AI Overviews Cut Organic Clicks 38%: What Google’s $60B Quarter Hides

The first randomized field experiment isolating the causal effect of Google AI Overviews just confirmed what publishers have been measuring correlatively for over a year: organic clicks drop 38% on queries where AI Overviews appear. At the same time, Alphabet posted $60.4 billion in Q1 2026 search revenue — up 19% year-over-year, accelerating from the prior quarter. These two data points are not contradictory. They describe the same phenomenon from two different vantage points: Google’s and everyone else’s.

What Happened

The Study That Changed the Conversation

Researchers Saharsh Agarwal and Ananya Sen from the Indian School of Business and Carnegie Mellon University published a working paper in April 2026 that marks a turning point in the AI Overviews traffic debate. Previous analyses were correlational — noticing that CTRs fell when AI Overviews appeared, without proving causation. This study used a randomized field experiment with 1,065 U.S. desktop Chrome users recruited via the Prolific platform, split into three groups over two weeks in January–February 2026.

The three groups:
– Control: Normal Google search experience
– Hide AIO: Chrome extension silently removed AI Overviews from results — over 95% of participants never noticed the change
– AI Mode: Traffic redirected to Google’s AI Mode interface

The study was pre-registered with the AEA RCT Registry, meaning researchers locked in their methodology before seeing results — the gold standard for preventing selective reporting and p-hacking.

Key findings from Search Engine Journal’s coverage of the study:

• AI Overviews appeared on 42% of queries in the control group
• Organic clicks dropped 38% on queries where AI Overviews triggered
• Zero-click searches jumped from 54% to 72% when AI Overviews appeared — an 18-percentage-point swing
• Removing AI Overviews nearly doubled outbound clicks, consistent with the fact that 85% of AI Overviews appeared in the top position, directly blocking organic results below
• User satisfaction, perceived quality, and ease of finding information were statistically identical across the control and Hide AIO groups
• Sponsored click rates remained stable regardless of AI Overview presence

That last pair of findings is the crux of the entire story. Google’s public argument — articulated by Search head Liz Reid at least three times in public appearances since August 2025 — is that AI Overviews eliminate “bounce clicks”: low-value visits where users grab a quick fact and immediately return to search. The implication is that the traffic being lost was never worth having.

The ISB/CMU study directly tests this framing. If the clicks being lost were primarily low-value bounce traffic, then removing AI Overviews should produce a worse user experience — users are now clicking through to publisher pages to get facts they could have gotten on the SERP. Instead, the study found satisfaction was statistically identical. Users who got results without AI Overviews reported no degradation in experience. They weren’t bouncing out of frustration. They were simply clicking through to publishers instead of getting answers on Google’s results page.

Google’s Earnings: The Platform View

As reported by Search Engine Journal, Alphabet’s Q1 2026 earnings showed Google Search and Other revenue at $60.4 billion, up 19% year-over-year — an acceleration from 17% growth in Q4 2025. Total Alphabet revenue hit $109.9 billion, up 22%. CEO Sundar Pichai told analysts: “Queries are at an all-time high,” crediting AI Mode and AI Overviews for driving increased Search usage. He also noted that Google had reduced the cost of core AI responses by more than 30% since upgrading to Gemini 3, and that AI Mode had reached approximately 100 million monthly active users and 75 million daily active users.

Microsoft’s Earnings: A More Complex Picture

Microsoft’s FY26 Q3 earnings, covered by Search Engine Journal’s earnings analysis, told a more nuanced story. Bing crossed 1 billion monthly active users for the first time — a genuine milestone for a search engine that has spent the last three years in aggressive AI-enabled growth mode. But search ad revenue grew only 12% year-over-year (excluding traffic acquisition costs), and the broader Microsoft search and news segment was actually down 1% at $13.2 billion. CEO Satya Nadella acknowledged the consumer business is doing “the foundational work required to win back fans.” Bing’s global search market share remains approximately 5%, per StatCounter data from March 2026 — a number that illustrates how far the monetization runway still has to grow before Bing can extract platform-level economics from AI search.

Why This Matters

The Framing Problem Is Enormous

The narrative gap between platform earnings and publisher data is not a matter of different interpretations. It reflects structurally different economic outcomes from the same set of searches.

