Google's AI Overviews are answering questions before users click. Top-3 ranked pages saw 18-34% CTR drops in May 2026. What this means and how to adapt.
Something strange happened to a lot of website owners in May 2026. Their Google Search Console numbers looked fine at first glance. Rankings were holding. Impressions were steady or even growing. But clicks were down — in some cases dramatically. Not 5% or 10%, but 18%, 25%, 34% drops for pages that hadn't moved from their top-3 positions.
The culprit wasn't a manual penalty or an algorithm update that reshuffled the rankings. It was AI Overviews — Google's AI-generated answer summaries that appear above organic results and answer the user's question before they ever see a link to click.
When you search for most informational queries on Google now, you'll see an AI-generated answer block at the top of the results page. Google calls these AI Overviews. They pull information from multiple sources across the web, synthesize it into a direct answer, and present it in a format that often makes clicking through to a source unnecessary.
The query "how to remove a stripped screw" used to send users to WikiHow or a popular DIY blog. Now they get a step-by-step AI answer at the top of the page. The article that ranked first still shows up below it. It still gets impressions. But a large portion of users who would previously have clicked through to read the answer are now satisfied by the AI summary and don't click at all.
Google removed the AI Overview click-tracking signal in early May 2026, making it harder for publishers to directly measure the impact. But the pattern is visible in the data: pages in the top 3 for high-volume informational queries are seeing CTR drops of 18-34% in categories where AI Overviews now appear consistently.
Every significant Google algorithm update over the past 20 years changed which pages ranked. Panda targeted thin content. Penguin targeted manipulative link building. Core updates reshuffled quality assessments. In every case, the underlying model was the same: earn a higher ranking, get more clicks. The better your SEO, the more traffic you get.
AI Overviews break that model. The ranking itself still matters — Google's AI pulls from highly-ranked sources to generate its summaries, so being ranked #1 still increases the chance your content is cited. But being cited and being clicked are now different things. Your content can contribute to an AI answer while the user never visits your page. For ad-supported publishers and affiliate sites, this is an existential challenge. For businesses that use content to drive conversions, it's a significant shift in the relationship between organic visibility and actual traffic.
The content that holds up best in an AI Overview environment tends to share a few characteristics. First, it contains something that can't be easily summarized in a paragraph — original research, proprietary data, unique experiences, or nuanced analysis that requires context to understand. An AI can summarize the steps for changing a tire. It can't summarize a writer's 10 years of personal experience testing running shoes across different terrains and climates in a way that captures the value of that experience.
Second, it contains genuinely controversial or contested opinions. AI Overviews are designed to give factual answers, not opinionated ones. "What's the best project management software" can get an AI Overview. "Why I stopped recommending Notion after three years of daily use" probably won't, because it's framed as a personal opinion rather than a factual question.
Third, it requires action. Transactional content — booking pages, product pages, tool comparison pages where users need to sign up or purchase — is less vulnerable to AI Overviews because the AI can't complete the transaction for the user. Ranking for "best email marketing platform for e-commerce" is more durable than ranking for "how does email marketing work."
One category of search that remains largely protected from AI Overview impact is branded and navigational queries. When someone searches for your brand name specifically, or types your domain to navigate to your site, AI Overviews don't typically appear. These users already know what they want and where to find it.
This makes building genuine brand recognition increasingly valuable as a traffic strategy. An audience that comes back to your specific site — because they trust your perspective, follow your work, or know your brand — is more valuable than anonymous organic traffic that can be intercepted by AI summaries. Email lists, direct subscribers, and communities matter more than they did when organic search was the primary growth channel.
The sites adapting fastest to the AI Overview era share a few tactical approaches. They're publishing more original data: surveys, experiments, proprietary analyses that AI can't generate because the underlying research doesn't exist anywhere else. They're structuring content around first-person experience: "I tested this for 90 days and here's what actually happened" is harder to flatten into a generic AI answer than "here are the best options according to experts."
They're also shifting investment toward conversion-focused content over purely informational content, recognizing that informational traffic is increasingly being absorbed at the search results page rather than on their site. And they're doubling down on email and community channels to build audiences that return directly, insulated from changes in organic search behavior.
SEO isn't dead, but the rules changed faster in 2026 than at any point in the past decade. Ranking still matters — it influences which sources AI cites and which users find when they scroll past the AI summary. But ranking alone no longer translates to clicks the way it used to. The sites that adapt fastest are building content that AI references but can't replace, and audiences that return directly rather than discovering through search every time.
