diff --git a/apps/seo-www/content/one-year-ai-overviews-google-search.mdx b/apps/seo-www/content/one-year-ai-overviews-google-search.mdx index 59f3089c..378d6e9e 100644 --- a/apps/seo-www/content/one-year-ai-overviews-google-search.mdx +++ b/apps/seo-www/content/one-year-ai-overviews-google-search.mdx @@ -24,7 +24,11 @@ When ChatGPT entered the mainstream, it undeniably altered expectations. For the ### Google's response: AI Overviews +<<<<<<< aaron/landing-page-refresh +On top of introducing their own ChatGPT-esque platform, Gemini, Google recognised that it had to respond with a solution integrated within the existing architecture of search that they pioneered and propagated for the last two decades. What began as Search Generative Experience evolved into AI Overviews, and within a year expanded to over 100 countries and territories with more than one billion monthly users. By May 2025, it reported availability in 200+ countries and 40+ languages, describing over 1.5 billion users engaging with AI Overviews globally (Google’s October 2024 rollout and May 2025 expansion updates). +======= On top of introducing their own ChatGPT-esque platform, Gemini, Google recognised that it had to respond with a solution integrated within the existing architecture of search that they pioneered and propagated for the last two decades. What began as Search Generative Experience evolved into AI Overviews, and within a year expanded to over 100 countries and territories with more than one billion monthly users. By May 2025, it reported availability in 200+ countries and 40+ languages, describing over 1.5 billion users engaging with AI Overviews globally ([Google, Oct 2024 rollout announcement](https://blog.google/products/search/ai-overviews-expansion/)) ([Google, May 2025 expansion update](https://blog.google/products/search/google-search-ai-overviews-update/)). +>>>>>>> main ### 2026: The data is here @@ -38,7 +42,11 @@ Google may have prevented an exodus of users by introducing AI Overviews, and ne AI Overviews do not appear on every query, and it seems Google is still figuring out how often AIOs should appear, with the percentage shifting inconsistently. +<<<<<<< aaron/landing-page-refresh +SE Ranking reported AI Overviews appearing on 7.47 percent of sampled US keywords in July 2024, rising to 18.76 percent by November 2024, with UK prevalence at 19.23 percent in the same period. [Semrush](https://www.semrush.com/blog/ai-overviews-study/), analysing more than 10 million keywords, reported 6.49 percent prevalence in January 2025, peaking at 24.61 percent in July 2025 before stabilising at 15.69 percent in November 2025. A real-user browsing panel from Pew Research Center observed AI summaries in 18 percent of Google searches in its March 2025 dataset. +======= SE Ranking reported AI Overviews appearing on 7.47 percent of sampled US keywords in July 2024, rising to 18.76 percent by November 2024, with UK prevalence at 19.23 percent in the same period ([SE Ranking SERP study, 2024](https://seranking.com/blog/ai-overviews-study/)). Semrush, analysing more than 10 million keywords, reported 6.49 percent prevalence in January 2025, peaking at 24.61 percent in July 2025 before stabilising at 15.69 percent in November 2025 ([Semrush AI Overview analysis, 2025](https://www.semrush.com/blog/ai-overviews-study/)). A real-user browsing panel from Pew Research Center observed AI summaries in 18 percent of Google searches in its March 2025 dataset ([Pew Research Center, AI summaries user panel study, 2025](https://www.pewresearch.org/internet/2025/ai-summaries-in-google-search/)). +>>>>>>> main --- @@ -46,7 +54,11 @@ SE Ranking reported AI Overviews appearing on 7.47 percent of sampled US keyword With regard to when AI Overviews are triggered, data shows that the structure of the user's query is a strong influence. +<<<<<<< aaron/landing-page-refresh +Pew’s behavioural analysis found that searches containing ten or more words triggered AI summaries 53 percent of the time, compared to just 8 percent for one- to two-word searches. SE Ranking reported that four-plus word queries triggered AI Overviews 60.85 percent of the time in its November 2024 dataset. [Semrush](https://www.semrush.com/blog/ai-overviews-study/) similarly observed that longer, more specific queries are more likely to produce AI modules. +======= Pew's behavioural analysis found that searches containing ten or more words triggered AI summaries 53 percent of the time, compared to just 8 percent for one- to two-word searches ([Pew Research Center query-length findings, 2025](https://www.pewresearch.org/internet/2025/ai-summaries-in-google-search/)). SE Ranking reported that four-plus word queries triggered AI Overviews 60.85 percent of the time in its November 2024 dataset ([SE Ranking