A free scan shows the junk files, broken settings and background clutter dragging Windows down - then fixes them in one click.Free scan · Windows 10 & 11What does publisher traffic data show?
Publisher-side data shows a meaningful decline in Google-referred traffic for some sites and periods, but the size of the decline depends heavily on the dataset and channel being measured.
According to Chartbeat’s June 18, 2026 analysis, average weekly pageviews across its publisher network fell from approximately 7.64 billion in 2024 to 7.19 billion in 2025, a decline of about 6%. Chartbeat cautioned that 2024 was an unusually large news year because of the global election cycle, so the total-pageview comparison is not a clean test of the AI-search effect alone.
The channel-specific result was considerably sharper. Chartbeat reported that Google Search pageviews fell 34% between December 2024 and December 2025, while Google Discover pageviews fell 16% over the same period. Search and Discover should not be combined: both are Google distribution channels, but they expose publishers through different user journeys and ranking systems.
Chartbeat also reported different results by publisher size within its own network. Google Search pageviews declined 22% for large sites, 47% for medium sites, and 60% for small sites. These are Chartbeat-network figures, not a universal estimate for all large, medium, or small publishers. The results nevertheless indicate why a stable web-wide average can coexist with a severe traffic problem for an individual site.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Similarweb reported a separate directional signal. According to Similarweb’s July 2, 2025 publisher report, organic traffic to news sites had declined 26% since Google launched AI Overviews, while ChatGPT referrals to publishers had increased 25-fold. Similarweb’s figures are cross-site estimates rather than a publisher’s own analytics ledger, so they are useful for market-level comparison and corroboration, not for replacing first-party revenue and session data.
| Source and date | Reported measurement | Reported result | Scope and limitation |
|---|---|---|---|
| Pew Research Center, 2025 | Observed browsing by 900 U.S. adults across 68,879 unique Google searches | 12,593 searches displayed an AI-generated summary; users clicked less often when a summary appeared | User behavior in the United States during March 1–31, 2025; not a complete publisher-revenue measurement |
| 2026 research preprint | Panel analysis of Google result pages with AI Overviews | About 1% of AI-Overview visits produced clicks to cited sources | Preprint evidence; the result is not a universal rate for every market or query |
| Chartbeat, 2026 | Average weekly pageviews across a publisher network | 7.64 billion in 2024 versus 7.19 billion in 2025, or about a 6% decline | Network data; 2024 news-cycle effects make the year-over-year comparison imperfect |
| Chartbeat, 2026 | Google Search pageviews in its publisher network | 34% lower in December 2025 than in December 2024 | Specific channel, months, and Chartbeat network; not all web traffic |
| Similarweb, 2025 | Estimated cross-site traffic to news publishers and ChatGPT referrals | 26% lower organic news-site traffic and a 25-fold increase in ChatGPT referrals | Modeled or estimated market data; not first-party analytics for every publisher |
Why do the numbers disagree?
The numbers can differ without one source necessarily being wrong because Google, Pew, Chartbeat, Similarweb, and individual publishers measure different populations, channels, and outcomes.
| Measurement lens | What it can show | What it cannot establish by itself |
|---|---|---|
| Google’s aggregate Search data | Total organic click volume across websites, click-quality trends, and changes in query behavior | Whether a particular publisher, news category, country, or query group lost traffic |
| Pew’s observed U.S. panel | Whether people click less often when an AI summary appears in an actual browsing session | How much revenue a publisher lost or whether AI caused a site-wide annual decline |
| Chartbeat’s publisher network | Pageview changes by channel, period, and publisher-size group within its network | Traffic changes for publishers outside the network or the exact causal contribution of AI Overviews |
| Similarweb’s cross-site estimates | Market-level organic traffic trends, referral changes, and competitive direction | A site’s precise first-party sessions, conversions, ad revenue, or subscription revenue |
| Individual publisher analytics | Landing pages, referrers, engagement, conversions, revenue, and audience behavior for that publisher | The total number of clicks Google sends to every site or the full visibility of a page inside an AI answer |
Google describes the broadest denominator. In its August 6, 2025 statement, Google said that overall organic click volume from Google Search to websites had been relatively stable year over year. Google also said average click quality had increased, that it continued to send billions of clicks to websites every day, and that AI features were encouraging longer and more complex queries.
