🎧

​An AI-generated audio show based on your search history is rolling out to Discover users in the US. 'Daily Listen' can be enabled in Search Labs and will display at the top of the Google Discover page 24 hours after enabling.

🔍

​Brands ranking on page 1 of Google are more likely to be mentioned ChatGPT Search. The study from Seer Interactive, also found that there was very little impact from backlinks and content variety for ChatGPT mentions.

🥸

​Microsoft is tricking users into thinking they're using Google if they search 'Google' on Bing. The search page looks A LOT like Google and isn't the first time Microsoft has used similar tactics to retain Bing users.

📉

​Stopping or reducing SEO activity can create a yearly 10-20% depreciation in traffic. This may seem obvious to many but some businesses believe that reducing/stopping SEO work would result in maintaining rankings and traffic.

✅

​In some instances, the 'Validate fix' function in Google Search Console can expedite crawling. It does not automatically trigger a recrawl (you can see how that would be abused!) but can be useful for specific issues.

🥙

​Restaurants can now 'add to menu' in Google Business Profile rather than replacing the entire menu each time. This simplifies adding a new item to a menu and can be done by simply uploading additional PDFs.

☁️

​Using AI to create content aimed at getting backlinks is "almost certainly against Google's spam policies". Google regularly advises against creating content solely to acquire backlinks and this is exacerbated with AI.

📊

​The data from the 24-hour view in Google Search Console can now be exported. The export process is exactly the same as other views and allows us to export more granular GSC data than ever before.

Each week in The Digital PR Observer Newsletter you'll get:

• The latest Digital PR news and resources
• 5 quick fire tips to enhance your Digital PR activity
• 5 data sources you can use for Digital PR campaigns
• 5 successful campaigns from the archives
• The latest Digital PR jobs

Are LLMs AI?

🤔 “Is ChatGPT AI?”. The answer is YES! Although I often hear people argue about this, let’s break it down and take some advice from a true expert on the topic, instead of randos on LinkedIn ⤵️
​
⚖️ The main confusion is when people argue, “well, it’s not actually intelligent, so it’s not AI”. Generally, what these people are thinking of is GAI: General Artificial Intelligence, which would be an AI that can understand, learn, and apply intelligence across a broad range of tasks at a human level or beyond - we’re not there yet!
​
🤖 The generally accepted broad definition of AI is generally an automated system that perform tasks commonly associated with the intellectual capability of humans - or there abouts, which means, as we have had for decades, we have lots of “specialised AI”.
​
⭕ What about things like machine learning and deep learning? The Venn diagram on this post is a recreation of the same image I found in an excellent book I am revisiting called “Deep Learning with Python”, which is by François Chollet.
​
🔄 Machine learning is a branch of AI that enables systems to automatically learn and improve from experience without being explicitly programmed, and within this you have Deep Learning, which are multi-layered neural networks to automatically learn representations of data with multiple levels of abstraction.
​
💡 LLMs like ChatGPT are examples of a type of Deep Learning model.
​
Don’t agree? François, apart from being the author of this book, the creator of Keras (high-level neural network library that runs on top of TensorFlow to make all this accessible to laymen like me), working at Google on AI, and writing world-renowned papers on the subject, also won the Global Swiss AI Award for breakthroughs in AI.
​
So… this is who you will have to take it up with. 😅
​
Side note: In 2025 I will be doing some unsolicited SEO tips specifically around the topic of AI, as it is completely inseparable from our industry - and the misinformation and hype can be damaging. This is still 'on mission' to me for my unsolicited SEO tips, which is to provide free advice to help educate businesses around what 'good' SEO looks like and avoid bad actors.

​

How does Google's Source Rank work?