When Google says “queries are at an all-time high,” that is almost certainly true. More people are searching. But increased search volume generating increased revenue for Google does not translate to increased referral traffic for publishers. The mechanism has changed: AI Overviews now answer queries directly on the results page, allowing Google to monetize the search event without sending users anywhere. The user satisfies their intent. The publisher receives no visit. Google records the ad impressions.

Liz Reid’s bounce-clicks argument has been made publicly at least three times: an August 2025 blog post, an October 2025 Wall Street Journal interview, and an April 23, 2026 appearance on Bloomberg’s Odd Lots podcast. Her specific framing: “Users who quickly click and return to search no longer need to visit the page because they get the fact from the Overview. Those wanting to read longer still click through.” This is an internally consistent argument — but it has never been supported by data from Google, and the ISB/CMU study was designed specifically to test whether the lost clicks were low-value bounces or legitimate referrals.

Who Gets Hit Hardest

Informational content publishers face the steepest cliff. The queries where AI Overviews appear most frequently — definitional (“what is X”), procedural (“how to Y”), and comparative (“X vs Y”) — are exactly the high-volume informational content types that drove traffic-based business models for over a decade. Health publishers, legal information sites, how-to content libraries, educational resources, and any site built around answering factual questions are in the primary impact zone.

Agencies managing content programs on traffic KPIs are exposed. If your retainer deliverables are measured in sessions or pageviews from organic search, you are managing toward a metric that is structurally declining for the most common content types. The conversation with clients needs to happen proactively, before Q2 reports surface the numbers.

E-commerce brands relying on informational top-of-funnel content face a narrowing acquisition channel. Traffic-to-subscriber pipelines built on informational blog content are compressing as referral volume drops. Every email subscriber acquired through this channel now costs more than it did 18 months ago, because fewer people are reaching the content in the first place.

The Google Network revenue data confirms the distribution shift is structural. As noted in SEJ’s earnings analysis, Google Network revenue — the ad product that monetizes publisher inventory — fell below $7 billion for the first time in Q1 2026, declining from approximately 12% to 9% of total Alphabet ad revenue over several quarters. Search revenue expanded 31% within Alphabet’s overall ad segment while Network revenue contracted. More value is being captured on-SERP; less is being distributed outward to publishers. This is not a one-quarter anomaly. It is a directional trend that has held for six consecutive quarters.

The Data

Multiple research efforts and platform disclosures are now producing a coherent picture. The divergence between platform-level metrics and publisher-level data is not noise — it is the signal.

Click Impact by Research Source

Platform Revenue vs. Publisher Distribution, Q1 2026

What these tables reveal in combination: Google is successfully monetizing more queries while sending fewer users to external websites. The platform economics are working. The publisher economics are not.

Real-World Use Cases

Use Case 1: SaaS Brand Auditing Informational Content for AI Overview Risk

Scenario: A B2B SaaS company has 350 blog posts driving approximately 65,000 monthly organic sessions. About 55% of that traffic comes from informational queries — “what is [category term],” “how to [workflow task],” “[feature] vs [feature]” — exactly the query types where AI Overviews appear at highest frequency. Leadership is questioning the content investment ROI.

Implementation: Export the top 150 organic traffic queries from Google Search Console. For each, run the query and note whether an AI Overview appears. Flag every query where the AI Overview answers the user’s full intent without requiring a click. For flagged articles, evaluate whether the content can be elevated to serve mid-funnel intent: add proprietary benchmark data, specific customer case studies, expert implementation commentary, or interactive comparison tools that AI cannot replicate on-SERP. For articles that cannot be upgraded, reallocate production budget toward commercial-intent content — use cases, ROI guides, implementation walkthroughs — that AI Overviews are structurally less likely to fully answer.

Expected Outcome: Most B2B SaaS content libraries have 30–50% of high-traffic queries concentrated in AI Overview-heavy informational territory. Identifying this proactively — before a client asks why sessions are down — lets you lead with a strategic reframing rather than defending declining numbers. Presenting the audit as a “portfolio optimization for AI-era search” is a better conversation than explaining why traffic fell 20%.

