{"id":110374,"date":"2026-01-14T00:42:48","date_gmt":"2026-01-14T00:42:48","guid":{"rendered":"https:\/\/mailinvest.blog\/index.php\/2026\/01\/14\/google-downplays-geo-but-lets-talk-about-garbage-ai-serps\/"},"modified":"2026-01-14T00:43:59","modified_gmt":"2026-01-14T00:43:59","slug":"google-downplays-geo-but-lets-talk-about-garbage-ai-serps","status":"publish","type":"post","link":"https:\/\/mailinvest.blog\/index.php\/2026\/01\/14\/google-downplays-geo-but-lets-talk-about-garbage-ai-serps\/","title":{"rendered":"Google Downplays GEO – But Let’s Talk About Garbage AI SERPs"},"content":{"rendered":"

<\/a>\n<\/p>\n

\n

Google\u2019s Danny Sullivan and John Mueller\u2019s Search Off The Report podcast provided steering to SEOs and publishers who’ve questions on rating in LLM-based search and chat, debunking the generally repeated recommendation to \u201cchunk your content material.\u201d However that\u2019s actually not the dialog Googlers needs to be having proper now.<\/p>\n

web optimization And The Subsequent Era Of Search<\/h2>\n

Google used to rank content material primarily based on key phrase matching and PageRank was a approach to prolong that paradigm utilizing the anchor textual content of hyperlinks. The introduction of the Data Graph in 2012 was described as a step towards rating solutions primarily based on issues (entities) in the true world. Google referred to as this a shift from strings to issues.<\/p>\n

What\u2019s taking place in the present day is what Google in 2012 referred to as \u201cthe following era of search, which faucets into the collective intelligence of the net and understands the world a bit extra like folks do.\u201d<\/p>\n

So, when folks say that nothing has modified with web optimization, it\u2019s true to the extent that the underlying infrastructure remains to be Google Search. What has modified is that the solutions are in a long-form format that solutions three or extra extra questions past the consumer\u2019s preliminary question.<\/p>\n

The reply to the query of what\u2019s completely different about web optimization for AI is that the paradigm of optimizing for one key phrase for one search result’s shattered, splintered by the question fan-out.<\/p>\n

Google\u2019s Danny Sullivan and John Mueller took a crack at providing steering on what SEOs needs to be specializing in. Do they hit the mark?<\/p>\n

How To Write For Longform Solutions<\/h2>\n

Provided that Google is surfacing multi-paragraph lengthy solutions, does it make sense to create content material that\u2019s organized into bite-sized chunks? How does that have an effect on how people learn content material, will they prefer it or go away it?<\/p>\n

Many SEOs are recommending that publishers break up the web page up into \u201cchunks\u201d primarily based on the instinct that AI understands content material in chunks, dividing up the web page into sections. However that\u2019s an arbitrary method that ignores the truth that a correctly structured internet web page is already damaged into chunks via the usage of headings, HTML components like ordered and unordered lists. A correctly marked up and formatted internet web page ought to already be formatted into logical construction {that a} human and a machine can simply perceive. Duh\u2026 proper?<\/p>\n

It\u2019s not stunning that Google\u2019s Danny Sullivan warns SEOs and publishers to not break their content material up into chunks.<\/p>\n

Danny stated:<\/em><\/p>\n

\n

\u201cTo go to one of many issues, , I talked concerning the particular issues folks like, \u201cWhat’s the factor I want to enhance.\u201d One of many issues I hold seeing again and again in a number of the recommendation and steering and persons are making an attempt to determine what will we do with the LLMs or no matter, is that flip your content material into bite-sized chunks, as a result of LLMs like issues which can be actually chew measurement, proper?<\/p>\n

So we don\u2019t need you to try this. I used to be speaking to some engineers about that. We don\u2019t need you to try this. We actually don\u2019t. We don\u2019t need folks to must be crafting something for Search particularly. That\u2019s by no means been the place we\u2019ve been at and we nonetheless proceed to be that manner. We actually don\u2019t need you to assume it’s worthwhile to be doing that or produce two variations of your content material, one for the LLM and one for the online.\u201d<\/p>\n<\/blockquote>\n

Danny talked about chunking with some Google engineers and his takeaway from that dialog is to suggest towards chunking. The second takeaway is that their techniques are set as much as entry content material the way in which human readers entry it and for that cause he says to craft the content material for people.<\/p>\n

Avoids Speaking About Search Referrals<\/h2>\n

However once more, he avoids speaking about what I believe is the extra essential aspect of AI search, question fan-out and the impression to referrals. Question fan-out impacts referrals as a result of Google is rating a handful of pages for a number of queries for each one question {that a} consumer makes. However compounds this example, as you will note additional on, is that the websites Google is rating don’t measure up.<\/p>\n

Focus On The Huge Image<\/h2>\n

Danny Sullivan subsequent discusses the draw back of optimizing for a machine, explaining that techniques ultimately enhance that often implies that optimization for machines cease working.<\/p>\n

He defined:<\/em><\/p>\n

\n

\u201cAfter which the techniques enhance, most likely the way in which the techniques at all times attempt to enhance, to reward content material written for people. All that stuff that you simply did to please this LLM system which will or might not have labored, might not carry via for the long run.<\/p>\n

\u2026Once more, it’s a must to make your individual selections. However I believe that what you are likely to see is, over time, these little or no particular issues are usually not the issues that carry you thru, however , you make your individual selections. However I believe additionally that many individuals who’ve been within the web optimization house for a really very long time will see this, will acknowledge that, , specializing in these foundational targets, that\u2019s what carries you thru.\u201d<\/p>\n<\/blockquote>\n

Let\u2019s Discuss About Rubbish AI Search Outcomes<\/h2>\n

I’ve recognized Danny Sullivan for a very long time and have a ton of respect for him, I do know that he has publishers in thoughts and that he really desires for them to succeed. What I wanted he would speak about is the declining site visitors alternatives for subject-matter specialists and the seemingly arbitrary rubbish search outcomes that Google persistently surfaces.<\/p>\n

Topic Matter Experience Is Lacking<\/h3>\n

Google is deliberately hiding knowledgeable publications within the search outcomes, hidden away within the Extra tab. With the intention to discover knowledgeable content material, a consumer has to click on the Extra tab after which click on the Information tab.<\/p>\n

How Google Hides Knowledgeable Net Pages<\/h4>\n

\"How<\/p>\n

Google\u2019s AI Mode Promotes Rubbish And Websites Missing Experience<\/h2>\n

This search was not cherry-picked to indicate poor outcomes. That is actually the one search I did asking a legit query about styling a sweatshirt.<\/p>\n

Google\u2019s AI Mode cites the next pages:<\/strong><\/p>\n

1. An deserted Medium Weblog from 2018, that solely ever had two weblog posts, each of which have damaged pictures. That\u2019s not authoritative.<\/p>\n

2. An article revealed on LinkedIn, a enterprise social networking web site. Once more, that\u2019s not authoritative nor reliable. Who goes to LinkedIn for knowledgeable fashion recommendation?<\/p>\n

3. An article about sweatshirts revealed on a sneaker retailer\u2019s web site. Not knowledgeable, not authoritative. Who goes to a sneaker retailer to learn articles about sweatshirts?<\/p>\n

Screenshot Of Google\u2019s Rubbish AI Outcomes<\/h4>\n

\"\"<\/p>\n

Google Hides The Good Stuff In Extra > Information Tab<\/h3>\n

Had Google defaulted to precise knowledgeable websites they could have linked to an article from GQ or the New York Occasions, each respected web sites. As an alternative, Google hides the prime quality internet pages beneath the Extra tab.<\/p>\n

Screenshot Of\u00a0 Hidden Excessive High quality Search Outcomes<\/h3>\n

\"\"<\/p>\n

GEO Or web optimization \u2013 It Doesn\u2019t Matter<\/h2>\n

This entire factor about GEO or AEO and whether or not it\u2019s all web optimization doesn\u2019t actually matter. It\u2019s all a bunch of hand waving and bluster. What issues is that Google is not rating prime quality websites and prime quality websites are withering from a scarcity of site visitors.<\/p>\n

I see these low high quality SERPs all day lengthy and it\u2019s miserable as a result of there isn’t any pleasure of discovery in Google Search anymore. When was the final time you found a very cool website that you simply wished to inform somebody about?<\/p>\n

Rubbish on rubbish, on rubbish, on high of extra rubbish.\u00a0Google wants a reset.<\/p>\n

How about Google brings again the unique search and we are able to have all of the hand-wavy Gemini stuff beneath the Extra tab someplace?<\/p>\n

Take heed to the podcast right here:<\/strong><\/p>\n