Spam Update Done, GEO vs SEO, CTR Split & ChatGPT Sources | HighDegree* Marketing – Issue 26

Subscribe to HighDegree* Marketing Newsletter!


Spam update done. GEO is good SEO. Desktop CTRs rising, mobile falling. ChatGPT picks sources by scraper tier. Here’s what this week’s five stories mean for you.


Hi Marketers,

Welcome to the new issue of HighDegree*: Cutting Through the Noise in SEO & Digital Marketing.

Google’s June spam update wrapped up in two days flat and it hit harder than most. Meanwhile, Google’s own VP of Search is telling CMOs to stop chasing new acronyms and get back to basics. CTR data from Q1 2026 shows desktop and mobile are now moving in opposite directions. And if you’ve ever wondered how ChatGPT actually decides which sources to cite, this week someone finally read the network traffic to find out.

Let’s jump in.


➞ Digital Marketing This Week

  • Google’s June 2026 spam update finished rolling out in under 48 hours and based on the forum chatter, it felt bigger than a typical spam run
  • Google’s VP of Search says good SEO is good GEO, stop optimizing for bots and start creating content only your brand can produce
  • Google’s head of Search, Liz Reid, says the path to AI visibility is simple: make content people actually want to read
  • Advanced Web Ranking’s Q1 2026 CTR study found desktop click rates climbing across 22 industries while mobile rates dropped, the device split is now a real strategy concern
  • An SEO researcher spent two days reading ChatGPT’s raw network traffic to map exactly how it picks sources, and what he found changes how you should think about AI SEO

➞ Subscribe

If you are not subscribed to HighDegree* yet, please subscribe to us by clicking the “Subscribe” button below. I will post every issue to the HighDegree* Substack version, so you can get it instantly in your inbox every time I hit publish!

Thanks for reading HighDegree* Marketing! Subscribe for free to receive new posts and support my work.


➞ Google’s June 2026 Spam Update Is Done, But It Felt Bigger Than Usual

Google launched its June 2026 spam update on June 24 at around noon ET and wrapped it up two days later on June 26. That’s fast, even by Google’s standards for spam rollouts. What made it stand out wasn’t the timeline; it was the reaction. Site owners across niches reported drops ranging from 10% to 80%, and WebmasterWorld threads filled up with accounts of traffic swings every 20 minutes, spam sites surfacing in results, and losses on sites with no spam at all.

Google called it a “normal spam update” targeting sites that violate its spam policies, global across all regions and languages. It doesn’t target link spam, site reputation abuse, or several other specific policy areas. If your site took a hit, Google’s guidance is to review its spam policies; recovery can take months and depends on periodic refreshes, not a quick fix.

The key takeaway: volatility trackers showed significant movement, and sites that hadn’t knowingly spammed still saw losses. Audit your site against Google’s current spam policies if you’re seeing unexplained traffic drops from late June.

Read the full report from seroundtable.com ➞

The spam update timing connects directly to what Google’s own leadership has been saying about what actually survives these sweeps.


➞ Google to CMOs: Stop Chasing GEO, AEO, and LLM SEO; Just Do Good SEO

Brendon Kraham, Google’s VP of Search and Commerce, published a piece on Think with Google this month with a clear message: good GEO is good SEO. AI Overviews and AI Mode run on the same core ranking systems as traditional search; the AI features retrieve content from the existing index, so if your SEO is solid, you’re already positioned for AI search.

His practical points are worth keeping. Don’t optimize for bots, Google’s systems understand language the way a human does, so keyword-stuffed copy and artificial snippets don’t help. Stop worrying about LLMs.txt, Google doesn’t use it for ranking. Your biggest edge is content only your brand can create: first-hand experience, internal expert voices, specific use cases over generic topics.

Kraham also pushed back on third-party AI measurement tools, Google doesn’t give them internal metrics access. Your own business outcomes (leads, sales, sign-ups) are the most reliable signal. Search Console’s new AI impression reports give you a useful baseline if you want to track AI visibility directly.

Read the full report from business.google.com ➞

Google’s head of Search backed this up in a separate interview, but put the message in even simpler terms.


➞ Google’s Liz Reid: The Path to AI Visibility Is Content People Actually Want to Read

When asked what publishers can do to get their content seen in AI search, Google’s VP of Search Liz Reid gave a two-part answer. First: technical accessibility, Google needs to crawl and index your pages without friction. Second: quality. Make content people actually want to read.

The second point is more specific than it sounds. AI features on Google surface content based on the same signals that drive organic rankings: helpfulness, authority, and relevance. There’s no separate optimization layer for AI visibility. If your pages rank well on informational queries and people find them useful, they’re more likely to show up in AI Overviews and AI Mode.

The interview came as publishers grow frustrated about declining referral traffic, and Reid’s answer didn’t offer a workaround. It reinforced what data keeps confirming: the strategy for AI search and for organic search is the same. Build content that’s genuinely useful to a real reader, make it technically clean, and give your site time to build authority in a specific topic area. Writing specifically for AI snippet formats doesn’t move the needle independently.

