Heads up!!! Google stripped destination URLs from the signed-out SERP HTML. So /goto links can no longer be resolved without a GET to /goto?url=<token> (btw: don't follow the redirect, just read Location). The "Translate this page" trick is dead too 💀

ChatGPT *already knows* what brands it's going to recommend before it searches, according to data shared by @suganthan Even more relevant when you learn 95% of users use the "free" version of ChatGPT, which barely searches at all 😳

agents use llms.txt more and more now i am so fcking tired of people not understanding the difference between SEO and agentic/agent use no google will not use llms.txt, claude/codex/etc will same with markdown etc.

Patrick goes deep on all of it — indie tools, the death of generic AI content, and building for the GEO era. 🎧 Full episode with @patrickstox on @unscriptedseo: unscriptedseo.com/patrick-st…

I think the next frontier of spam fighting from the search/AI companies is going to hit one of the most popular page types in SEO/GEO right now: comparison and alternative pages. These are pages on your own site that pit your brand against competitors, across the whole brand or a specific product or service (Brand X vs. Y, Product X vs. Y, [Competitor] Alternatives). They've been making the rounds as an SEO/GEO goldmine lately, and "build a lot of these" is advice I keep hearing at conferences and reading in industry blogs and social posts. This is similar to what I've been sharing all year with "listicles" - and I've got a new Substack piece coming out (ASAP!) on how Google and OpenAI are tuning their AI answers to lean less on brands promoting themselves. Stay tuned for that. Comparison pages are trickier and more nuanced. There are genuine use cases for doing them well, especially at a small scale, and I've occasionally recommended them to clients in certain situations. But like everything in SEO: once a tactic works, it becomes popular and people scale it. (File under: "this is why we can't have nice things" 🫠) The "build these pages" advice inevitably leads folks to use AI to spin up as many as possible. I've already advised several companies that launched dozens, even hundreds of comparison/alternative pages (AKA an "SEO pattern") and ran into trouble. For example, A couple sites got hit by Google's late January update this year. The fundamental problem: like any other "review" page, Google already has strict criteria for product reviews. The core requirement is that you prove you actually tried and tested the product, with evidence. For a brand comparing itself to competitors, meeting that bar would mean you've hired your competitor(s) and your own company, and are authentically reviewing your honest experience using their services. How many brands can say they've actually done that? My guess is probably close to 0. What usually ends up happening instead is brands do "research" about their competition, leaning on negative reviews of their competitors or worse - making up incorrect or untrue information to make your own offering look better than theirs. I believe we're already seeing - and will keep seeing - Google, OpenAI and other AI companies look for "objective" 3rd party reviews* instead of leaning on biased brand content when they generate answers. This is ultimately why I think search/AI companies lean so heavily into Reddit discussions, YouTube reviews, Trustpilot, G2, and other major review sites. (*Yes, I know these 3rd party sites are often influenced / manipulated / pay-to-play as well... so that's a big challenge for them too.) So, as always: test, experiment, learn. But my spidey senses say that building these pages at scale - and doing it inauthentically - like trashing your competitor while calling yourself the best - is risky. When too many sites adopt these approaches, they become a liability long-term.