long-tail findings, 2024](https://seranking.com/blog/ai-overviews-study/)). Semrush similarly observed that longer, more specific queries are more likely to produce AI modules ([Semrush AI Overview analysis, 2025](https://www.semrush.com/blog/ai-overviews-study/)). +>>>>>>> main As users increasingly phrase searches as natural-language prompts, AI synthesis becomes more common. @@ -56,7 +68,11 @@ As users increasingly phrase searches as natural-language prompts, AI synthesis AI Overview's early rollout was heavily skewed toward appearing in informational queries, but that pattern is shifting. +<<<<<<< aaron/landing-page-refresh +Informational queries accounted for 91.3 percent of AI-triggering keywords in January 2025, declining to 57.1 percent by October 2025. During the same period, navigational queries triggering AI Overviews increased from 0.74 percent to 10.33 percent, with commercial and transactional categories also rising ([Semrush](https://www.semrush.com/blog/ai-overviews-study/), 2025). +======= Informational queries accounted for 91.3 percent of AI-triggering keywords in January 2025, declining to 57.1 percent by October 2025. During the same period, navigational queries triggering AI Overviews increased from 0.74 percent to 10.33 percent, with commercial and transactional categories also rising ([Semrush intent shift analysis, 2025](https://www.semrush.com/blog/ai-overviews-study/)). +>>>>>>> main This means that organic results are not only under threat for informational searches, but also with branded and purchase-adjacent search queries. @@ -74,7 +90,11 @@ This is even more prevalent on mobile, where screen real estate is even more sca This visual displacement's impact has been felt with significant user behavioural shifts. +<<<<<<< aaron/landing-page-refresh +Seer Interactive analysed 3,119 informational queries across 42 organisations and found that organic CTR declined by 61 percent on queries with AI Overviews, with paid CTR declining by 68 percent over the same window. Ahrefs independently re-ran analysis using December 2025 data and observed a 58 percent lower average CTR for the top-ranking page when an AI Overview was present. +======= Seer Interactive analysed 3,119 informational queries across 42 organisations and found that organic CTR declined by 61 percent on queries with AI Overviews, with paid CTR declining by 68 percent over the same window ([Seer Interactive CTR analysis, 2024–2025](https://www.seerinteractive.com/blog/ai-overviews-impact-ctr/)). Ahrefs independently re-ran analysis using December 2025 data and observed a 58 percent lower average CTR for the top-ranking page when an AI Overview was present ([Ahrefs AI Overview CTR study, 2025](https://ahrefs.com/blog/ai-overviews-ctr/)). +>>>>>>> main These figures do not indicate that search is vanishing, but instead the way users interact with search is - if an AI Overview appears, the need for clicking into an article decreases dramatically. @@ -84,6 +104,12 @@ These figures do not indicate that search is vanishing, but instead the way user The decline in click-through rate, however, cannot be purely attributed to the rise of AI Overviews. The decline in clicks aligns with broader zero-click trends that predate the introduction of AIOs. +<<<<<<< aaron/landing-page-refresh +Clickstream research conducted using Datos and SparkToro estimated that 58.5 percent of US searches and 59.7 percent of EU searches in 2024 ended without any click, meaning no ad, no Google-owned property, and no external website was visited. Only 360 per 1,000 US searches resulted in a click to the open web. Desktop-only figures are lower but rising: in the US, desktop no-click searches increased from 24.4 percent in March 2024 to 27.2 percent in March 2025, with EU and UK figures rising from 23.6 percent to 26.1 percent in the same period. + +AI Overviews did not create zero-click behaviour, but it is reinforcing the trajectory that was already underway. + +======= Clickstream research conducted using Datos and SparkToro estimated that 58.5 percent of US searches and 59.7 percent of EU searches in 2024 ended without any click, meaning no ad, no Google-owned property, and no external website was visited. Only 360 per 1,000 US searches resulted in a click to the open web ([SparkToro & Datos zero-click study, 2024](https://sparktoro.com/blog/zero-click-search-study-2024/)). Desktop-only figures are lower but rising. In the US, desktop no-click searches increased from 24.4 percent in March 2024 to 27.2 percent in March 2025, with EU and UK figures rising from 23.6 percent to 26.1 percent in the same period ([SparkToro desktop search analysis, 2025](https://sparktoro.com/blog/desktop-search-zero-click-update/)). @@ -92,6 +118,7 @@ AI Overviews did not create zero-click