Read the full report from searchenginejournal.com ➞

The “content quality” argument only goes so far if your clicks are dropping even when you rank, which is where Q1 2026 CTR data gets uncomfortable.


➞ Desktop CTRs Are Rising. Mobile CTRs Are Falling. Q1 2026 Broke the Two Apart.

Advanced Web Ranking’s Q1 2026 CTR report, pulled from 22 industries, landed one hard conclusion: desktop and mobile click-through rates are now moving in opposite directions, and the gap is wide enough to change how you plan.

Across industries where search demand grew, Careers, Health & Fitness, Personal Finance, Technology, desktop CTRs climbed sharply while mobile either held flat or dropped. In the Careers industry, the top four desktop positions gained a combined 10.96 percentage points in CTR, while mobile first-position CTR fell 3.23 pp. In Technology & Computing, desktop top-five positions gained 12.51 pp combined; mobile first position lost 2.30 pp.

The pattern held even where demand fell. In Law, Government & Politics, the first mobile position shed 9.03 pp, the steepest single-position loss in the entire dataset. Desktop top-six positions in the same category gained 21.00 pp combined.

Style & Fashion and Real Estate saw desktop top-six and top-ten positions gain over 20 pp combined. The report concludes: click behavior has “fundamentally decoupled by device type.” If your traffic is mobile-heavy and it’s been dropping, this is the structural reason, not just algorithm noise.

Read the full report from advancedwebranking.com ➞

Knowing how clicks behave on Google is one thing. Understanding how ChatGPT decides what to cite is another and this week, someone actually dug into the mechanics.


➞ Someone Read ChatGPT’s Network Traffic for Two Days. Here’s What They Found.

Suganthan Mohanadasan spent two days reading raw JSON from ChatGPT’s browser network panel, not the outputs, the actual machinery underneath. What he documented changes how you should think about AI source optimization.

Every web result ChatGPT pulls carries a result_source field with one of four values. labrador is a licensed tier covering Reuters, the WSJ, Wikipedia, essentially closed to independent sites. serp handles news. bright (Bright Data) and oxylabs (Oxylabs) are commercial scrapers doing the heavy lifting for commercial, shopping, and regional queries and that’s the tier most content creators are competing in.

A few other findings matter. Queries tagged text skip the web entirely and answer from training, no page gets into that response. The thinking model fires dozens of sub-queries per question, including direct site: probes against pricing pages. If your pricing is behind JavaScript, the model can’t read it and falls back to G2. YouTube gets fetched heavily but almost never cited, because citations bind to text and YouTube pages don’t expose transcripts in HTML. Reddit does, which is why it leads citation counts.

The practical upshot: put your facts in plain HTML. One strong page per topic beats ten thin ones.

Read the full report from suganthan.com ➞


➞ From Google

Everything from Google Search this week —

Google Begins Rollout Of Top Stories Carousel In AI Overviews (seroundtable.com)

Google Moves Recipe Links To Top Of AI Mode Results (searchenginejournal.com)

Google Makes Recipes in AI Mode More Publisher Friendly (searchengineland.com)

Google: Meta Descriptions Not Required For SEO. But They’re Worthwhile. (searchenginejournal.com)

Google Answers Question About SEO For AI Agents (searchenginejournal.com)

Google Search Console Has First Generative AI Reporting Bug On June 24 (seroundtable.com)

Google Ads Tests Open In New Window Icon (seroundtable.com)

Google Removes Shocking Content From Ads Not Allowed For Young Users (seroundtable.com)

Google Ads / Merchant Center Customer Match Uploaded Via Merchant API (seroundtable.com)

Google Breaks Out Where Local Inventory Ads & Free Local Listings Are Available (seroundtable.com)

Google Ads PMax Product Reporting By Asset Group & View Audience Segments (seroundtable.com)

Google AI Overviews Tests Button To Go To Only Web Results (seroundtable.com)

Google Search Console Page Indexing Report Delayed By Two Weeks Again (seroundtable.com)

Google AI Mode Ask Anything Box With Autocomplete (seroundtable.com)

Search Generative AI Control (support.google.com)

Google’s Mueller Explains How AI Search Impressions Get Counted (searchenginejournal.com)

Google’s LLM Patent Suggests a New Goal for SEO: Teaching AI Who You Are (searchengineland.com)

Google Introduces New Merchant Center Agency Roles (searchengineland.com)

Google Merchant Center: How To Remove Found By Google Products (seroundtable.com)

Google Ads Will Soon Allow Some Final URLs To Redirect To A Different Domain (seroundtable.com)

Google Ads: Unique Search Categories With Clicks, Conversions & Impressions (seroundtable.com)


➞ AI + Social

Find out what’s happening in the social media and artificial intelligence world —

OpenAI Is Limiting Its New AI Models to ‘Trusted Partners’ at the Government’s Request (qz.com) – OpenAI will delay GPT-5.6’s release following a Trump administration request, limiting early access to approved government partners.

ChatGPT Is the Most Popular AI Tool for People in the US (socialmediatoday.com) – New survey data confirms ChatGPT leads all AI tools in the US by a wide margin.

