AI Overviews Cut Clicks by 58%. The Damage Isn't Spread Evenly.
The first major Ahrefs study in April 2025 put the AI Overviews click loss at 34.5%. The December update pushed it to 58%. If you run SEO for a health publisher or a B2B SaaS blog, that number probably matches what you've been watching in Search Console over the past year. If you run ecommerce, you might have barely noticed.
That gap is the part most coverage misses. The 58% figure is an average across 300,000 keywords, and averages in SEO are roughly as useful as knowing the average depth of a swimming pool before you dive in. Some categories are drowning. Others are standing in ankle-deep water wondering what everyone is yelling about.
This is a guide to figuring out which group you're in and what to do about it either way. (If you want the broader picture of how Google's ranking system interacts with AI features, we covered the full signal breakdown in our guide to how Google decides what ranks.)
What the 58% Number Actually Means (and What It Hides)
Ahrefs analyzed 300,000 keywords in December 2025: half triggered AI Overviews, half didn't. For keywords with an AI Overview, the average position-one CTR dropped from 7.3% to 1.6% over two years. That's the 58% figure making the rounds.
Seer Interactive's study, which analyzed 3,119 queries across 42 client organizations with 25.1 million organic impressions, found something arguably worse: organic CTR fell 61%, from 1.76% to 0.61%. Paid CTR dropped 68%, from 19.7% to 6.34%.
These are real numbers from real datasets. I'm not going to soften them. But the panic articles tend to make the same mistake: they treat the average as if it applies equally to every query type and every business.
It doesn't. Not even close.
Branded searches actually see an 18% increase in CTR when an AI Overview is present, according to research from Amsive. Ecommerce queries trigger AI Overviews on roughly 3 to 14% of searches, depending on whose data you trust, down from 29% when the feature first rolled out. Meanwhile, 82% of health-related queries now show an AI Overview.
The headline number is scary. The category-level breakdown tells a more useful story.
Health, Finance, and Tech Are Taking the Worst of It
The categories getting hit hardest share one trait: their most-searched queries are answerable in a paragraph.
82% of health-related searches now trigger an AI Overview. "Symptoms of iron deficiency," "how long does strep throat last," "is ibuprofen safe with blood thinners." Google can summarize these from authoritative sources without the user ever needing to click. For healthcare publishers and health-focused content sites, this is close to existential.
Tech and finance publishers are in a similar spot. Several have reported 40% or more traffic losses over the past 18 months, and reporting from Search Engine Land suggests the losses are accelerating rather than flattening. The queries that built their audience ("what is a Roth IRA," "how to set up a VPN," "best budgeting app") are exactly the kind Google's AI handles comfortably.
The mechanism is straightforward: these queries have consensus answers that can be sourced from a handful of authoritative pages. When Google can assemble a reliable summary from WebMD, Investopedia, or a major tech publication, there's less incentive to send users elsewhere. The sites getting cited in those summaries are doing fine. Everyone else is watching their traffic charts slide.
Informational queries starting with "best" now have an 83% AI Overview trigger rate.
That distinction matters more than industry labels. It's not really about whether you're in health or tech or finance. It's about whether the answer to your target query fits in three paragraphs. If it does, Google will probably try to answer it without sending anyone your way.
We covered the country-level data on this recently, where Sistrix found AI Overviews had taken 265 million monthly clicks in the UK alone. The category breakdown there told the same story: informational and educational content absorbed the bulk of the damage while product and local searches held up.
Ecommerce and Local Search Got Lucky (for Now)
If you're running an ecommerce operation, the numbers are more reassuring than the headlines suggest.
Only 3 to 14% of shopping queries trigger AI Overviews in early 2026. That's actually down from 29% during the initial rollout. Google figured out pretty quickly that summarizing product listings doesn't serve users, or its own ad revenue, particularly well.
Transactional intent is the moat. "Buy running shoes" doesn't get an AI Overview. "Best running shoes for flat feet" increasingly does, because it's informational first and transactional second. The line between safe and exposed comes down to intent, not industry.
Local searches follow a similar pattern. "Plumber near me" still sends clicks. "How to fix a leaking faucet" mostly does not. If your SEO strategy leans heavily on top-of-funnel informational content to feed a conversion path, that top of the funnel is the part AI Overviews are compressing.
I wouldn't get too comfortable, though. Current projections suggest AI Overviews will appear on 40% of shopping queries by end of 2026 if the expansion rate holds. That's the number I'd be watching if I were running ecommerce SEO right now.
Getting Cited Changes the Math, but Don't Count on It
Seer Interactive's data shows that being cited in an AI Overview is worth 35% more organic clicks and 91% more paid clicks compared to not being cited at all. That sounds promising until you look at the citation trend.
AI Overview citations from top-10 ranking pages dropped from 76% to 38% over the past year. Google is pulling from a wider and less predictable set of sources for these summaries. Ranking on page one used to be the whole game. Now it gets you considered for citation. Not guaranteed.
From what I've seen across the studies, the sites that consistently get cited tend to share a few patterns: they structure content with clear, direct answers near the top of the page, they have depth beyond the surface-level response, and they've built topical authority across clusters of related content. Which is, honestly, what Google's own documentation has been nudging people toward for years. The difference is that it actually matters now instead of being a nice-to-have.
We also looked at a related problem recently: pages that rank in the top 10 but get completely ignored by AI Overviews. The retrieval gap between ranking and citation is real, and it seems to be widening.
The Diagnostic Sequence That Tells You Where You Stand
Most advice about AI Overviews boils down to "create better content," which is about as useful as telling someone who lost their job to "make more money." A more productive approach starts with figuring out exactly how exposed you are.
Open Search Console. Filter for queries with more than 100 impressions over the past 90 days. Sort by CTR ascending. The bottom 20% of that list are almost certainly your AI Overview casualties. If you see queries where impressions are stable or growing but clicks are declining, that's the signature pattern.
Categorize the damage. Pull that bottom 20% into a spreadsheet and tag each query as informational, navigational, or transactional. If most of your casualties are informational ("what is," "how to," "best X for Y"), you have a content strategy problem to solve. If transactional queries are losing clicks too, check whether you're appearing in the AI Overview citation. Different problem, different fix.
For informational casualties, go deeper than the summary can. Google's AI Overview answers the surface question. Your content needs to answer the question behind the question. "How to set up Google Analytics" gets summarized in three sentences. "How to set up Google Analytics event tracking for a multi-step checkout with server-side tagging" does not, because the answer is too complex and too specific to compress. The content that survives AI Overviews is the content that's too useful to summarize.
For citation gaps, restructure. Put a clear, direct answer in the first 100 words of your page. Use structured headers. And build clusters of related content that establish authority on the topic. Isolated blog posts rarely get cited. Interconnected topic clusters do. (This part seems to matter more than most people realize.)
Build brand queries. From what I've seen, this is the most durable play and the one most teams skip because it's slow. Branded searches with AI Overviews actually see higher CTR. People searching for your brand by name are going to click through regardless of what Google puts at the top of the page. Every dollar you spend on brand awareness feeds this channel, whether or not you're thinking about it in SEO terms.
And there's a complicating factor beyond just AI Overviews: even queries without an AI Overview saw a 41% organic CTR decline in Seer's data. Users are going to ChatGPT, Perplexity, and social platforms before they ever reach Google. The traffic erosion is a multi-front problem. AI Overviews are just the most measurable piece of it.
The sites treating this as a content quality issue are mostly right. The ones treating it as someone else's problem are probably in for an unpleasant Q4.