Publisher studies often describe narrower denominators: news sites, a particular analytics network, Google Search rather than Discover, or a before-and-after period. Traffic can therefore remain stable across the web while falling sharply for a publisher that depends on informational search, non-branded queries, or Google referrals. Google’s official explanation of AI Search traffic and click quality is a necessary counterpoint to third-party publisher data, not a reason to discard that data.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Clear out junk files and repair common Windows errors3Scan for outdated or missing drivers - takes under a minuteSeveral non-AI factors can also move the numbers: seasonal demand, major news events, election-year comparisons, Google algorithm updates, changes in publisher publishing volume, shifts in brand demand, and changes in the mix between Search, Discover, social, email, and direct traffic. A falling click-through rate with stable average rankings is consistent with a changed results page, but it does not prove that an AI Overview was the only cause.
Are AI Overviews and AI Mode the same?
AI Overviews and AI Mode are related Google AI-search features, but they create different user journeys and should be measured separately whenever the available data permits.
| Feature | Where it appears | How the journey works | Publisher implication |
|---|---|---|---|
| AI Overviews | An AI answer layer inside a conventional Google results page | Google summarizes information and may cite publisher links before the user chooses a conventional result or citation | A citation or impression may occur without a publisher referral click |
| AI Mode | A more conversational Google Search experience | Users can ask follow-up questions; Google described advanced reasoning, multimodality, and query fan-out that breaks a question into subtopics and runs multiple searches | The discovery process can last longer and can be harder to attribute to a single conventional query or referrer |
| Traditional results | The standard results page without relying on an AI answer as the primary interaction | Users select a listed result and proceed directly to the publisher page | Search impressions, rankings, clicks, and landing sessions are more familiar, although zero-click behavior still exists |
Google announced on May 20, 2025 that AI Mode was rolling out in the United States without a Labs sign-up requirement. Google described AI Mode as a more complete AI-search experience with advanced reasoning, multimodality, follow-up questions, and query fan-out. The Google AI Mode announcement describes the product rollout, but the announcement does not establish that AI Mode has identical availability or traffic effects in every country.
AI Overviews and AI Mode can both expose publisher links, but a publisher should not count a link appearance as a conventional referral. AI Overviews may answer the immediate question on the results page; AI Mode may mediate a longer chain of questions before a user reaches a site, another search service, or no site at all.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Why do AI search features threaten publisher revenue?
The commercial risk is not simply fewer visits. A lost referral can remove several downstream opportunities: an advertising impression, an affiliate interaction, an email signup, a subscription conversion, a repeat visit, or a chance to build a direct audience relationship.
Publisher economics commonly follow a chain: search impression, click, pageview, engagement, monetization, and sometimes conversion. An AI summary can answer a query before the user reaches the publisher, reducing the number of opportunities in that chain. A citation can preserve some visibility while still failing to produce the pageview needed for advertising, affiliate, or subscription revenue.
Penske Media Corporation’s February 12, 2026 filing in Penske Media Corporation v. Google makes arguments about this publisher-revenue mechanism. The filing is a party’s court submission responding to a motion to dismiss, so its allegations should not be presented as final judicial findings. The Penske Media court filing is evidence of the legal argument and the stakes, not independent proof that every AI-generated answer caused a specific dollar loss.
Not every lost click becomes a lost dollar. Some users may return directly, search for the brand, subscribe later, or click a different page. Conversely, a publisher may lose valuable high-intent visitors even when total sessions look stable. The correct business question is therefore not only how many Google clicks disappeared, but which audience segments, pages, conversions, and revenue opportunities changed.
Free tools Windows power users keep installed
One-click scans. No signup required.
What does Google say in response?
Google’s response is that AI Search is changing the shape of search rather than destroying the web-referral system at aggregate scale.
Google states: “Overall, total organic click volume from Google Search to websites has been relatively stable year-over-year.” Google also states: “We continue to send billions of clicks to websites every day.” Both statements come from Google’s official August 6, 2025 Search statement, which says AI features create more queries, show more links, and send users to sites when they want deeper information, reviews, original perspectives, or purchases.
Google’s position can be true alongside publisher losses because aggregate click volume and publisher-level traffic are different measurements. A click can move from one site to another, from a news publisher to a large platform, or from a traditional result to a different kind of interaction. Google’s aggregate stability claim does not tell a small publisher whether its non-branded search clicks fell 60%, just as Chartbeat’s network decline does not measure every website on the internet.
