Position-1 organic click-through rate (CTR, the share of people who click a result after seeing it) on queries where AI features appear has collapsed from 27% to 11% (SISTRIX, March 2026). Brands cited inside the AI Overview see +35% organic clicks and +91% paid clicks versus uncited competitors on the same query (Digital Applied, March 2026). Those two numbers, side by side, are the operational truth of SEO in 2026.
This is one of seven spokes under What Google I/O 2026 Just Changed. It exists because the data is scattered across a dozen reports of varying quality, and most small-business owners I talk to are still operating on traffic models from before AI Overviews shipped. Let me lay out what we know, where the numbers come from, which ones to trust, and what they mean for what you should do next.
The single most useful chart in 2026 SEO
If you read nothing else, read this: position-1 CTR on AI-impacted queries fell from 27% to 11%, while brands cited inside the AI Overview (AIO, the synthesized answer box Google now puts above the links) gained +35% organic and +91% paid clicks against their uncited competitors.
The headline says CTR collapsed. The fine print says it collapsed for the uncited. If you are cited inside the AI answer, the math reverses. The +35% organic and +91% paid numbers from Digital Applied are post-AIO measurements on the same queries: your competitors saw a drop, you saw a gain.
This is why I keep telling clients: rank tracking, as a single metric, is becoming partial fiction. The question is no longer "where do I rank?" The question is "do I get cited?"
The data sources, sorted by quality
Not all of these numbers are equally trustworthy, and methodology is what separates the load-bearing figures from the noise. The table below ranks every source I'm working from, highest to lowest confidence:
| Tier | Source | Headline figure | Why this confidence level | |---|---|---|---| | 1 | Pew Research Center | 46.7% CTR decline (15% without AIO, 8% with), across 68,000 queries | Largest sample, published methodology, no conflict of interest | | 1 | DMG Media (UK CMA filing) | Desktop CTR fell 89% (25.23% to 2.79%), mobile fell 87% | Public-record, sworn regulatory submission | | 1 | Chegg quarterly financials | 49% educational-traffic loss (Jan 2024 to Jan 2025) | Audited public filings; cleanest informational-publisher case study | | 2 | SISTRIX (March 2026) | Position-1 CTR on AI-impacted queries fell 27% to 11% | Broad index methodology; described but not fully published | | 2 | Ahrefs | Informational-keyword CTR decline of 34.5% | Established firm, named method | | 2 | Similarweb | Zero-click searches rose 56% (May 2024) to 69% (May 2025) | Large-scale aggregation | | 2 | Amsive | Branded-search CTR rose +18% | Established firm, named method | | 2 | Authoritas | AIO placement volatility ~70% over 2-3 months | Established firm, named method | | 2 | Advanced Web Ranking | AIO word count ~169; first organic result pushed ~1,674 pixels down on an 8,000-keyword sample | Established firm, named method | | 3 | Digital Applied (March 2026) | +35% organic and +91% paid clicks for cited vs uncited brands | Single-source, but consistent with our own client data | | 3 | 12AM Agency | YoY referral traffic down 33%, informational-query CTR down 61% | Mostly unattributed; treat as directional | | 3 | Dr. Matt Lynch | 20-40% traffic decline range | Unattributed; supplementary only |
The pattern in that table is what matters: the tier-1 rows survive cross-examination, the tier-3 rows are vibes until someone confirms them.
The magnitude range: what your actual decline might look like
If you're trying to figure out what to expect on your own site, the honest answer is "it depends, but here are the rails":
- Common case for informational content with AI Overview triggering: 35-50% CTR decline. Pew (46.7%), Ahrefs (34.5% informational), and SEJ Jan-Feb 2026 testing (38% organic-click reduction) all land in this band.
- Worst case, mainstream-news publishers: 87-89% (DMG Media). When the AIO summarizes an entire news story, the click stops happening.
- Best case, branded queries: +18% CTR lift (Amsive). Brand recognition compounds in this environment.
- Cited-brand bonus: +35% organic and +91% paid for those quoted inside the AIO (Digital Applied).