Use Case 2: E-Commerce Brand Pursuing AI Overview Citation Presence

Scenario: A consumer goods brand finds that category queries like “best [product type] for [use case]” are triggering AI Overviews that cite competitor product pages but not theirs. They want brand visibility in the zero-click experience, even if they can’t always generate the referral click.

Implementation: Map the top 25 category queries where AI Overviews appear and identify which competitor pages are being cited. Analyze the structure of those pages — most AI Overview citations pull from content with direct, quotable answers in the first 200 words, complete FAQ and Product schema markup, and structured specification data. Update your product and category pages to match: implement FAQ schema, ensure key user questions are answered specifically and concisely near the top of each page, and add full product specification tables. Build supporting topical content that establishes category authority across adjacent informational queries.

Expected Outcome: Pages optimized for AI Overview citation typically see citation rate increases within 60–90 days of structured content updates. The structural improvements — clear answer formatting, complete schema, topical depth — also lift traditional organic rankings simultaneously. In zero-click scenarios, brand citations create awareness that matters when users advance to active evaluation. Being cited is not a traffic referral, but it is structurally better than being invisible.

Use Case 3: Content Agency Restructuring Client Retainers

Scenario: A content marketing agency has seven clients on monthly blog retainers built around informational keyword strategies. As Google-referred traffic correlates with AI Overview adoption on their target queries, clients are questioning whether the retainer ROI is holding up.

Implementation: Conduct a per-client query audit identifying which keywords in their current rankings trigger AI Overviews at high frequency. Reframe retainer deliverables away from “X posts per month” toward a deliverable mix that includes original research assets (client-survey data, proprietary benchmarks), expert-sourced commentary pieces with specific practitioner perspectives, customer case studies with concrete results, and conversion-focused comparison content. Introduce reporting that tracks AI Overview citation rate alongside traditional traffic metrics — this demonstrates brand value even when referral click volume declines. For clients in high-AIO verticals (health, finance, legal, home improvement), proactively propose reducing commodity informational volume in favor of depth and specificity.

Expected Outcome: Agencies that restructure retainers proactively tend to retain accounts longer than those defending traffic-decline narratives. Clients can see their brand cited in AI Overviews, their commercial-intent content holding traffic, and their production resources allocated toward content types with the best structural protection in the AI search era.

Use Case 4: Independent Publisher Diversifying Away from Google-Referred Traffic

Scenario: An independent media publisher generating revenue from display advertising and affiliate commissions has watched Google-referred traffic fall approximately 30% over 18 months, consistent with the Chartbeat/Reuters finding covered in SEJ’s analysis. They need a new acquisition engine.

Implementation: Audit current traffic source concentration — if more than 60% of sessions originate from Google organic search, that is existential concentration risk. Launch a free weekly email newsletter anchored to original editorial perspective, not repurposed blog posts. For every informational article that still ranks and drives referral traffic, install a contextually specific email capture offer tied directly to the article topic. Build distribution on channels where AI Overviews have no reach: LinkedIn newsletters, YouTube, podcast appearances, Reddit community engagement. Reduce editorial calendar allocation to commodity informational queries; increase original reporting, proprietary data analysis, and expert-led content that cannot be sourced by AI from existing web content.

Expected Outcome: Publishers who complete this channel diversification find that email subscribers convert to revenue at significantly higher rates than Google-referred anonymous visitors. The owned channel also eliminates existential dependence on a referral source that has now structurally changed its distribution model and shows no sign of reversing the change.

Use Case 5: Paid Search Team Recalibrating Budget Allocation

Scenario: A performance marketing team notices impression volume is up on informational keywords but CPA has risen and conversion rates have declined. AI Overviews appear on a significant percentage of their target queries. They are evaluating whether to defend or redeploy budget.

Implementation: Segment Google Ads campaigns by query intent: informational, commercial, and transactional. Pull CTR and CPA data by intent segment over the past six months. The ISB/CMU study found that sponsored clicks remained stable even when organic clicks collapsed — paid traffic has structural insulation from AI Overview displacement. However, if user intent on an informational query is satisfied by the AI Overview, a click to a top-of-funnel landing page is unlikely to convert efficiently. Identify queries where CPA is highest and AI Overview presence is highest — these are the candidates for budget reallocation toward commercial and transactional terms where AI Overviews appear less frequently and user intent is purchase-ready.

