🔥 SEOs Diners Club #210: Why LLM Referrals Are Your Most Valuable Audience?

Hey everyone, and welcome to the new issue of SEOs Diners Club!

In the world of digital marketing, sometimes the biggest opportunities come from the most unexpected places. This week, we're going to look beyond the massive traffic numbers we all chase and focus on a small but incredibly powerful audience: visitors coming from AI engines.

A 13-month analysis published in Search Engine Land proves why this "small but mighty" traffic should be your top priority in 2026. Fasten your seatbelts, because the numbers are truly game-changing. If you're ready, let's dive in!

🎯 LLM Referral Traffic Hits a Staggering 18% Conversion Rate!

The most striking data of the week undoubtedly came from a comprehensive 13-month analysis published in Search Engine Land. The study revealed that referral traffic from Large Language Models (LLMs) like ChatGPT and Gemini has an incredible average conversion rate of 18%. This rate blows all other digital marketing channels—paid search, organic search, social media, you name it—out of the water!

Why So High?

Although the total traffic volume from LLMs is still low for now (under 2% on average, according to the analysis), the quality of this traffic is on another level. The reason is simple: a user who asks a question to an AI, reads the synthesized answer, and clicks on a reference link to your site has completed a much more qualified journey than a classic search user. This user is:

• More Informed: They have already received basic information on the topic from the AI and know what they are looking for.

• More Focused: They come to your site not out of general curiosity, but to find the answer to a specific question or the details of a product.

• Closer to Conversion: They are at the bottom of the sales funnel. We call this the "Educated Click."

Actionable Takeaway: This data proves once again how critical GEO (Generative Engine Optimization) strategies are. It's no longer just about getting a lot of traffic; it's about attracting the right, high-quality traffic. In your Google Analytics or whatever analytics tool you use, immediately define and start tracking referral traffic from AI platforms (e.g., t.co, google.com/search?q=...&source=gemini) as a separate segment. Identify which of your content is being referenced more by AI and invest more in these types of content (especially "How-to," "What is," comparison, and in-depth analysis content). Don't be fooled by the low traffic volume; this audience could be the most valuable on your site. This, combined with the topic we covered last week and published a detailed guide on at Stradiji, the Google AI Overviews Turkey Preparation Guide, shows how holistic an approach GEO requires. It's not just about being a source referenced by AI, but also about being correctly positioned in new-generation search results like AI Overviews, which is an integral part of the strategy.

⚙️ 2. Algorithm & Search Engine Updates

Google Completes Its First-Ever Discover-Only Core Update

Another development as important as our main topic came from Google. Google made history with its February 2026 Discover Core Update, which it started on February 5 and announced as complete on February 27. This was the first core update in history to target only the Discover platform.

We've been viewing this topic as a "black box" for a while. But this week, an incredible study by a Turkish researcher, Metehan Yeşilyurt, who delved into Google's own SDK (Software Development Kit), has cracked open that black box. I want to express how proud I am of him as a fellow Turk.

In his blog post, Metehan reveals that Discover has a 9-stage content pipeline and works in ways beyond our predictions. Here are some of his most striking findings:

• The Publisher Filter Comes Before Everything Else: A publisher can be filtered out right at the beginning, before user interests or click-through rate predictions (pCTR) even come into play. If a domain receives enough negative feedback, even from a single article, the entire domain can be suppressed for that user. This means the penalty is much easier and broader than any reward.

• og:title is Not Just a Label: The og:title tag is more than just a headline; it directly feeds the click-through rate prediction model.

• Freshness and Decay: The first 7 days of content have the highest freshness weight. After 30 days, a "decay" process, tracked hourly, begins.

• Dismissed Content Never Returns: Content that a user dismisses from their feed is permanently recorded on the device and is never shown again.

Metehan's in-depth analysis shows how sensitive and careful we need to be when creating content for Discover. I strongly recommend that you check out his blog, his control panel with the free pCTR tool, and subscribe to his newsletter.

🤖 From the World of AI

Claude Automates Recurring Tasks

This week, Anthropic announced the scheduling of recurring tasks feature in Claude's "Cowork" desktop application. You can now have Claude automatically perform tasks like preparing reports, summarizing emails, and conducting competitive analysis on a daily, weekly, or custom-period basis. This feature holds immense potential for increasing efficiency, especially for agencies and marketing departments. I will be covering this topic and its practical use cases in a detailed video on my YouTube channel soon. If you haven't subscribed yet, do it now so you don't miss out!

ChatGPT's SEO Success and Other Platforms

An analysis published in Search Engine Land reveals how effectively ChatGPT uses SEO for its own growth. According to the analysis, ChatGPT's 76.5 million monthly organic visits, against an SEO investment of $600,000, create an annual revenue potential of about $92 million. That's an incredible ROI of 15,200%! The article also compares the SEO performance of Claude and Perplexity, showing how even AI companies are dependent on SEO for growth.

Claude Rises to #2 in the App Store

According to TechCrunch, Anthropic's chatbot Claude has risen to the #2 spot among free apps in the US App Store. It's noteworthy that this rise comes after the company's dispute with the Pentagon, where it tried to draw ethical lines regarding the military use of AI. It seems that an ethical stance and transparency are resonating with users.

Perplexity's New Move: The Multi-Model Computer

Perplexity has announced a new tool called "Computer" that combines 19 different AI models into a single system. Offered with a $200/month "Max" subscription, this tool is described as a "computer user agent" that can autonomously execute complex workflows. Perplexity's move reinforces the argument that the future will not be a single super model, but an ecosystem where multiple models specialized for different tasks work together.

👋 Until Next Time!

That's all for this week! As you can see, the digital world never stands still. AI is fundamentally changing search engines and marketing strategies. Those who keep up with this change will be the winners of the future.

If you want your brand to be one step ahead in this new order and make the most of the opportunities of the AI era, you can check out Stradiji's SEO and GEO consulting services. Let's shape the future of your brand together.

See you next week with new notes!

Best,

Mert Erkal

About Mert Erkal

Mert Erkal is an expert with 15+ years of experience in digital marketing and SEO. As the founder of Stradiji, he provides consulting services to corporate companies on SEO strategies, conversion rate optimization, and AI integration.

More Information:

• Company: https://www.stradiji.com

• Digital Marketing Notes: https://www.merterkal.com/

• Twitter: https://twitter.com/merterkal

• LinkedIn: https://www.linkedin.com/in/merterkal

• YouTube: https://www.youtube.com/@stradiji

Want to Support My Newsletter?

I love sharing the latest developments and strategies in the SEO world with you. If you find my content helpful, you can support me by buying a coffee. ☕ Click 'Buy Me a Coffee' to contribute to knowledge sharing. Let's achieve more together!