✅

​Google explains how Search Quality raters define page quality. As well as E-E-A-T, quality is determined by how effectively the main content helps the page to achieve its purpose and provide a satisfying user experience.

🧑‍⚖️

​Judge dismisses publisher lawsuit against Google's AI Overviews. The plaintiffs argued that there is an unwritten agreement for Google to send traffic to publisher websites. The judge dismissed these as 'expectations not agreements'.

🔍

​Google officially recommends fact-checking AI-generated content. The AI generative content documentation now includes "It is critical to manually fact-check and review all AI-generated content".

🏷️

​Gemini adds UTM parameters to referral links. These UTMs have been used on other AI search engines, such as ChatGPT, and should make attributing traffic from Gemini in GA4 easier than before.

📊

​Microsoft Clarity adds customisable brand terms to its reporting. Users can now edit and add to the pre-populated list as well as adding localised terms and common typos.

📈

​AI Overviews appear for over 90% of branded terms in September 2026. This has tripled in the last few weeks from around 30% to over 90%. AIOs have been confirmed for both small and large brands.

🛍️

​Google adds merchants' loyalty features to AI Mode and Gemini. This feature allows merchants to show their loyalty programs and rewards in search results. These can include discounts, free shipping and reward cards.

When the stakes are high, AI citation patterns change

In everyday or commercial queries, brand-owned content earns citations. But when users ask AI about their health or finances, the rules change. Government and official sources dominate, Citation Cores shrink, and brand content struggles to cut through.

Semrush's latest YMYL Mini Study maps the trends in AI sources when the stakes are highest, across 22 topics, four platforms, and millions of prompts.

If your brand operates in finance, health, or any high-trust category, this changes how you need to approach your AI visibility.

You make searching fun

I have become a better SEO by looking at Chrome network data to get hints on how systems from Google and OpenAI are working in the background. I have been doing this for years, and it genuinely gives you an edge over competitors.

If you want to know more about this (you can do it too), along with real-world examples, then drag yourself out of bed for the first brightonSEO sessions on Friday; I have 20 minutes of nerd magic for you. ✨

Never going RAG again

Is RAG happening? is an important question for search in 2026. RAG stands for Retrieval Augmented Generation, and it means that answers being generated by AI platforms are being checked against, and improved by, data that is fetched from outside the model.

🤔 Why is this important? If you determine RAG is not happening, then the entire answer comes from parametric memory; that is, weightings made from the training data. The only way you can impact this answer significantly is by being so heavily present in the next set of training data that you shift those weights; which could take months or years. It is a great long-term goal, but it isn't something you can impact and measure quickly.

🔃 One of the most common forms of RAG uses traditional search engines like Google or Bing to conduct multiple searches in the background, where sites are selected and parts of page content are retrieved. In some cases, LLMs use other specific data sources, both internal and external.

🛠️ You can use tools like QueryFan.com to see if RAG is happening within platforms such as Gemini and ChatGPT, as well as the associated fan-outs that happen based on personalised prompts.

I've got some more detail about that in my talk from Search 'n Stuff Antalya 2025, but as a very first step, it is important to know if grounding is happening at all!

Link your own way

Add more internal links is NOT an optimisation by default (despite what your AI tool is telling you). 🤡

I ran the same page through an AI SEO tool 5 times, and every time it told me to add more internal links from pages it was gleefully fetching from Search Console. WHEN DOES IT END? 🫠

You can also throw use vectors to determine internal linking on that pile. Just because a page looks similar to another one does not automatically mean it is a good idea to link to it.

Links are for navigation. Links are for users. Link them to where they might want to go next based on your understanding of the journey. That can come through intentional design, through qualitative user testing and feedback, through quantitative analytics data and outcomes, and yes, you can even get suggestions from vector similarity.

Yes, search engines will use logic that is roughly if lots of pages, or important pages link to a page, it is probably relevant or important, but for the love of cats.txt, don't overcook your whole site chasing optimisation ghosts.

Second hand news

The alignment between “brand values” vs actually doing those things and making them known with digital PR has becoming more relevant to SEO than ever.

I saw this screenshot from Ann Smarty showing that Google AIOs are appearing for straight up brand searches, with a particularly damning report of GoDaddy. Nuturing and controlling brand narrative means the difference between getting the click or not.

Gold dust position

Even if you're doing 'traditional' SEO, Tom Capper at Search 'n Stuff is showing that "position 1" can mean very different things, with <1% of SERPs having organic position 1 and 2 visible on a standard device.

Tom makes a good argument around visibility being more about visual prominence.

Top Core Updates referrer leaderboard

A big thank you to our top referrers, who have signed up over 10 people to the Core Updates newsletter, go follow them!

🥇 MJ Cachón Yáñez (LinkedIn / Bluesky)

🏅 Nikki Pilkington (LinkedIn / Bluesky)

🏅 Lidia Infante (LinkedIn / Bluesky)

🏅 Saijo George (LinkedIn / Bluesky)

🏅 Alice Rowan (LinkedIn)

🏅 Martijn Jeurissen (LinkedIn)

🏅 Chester Beard (LinkedIn)