ChatGPT Users Are Now Mostly Non-English (searchenginejournal.com) – A significant shift: the majority of ChatGPT’s active user base now communicates in languages other than English.

Anthropic Launches Cheaper Claude Sonnet 5 Model (finance.yahoo.com) – Anthropic is cutting costs with a more affordable Sonnet 5 as enterprises push for better AI value.

Anthropic Unveils ‘Claude Science’ for Scientific Research (reuters.com) – A dedicated AI platform for scientific use cases, signaling Anthropic’s push into specialized vertical markets.

AI Agents Are Failing To Read B2B Pricing, Study Warns (searchenginejournal.com) – A new report finds AI agents routinely fail to extract accurate B2B pricing — mostly because the data is behind JavaScript or buried in PDFs.

Meta Plans to Replace 90% of Content Review Staff With AI (socialmediatoday.com) – Meta is moving toward AI-driven content moderation at scale — a major operational and trust signal for the platform.

TikTok Launches Agentic Hub (socialmediatoday.com) – TikTok’s new Agentic Hub lets brands build and deploy AI agents for automated marketing workflows on the platform.

TikTok Launches Branded Microdramas (socialmediatoday.com) – Short serialized branded content is now a formal TikTok ad product — worth watching if short-form video is part of your strategy.

Instagram Tests More Algorithm Controls (socialmediatoday.com) – Instagram is experimenting with giving users more control over feed content, which could affect creator reach and distribution patterns.

Threads Adds Live Chat Co-Hosts (socialmediatoday.com) – Threads is expanding live features with co-host functionality, pushing further into creator engagement territory.

WhatsApp Now Lets You Reserve Usernames (techcrunch.com) – A significant platform move: WhatsApp is finally letting users claim unique handles.

LinkedIn Tests Suggested Feeds (socialmediatoday.com) – LinkedIn is experimenting with algorithmically surfaced content outside your network, similar to what Instagram and TikTok already do.

LinkedIn Automates Job Application Process for Premium Users (socialmediatoday.com) – LinkedIn Premium users can now have AI handle parts of the job application process automatically.

OpenAI, Microsoft Sued by Publishers for Scraping Articles (bloomberglaw.com) – A growing coalition of publishers is taking legal action over unauthorized use of content for AI training.

Microsoft Clarity Now Flags Bots That Ignore Robots.txt (searchenginejournal.com) – Microsoft’s analytics tool now surfaces when non-compliant bots are crawling your site in violation of robots.txt rules.

Reddit Launches ‘People are the Best’ Campaign (socialmediatoday.com) – Reddit’s latest brand campaign leans into user-generated authenticity as a differentiator from AI-generated content floods.

DuckDuckGo’s AI Feature Is Telling Users That Trump Died of Rabies (futurism.com) – A stark reminder that AI hallucinations in search products are still very much a live problem, even from established players.


➞ Worth Reading

These are the articles that will help you refine your marketing knowledge —

How To Define & Report SEO KPIs That Actually Move The C-Suite (searchenginejournal.com)

The Shift to Agentic AI: Evidence From Codex (openai.com)

The Downstream Impact of AI Visibility (similarweb.com)

6 Content Audit Workflows to Build in Claude (searchengineland.com)

Clarity Now Surfaces Robots.txt Violations in Bot Analytics (clarity.microsoft.com)

Translated Reddit Is Winning AI Citations (peec.ai)


➞ What This All Means for Your Strategy

This week’s five stories share a common thread: the rules haven’t changed, but the stakes for ignoring them have.

Google’s spam update and Liz Reid’s message say the same thing from different angles; Google clears out what isn’t useful, and the path back is the same as the path forward. Kraham’s GEO piece adds the strategic frame: stop building for AI acronyms and start building content no one else can create.

The CTR data makes it practical. If you’ve treated mobile and desktop as one audience, Q1 2026 says they’re not. Check your device-level Search Console data, you may find two different traffic stories.

And Suganthan’s network traffic analysis hands you something most AI SEO guides can’t: actual mechanics. Your facts belong in plain HTML. Reddit earns citations because its content is readable text. One strong page per topic beats ten thin ones.

Put it together and the playbook is the same across all five stories: useful, specific, well-structured content from a credible source, on a site that loads cleanly and doesn’t hide its data from bots.

Until next week,

Nishat from HighDegree* Marketing


P.S. Have a question about implementing these strategies? Hit reply – we read every email and often feature reader questions in future issues.


➞ Who is Nishat Shahriyar?

I’ve been a Digital Marketing Strategist since 2007. Today I’m the Product Marketing Lead at Fluent Forms (a WordPress form and lead-generation plugin used by 700K people), FluentPlayer (a WordPress Video Player Plugin) and Ninja Tables (a WordPress table builder plugin used by 80K people). Before that, I led marketing at Fluent Support (a WordPress helpdesk plugin used by 10K people).

Connect with me on LinkedIn or follow me on X/Twitter at @rednishat for SEO & Marketing updates.

Subscribe to HighDegree* Marketing Newsletter!


Posted

in

by

Tags: