SEO Recap covering October 27, 2026: Google downplays frozen indexing reports amid ranking volatility

Built by Stephanie Chung·6 min read·8 stories

Daily summary of what matters in SEO, GEO, AEO, and AI search generated with Claude Code (beware of hallucinations)


Generative Engine Optimization (GEO) & Answer Engine Optimization (AEO)

The Risk of Building a Separate "Shadow Web" for AI
Source: Search Engine Journal

  • Creating content that exists only to influence AI answers, detached from the public website, echoes old SEO manipulation tactics like cloaking and doorway pages that search engines eventually penalized.
  • Google's enforcement history maps the pattern:
    • 2011 Panda targeted low-quality content
    • 2012 Penguin tackled link spam
    • 2022 Helpful Content rewarded people-first content
    • 2024 spam policy updates clarified scaled content abuse regardless of AI or human authorship
  • Serving the same information in different formats (llms.txt, Cloudflare's Markdown for Agents, structured data, feeds) is legitimate. Publishing different information to different audiences creates trust and verification problems.
  • AI providers have little incentive to trust sources a brand deliberately keeps separate from what users can independently verify, so maintain one consistent, authoritative source useful to both people and machines.

AI in Search / AI Overviews

Google Says Disabled Discover Source Taps Were a Bug
Source: Search Engine Journal

  • Rajan Patel, VP of Engineering for Search, said on X that disabled source taps on Discover's AI summary cards were "a bug in a Discover experiment" and that Google turned the experiment off.
  • Researcher Damien Andell reported the issue on Oct. 23, tracing it to a flag called "disable_aio_sources_button_tap" that stopped taps on sources from opening the article feed.
  • Andell checked again on Oct. 27 and found the experiment was removed and replaced with another one called "aio tap falls through to hero."
  • The Discover generative AI report in Search Console only measures Impressions, when users see links to your pages via Discover's generative AI, not source taps.

OpenAI Signals Data Shows Information Seeking Rivaling Writing on ChatGPT
Source: Search Engine Journal

  • Information seeking hit 20.7% of consumer ChatGPT messages in August, nearly matching writing at 20.6%, the first month seeking edged ahead in OpenAI's Signals V3.0 series.
  • Writing fell from 34.1% of messages in July 2024 to 21.4% in September, while seeking information rose from 16.7% to 20.2% over the same period.
  • Practical guidance has been the largest topic every month since March 2025, reaching 28.1% in September. Work-related messages made up 32.2% of all messages.
  • Seeking information covers facts, product and service questions, and recipe help bundled into one category, so product queries can't be isolated. Signals does not report how often ChatGPT cites external sites.
  • Signals runs lower than a 2025 OpenAI and Harvard paper that put seeking information at 24% by mid-2025, a near 6 point gap, so always cite the version and time frame.

Technical SEO

Mueller Downplays Frozen Search Console Indexing Report
Sources: Search Engine Journal · Search Engine Roundtable

  • A Redditor reported their Page Indexing report had not updated since October 16 and flatlined, asking whether Google's data pipeline had stalled.
  • John Mueller said the site's indexing numbers were stable and asked what change they expected, noting indexing doesn't change quickly and shows up in Performance reports first.
  • Mueller told SEOs to "save your energy" on report freshness, saying Google puts dates in the reports and prioritizes faster updates where they are actually useful.
  • He pointed to a 16 year old Matt Cutts video arguing Google monitors its own pipelines and catches failures before most users notice, treating crawling/indexing and the reporting of it as 2 separate processes.

Google Updates Favicon Docs to Require Following Homepage Redirects
Source: Search Engine Roundtable

  • Google's favicon documentation now specifies Googlebot must be able to crawl the homepage, follow any redirects to the page where the favicon is specified, and that these URLs must not be blocked by robots.txt or other means.
  • Mueller said Google made the change because of AI, after seeing that AI systems didn't understand how to investigate favicon problems, and that testing the new wording as a prompt worked much better.
  • Double-check that Google can reach the homepage, follow its redirects, and fetch the favicon file via Googlebot-Image.

Organic Search & Algorithm Updates

A Framework for Allocating the Next $100,000 of Marketing Budget
Source: Search Engine Journal

  • Map the full funnel from first visit through month-12 retention using 90 days of data, pulling retention from billing rather than the CRM, because the stage limiting growth often sits after the sale.
  • Price each option against the constraint. In the hypothetical example, CRO offered the best ratio (~$1.1 million in new ARR for ~$15,000) while an extra $60,000 in paid search returned only ~$36,000 in ARR.
  • AI search visibility shows up in win rate, not traffic. When buyers build shortlists in ChatGPT, Gemini, Perplexity, or AI Mode and don't see you, every point of win rate was worth ~$260,000 a year at the example's volume.
  • Spend ~10% on research first, fund the constraint, hold back 15% until day 45, and give each line one kill metric so money follows evidence instead of the prior quarter's plan.

Ziff Davis's Jonathan Berthold on Testing When Traffic Is Flat and Budget Is Tight
Source: Search Engine Journal

  • Start with a full site analysis, then target commercial and transactional intent, since top-of-funnel traffic is being siphoned by LLMs, AI Overviews, and AI Mode and often doesn't convert.
  • Keep experiments small so a bad schema or content test can't tank the whole site, and replicate results before scaling a pattern across the rest of your work.
  • Give an SEO test ~30 days to show movement in Search Console (60 at most), while paid media runs until it spends 2-3 times the target cost per acquisition.
  • Use AI citation data to find link building opportunities by reaching out to publications that cite competitors for bottom-of-funnel keywords, a step most brands skip.
  • Target how your ICP actually searches rather than keyword volume, since a low-volume high-intent query like "I need a plumber to fix my toilet" converts far better than a high-volume informational one.

Another Unconfirmed Google Ranking Update Spikes Volatility
Source: Search Engine Roundtable

  • Barry Schwartz reported a larger volatility spike and increased community chatter around an unconfirmed update over the past 24 hours, following earlier unconfirmed updates on October 24, 20-21, and 16.
  • Tracking tools ran hot, with Semrush reading 9.4, though Schwartz noted the tools are struggling to track accurately.
  • SEO chatter described huge keyword gains without traffic gains, sharp Monday traffic drops with conversions down, and complaints that Google's News and Discover surfacing feels broken.
  • The September 2026 spam update ran September 24-October 8, and Google has not confirmed this latest movement.
Published on