Prompt volumes keep increasing - people keep using search and their behaviors keep changing Search is the plumbing that connects users to the information they are looking for Users change, search changes. It’s normal, don’t panic. Look at the big picture and figure out a way to help users @methode @shenzhenseoconf

I can't share the URL, but I just found an example of ChatGPT using fan-out queries that contained site: for the *old domain* for a well-known brand that migrated its domain over 6 months ago. 👀 The migration was handled well from an SEO perspective, but the old domain name must be stored in ChatGPT's memory. The worst part? When you use Google or Bing to do a site: search that is limited to the old version of the domain, the new domain's URLs don't show up in the search results. So your shiny new URLs likely won't be part of the response unless they are discovered through another fan-out query or another part of the search 🫠

This is why we can’t have nice things and why some SEOs (really “GEOs” in this case) give the broader industry such a bad reputation.

One of the biggest mistakes I see site owners make when chasing SEO/GEO performance: they don't realize that certain page types - especially when published at scale - can be more of a liability than an asset. This is especially true right now with AI tools that recommend new pages to build for rankings or citations, without understanding which types of pages can actually be risky for overall performance. For a long time, Google and other search engines have been able to understand the core focus of a website and identify pages that don't fit into that focus. Or worse: pages built for the sole purpose of driving SEO (and now AI) traffic/visibility, without providing a genuinely helpful, objective, truthful experience for the user. Something I see often: many folks who create these pages think they *are* helpful. I get it - it's natural to feel this way when you put time and effort into creating content. But what they often don't stop to ask is whether an objective third party would also truly see the page as helpful, truthful, and unbiased - or whether they'd see it as promotional content pretending to be "educational." When you create biased or self-promoting pages at scale - content that hooks the reader with the promise of educating them, but ultimately steers them toward your product as the single solution to their problem - that pattern can become a major SEO liability for your domain. A core theme of many Google updates over the years: they want objective, E-E-A-T-driven content that actually helps users - not sponsored or biased information meant to persuade people to take action on behalf of the brand. With AI search, I'm seeing companies now produce these types of pages faster than ever. It seems like the obvious way to earn more mentions/citations, and AI makes it easier than ever to scale templates programmatically. But this is a classic SEO mistake: scaling this content often backfires within ~6 months, and it can ultimately drag down the performance of the entire domain. It's the single most common pattern I see among sites affected by algorithm updates. I've now identified over *70 companies* impacted by Google's January 20th update, where this exact situation played out: many using highly similar SEO/GEO-driven content approaches. When Google can spot a footprint like that at scale, it becomes very easy for them to demote dozens of companies using the same tools and techniques to drive self-promoting SEO and GEO visibility. And when so many companies are using the similar tools that make similar recommendations, we're going to keep ending up with this same problem. IMO, the best you can do right now is spend that time, energy, and budget creating something truly unique. Not "unique" in the sense of asking a language model to "add information gain" to your otherwise AI-generated content; but unique because it's made of original research, opinions, ideas, and genuine expertise.