Ryan Carrigan

@jbobbink Maybe you filter your own prompts but competitors also track, general research tools track… hard to filter it away.

LLM / AI trackers are quietly breaking their users' own analytics. And nobody in SEO is talking about it yet. Here is the problem. When you run daily prompt tracking across ChatGPT, Perplexity, Google AI Mode and others, the LLM sometimes decides it needs fresh information to answer. It fires off a RAG request and your pages get crawled. Not every prompt triggers this, but enough of them do when you are tracking hundreds or thousands of prompts per month. Those crawls land in your server logs as bot hits. They inflate your crawl data. They dirty your content performance analysis. And they are caused by the tool you bought to measure your AI visibility. This matters more than it would have two years ago. GA4 strips known bot traffic automatically, so the pollution does not show up there. But log file analysis is where the real GEO work is happening. Tools like botsanalyzer(.)com have made server log parsing accessible to any marketer, and practitioners are increasingly using logs to track which AI crawlers visit, how often, and which pages they fetch. That is where the signal lives for AI search optimization. And that is exactly where LLM tracker pollution is hiding. This is not a new pattern. Rank trackers did the same thing to Google Search Console years ago. Every time your rank tracker checks position 37, Google counts an impression. The more keywords you track, the noisier your GSC data gets. Most SEOs quietly accepted it. The log file version is worse. LLM tracker requests arrive with rotating user agents, sometimes through Bing, sometimes through Google, sometimes direct from OpenAI or Anthropic infrastructure. You cannot filter them out after the fact. There is a potential fix. LLM tracking vendors should publish a timestamped feed of when prompts were processed and which pages were fetched in response. A simple API or CSV export would be enough. We can overlay that feed against our logs and subtract the bot traffic our own tool caused. The only question remains: do we know when a LLM requests a URL or not? Its valuable data on it owns. It is not technically difficult but may be impossible to do correctly. It is a product decision worth investigating. And until one vendor ships it, every AI visibility report we publish is sitting on top of log data contaminated by the

@GrindstoneSEO @brucebiz2 @myeyesshine_ @wearecited sick, how did you determine the prompts worth tracking?

@brucebiz2 @myeyesshine_ @wearecited This is what mine looks like currently but I'm about to put in a good session with some logic upgrades/enhancements right now

I have sites Google refuses to index beyond the home page yet they're getting more than 100% coverage rate in GPT across a huge query fan. Careful who you read, the AIEO landscape is worse than the blackhat SEO landscape of ~20 years ago.

paid a developer $10k only to realize distribution is all that matters spent $200 in credits only to realize distribution is all that matters same lesson but now cheaper

@dejanseo This is awesome. How are you determining the chunks?

This tool shows which parts of the page will represent it in Google's AI search: gs.dejan.ai Google takes bits and pieces of text from pages and assembles a small query-specific extractive summary for each grounding source and this forms a broader model context together with personalization, user prompt and any attached media. Apart from your branding this is all you have to influence the model.

@nateliason As someone who does a lot of programmatic SEO, I wouldn't do this. Very likely, you will nuke your site.

This Programmatic SEO listing on Claw Mart looks SICK. Just purchased.

@gilshafir @foundmyfitness @grok can you answer this

@foundmyfitness What are the two brands?

Beware of most creatine gummies. Recent third-party testing found that 2/3 of creatine gummy brands contained virtually no creatine, despite label claims. Several brands advertised 5 g of creatine per serving, but testing found most delivered <0.5% of the stated amount. Only two brands actually matched (or exceeded) what they claimed. Gummies are often a bad delivery system in general, not just for creatine. Manufacturing heat and processing can degrade sensitive compounds and reduce biological activity. You’re also paying for sugar, fillers, and convenience, not potency. Plus, if you’re aiming for a meaningful daily dose (e.g., 10–20 g/day), that’s a lot of gummies. Creatine monohydrate powder is still the gold standard for those who want to supplement.

@nickgraynews #8 is something so many brands are sleeping on. Too focused on CAC and not considering the overall impact of these placements.

I went to an SEO meetup yesterday Here are 10 interesting things that I learned 1. Stop optimizing for keywords. Instead, optimize for the Agentic Research Process AI (LLMs) performs web searches just like humans do to verify facts. Instead of targeting keywords users type, target the queries an AI agent would search to build an answer about your topic. "In this new AI search world, how do you actually get brand mentions and citations? You have to look at what the agents are searching." 2. The Entity is the new Keyword Google and AI models no longer just look for string matches. Instead they look for "Entities" which is basically a known person, place, or thing. You must establish your brand or yourself as the Entity that is the authority on a specific topic. 3. Reputation Management is now 50% of SEO Because AI summarizes what is known about a brand, having a clean, authoritative digital footprint is essential. If someone searches your name, you need to control the top results to influence what ChatGPT says about you. 4. The Listicle strategy for AI visibility To get recommended by an AI as a software solution, you must appear on the "Best X Software" listicles (e.g., Efficient App, G2, TechRadar, etc). LLMs crawl these lists to determine who is a market leader. If you aren't on the list, you don't exist to the AI. 5. Shift from SEO Keyword Philosophy to Insight-Led Expertise Don't just write content to fill a keyword gap. Write content based on first-party data and unique subject matter expertise. That is the only moat left against AI-generated slop. 6. YouTube doesn't need high production We talked a lot about how people are using YouTube to boost rankings and get more traffic. A Loom video walkthrough often converts better than a highly produced commercial because it feels more authentic and helpful. 7. The Shorts Algo Hack YouTube Shorts and TikToks can and do rank well in Google Search. If you want to rank for a software tool, sometimes a 60-second vertical video will leapfrog a 2,000-word blog post. 8. Affiliate expenses are a Frenemy Game You have to pay the tax to the affiliates (lists of "Best Tools") to get the brand mentions required to rank organically yourself. Think of this more like a sunk cost and focus less on your CAC for these placements. 9. The Cluster Strategy I’m not sure where we landed here. But the idea was: Don't just write one article on a topic. You need to surround the topic with 10-20 articles to prove to Google you are a subject matter expert. However, one person said his website had very few articles but they were all extremely high quality - and he was ranking well. 10. Brand Mentions are the new Backlinks In the LLM era, a clickable link matters less than the text context. If an AI reads your brand name associated with "reliable," "affordable," or "best" across the web, this is good. Even without a link it will strengthen your Entity status. You need to get people talking about you, not just linking to you. THE END Thanks to Chandler from SelfPublishing.com for hosting us and for everyone who replied to my post here asking for people to invite.

