🏷️

​AI-generated content must be labelled clearly according to new EU guidelines. The EU is pushing digital transparency that expands outside of just images and includes published text that doesn't receive human review.

📊

​Social media properties rollout globally for Google Search Console. These new platform properties were announced a few weeks ago, and users can now connect YouTube, Instagram, TikTok and X to GSC.

🔍

​52.3% of Gen Z searches end with zero clicks, higher than any other age group. Searchers aged 55-64 are less likely to search without clicking, with 42.7% zero-click searches.

📰

​Opting out of Google's AI features may also opt out of Top Stories. Top Stories carousels now appear in AI Overviews, so publishers who wish to opt out of AIOs will need to weigh the risk of losing coverage in Top Stories.

🤝

​Google's Open Knowledge Format adds 5 new trust signals. These new trust signals for agents are categorised using the following types: provenance, trust, freshness, lifecycle and attestation.

⏱️

​Google's 5-second rendering limit is a myth. This oft-cited myth was tested, and found that some rendering calls can run as long as 54 seconds. Of course, no sane user would wait that long for a page to render.

📅

​Google may not notice updates to the unavailable_after attribute. This attribute is used to tell Google not to show a page after a specific date, but Google may not pick up if that date is changed, as it may not try to crawl that page again.

🚀 Help SEOFOMO reach 50K

The SEOFOMO newsletter is on the road to reach 50,000 subscribers, and to celebrate, they're giving back to the search community helping them get there.

Subscribe to the free SEOFOMO newsletter, share your unique referral link, and each verified new subscriber you refer gives you one additional eligible entry to win:

🎟️ Grand Prize:

A ticket to one eligible SEO conference worldwide, plus SEOFOMO swag and a private SEO / AI Search Q&A session with Aleyda Solis. Conferences to choose from include MozCon, brightonSEO, Search n' Stuff, Women in Tech SEO Fest, SERP Conf, SearchNorwichXL, and more!

🎁 Plus:

5 Community Experience winners will get access to an exclusive virtual SEO & AI Search group Q&A with Aleyda, plus a WTSFest London 2027 Recording Ticket.

The giveaway runs until SEOFOMO reaches 50,000 verified subscribers, or December 31, 2026, whichever comes first.

Apply for a FREE SearchNorwichXL ticket

We're giving away some free tickets for our upcoming conference, SearchNorwichXL.

So, if you've seen the amazing line up of speakers and want to visit Norwich in September, now is your chance to apply!

Fill out the form below with your details and we'll be announcing the winners in August.

Search is like a box of semantics

More free local model SEO fun: You can analyse the semantic similarity of pages to detect duplicate, similar and potentially off-topic, less relevant content on a site with Screaming Frog & Ollama.
​
Using their existing AI provider integrations via ‘Config > API Access > AI’ (you can use cloud models like OpenAI, Gemini) and you can capture vector embeddings of pages.

You can't keep running SEO with LLMs

Reason number 742 you can't get your LLMs to run your SEO for you: even when there is clear documentation, they still can give the wrong information.
​
This was a simple query about whether you can use the Google Search Console 'Change of Address' tool, when you are migrating part of your website from a subdomain to a subfolder.
​
My colleague Seb Atkinson (very good freelance SEO) was discussing this, and I checked the docs to make sure I was correct (things can change).
​
However, I thought it would be a nice question to ask an LLM, and disappointingly, *despite itself referencing the official docs* it was generating answers from a 13-year-old forum conversation 🤡
​
It's wild to me that people are putting Claude et al into the driving seat for their SEO (or anything for that matter).
​
I can't stress enough, I want my competitors to keep doing this, as all of those dents, scrapes, inefficiencies, little bad decisions are going to add up over time.
​
It may seem like a small thing, but this is the 'O' in SEO. It is the culmination of dozens and dozens of small wins over a long time period. Using the CoA tool sped up my migration instead of yours, so I got discovered and won that big link, which got us that ranking quicker, which meant we won that new bit of business, who then referred us.... etc...

I'm here to try out my search legs

🚨 Big update for SEOs and Content Creators! Google Search Console just rolled out "Platform properties" globally. 🌍
​
This means you can now link your social media profiles and video channels (like YouTube, TikTok, Instagram, etc.) directly to GSC as their own distinct properties.
​
For the first time, you can get a clear, holistic view of exactly how your social and video content is performing in Google Search results, completely separate from your main website's metrics. 📈
​
If you check your search console, you should get a pop up, offering you to link up these channels 👇

AI is the fruit of SEO

Google does not care if you are using AI to write your content, but this confuses most people ⤵️
​
Google specifically states they don't care if you use AI to produce, or help produce your content.
​
⭐ While I've called out a lot of bad uses for AI, I've been writing since 2023 that AI-produced content is not the problem.
​
⭐ Ahrefs Ryan Law just published a study that seems to confirm this from a data lens as well; AI content is ranking. I can only think this was needed as the 99.9% of AI content that is bad has got people into a mindset that all AI content = bad.
​
☝ However: This does not mean you can just tell AI to write you an article on a subject. You won't be producing anything new (amusing hallucinations notwithstanding), and you will make what Google calls 'commodity content', which will likely end you up with some kind of site suppression and loss of traffic.
​
It is absolutely possible to get AI to produce content that meets Google's guidelines. It's just incredibly rare, because >99.9% of it is people reading "Google says it's okay to make content with AI" and that's all they want to hear before making tokens go brrrrr.
​
A great example is Jono Alderson's blog. He is one of my favourite writers, and there is no way an AI is making the points Jono makes on autopilot, but a lot of the text will come back as '100% AI'.
​
The AI detection is not the point, the content is.
​
"Google doesn't mind AI content" does not equal "all AI content is fine", and it needs to be more than "human in the loop" with some vague checking after the fact.

Mama said LLMs don't know how they work.

Please stop asking LLMs how they work, they DON'T KNOW, and please for the love of all things non-slop, please stop writing articles based on the LLM output 😭
​
This reinforcing loop is so dangerous:

1. User A asks LLM how it works
2. Predictive text backed with some spicy forum speculation
3. User A writes long-form article on how LLM works
4. User B asks LLM how it works
5. LLM reinforces answer based on article it finds from User A
6. User B also writes a similar article
7. Misinformation spreads, grows, solidifies as consensus
8. Everyone dumberer

Happy Friday, have a think about the societal impact of the poisoning of generations of knowledge from unilaterally-controlled tools primarily used to enrich billionaires.

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)