behaviour, but it is reinforcing the traj Search and data on a laptop +>>>>>>> main ## What the Data Shows Has Not Fundamentally Changed With so much change in the structure of search results, the behaviour of Google, and the behaviour of search users, some continuity lies in the composition of AI Overviews. @@ -100,7 +127,11 @@ With so much change in the structure of search results, the behaviour of Google, Despite visual shifts, citation patterns remain anchored in traditional ranking strength. +<<<<<<< aaron/landing-page-refresh +75 percent of sites cited within AI Overviews came from the top 12 organic rankings, and 90 percent from position 35 or higher (Botify & DemandSphere, 2025). SE Ranking reported that 92.36 percent of US AI Overviews linked to at least one domain ranking in the top 10, with UK figures at 93.78 percent. +======= 75 percent of sites cited within AI Overviews came from the top 12 organic rankings, and 90 percent from position 35 or higher ([Botify & DemandSphere AI citation study, 2025](https://www.botify.com/blog/ai-overviews-study/)). SE Ranking reported that 92.36 percent of US AI Overviews linked to at least one domain ranking in the top 10, with UK figures at 93.78 percent ([SE Ranking citation overlap findings, 2024](https://seranking.com/blog/ai-overviews-study/)). +>>>>>>> main This substantial overlap suggests that AI synthesis layers draw primarily from the same authority pool that traditional organic rankings have prioritised for years. AI Overviews, and LLM search responses in general, are still drawing their results from Google's rankings, and have not changed that playbook. @@ -108,7 +139,11 @@ This substantial overlap suggests that AI synthesis layers draw primarily from t ### Semantic Relevance Remains Central +<<<<<<< aaron/landing-page-refresh +Botify and DemandSphere also observed that semantic similarity between AI summaries and cited pages strongly correlates with citation likelihood. [Google’s documentation](https://developers.google.com/search/docs/appearance/ai-overviews) explains that AI Overviews use “query fan-out” techniques and do not require additional technical configuration beyond standard indexing and snippet eligibility. +======= Botify and DemandSphere also observed that semantic similarity between AI summaries and cited pages strongly correlates with citation likelihood ([Botify & DemandSphere semantic findings, 2025](https://www.botify.com/blog/ai-overviews-study/)). Google's documentation explains that AI Overviews use "query fan-out" techniques and do not require additional technical configuration beyond standard indexing and snippet eligibility ([Google Search documentation on AI Overviews](https://developers.google.com/search/docs/appearance/ai-overviews)). +>>>>>>> main This means that foundational SEO principles such as structured clarity, semantic alignment, and topical authority remain central to visibility, and articles that answer the user's search intent using natural language will still be regarded as authoritative sources and cited. @@ -116,6 +151,12 @@ This means that foundational SEO principles such as structured clarity, semantic ### Dominant Domains Continue to Dominate +<<<<<<< aaron/landing-page-refresh +Pew’s study found that Wikipedia, YouTube, Reddit, and .gov domains frequently appear within AI summaries. These are domains long recognised for strong authority signals within traditional search. + +The continuity indicates that while presentation evolves, the hierarchy of credibility remains largely intact. + +======= Pew's study found that Wikipedia, YouTube, Reddit, and .gov domains frequently appear within AI summaries ([Pew Research Center citation analysis, 2025](https://www.pewresearch.org/internet/2025/ai-summaries-in-google-search/)). These are domains long recognised for strong authority signals within traditional search. The continuity indicates that while presentation evolves, the hierarchy of credibility remains largely intact. @@ -124,6 +165,7 @@ The continuity indicates that while presentation evolves, the hierarchy of credi AI and digital future concept +>>>>>>> main ## So What Is Next for Search? Two prevailing narratives dominate discussions about search's future. The first declares SEO dead—that generative interfaces will replace ranked links entirely, and that optimising for Google will become as relevant as optimising for Yahoo. The second insists that the bulk of the change has already occurred, and that this is what search will look like for the foreseeable future. @@ -134,9 +176,15 @@ We at Fluid Posts believe in neither of these extremes (which might have been un The popularised, overly simplified "SEO is dead" argument goes something like this: users are abandoning Google for ChatGPT and Perplexity, AI will synthesise