Google is shaking up search results again, and this time several structured data types are getting the axe. The company is discontinuing Rich Results support for Book, Course, ClaimReview, and SpecialAnnouncement markup. Meanwhile, niche SERP elements are vanishing: “Today’s Doodle” boxes, nutrition facts panels, “nearby offers” snippets, local bike-share status, and practice problem results. Changes roll out over the coming weeks.
What This Actually Means
If you spent months implementing markup for books, courses, or fact-checking — congratulations, that work just turned into a pumpkin. Unsupported code will simply be ignored, but this can cause minor traffic and CTR drops, especially if Rich Results were your main click source.
Google doesn’t explain the reasons, but the logic is clear: the company is simplifying search results, keeping only formats that are massively used and genuinely improve user experience. Niche markup used by thousands of sites out of millions becomes a maintenance burden.
Gemini Takes Over Google Maps
While Google removes some features, it adds others — and here’s where things get interesting. Google Maps is getting Gemini integration. Users can now point their phone camera at a store or restaurant and ask “What is this place and why is it popular?” — the system delivers instant AI answers based on 250 million places and Street View data.
Navigation is changing too. Instead of bland “turn right in 200 meters,” the system references landmarks — “turn at the red building with the coffee shop.” For local SEO, this means Google Business Profile descriptions become even more important. AI will digest this content and use it in user responses.
Ironwood TPU: Google Challenges NVIDIA
In hardware news, Google deployed new Axion CPUs and seventh-generation Ironwood TPUs. The company claims these processors outperform NVIDIA GB300 in performance and energy efficiency. Ironwood achieves a 10x increase in peak performance compared to TPU v5p.
Why does this matter to SEO? More powerful hardware means more sophisticated AI models in search. Google can process more context, understand more complex queries, and deliver more relevant results. Simple SEO tricks will work progressively worse.
Gemini API and Structured Outputs
Google significantly expanded JSON Schema support in Gemini API for structured outputs. Developers can now use schema libraries like Pydantic and Zod directly without conversion layers. The update supports key JSON Schema keywords: anyOf for union types, $ref for recursive structures, minimum/maximum for numeric constraints.
For those integrating AI into services, this simplifies life. For everyone else — it’s a signal that Google is betting on structure and machine-readable content. Schema.org isn’t going anywhere; priorities are just shifting.
URLs Are Case-Sensitive: 2025’s Surprise
Google confirmed what many forget: canonical URLs, robots.txt, and sitemap entries are case-sensitive. Inconsistency between `<link rel=canonical>`, sitemaps, and internal links can fragment indexing signals.
Simply put: if your sitemap has `example.com/Page`, canonical has `example.com/page`, and internal links point to `example.com/PAGE` — Google may treat these as three different pages and dilute link equity. Solution is simple: standardize everything to lowercase and move on.
Algorithm Fluctuations in Late October
Between October 28-31, major rank tracking tools recorded unconfirmed Google algorithm fluctuations. Sites with thin content, aggressive affiliate clusters, and duplicate pages took hits. Google officially confirmed nothing, but the pattern is obvious: content quality remains the priority.
YouTube SEO Gets AI Assistant
YouTube launched Ask Studio AI — a natural language analytics assistant. You can now ask “Which of my videos perform best in Germany?” and get answers without digging through dashboards. Built-in A/B testing for video titles and thumbnails also launched.
This directly impacts CTR and watch time, making video SEO more data-driven. If you work with YouTube, these tools become mandatory for optimization.
Brand Mentions Beat Backlinks
The most important news for old-school SEO: Ahrefs analysis shows sites frequently mentioned (even without backlinks) are far more likely to be cited in AI search answers. Large language models reindex known entities, so strong brand footprint is crucial.
This flips classic link-building strategy. Before: get a link from an authoritative site = profit. Now: let people talk about you on forums, social media, comments, Reddit — even without links. AI notices and mentions your brand more often in answers.
What SEO Pros Should Do
- Urgent structured data audit. Check for Book, Course, ClaimReview, and SpecialAnnouncement markup. Remove or replace with supported types. Prioritize FAQPage, HowTo, and Organization — they work with AI Overview.
- Redefine off-page strategy. Move from exclusively link-focused tactics to brand building through PR, community participation, content that gets mentions. Brand visibility is now critical for AI Overview appearance.
- Optimize for Google Maps with AI. Enrich Google Business Profile with detailed content. Focus on reviews, service descriptions, and clear landmark positioning near your location. AI will use this in user responses.
- Standardize URLs and metadata. Implement a lowercase-only policy for all URLs, canonical links, and sitemaps. Use Search Console to verify consistency.
- Prepare for smarter search. Ironwood TPU means Google can run more sophisticated models. This doesn’t mean “end of SEO,” but tactics must become more sophisticated. Simple tricks like keyword stuffing and PBN networks will work even worse.
Reality
Google is simultaneously simplifying and complicating search. Simplifying for users by removing excess SERP elements and implementing AI assistants. Complicating for SEO pros by changing the rules: from links to mentions, from markup to content, from technical tricks to real value.
Good news: quality content and strong brand have always been the foundation of successful SEO. Bad news: now there’s absolutely no way around it. AI models are smarter than any past algorithm, and deceiving them is much harder. Either play fair or prepare for constant updates that will wipe out positions.
No comment