Expected Outcome: Performance teams that segment by AI Overview prevalence typically find CPA on informational keyword clusters running materially higher than on commercial-intent queries. Rebalancing budget toward commercial-intent terms — even at higher CPCs — commonly improves overall campaign efficiency when measured against pipeline-stage conversion events.

The Bigger Picture

This is not a story about a single algorithm change or a product feature that might be rolled back. It is the story of what happens when the web’s largest referral infrastructure decides to answer queries rather than route them — and when that decision is validated by record revenue growth.

Google built its position over 25 years by being the world’s best router: find the best answer, send the user there, monetize the connection through advertising. Publishers built businesses on that routing mechanism. SEO emerged as a discipline to optimize for it. Content marketing as a channel — the high-volume blog strategy, the long-tail keyword approach, the informational content flywheel — was architected for a world where Google’s incentive aligned with publisher interests. Google wanted users to click through to quality content because that kept users trusting Google. Publishers wanted to receive those clicks. The interests were aligned.

AI Overviews change the incentive structure fundamentally. Google can now monetize a query by answering it on-SERP: the user gets their answer, Google captures advertising impressions, and the referral click becomes optional. The ISB/CMU study’s finding that user satisfaction is identical with and without AI Overviews is critical not because it validates or invalidates the bounce-clicks argument, but because it demonstrates that Google has successfully inserted itself between user intent and publisher fulfillment without degrading the experience users associate with Google. That is a stable equilibrium for the platform. It is an unstable one for publishers.

Alphabet’s Q1 2026 earnings confirm the transition is working at the platform level: 19% search revenue growth, all-time high query volumes, AI Mode at 100 million monthly users, core AI response costs down 30% since Gemini 3. Meanwhile, Google Network revenue — the product that monetizes publisher inventory — fell below $7 billion for the first time, continuing a multi-quarter decline while total search revenue expanded. More value is being captured on-SERP, less is being distributed outward.

Microsoft’s position adds a second dimension. Reaching 1 billion monthly active Bing users is a real achievement, but the -1% segment revenue tells the monetization story: user reach does not automatically translate to platform economics without the market-dominant position Google holds. Bing at 5% global share, even with strong AI features, cannot extract the same per-query value as a 90%-share provider. The search advertising market rewards concentration at a level that makes the AI arms race only partially relevant to competitive dynamics.

For marketers, the big picture points toward accepting a secular, not cyclical, shift. Informational content as a primary traffic channel strategy is in long-term decline. The algorithm is designed to answer queries, not to route traffic. Businesses that reallocate earliest toward owned channels, original proprietary research, and commercial-intent content will be structurally better positioned for the next five years. Those defending declining informational traffic metrics are managing toward a model the platform has departed.

What Smart Marketers Should Do Now

1. Audit your traffic for AI Overview exposure before the next client review.

Pull your top 100 organic traffic-driving queries from Google Search Console and check which ones currently trigger AI Overviews. Most informational content libraries have 35–50% of high-traffic queries in AI Overview-heavy territory. Identifying this proactively — before declining metrics surface in a quarterly review — lets you lead the strategic conversation rather than respond to it. Frame it as a portfolio optimization for the AI-era search landscape, not as a performance problem you’re reacting to.

2. Shift success metrics away from sessions toward owned-audience conversions.

Traffic from informational search is structurally unreliable going forward. Add email subscriber acquisition rate, newsletter sign-up volume, and direct return visits as primary KPIs alongside sessions and organic traffic. Every informational article that still ranks should include a contextually specific email capture offer — not a generic pop-up, but a direct continuation of the value the visitor just consumed. Converting Google-referred visitors into owned-audience members before the referral volume narrows further is the highest-leverage action available right now.

3. Invest in content that AI cannot source from the existing web.

Original research, proprietary data, real customer case studies with specific results, practitioner perspectives built from actual deployment experience — these are structurally resistant to AI Overview commoditization because they don’t exist anywhere else on the web for AI to surface. A survey of your customer base, an analysis of platform data, an interview with a practitioner who shares concrete deployment numbers: this content serves informational queries with a specificity and originality that AI Overviews cannot replicate without citing your source directly. When your source is cited, you maintain visibility in zero-click scenarios.

