Organic search clicks are down 42% from their pre-AI Overviews baseline. That’s not a bad quarter or a penalty recovery. That’s a structural shift in how Google delivers information, and it’s been accelerating since AI Overviews expanded at scale in 2025.
The number comes from Define Media Group’s analysis of Google Search Console data across 64 sites, published this week by Search Engine Land. Traffic fell 16% immediately when AI Overviews launched. It never recovered. By Q4 2025, the drop reached 42% from the pre-launch baseline.
Most agencies are treating this like a traffic problem. It’s actually a content format problem. The brands earning citations inside AI Overviews are not losing traffic; they’re gaining it. The brands optimizing for a version of Google that no longer exists are the ones seeing the declines.
Here’s what the data shows about what’s actually working, and why the format of your content matters as much as the topic.
The Scale of the Shift (And Why It’s Not Slowing Down)
Zero-click searches now account for 65-70% of all Google queries in early 2026. Google places an AI-generated summary at the top of results for a growing percentage of informational queries. Before a user sees your link, they may already have their answer.
Seer Interactive’s study across more than 3,100 queries and 42 organizations found that organic click-through rates dropped 61% for searches where AI Overviews appeared, falling from 1.76% to 0.61%. For informational and evergreen content, the losses are even sharper.
The categories hardest hit are the ones most agencies rely on: how-to content, definition posts, comparison guides, and listicles targeting high-volume informational queries. These keywords generated traffic for years because they ranked well. Now Google answers them directly, often without sending a single click.
That same Define Media Group data shows one category growing: breaking news. Traffic from news content grew 103% from November 2024 through early 2026. Google Discover, meanwhile, grew 30% across the same portfolio as web search traffic fell. For the first time in that dataset, Discover and web search now drive roughly equal traffic.
The message is clear: Google is not going back. Informational content that doesn’t earn a citation in the AI Overview is becoming invisible. The question is what you do about it.
Why Most Content Is Not Built to Get Cited
AI models cite sources that make their job easy. That sounds obvious, but the implications are specific.
GenOptima’s March 2026 analysis of Gemini’s citation and web search triggering behavior identified several concrete patterns. Informational prompts containing phrases like “how to,” “best practices,” and “techniques” trigger Gemini web search 100% of the time. When Gemini pulls content for those queries, it uses text fragment anchoring: it locates specific sentences that function as standalone definitions or complete answers and extracts them directly.
Content that opens sections with contextual transitions (“Now that we understand the basics…”) or questions (“But what does that mean for your strategy?”) is structurally harder to cite. The model has to parse context to find the answer. Content that opens every major section with a single, self-contained definition or direct answer gives the model exactly what it needs.
GenOptima calls this definition-first writing. Every section opens with one sentence that defines or answers the core concept completely, with no prior context required. If someone read only that sentence, they’d have a useful answer.
Second factor: evidence grounding. AI models increasingly deprioritize content with unsupported assertions. Every factual claim needs a traceable source. “Studies show…” without a citation is a red flag in AI model weighting. Specific data tied to named organizations is significantly more likely to appear in AI-generated answers.
Third factor: content format. GenOptima’s analysis found that listicle-format content drives 74.2% of AI citations. Numbered lists, step-by-step formats, and clearly enumerable structures are what AI models prefer when synthesizing answers. They’re extractable. They’re scannable. They map directly to how AI generates responses.
The irony: the content format that was falling out of SEO favor (the “thin listicle”) is the structure AI models rely on most for citations, provided the content behind each item is substantive and evidence-grounded.
The Three Format Moves That Change How AI Reads Your Content
1. Rewrite Your Section Openers
Go through your highest-traffic informational posts. Every H2 section that opens with a transition sentence, a question, or a narrative hook needs to be rewritten. The first sentence of each section should define the concept or answer the question directly.
Before: “Understanding why this matters requires looking at how search has changed over the past few years.”
After: “Definition-first writing is a content structuring technique where every major section opens with a self-contained sentence that defines the topic without requiring prior context.”
The second version is citable. The first is not. Apply this to your entire back catalog, not just new posts. Content you published two years ago is still being crawled and evaluated by AI models. A format update with no change to the core content can shift a page from never-cited to frequently-cited.
2. Add Structured Evidence to Every Major Claim
Every statistic, trend, or assertion in your content should trace to a named source. Swap vague references for specific attribution: instead of “research shows AI search is growing,” write “Google Search Console data across 64 sites shows organic clicks fell 42% from the pre-AI Overviews baseline, according to Define Media Group’s March 2026 analysis.”
This also means linking to your sources inline, not just collecting them in a bibliography section. Inline citations signal credibility to both AI models and human readers.
3. Add Quick-Answer Blocks After Every H2
After your section opener, before diving into the narrative, include a brief structured summary of what the section covers. Two to four bullet points. This gives AI models a scannable, extractable version of your content that maps to how they generate responses.