Did you know for news, Google has a patent describing "Source Rank" that aims to asses how trustworthy and important a news source is?
​
📰 It looks as basics such as how many articles a source publishes, how long those articles they are - they even look at writing style, spelling and grammar errors.
​
🎙️ If you're interested in knowing more, this week's The SEO Patent Podcast covers the patent: "Systems and Methods for Improving the Ranking of News Articles".
​
🤖 If you haven't heard the podcast before, it's a podcast I use personally to learn, that generates conversations based on source patent PDFs. It's an easy way to get a basic understanding of patents to decide if they are worth taking a deeper read into.

​

Don't forget to check HSTS

Having an HSTS enabled site means that browsers won't even try and access the non-https version, you'll simply see a "307" redirect (which isn't a true server redirect). This can be tricky if you're trying to diagnose issues.
​
🐸 You can actually use Screaming Frog to test what is happening behind the 307 redirects by disabling "Respect HSTS" policy in Configuration > Spider > Advanced.

​

How long does SEO take?

"How long does SEO take?" - Have you seen all these engagement bait posts trying to trick you? 🎣
​
🦈 I've seen posts recently saying people saying you should see SEO results within weeks, or people don't know what they are doing.
​
📅 I've seen others saying it always takes 6 months plus to get results.
​
What's the truth? ⤵️
​
Likely these people all have different ideas of "results" are, and they probably have limited experience. In the 20+ years I have been doing SEO, I have seen what I define as "results" within a month, and other times I've worked for a a year, tripled revenue, but nothing had happened for the first 7 months.
​
Why? Here's a useful analogy for you to explain this to your stakeholders:
​
"The question 'how long does it take to get fit?' is a very similar question to 'how long does SEO take?', first we need to know:
​
🏋‍♂️ What is the end goal we are measuring to? Both SEO and fitness are never "finished", so you have to pick milestones. Lift a certain weight or get a specific amount of traffic. Beat a friend in a race or outrank a competitor. Or maybe we're just happy to make some progress?
​
👣 What is our starting point? Is this our first time in the gym training, or did we used to play sports? Is this the first time you've done SEO or did you build some links before?
​
🏆 What are our competitors doing? How often are they training? How long have they been training for? Progress measured against competitors is by definition competitive, so you need to know what they are putting in vs you.
​
⌛ What resources do we have? Do we have time to train every day? Do we have money for dieticians, personal trainers, S&C coaches? Do you have in-house skills to help with SEO? How much budget do you have for specialists?
​
All of these are things you need to consider and have answers for before you can honestly answer "how long SEO takes".

Does Google use CTR as a ranking factor?

Google does not use the individual CTR of a result in “ranking” for that URL. However ⤵️

🧪 They do appear to use a predicted CTR, which helps determine if your result is going to be useful to the query (like in Google Ads)

❌ This model is not going to be affected in real-time or long-term in any meaningful way just by clicking on your site result.

🔃 Google do have a re-ranking phase before displaying the SERP, that does take into account SERP interaction behaviour.

👩‍⚖️ This was detailed in the DOJ trial, showing how the SERP for “nice pictures” would change when there was a terrorist attack in Nice, France - the user behaviour on the SERP drastically changed.

🦹‍♂️ The effect is why “CTR manipulation” case studies show results, they are triggering this re-ranking caused by a discrepancy between the predicted and actual CTR.

⏰ However, because this CTR manipulation is not being attributed to the site in any kind of “scoring” way (in ranking), as soon as the manipulation stops, the SERPs go back to “normal” (as shown in such experiments).

📰 This is one of the systems, likely in Google’s “Twiddler framework” that acts on top of the main ranking algorithm to adjust to quick intent changes, such as things coming up in the news.

💡 One of the best ways to track changes in search intent is with People Also Ask data: When something comes up in the news, you’ll see new queries within hours - which is why we built AlsoAsked 😎

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 / X)

🥈 Nikki Pilkington (LinkedIn / X)

🥉 Lidia Infante (LinkedIn / X)

Would love to hear your feedback!

If you have some thoughts, a tip for the next edition, or just want to say hi, you can reply to this email.

I read all the feedback as I geniunely want this to be the most helpful SEO newsletter in the industry.

Have a great week!
Mark & Jack