Power outages and radar reflectivity over the past 72 hours in 1 hour intervals. Nothing spookier than losing electricity. Happy Halloween. Stay safe. Hopefully everyone affected gets their power back on soon.

@bhalligan Big companies regularly use M&A as a way to look inside companies to copy them. I always take the conversations but wouldn’t go down the sale rabbit hole until you’re ready to run a process.

Dear Founders, If you have acquisition interest from a bigger co, take it very seriously. If you don’t sell, that bigger co will buy your competitor or build. Not great for you. Also, valuations are sky high right now, so it’s not a terrible time to sell. ❤️, Brian.

@gaetano_nyc Did the same during COVID and might as well have lit the money on fire. If it’s a SAFE, I’m out. Only deals I’ll do now are ones that have opportunity to cashflow.

@carriganisms Several of them, yes.

Since 2020, I’ve angel invested a total of 6 figures into various startups and venture funds. To this day I still have not seen a return. Here’s what they don’t tell you about angel investing: 1. It takes a long ass time before a liquidity event happens. Easily 5 years minimum. But likely longer. 2. There’s an opportunity cost. We are in precious metals / crypto bull run. Stocks are running phenomenal. $100K in those assets will see faster and probably better returns. 3. Angel investments are illiquid. There’s no changing your mind once it’s done. 4. It’s a pretty big gamble. You’re hoping for outsized returns but chances are low. 5. You have to be comfortable with losing 100% of your investment. 6. Generally you are betting more on the founders than the actual idea. This is true for half of my investments. 7. You can get diluted later on and see much smaller returns than expected. I’m way more selective now and less likely to continue angel investing given how well other asset classes are performing, and the opportunity cost involved.

@nickeubanks Yeah, I'm sure it would need to be an official ad marketplace announcement.

@carriganisms guessing this bet is specific to ad blocks (with disclosure)

Anyone want to place a bet?

@nickeubanks Seems like you should place a bet then...haha

@carriganisms Technically they already have ads. They’re just not disclosed 😬

@JamesonCamp Tempting but long term it’s a losing strategy.

I haven’t touched anything blackhat or grey hat marketing related in maybe 15 years But my god would it be easy to print money these days with AI and aff or Ecom if you were willing to cross that line

@boringmarketer What is the transcript you’re passing to Claude? Notes from the screen recordings about user behavior?

Take two weeks and accomplish/learn more than most do in 6 months: Use Claude Code: - Build a simple SaaS or boring service with existing demand - Build ~20 keyword focused landing pages - Optimize the site for technical SEO - Add search console and google analytics - Reduce onboarding friction - Fast time to value user experience - Harden for security - Watch 5 people use it, record - Pass transcript to Claude, optimize Put in 10-12 hour days over two weeks, you’ll come out the other side a different person

@oliviaakory Is this ecom?

I was tagged into a chart that another inc measurement vendor published asking for my thoughts on Pinterest. Is it the most incremental of all the social channels? The data showed iROAS by partner: Pin $2.65 Meta $2.18 Snap $2.10 TikTok $1.64 Reddit $1.26 X $.30 My thoughts: Pinterest is looking really good @ Haus as well, and we have some case studies coming soon. A lot of exciting opportunity. That said, it's really hard for me to compare channel performance without talking about scale. I think it's misleading to compare the iROAS of different social channels without overlaying spend levels. My guess is that investment in Meta is 10x Pin, and that should be noted in these types of comparative analyses.

@gaetano_nyc Can you give an example? So much of the "win in LLMs" advice is just repackaged SEO strategy. Digital PR, updating content, etc.

I’m hearing a lot more “oh we just wanna do this for LLMs” now on client calls. There is BAD ADVICE everywhere and people are under the influence. My response is always: “If LLMs didn’t exist, would you still be doing this?” And the answer is almost always… No.

@gaetano_nyc If it's an emerging category, how big is the search opportunity? And is there much competition around these searches?

I have a client with ZERO authority in an emerging category. They think just by publishing “thought leadership” it will make them an authority. Of course it won’t. They need external coverage. So here’s the question for you guys: 1. Do they need to hire a link building company? 2. Do they need to do “digital PR”? 3. Do they need to hire a mix of both? 4. Should it be an agency or a “sniper” consultant? 5. How much do they need to invest before seeing results? 6. How long do they need to invest before they see results? Best answer gets an intro.

@lilyraynyc @dejanseo Strong argument. GPT plug-ins will likely also be a big opportunity.

Fantastic new article by @dejanseo that explains why GPT-5 is actually a good thing for SEOs: It relies on grounding (web search) and other tools to be accurate Worth a read dejan.ai/blog/gpt-5-mad…