The Great Decoupling: Why SEO Clicks Are Vanishing in 2026
Your impressions are up and your clicks are flat or falling because AI answers now resolve the query on the results page, so people read your content without ever visiting your site. This is the great decoupling: visibility and traffic, once locked together, have come apart. The fix isn't to chase the lost click. It's to measure whether AI engines cite you, and to engineer your content so they do.
If you run growth or marketing and your Search Console graph shows the impressions line climbing while the clicks line goes sideways, you're not imagining it. The gap is real, it's widening, and it changes how you should report performance to your board.
What is the great decoupling in SEO?
For most of search history, impressions and clicks moved together. More visibility meant more visits. Pew Research Center broke that link with hard data in July 2025. Across 68,879 Google searches from 900 US adults tracked in March 2025, people clicked a traditional search result on 8% of visits when an AI summary was shown, against 15% when it wasn't. (Pew Research Center)
So when an AI summary appears, the click rate roughly halves. Your content can still rank, still earn the impression, still feed the answer, and still get no visit. That's the decoupling in one number.
It gets starker. Pew found people clicked a source link inside the AI summary itself on just 1% of visits. The summary answers the question; the citation sits there unclicked. People were also more likely to end the session entirely after a page with an AI summary (26% of those pages) than after a standard results page (16%).
Why are my search clicks disappearing?
Two forces, both pointing the same way.
AI answers absorb the click. Google's AI Overviews read several sources and write the answer on the results page. The query is resolved before anyone scrolls. Seer Interactive, analysing roughly 25 million impressions, found organic click-through rate dropping from 1.62% to 0.61% when an AI Overview was present. (Search Engine Journal)
Zero-click is now the default. Semrush's 2025 study put the share of US searches ending without any click to an external site at 58.5%. (Semrush) The behaviour was already shifting before AI. AI Overviews poured fuel on it.
How widespread are AI Overviews? More volatile than the scare headlines suggest. Semrush tracked them across 10 million-plus keywords: 6.49% of queries in January 2025, a peak of 24.61% in July, settling to 15.69% by November. (Semrush) Google has been pushing them down the funnel, into commercial and comparison queries, the ones closest to a purchase. That's the part that should worry a CMO.
How much are AI Overviews really cutting clicks?
The verified figures, by source:
| Source | Metric | Finding |
|---|---|---|
| Pew Research Center (Jul 2025) | Traditional link clicks, AI summary present vs absent | 8% vs 15% of visits |
| Pew Research Center (Jul 2025) | Clicks on a source inside the AI summary | 1% of visits |
| Seer Interactive | Organic CTR, AI Overview present vs absent | 0.61% vs 1.62% |
| Semrush (2025) | US searches ending with no external click | 58.5% |
| Semrush (2025) | AI Overview prevalence, Jan to Nov 2025 | 6.49% rising to 24.61%, then 15.69% |
One caveat worth flagging: Google publicly disputed the Pew study, arguing the methodology over-counts the effect. (ppc.land) Read the numbers as a strong directional signal, not a precise universal constant. The direction isn't in question. Every independent dataset points the same way.
Which industries feel the decoupling first?
The pattern hits informational and advisory content hardest, because that's exactly what an AI summary is good at compressing.
- Professional and advisory services. Legal, consulting, and financial guidance get synthesised into the answer. The expertise that pulled visitors now feeds the summary, often without a visit.
- Healthcare and life sciences. Medical and symptom queries are prime summarisation targets. People read the answer and stop.
- Finance and fintech. Regulatory explainers, product comparisons, and planning advice increasingly resolve in the overview.
If your funnel starts with someone Googling a question your content answers, this is your problem now, not a future one.
What should marketers do about the great decoupling?
Stop optimising for a metric that's structurally shrinking. Start optimising for the one that's growing: whether AI engines pick you as a source. This is the discipline behind GEO and how it differs from traditional SEO.
The academic groundwork is already done. The 2023 GEO paper from researchers at Princeton, Georgia Tech, the Allen Institute for AI, and IIT Delhi showed that specific content changes, like adding citations, statistics, and quotations, lifted a source's visibility in generative engine answers by up to 40%. (GEO: Generative Engine Optimization, arXiv) Citation, in other words, is engineerable.
Four moves that follow from the data:
- Track citations, not just clicks. Add AI visibility to your reporting. Are ChatGPT, Perplexity, Gemini, and Google's AI Overviews naming you for the queries that matter? That's the new front rank.
- Write so engines can quote you. Lead with a clear answer. Back claims with figures and named sources. Structure content so a model can lift a clean, attributable passage. The GEO research is explicit that this is what gets picked up.
- Reframe the metric for the board. A cited brand still wins. Semrush data shows brands cited inside AI Overviews earning meaningfully higher click-through than uncited ones. Being the source beats being result number four.
- Build entity authority. AI engines trust sources they can recognise and corroborate. Consistent structured data, a coherent topical footprint, and third-party validation make you a safer source to cite.
This is the heart of generative engine optimisation as a service: engineering your content to be the source AI answers are built from, then measuring it.
A word of caution, because it matters. Plenty of agencies are already chasing AI citations with manipulation tactics, and search engines are starting to penalise them. The durable approach is to earn the citation by being genuinely the best, most verifiable answer. That's the only version that survives the next algorithm update.
Frequently asked questions
What does the great decoupling mean for SEO?
It means impressions and clicks no longer move together. Your content can rank and appear more often while sending fewer visits, because AI summaries answer the query on the results page. Pew Research found clicks roughly halve when an AI summary is present, from 15% to 8% of visits.
Are AI Overviews really reducing website traffic?
Yes, by every independent measure, though the exact size is debated. Pew found a near-halving of clicks. Seer Interactive measured organic CTR falling from 1.62% to 0.61% with an AI Overview present. Google disputes the methodology, but not one dataset shows traffic holding steady.
How do I measure success if clicks are falling?
Track AI citations and share of voice inside AI answers, alongside clicks. Measure whether engines like ChatGPT, Perplexity, Gemini, and Google AI Overviews name you for your priority queries, then watch for downstream effects: branded search, direct visits, and inbound enquiries.
What is GEO and how is it different from SEO?
SEO optimises to rank in a list of links. GEO, generative engine optimisation, optimises to be cited inside an AI-generated answer. The 2023 GEO research paper showed that adding citations, statistics, and quotations can raise a source's visibility in generative answers by up to 40%.
The bottom line
The great decoupling isn't a glitch to wait out. It's the new shape of search, and the click you used to count is being absorbed into the answer. My view: marketers who keep reporting clicks as the headline number are measuring the wrong thing and will look out of touch in front of their board within a year. The metric that matters now is whether AI engines cite you. Measure that, engineer for it, and earn the citation honestly. The visit was always a proxy for influence anyway. Influence is what's left, and it's measurable.
References:
- Pew Research Center: Google users are less likely to click links when an AI summary appears (July 2025)
- Search Engine Journal: Pew Research confirms AI Overviews eroding the web ecosystem
- Semrush AI Overviews Study
- GEO: Generative Engine Optimization (arXiv 2311.09735)
- Google disputes the Pew study (ppc.land)
More on how we approach it: our AEO practice.
Sotiris Spyrou
Sotiris Spyrou is the founder of VerityAI, a Responsible AI advisory for boards and AI-deploying businesses. With 27 years across agencies, global in-house roles, and the C-suite, he advises leaders on AI governance and risk, and on answer-engine visibility engineered without the dark patterns the rest of the industry is getting penalised for. He is the author of TRANSFORM, AI Moats, and Ethical AI.
Founder at VerityAI