SEO Recap covering July 2, 2026
8 min read·11 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)
Microsoft's Web IQ and Bing Webmaster Tools Split Ranking From Citation
Source: Search Engine Journal
- Bing Webmaster Tools now holds two separate reports: Search Performance (clicks, impressions, CTR, position) and AI Performance, which in June added a Citation Share metric showing your share of citations for a grounding query.
- Web IQ is a suite of grounding APIs built on the Bing index that returns passage-level evidence objects instead of ranked documents, and Microsoft says it powers grounding in both Copilot and ChatGPT.
- Web IQ scores retrieved passages on a metric Microsoft calls GDSAT across completeness, freshness, and authority, meaning a page can rank well yet have sections passed over as evidence.
- The author recommends making each section self-contained, front-loading the claim, naming entities instead of pronouns, and reading pages the way a passage selector would.
- Web IQ inherits Bing's existing robots.txt and BingBot controls with no new user-agent, so old crawl directives may now decide whether content can be used as AI evidence.
Bing Backfilled AI Performance Report Data on June 1
Source: Search Engine Roundtable
- Glenn Gabe noticed large surges in impressions and citations across nearly all his Bing Webmaster Tools profiles starting June 1.
- Microsoft's Krishna Madhavan said the jumps are "just regular data backfilling" with "no anomalies," describing it as a routine data pipeline artifact.
- The implication is that Bing's AI Performance reports did not include all data sources at launch and added more retroactively.
- Madhavan advised focusing on trends rather than absolute citation counts.
AI Visibility Depends on Who Writes About Your Brand
Source: MarTech
- BrightEdge found only about 16.5% of sources cited in AI Overviews also rank in Google's organic top 10, and Moz's analysis of 40,000 AI Mode queries found 88% of citations came from pages outside the organic top 10.
- Muck Rack's Generative Pulse study of more than 25 million links across ChatGPT, Claude, and Gemini found earned media accounts for 84% of AI citations and paid content just 0.3%, with journalism at 27%.
- Ahrefs data shows sites with author schema are nearly three times as likely to appear in AI answers, underscoring the weight of named, credentialed authors.
- A Semrush experiment found Google AI Mode cited 36% of new pages within 24 hours (dropping to 26% by day 30), while ChatGPT started at 8% and reached 42% by day 30.
- The author recommends leading with the claim, putting a named author on everything, prioritizing steady distributed placement, refreshing content quarterly, and tracking citation share monthly across engines.
AI in Search / AI Overviews
Google Data Shows AI Mode Users Have Moved Past Keywords
Source: Search Engine Journal
- Google's report "How People Are Using AI Mode in the U.S." says the average AI Mode query is now triple the length of a traditional search query.
- Follow-up queries have grown more than 40% per month on average, and multimodal interactions account for more than one in six AI Mode searches, with image-input searches up more than 40% month over month.
- Planning queries have grown 80% faster than overall AI Mode volume, brainstorming queries 30% faster, and "which" queries 40% faster over six months.
- Google reports more than 1 billion monthly active AI Mode users globally, with query volume doubling every quarter since launch.
- The author recommends auditing top pages against natural-language prompts, building a follow-up question inventory, and preparing visual assets for multimodal indexing.
Technical SEO
John Mueller Flags Case Study on Why LCP Fixes Miss the Real Element
Source: Search Engine Journal
- Nuvemshop's year of Core Web Vitals work, published on web.dev June 24, found the browser was often selecting the wrong LCP element because CSS transitions on carousels and banners delayed visibility detection.
- The team removed CSS transitions from top sections, pulled loading="lazy" off the first above-the-fold image, and added fetchpriority="high" to the likely LCP image with validation logic.
- The share of stores with good LCP scores rose from 57% to 96%, and overall Core Web Vitals pass rate improved from 48% to 72%.
- Nuvemshop reports an 8.9% increase in mobile organic conversion rate and 8.4% rise in cart engagement year over year, though these are self-reported and not a controlled test.
- Mueller's takeaway: verify which element the browser actually treats as LCP before optimizing, especially on template-driven or carousel-heavy layouts.