The strongest conclusion is therefore narrower than the headline claim. Google AI search features appear to reduce clicks for many searches that show AI answers and to put pressure on publishers that depend on search referrals. The available evidence does not establish that Google AI search has uniformly destroyed traffic across all websites.
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →What are regulators and publishers asking Google to change?
The regulatory dispute concerns transparency, control, attribution, and bargaining power as much as traffic volume.
At a July 7, 2025 UK Competition and Markets Authority publisher roundtable, one participant reported internal data showing a 70% desktop click-through-rate decline and a 52% mobile click-through-rate decline on searches displaying an AI Overview. Those figures were the internal data of a participating publisher, recorded in the CMA roundtable summary; they were not a CMA-wide finding and should not be generalized to all publishers.
The same CMA document records publisher requests for clearer information about when content appears in AI Overviews, how often content is used without a click, and how AI-related traffic is represented in Search Console. These requests address a basic measurement problem: a publisher cannot reliably manage a distribution channel that it cannot clearly observe.
On June 3, 2026, the CMA imposed a Google search publisher conduct requirement. The requirement calls for effective controls over the use of search content in generative AI, clear explanations of how content is used, detailed engagement metrics, and clear attribution and access mechanisms. The CMA’s Google search publisher conduct requirement is a concrete policy response to the transparency and control concerns recorded in the roundtable.
The policy does not settle the causal debate over every traffic decline. It does, however, recognize that publishers need enough information and control to distinguish visibility from usage, citations from referrals, and AI-mediated discovery from conventional search traffic.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How should publishers measure Google AI search traffic?
Publishers should build a measurement system that separates distribution channels and follows the reader beyond the initial click. No single dashboard can reliably answer whether an AI Overview caused a specific revenue change.
- Establish comparable periods. Compare equivalent dates, weekdays, countries, devices, and content categories rather than comparing an unusually large news year with an ordinary year. Record major Google updates, election cycles, breaking-news events, and publisher strategy changes alongside the traffic data.
- Separate Google channels. Report Google Search, Google Discover, direct, social, email, and other referrals independently. A 34% Search decline cannot be inferred from a 6% total-pageview decline, and a Discover change should not be used as a proxy for Search.
- Segment query and page performance. Compare branded and non-branded demand, informational and commercial queries, article templates, evergreen pages, breaking-news pages, country, device, impressions, average position, clicks, and click-through rate. Stable rankings with lower CTR identify a results-page or query-mix problem, but they do not prove which Google feature caused it.
- Connect traffic to business outcomes. Track engaged sessions, ad impressions, affiliate interactions, email registrations, subscriptions, returning users, and revenue by landing page or content group. A citation without a session may have awareness value, but it should not be counted as a monetized referral.
- Look for AI referrals separately. Preserve referrer information and campaign parameters where possible, and inspect whether AI products send identifiable sessions. Do not assume that all AI-mediated discovery appears as a clean, separately labeled source.
- Benchmark outside the site. Use cross-site data to test whether a decline is specific to one publisher or visible across a category. Treat modeled estimates as directional context, not as a replacement for the publisher’s own analytics.
- Track direct audience health. Monitor branded search demand, direct visits, email traffic, app usage, memberships, subscriptions, and repeat visits. A publisher that loses some search referrals but grows direct relationships may have a different business outcome from a publisher that loses both search and returning audiences.
Publishers should not assume that optimizing content to appear in AI answers will restore every lost click. AI visibility can be worth measuring, but the commercial objective may be a qualified visit, a subscription, a direct audience relationship, or a defensible brand—not merely a citation count.
Which tools answer which measurement question?
Different tools answer different parts of the problem. First-party analytics and Search Console are the starting point for a publisher’s own traffic and query data; network and market tools add context but should not be confused with first-party measurement.
