Mitasys Technologies @mitasys

This is a cool one to watch: this site had a ton of low-quality programmatic content that it cleaned up late last year, and it's seeing the benefits now, ~8 months later

I wrote a thing: iloveseo.net/the-arrival-ef…

SEO Tip! Use the following in Google Search Console to apply regex for filters: ➡️ QUERY LENGTHS: Single-word queries only ^\w+$ Exactly 2-word queries ^\w+\s\w+$ Exactly 3-word queries ^\w+\s\w+\s\w+$ 5+ word queries ^\w+(\s\w+){4,}$ 7+ word queries ^\w+(\s\w+){6,}$ Queries over 60 characters ^.{60,}$ Short queries under 20 characters ^.{1,20}$ ➡️ QUESTION/INTENT All question-style queries ^(how|what|why|when|where|who|which|can|does|is|are|will|should|do)\s How to / how do / how much" etc ^how\s(to|do|does|can|much|many|long)\s Definitional "what" queries ^what\s(is|are|does|do|was|were)\s Yes/no style questions ^(is|are|can|does|do|will|should|has|have)\s Why is / why does / why won't ^why\s(is|are|does|do|won.t|can.t)\s Educational/informational intent .*(guide|tutorial|tips|learn|explained|examples|ideas).* Definitional queries .*(what is|what are|definition of|meaning of|overview of).* "How to" specifically (anchored) ^how\sto\s Beginner/intro intent .*(for beginners|step by step|getting started).* Commercial investigation .*(best|top|vs|versus|compare|comparison|alternative|alternatives|review|reviews).* Direct comparison queries .*\svs\s.* "Best X for Y" pattern .*(best|top)\s.*\s(for|to)\s Evaluation intent .*(worth it|should i|is it good|recommended).* ➡️ TRANSACTIONAL Transactional signals .*(buy|purchase|order|shop|price|pricing|cost|costs|hire|get).* Price-sensitive queries .*(cheap|affordable|budget|discount|deal|coupon|offer).* Acquisition intent .*(free|download|trial|sign up|register|get started).* Local transactional intent .*(near me|nearby|close to).* High-intent service queries .*(quote|demo|consultation|book|booking).* ➡️ NAVIGATIONAL Navigational/login queries .*(login|log in|sign in|account|dashboard|portal).* Brand + product-type queries .*(website|site|app|software|tool|platform)$ ➡️ SINGULAR/PLURAL Plural Forms .*(tip|tool|idea|strategy|example|way|option|method|step)(\s|$) Singular forms .*(tips|tools|ideas|strategies|examples|ways|options|methods|steps)(\s|$) ➡️ LOCATIONS .*\sin\s[a-z]+$ ➡️ YEAR SPECIFIC Year-specific queries .*(2024|2025|2026).* Freshness-seeking queries .*(latest|new|updated|recent|this year|now).* "Is X still working / relevant" queries .*(still|anymore|yet).* ➡️ PROBLEM/TROUBLESHOOTING Troubleshooting queries .*(not working|broken|fix|issue|problem|error|slow|wrong|failed).* Solution-seeking queries .*(how to fix|how to stop|how to avoid|how to prevent).* Frustration/failure queries^why\s.*(not|won.t|doesn.t|can.t).* ➡️ TOFU/BOFU/MOFU TOFU — question-led awareness queries ^(what|how|why|who|when)\s MOFU — consideration queries .*(best|vs|review|compare|alternative).* BOFU — decision queries .*(buy|price|pricing|cost|hire|quote|demo|trial).* Find this useful? did you know you can do ALL of this in SEO Stack without having to use Regex! Try the powerful filtering in SEO stack for free! seo-stack.io

Hi Greg. I really enjoy your videos and have learned a lot from you. But as someone who has worked with businesses impacted by Google manual or algorithmic actions for 14 years now, I really strongly caution against what you're recommending here. You're describing several things outlined in Google's spam policies: 1. Doorway abuse 2. Scaled content abuse It's not enough to have unique wording. The rater guidelines that describe what Google wants to consider high and low quality content talk extensively about paraphrased content and whether real effort is put into making content original and insightful. The worst part about this is that it often works...for a while...and then when traffic plummets across your entire site some weeks or months later, it's impossible to know what to do. Do you remove that content? Is there something else to blame? You rarely get any information from Google on why your traffic dropped. And it's often across your entire site, not just the pages you scaled for SEO.

keyboard_arrow_left Previous keyboard_arrow_right Next

want AI content SEO to rank? do not generate endless slop everything you can generate directly using an LLM without unique context is already in the training data => useless and does not rank it's a content black hole; the question has been answered focus on anything not inside the training data; you can get thousands of chatgpt citations without a single backlink or "domain rank" everyone will destroy their website in the coming months with "pSEO" big opportunity for the rest of us

The most highly linked to content in B2B isn't thought leadership content... It's data round ups. Even though thought leadership and ultimate guides are what every CMO tells their team to publish. More here: foundationinc.co/lab/backlink-a…

One of the worst things you can do for AI search visibility? Destroying your SEO performance with shiny new AI search tactics that are ultimately dangerous for SEO - and by extension - AI search. In my latest Substack article, I wrote about the fundamental connection between SEO and GEO/AEO, along with some of the risky tactics I've seen sites employing, plus the early signs that these approaches are dangerous for SEO (and therefore AI search as well). lilyraynyc.substack.com/p/your-geo-str…