You’re not replacing the narrative; you’re adding a citation-ready extraction layer on top of it. Think of it as writing two versions of every section simultaneously: one for the human reading the full post, one for the AI model that needs a quick fragment to cite.
If you want to audit how your existing content performs against these standards, our AI Search Optimizer runs a citation-readiness check and identifies which pages are closest to earning placement in AI Overviews.
What You Should Be Tracking Now
The standard metrics, organic impressions, click-through rate, keyword rankings, are all telling you less than they used to. A page can lose 60% of its clicks while actually gaining visibility inside an AI Overview. You need a different measurement layer.
Track citation frequency across AI platforms. Are your pages being referenced inside ChatGPT responses, Perplexity answers, Gemini summaries, and Google AI Overviews? This is what the share of model metric captures. It measures your brand’s presence in AI-generated answers as a percentage of relevant queries, across each platform separately.
Brands investing in GEO (Generative Engine Optimization) report 30-40% higher AI referral traffic compared to traditional SEO-only strategies. The gap is widening as AI Overviews expand to more query types and more markets.
One platform worth watching: LinkedIn is currently the most-cited domain for professional queries across Google AI Overviews, ChatGPT, Microsoft Copilot, and Perplexity, according to Profound’s March 2026 analysis. For B2B agencies, this means your LinkedIn content, not your blog, may be your highest-leverage citation asset right now. Content published natively on LinkedIn is being surfaced by AI models more often than equivalent content on brand websites.
For clients in specialized verticals where AI search accuracy matters most, like healthcare, legal, and financial services, the AEO Checklist covers the platform-by-platform requirements for earning verified citations with source attribution.
What This Means for Agency Positioning
The 42% traffic drop is already in your clients’ data. They’re seeing it in Search Console. If you haven’t framed it for them, someone else will. The agencies winning new business right now are the ones walking in with a clear explanation of what happened and a specific plan for what to do next.
That plan is not “we’ll fix your SEO.” That plan is a content audit, a format restructure of your highest-value pages, a citation-tracking layer, and a strategy for appearing inside AI Overviews rather than below them.
AEO services are the packaging for that work. The agencies treating AEO as a separate, billable discipline from traditional SEO are the ones landing retainers based on a metric their clients didn’t know existed six months ago.
FAQ: AI Overviews and Content Strategy in 2026
What is causing the 42% organic click drop in 2026?
The 42% decline comes from Google’s expansion of AI Overviews, which place AI-generated answer summaries at the top of search results before any traditional organic links. When users get their answer from the AI Overview, they often don’t click through to the source. Define Media Group’s analysis of Google Search Console data across 64 sites confirmed the traffic never recovered after the initial 16% drop at launch.
Which content types are losing the most traffic to AI Overviews?
Informational and evergreen content is hardest hit. How-to posts, definition guides, comparison content, and listicles targeting high-volume informational queries are the primary casualties. These formats ranked well for years but now often trigger AI Overviews that answer the query directly. Breaking news content is the notable exception, growing 103% over the same period.
Can you get your content cited inside an AI Overview?
Yes. Brands cited inside AI Overviews often see net traffic gains even as overall organic clicks fall. The key is content format: definition-first section openers, inline evidence grounding with named sources, and structured listicle or step-by-step formatting. Gemini uses text fragment anchoring to pull specific sentences from source pages, so content that starts each section with a self-contained, citable definition has a significantly higher chance of appearing in AI-generated answers.
How does Google Discover fit into this?
Google Discover, which surfaces content in the Google app based on user interests rather than direct searches, grew 30% across Define Media Group’s dataset as web search traffic fell. It now drives traffic roughly equal to web search within that portfolio. Discover favors timely, topically relevant content from authoritative sources. For agencies, this means a diversified distribution strategy that includes Discover optimization, not just keyword targeting.
Is LinkedIn really the most cited domain in AI search?
For professional queries, yes. Profound’s March 2026 analysis found LinkedIn is the most-cited domain across Google AI Overviews, ChatGPT, Copilot, and Perplexity when users ask professional or B2B questions. Native LinkedIn articles, posts, and newsletters are being pulled into AI answers more frequently than equivalent content on brand websites. For B2B marketers and agency owners, this is a high-leverage channel that most teams are still treating as a social feed rather than a citation asset.
How quickly can format changes affect AI citation rates?
Most teams see measurable shifts within 4-8 weeks of restructuring high-value pages. AI models re-index content on a rolling basis. Definition-first rewrites and evidence grounding are the fastest wins because they don’t require new research, just restructuring of existing material. A full citation strategy takes longer to build, but the format changes to existing content can start moving the needle almost immediately.
The agencies that figure this out in March 2026 will have a meaningful advantage for the rest of the year. AI Overviews are not a feature Google will walk back. The query types they cover will expand. The only real question is whether your content is in the citation pool or outside of it.
Format is how you get in.