The spread is the point: a site that's 80% informational and 20% branded sees a very different aggregate than a site that's the other way around, and verticals shift the math again. 12AM Agency's data suggests B2B Tech and Education see AI Overview coverage above 80% of queries, while local-intent queries get AI coverage under 15%. Local businesses are partially insulated. National informational publishers are not.
What zero-click did
Zero-click searches (searches that end inside Google without an outbound click) rose from 56% in May 2024 to 69% in May 2025 (Similarweb). SparkToro and Datos put U.S. zero-click at 58.5%. Search Engine Journal's January-February 2026 testing measured zero-click rising from 54% to 72% within sessions where AI was actively used.
The trend line is consistent across every credible source: more than half of Google searches now end inside Google's interface. For some categories, more than two-thirds.
This is the "click-based economy collapsing" claim in concrete terms. The web didn't go away. The traffic that used to flow from the SERP to publishers did.
The Amsive +18%: why branded queries are a different game
The single most overlooked number in the I/O 2026 data set is Amsive's measurement that branded search CTR rose 18% in the AI Overview era. Every other number says CTR fell, which makes branded the exception worth understanding.
The mechanism: AI Overviews concentrate authority signals. When the AIO mentions a brand by name, the user often searches for that brand directly, so branded clicks went up while generic-keyword clicks went down. The middle of the SEO funnel got squeezed; the brand-recognition end got rewarded.
If you have spent the last decade chasing keyword rankings instead of building brand presence, this is the data point that tells you to invert the priority.
The methodological critique: which numbers to trust
A few patterns to keep in mind when you read SEO traffic-decline numbers in the wild:
- "33% YoY decline" with no source cited (12AM Agency): could be anything. Use as vibes, not analysis.
- Single-vendor self-reports without methodology: every SEO tool company has an incentive to publish dramatic numbers. Read SISTRIX, Ahrefs, Semrush data as directional, not definitive.
- Anything from Google itself about "click quality" or "user engagement" inside AI Overviews: unverifiable. Google does not let publishers validate the underlying engagement data.
- The strongest numbers are Pew Research (n=68,000, independent), DMG Media (regulatory filing), Chegg (audited financials), Similarweb (large-scale aggregation), and SISTRIX (broad index methodology).
If you're presenting this data to a CFO or a board, lead with Pew and DMG Media's Competition and Markets Authority (CMA, the UK regulator whose filings are sworn public record) submission. They will survive cross-examination.
What this means for what you do next
The data is consistent enough to drive three strategic moves:
Stop optimizing only for rank. A page that ranks first but doesn't get cited inside the AIO is now worse off than a page that ranks third but does. Old metrics measure the wrong thing.
Build brand presence wherever you can. The +18% branded-CTR lift is the cleanest "AI-era moat" signal in the dataset. It's also the longest-time-horizon investment, so start now.
Invest in citation engineering on your top-converting pages. This means the operational mechanics that get you quoted: single-sentence answers in the first 100 words, stable factual units, explicit entity references, schema that matches your prose claims. How to actually get cited in an AI Overview walks through the six mechanics.
The brands that figured this out in 2025 are the ones now showing up inside AI Overviews on the queries that matter to them. The ones that didn't are watching their traffic charts go down and blaming "the algorithm." The optimization target moved, and the sites still chasing rank are measuring a number that stopped mattering.
Sources
- Google AI Overviews impact on publishers (Matt G. Southern, Search Engine Journal; Pew n=68k, Ahrefs, DMG Media CMA filing, Chegg financials, Similarweb, Amsive)
- The search changes SEOs did not expect (Luis Rijo, PPC Land; SISTRIX March 2026)
- Google I/O 2026 SEO update (Tanner Medina, Launchcodex; Digital Applied, SparkToro/Datos, QuickSEO)
- SEO Pulse: core update amid I/O AI overhaul (Matt G. Southern, SEJ)
- Top Search Announcements - I/O 2026 (Alexander Soliman, What's On SEO)
- Search becomes an AI agent (Rich Sanger)
- 2026 Publisher's Guide to AI Overviews (Robert Portillo, 12AM Agency)
- The hidden threat (Dr. Matthew Lynch)