Study of 15K URLs shows that articles receiving Major Rewrites (30-100% of content) vastly outperform minor/moderate updates Even minor content updates helped, but content with no update performed the worst (1/2)

x.com/i/article/2033…

These guys from NYU and UV tested 35 LLMs and found something that matters for AI visibility. When similar content competes in a retrieval context, the model’s ability to pick the right source declines log-linearly toward zero. Not toward ‘good enough.’ Toward zero. The bottleneck isn’t context window length - they held input length constant, same result. It’s interference. More similar pages in the grounding pipeline = worse selection accuracy for all of them. What predicts resistance? Model size. Context window? No significant effect. They tried telling models to ‘ignore earlier info’ and ‘focus on recent updates.’ Barely moved the needle. CoT didn’t help either. For AI visibility this means: being structurally distinct from competing content matters more than being marginally better. Read that again. If your page looks like the other 8 in the grounding context, you’re adding to the interference problem, not solving it. Density beats length. Distinctiveness beats comprehensiveness. And flooding a category with similar placements may actually hurt citation rates, not help. Paper: arxiv.org/abs/2506.08184. Thank you @arxiv

This feels inevitable for all platforms.

John Ward @JohnWardHere

3 months ago

Facebook is updating their content originality policies

keyboard_arrow_left Previous keyboard_arrow_right Next

5 6 56 19K 23

"This GEO tactic gets you cited in LLMs!" "Interesting. How do you know that...?" "My research bro. I analyzed 10,000 prompts bro. All of the brands cited in ChatGPT have pages with expert quotes, statistics, inline citations, structured formatting, and topical depth…" "Ok…sure, but did the page get cited because it had those features? Or did it have those features because it was already high-authority content that performs well everywhere...and AI search just picked it up too?" Nobody peddling GEO fairy dust wants to answer this question. Right now, a huge chunk of GEO research is treating correlation like causation. Like these content traits are brand new levers you can pull to "get cited in AI search." Do X, get cited. Simple. Except: if everyone can just add charts and tables to a post and get cited, how does AI determine YOUR brand gets mentioned in a short output of 5 solutions in a space where thousands exist? It can't just be "do X tactics." That math doesn't work. What actually separates the brands getting cited? The same thing that's always separated them: real authority. Google's own "How Search Works" documentation states openly that links are a core signal for content quality. AI is very similar. HubSpot found that 92% of AI mentions come not from your own content, but third party sites mentioning you. So when someone says AI engines cite content with quotes, data, and good structure, they're simply observing a trait of authoritative content in general, not a "do x and get Y" causation factor. And even if these content tweaks do influence AI citation likelihood, SparkToro's recent research found there's less than a 1 in 100 chance that ChatGPT or Google AI will give the same list of brand recommendations in any two responses...even on the identical prompt run 100 times. The lists are different. The order is different. The number of results is different. Nearly every response is unique...so the agencies selling revolutionary GEO hacks are building on two broken assumptions: that these tactics directly cause citations (unproven), and that the outputs are consistent enough for those tactics to track reliably (they're not). AI search is real. It matters. Heck, we generate tons of leads from it. But the brands winning there are the same ones that have been winning everywhere... because they built actual authority over years, not because they added a data table to a blog post last Tuesday. Our own AI visibility increased the most by doing one thing: acquiring mentions + links on publications talking about our niche: being in more places that AI pulls from, increasing the likelihood AI mentions our brand.

Google is gaslighting the entire SEO industry about JavaScript. I have the receipts. This week Google removed the "Design for accessibility" section from their official JavaScript SEO documentation. Their reason? JavaScript rendering is "no longer a barrier" for Google Search. I nearly choked on my coffee. I manage SEO across hundreds of properties. I see what actually happens when Google tries to render JavaScript sites. Every single day. This is the fifth update to the same document since December 2025, part of a systematic campaign to replace broad cautions with specific technical guidance Here is what I see that Google apparently does not. Content behind tabs, accordions, and "load more" buttons? Completely invisible. Google does not click. Google does not scroll. Google does not interact. That content simply does not exist for them. Structured data injected via JavaScript? Random. Google's own documentation on developers.google.com admits that JS-generated Product markup makes shopping crawls less frequent and less reliable. I see it break constantly. Images loaded through JavaScript? Good luck getting those indexed. Lazy-loaded images behind interaction events are a black hole. Internal links rendered via JavaScript? Onely proved Google needs 9x more time to crawl JS pages than plain HTML. Nine times. Their experiment showed 313 hours for JS versus 36 hours for HTML to reach the same depth. A 2024 counterpoint study by Vercel/MERJ found most pages rendered within minutes. And here is the part nobody is talking about. AI crawlers cannot render JavaScript at all. GPTBot, ClaudeBot, PerplexityBot. They see raw HTML only. Onely's 2025 research estimated about 70% of modern websites may be completely invisible to AI search because of JavaScript dependencies. So Google removes the warning. Developers lean harder into client-side rendering. And what happens? Sites become more dependent on Google's proprietary rendering pipeline while going completely dark for every competing search and AI system. Convenient timing. Google is not telling you JavaScript is fine because it is fine. Google is telling you JavaScript is fine because it benefits Google. Server-side rendering is not optional. It never was. Do not let a documentation update convince you otherwise. Sources: 1. Barry Schwartz, "Loading Content With JavaScript Does Not Make It Harder For Google Search" Mar 5, 2026 2. Google Developers, JavaScript SEO Basics (the removed section) 3. Google Developers, Structured Data with JavaScript (the shopping crawl admission) 4. Vercel "How Google handles JavaScript throughout the indexing process" Jul 20224