| Tool or data source | Best use | Useful detail | Main limitation | Best fit |
|---|---|---|---|---|
| Publisher first-party analytics | Measuring actual sessions and business outcomes | Landing pages, referrers, device, country, engagement, conversions, and revenue | Usually cannot see every AI answer impression or citation that produced no visit | Every publisher making operational or revenue decisions |
| Google Search Console | Analyzing search impressions, queries, pages, clicks, average position, and CTR | Query- and page-level Search performance over selected periods and dimensions | Do not assume it provides a complete, clean AI Overview versus AI Mode split | Diagnosing Search visibility and click changes |
| Chartbeat | Real-time and historical publisher traffic analytics | Engaged time, referral sources, scroll depth, recirculation, benchmarking, headline testing, and reader-experience analysis | Its network research does not prove the cause of an individual site’s decline | Newsrooms and publishers managing content performance |
| Similarweb | Competitive traffic intelligence and market benchmarking | Estimated traffic volume, traffic share, changes, engagement, referral domains, and incoming or outgoing referral patterns | Cross-site estimates are not the publisher’s own server-side or analytics-platform data | Publishers, agencies, and marketers comparing categories or competitors |
| Semrush | Monitoring search and AI visibility alongside SEO activity | SEO, AI Visibility, content, social, local, and related search-marketing toolkits | Visibility monitoring is not proof of publisher sessions, citations, or revenue | Marketing teams and publishers investigating discoverability |
Chartbeat’s publisher analytics are designed around publisher-side performance, including engagement and referral analysis. Chartbeat’s materials describe real-time and historical analytics, benchmarking, headline and image testing, and reader-experience measurement. Chartbeat is especially useful when an editorial team needs to connect a traffic change with what happened after a reader arrived.
Similarweb is better suited to competitive traffic intelligence than to auditing a publisher’s own exact session count. Its documentation covers referral traffic, traffic volume, traffic share, changes, engagement, and referral domains. The Similarweb referral-traffic documentation and incoming-traffic documentation explain the market-level questions the platform can support.
Semrush is a natural option for teams that need an AI-search visibility platform or SEO traffic analysis alongside their own analytics. Semrush’s relevant toolkit covers SEO and AI visibility, but Semrush should be used to monitor discoverability rather than presented as evidence that Google AI features caused a particular traffic decline. A visibility score cannot substitute for first-party sessions, conversions, or revenue.
What should publishers do next?
Publishers should treat Google AI search as a distribution and attribution change, not as a single ranking problem.
Recommended Free Tools
- Protect the data layer. Keep consistent first-party analytics, referrer data, page-level revenue data, subscription events, and email-registration data before changing strategy.
- Find the exposed query classes. Identify pages with falling Search CTR despite stable rankings, then compare whether those pages answer simple factual questions that an AI summary can satisfy without a visit.
- Prioritize irreplaceable value. Original reporting, firsthand testing, unique datasets, expert analysis, reviews, local knowledge, and strong editorial perspective give users more reason to continue to a publisher page. These qualities do not guarantee a click, but they are more defensible than repeating information that a summary can reproduce.
- Build direct relationships. Email, subscriptions, memberships, apps, communities, and returning visitors reduce dependence on any one search interface. Diversification does not make search irrelevant; it limits the damage when a search channel changes.
- Measure citations and referrals separately. Record AI visibility where possible, but report cited appearances, identifiable AI referrals, conventional Google clicks, engaged sessions, and conversions as different metrics.
- Use external benchmarks carefully. Chartbeat can help a publisher compare network and engagement patterns; Similarweb can help assess market and referral changes; Semrush can help monitor search and AI visibility. None should be used to manufacture a causal claim that the data cannot support.
- Ask platforms for usable controls and metrics. The CMA’s requirements and publisher requests point toward the information publishers need: when content is used, how often it is used without a click, what engagement occurred, and how attribution works.
Bottom line: is Google AI search killing publisher traffic?
Google AI search features are not proven to be killing traffic across the entire web, but they are changing the economics of publisher discovery and reducing clicks for many AI-summary searches. Pew’s observed user behavior, the 2026 preprint, Chartbeat’s channel-specific declines, and Similarweb’s market estimates all support a real risk for many publishers. Google’s aggregate click-stability claim remains relevant because the losses are unevenly distributed.
The most defensible conclusion is that Google is reallocating attention. Publishers that depend on non-branded Google Search may lose the pageviews and monetization opportunities that once followed a search impression, even when rankings look stable. The response is not to treat every citation as success or every decline as proof of one cause; it is to measure each channel, connect traffic to revenue and audience ownership, and build relationships that do not end at the search results page.
The Bottom Line
Google AI search is reducing publisher clicks in many affected searches, but the evidence does not prove a universal web-wide traffic collapse. The practical risk is channel-specific: fewer Google referrals can mean fewer ad, affiliate, subscription, and audience-building opportunities, especially for smaller publishers. Measure Search, Discover, direct, social, email, AI referrals, engagement, and revenue separately before attributing a decline to AI.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