answers without needing to cite sources, and the era of click-based discovery is over. +<<<<<<< aaron/landing-page-refresh +The numbers tell a different story. Google still processes roughly 210 times more searches than ChatGPT (Ahrefs, 2025). More importantly, Google sends 190 times more traffic to websites than ChatGPT does — ChatGPT’s click-through rate is 96 percent lower than Google’s because the product is designed to keep users in conversation, not send them elsewhere. + +Even among users who have adopted AI tools, Google search usage has increased, not decreased (SparkToro, 2024). The “exodus” narrative is not supported by behavioural data. +======= The numbers tell a different story. Google still processes roughly 210 times more searches than ChatGPT ([Ahrefs, 2025](https://ahrefs.com/blog/chatgpt-has-12-percent-of-googles-search-volume/)). More importantly, Google sends 190 times more traffic to websites than ChatGPT does — ChatGPT's click-through rate is 96 percent lower than Google's because the product is designed to keep users in conversation, not send them elsewhere. Even among users who have adopted AI tools, Google search usage has increased, not decreased ([SparkToro, 2024](https://sparktoro.com/blog/new-research-google-search-grew-20-in-2024-receives-373x-more-searches-than-chatgpt/)). The "exodus" narrative is not supported by behavioural data. +>>>>>>> main Then there's the question of where AI tools get their answers. ChatGPT Search is powered by Microsoft Bing's index. Google's AI Overviews draw from their own organic rankings — as aforementioned, 92 percent of US AI Overviews link to at least one domain in the top 10. These systems are not replacing search indexes, but being built on top of them. @@ -170,7 +218,11 @@ Here at Fluid Posts, our experience confirms this interpretation. One of our cli ### The trajectory and our bet +<<<<<<< aaron/landing-page-refresh +This does not mean search will stay static. Google’s market share has dipped below 90 percent for the first time since 2010. Younger users show higher preference for ChatGPT. AI-driven traffic to retail sites surged 1,200 percent since mid-2024, and users arriving via LLMs convert at seven times the rate of Google traffic (Forbes, 2025). +======= This does not mean search will stay static. Google's market share has dipped below 90 percent for the first time since 2010. Younger users show higher preference for ChatGPT. AI-driven traffic to retail sites surged 1,200 percent since mid-2024, and users arriving via LLMs convert at seven times the rate of Google traffic ([Forbes, 2025](https://www.forbes.com/sites/kevinkruse/2025/08/14/seo-is-dead-3-strategies-to-win-in-the-age-of-ai-search/)). +>>>>>>> main The trajectory is clear: AI interfaces will continue to grow in prominence, but the systems determining what those interfaces surface remain anchored in traditional search infrastructure. We believe this tracks. Google has been the leading presenter of information for two decades, indexing and ranking more content than any system in human history. The algorithm they have devised, refined across trillions of queries and under constant adversarial pressure, has been king of search for two decades for a reason—it works. @@ -178,4 +230,9 @@ Could AI systems develop independent authority signals? Absolutely. But building In any case, even if LLMs manage to devise a completely independent infrastructure to evaluate authority signals, it will not happen overnight, and when that shift occurs, it will be noticed. Case in point: when Reddit started being cited heavily by ChatGPT, it did not occur in secret. +<<<<<<< aaron/landing-page-refresh +## For Now... + +======= +>>>>>>> main For now, the presentation of search results is new, but authority remains the central currency. Which theory ultimately proves correct—whether AI fully displaces traditional search or remains a layer on top of it—will unfold over the next several years, as everyone watches with three eyes open. diff --git a/apps/seo-www/src/routes/blog/$.tsx b/apps/seo-www/src/routes/blog/$.tsx index 34bfc4a5..ca4b5a3e 100644 --- a/apps/seo-www/src/routes/blog/$.tsx +++ b/apps/seo-www/src/routes/blog/$.tsx @@ -26,5 +26,19 @@ export const Route = createFileRoute("/blog/$")({ function Page() { const { data, postsOverview, tree } = Route.useLoaderData(); - return ; + return ( + + ); } diff --git a/packages/content/content-collections.ts b/packages/content/content-collections.ts index a913c824..76562641 100644 --- a/packages/content/content-collections.ts +++ b/packages/content/content-collections.ts @@ -36,9 +36,8 @@ const mdxTransformer = async < [ rehypeExternalLinks, { - target: "_blank", rel: ["noopener", "noreferrer"], - content: { type: "text", value: " 🔗" }, + target: "_blank", }, ], ],