4. Build a commercial-intent content moat.

The ISB/CMU study found that sponsored clicks remained stable even as organic clicks collapsed on AI Overview queries — commercial and transactional intent queries have more structural protection. Comparison content, ROI calculators, implementation guides, competitive breakdowns, and use-case-specific landing pages serve users in active evaluation mode. These are the people who need more than an on-SERP answer, whose sessions are longer and conversion rates are higher. Double down on this content tier. It is the most defensible segment of organic search traffic in the AI search era.

5. Track AI Overview citation rate as a content performance KPI.

Most enterprise SEO platforms added AI Overview citation tracking in 2026. If you haven’t set this up, do it now. Being cited in an AI Overview does not generate a referral click, but it creates brand visibility in zero-click searches — users see your domain or brand name in the context of answering their query. More importantly, pages that earn AI Overview citations tend to be well-structured, authoritative, and comprehensive — signals that also lift traditional organic rankings. Citation rate is a leading indicator of content quality alignment with where Google’s ranking signals are heading, and it gives you a metric to demonstrate content value even when click-based attribution declines.

What to Watch Next

The ISB/CMU peer review process (expected Q3 2026). The working paper’s methodology is strong — pre-registered, randomized, large sample size, blinded participants — but academic peer review may surface limitations or refinements. If the 38% organic click decline finding holds through peer review, it becomes a definitive reference point in every AI Overviews policy and strategy conversation. If reviewers identify methodology gaps, expect Google’s communications team to cite them immediately.

Whether Google releases supporting data for the bounce-clicks claim. Liz Reid has made the bounce-clicks argument three times across 10 months without producing a single supporting data point from Google’s own systems. The pressure is mounting from publishers, academic researchers, and regulators. If Google releases click-quality segmentation data — bounce rate, session duration, or conversion rate for clicks that would have occurred without AI Overviews — it could meaningfully shift the narrative. Continued silence increasingly functions as its own signal about what that data shows.

Google Network revenue trajectory through Q4 2026. The decline from approximately 12% to 9% of total Alphabet ad revenue is a multi-quarter trend, not a one-quarter anomaly. If Network revenue falls below $6.5 billion in any quarter through year-end, it confirms that the publisher ad ecosystem is being systematically deprioritized in favor of on-SERP monetization. Every publisher running AdSense and every programmatic advertiser buying Google Network inventory has a direct stake in this metric.

AI Mode expansion and Search Console transparency. Google AI Mode reached 100 million monthly active users in Q1 2026. As AI Mode becomes a more prominent default search option, the click patterns observed for AI Overviews may intensify — AI Mode is a more immersive on-SERP experience. Watch whether Google begins separating AI Mode query data from traditional Search Console reporting, and whether AI Mode queries generate any referral traffic data that site owners can access.

Regulatory action in the EU and UK. AI Overviews represent a new form of on-SERP self-preferencing: Google answers queries using content created by third-party publishers and distributes those answers to users without a referral click. Publisher lobbying groups in both markets — where Google already faces active antitrust proceedings — are building cases around this mechanism. A regulatory requirement to include click-through attribution or source links in AI Overviews could materially change the traffic picture within 12–18 months.

Bottom Line

The ISB/CMU randomized field experiment provides the first rigorous causal evidence that Google AI Overviews reduce organic clicks by 38% on the queries they appear on, with zero measurable benefit to user satisfaction or experience quality. Alphabet’s Q1 2026 search revenue of $60.4 billion — up 19% year-over-year — confirms the platform is successfully monetizing more queries while distributing fewer referral clicks to publishers. These are not contradictory findings: they describe a structural shift in how Google captures value from search, where more of that value stays on-SERP instead of flowing outward. For marketers, the correct response is not to wait for the traffic to recover — the mechanism that drove it has been redesigned. The strategic question is how fast you can restructure content programs toward owned channels, original data assets, and commercial-intent content before clients or leadership start asking why the informational traffic numbers look the way they do.