Google Ends Cache-Served AMP Pages in Search
Source: Search Engine Journal
- As of July 1, clicking an AMP result sends users to the domain's own AMP host page instead of a version served from Google's AMP cache.
- Google removed mentions of the AMP viewer, AMP Cache, and signed exchanges from its AMP documentation.
- Publishers no longer need to maintain the AMP cache or configure signed exchanges to display their own URL.
- Google confirmed the change affects only delivery, not ranking: AMP content will continue to rank like any other web page.
Cloudflare's New AI Crawler Rules Can Also Block Googlebot
Source: Search Engine Journal
- Cloudflare now sorts crawlers by behavior into Search, Agent, and Training rather than a single "block AI bots" switch, live now for all tiers.
- Starting September 15, Training and Agent crawlers will be blocked by default on ad-supported pages for new customers and unchanged free accounts, while Search stays allowed.
- Cloudflare will apply the strictest matching rule to multi-purpose crawlers, so blocking Training could also block Googlebot, Applebot, and Bingbot, which crawl for both search and AI training.
- Cloudflare's report says AI training now makes up the majority of crawler requests on its network, up from roughly 20% in spring 2025, with daily AI agent requests up more than 1,700% over the year.
- Sites on Cloudflare should review AI-blocking settings before September 15 to avoid cutting off search crawlers at the network level.
Ask an SEO: Should You Block AI Crawlers or Measure Their Value First?
Source: Search Engine Journal
- The piece distinguishes AI training bots (GPTBot), search indexing bots (OAI-SearchBot), and user-triggered fetchers (ChatGPT-User) by their differing value to a site.
- OpenAI's ChatGPT-User and Perplexity-User no longer commit to honoring robots.txt, so blocking them requires server- or WAF-level rules via tools like Cloudflare or AWS.
- Log file analysis over a 30-day sample is the most complete way to identify which AI bots crawl a site and which sections they target.
- Google Analytics added an "AI Assistant" channel that recognizes ChatGPT, Gemini, and Claude by referrer but does not capture Perplexity.
- The author warns that blocking all AI bots removes the ability to test which platforms cite you accurately and may make a site invisible if LLMs become a primary discovery method.
Google Routes Searchers to Publisher-Hosted AMP Pages, Dropping the Cache
Source: Search Engine Roundtable
- Google now sends Search users directly to publisher-hosted AMP pages instead of cached pages in an AMP viewer, effective July 1.
- Google said the change reduces maintenance for publishers, who no longer need to update the AMP cache or configure signed exchanges.
- Google updated its AMP documentation with new sections on how AMP pages appear in Search and revised three related help docs.
- There is no impact on how AMP pages rank or are served in Search and Discover; Glenn Gabe noted one AMP URL failed to load in the Google app during early testing.
Organic Search & Algorithm Updates
Fraudulent DMCA Takedowns Are Removing Legitimate Content From Google
Source: Search Engine Journal
- Malicious actors are filing fake DMCA notices with Google to get competitors' or unflattering content removed from search.
- The DMCA never requires Google to judge whether a notice is valid; its obligation under 17 U.S.C. 512(c) is to remove claimed material and later restore it after a counter notice.
- Google must wait 10 to 14 business days before reinstating contested content to give the claimant time to file a federal lawsuit, which explains slow restorations.
- Falsely accused parties can sue filers for knowing misrepresentation, but fraudsters use fake email and physical addresses, leaving victims little recourse.
- Ex-Googler Pedro Dias has raised the issue for months, arguing Google has a serious problem no one is fixing.
July 2026 Google Webmaster Report Recaps Spam Update and AI Reporting
Source: Search Engine Roundtable
- The June 2026 spam update rolled out in just two days and felt substantial, alongside several unconfirmed ranking updates through the month.
- Google said LLMS.txt files won't help or hurt rankings and reaffirmed HTML, not markdown, as the standard for SEO.
- Search Console expanded access to AI performance reports, published a deep dive on generative AI controls, and fixed the link report bugs.
- Google launched Search Profiles publisher pages with analytics and began rolling the Top Stories carousel into AI Overviews.
- The CMA ordered Google to disclose how its rankings work and enable data portability, while an updated report put zero-click searches to the open web at 27.6%.
Published on