A quick note up front, so we start on the right foot.<\/strong><\/p> The research I\u2019m about to reference is not mine. I did not run these experiments. I\u2019m not affiliated with the authors. I\u2019m not here to \u201cendorse\u201d a camp, pick a side, or crown a winner. What I am going to endorse, loudly and without apology, is measurement. Replication. Real-world experiments. The kind of work that teaches us in real time, in real life, what changes when an LLM sits between customers and content. We need more tested data, and this is one of those starting points.<\/p> If you do nothing else with this article, do this: Read the paper, then run your own test. Whether your results agree or disagree, publish them. We need more receipts and fewer hot takes.<\/p> Now, the reason I\u2019m writing this.<\/p> Over the last year, the industry has been pushed toward a neat, comforting story: GEO is just SEO. Nothing new to learn. No need to change how you work. Just keep doing the fundamentals, and everything will be fine.<\/p> I don\u2019t buy that.<\/p> Not because SEO fundamentals stopped mattering. They still matter, and they remain necessary. But because \u201cnecessary\u201d is not the same as \u201csufficient,\u201d and because the incentives behind platform messaging do not always align with the operational realities businesses are walking into and dealing with.<\/p> If you\u2019ve paid attention to public guidance coming from the leading search platforms lately, you\u2019ve probably heard a version of: Don\u2019t focus on chunking. Don\u2019t create \u201cbite-sized chunks.\u201d Don\u2019t optimize for how the machine works. Focus on good content.<\/p> That\u2019s been echoed and amplified across industry coverage, though I want to be precise about my position here. I\u2019m not claiming a conspiracy, and I\u2019m not saying anyone is being intentionally misleading. I\u2019m not doing that.<\/p> I am saying something much simpler. It\u2019s my opinion and happens to be based on actual experience \u2013 when messaging repeats across multiple spokespeople in a tight window, it signals an internal alignment effort.<\/p> That\u2019s not an insult nor is it a moral judgment. That\u2019s how large organizations operate when they want the market to hear one clear message. I was part of exactly that type of environment for well over a decade in my career.<\/p> And the message itself, on its face, is not wrong. You can absolutely hurt yourself by over-optimizing for the wrong proxy. You can absolutely create brittle content by trying to game a system you do not fully understand. In many cases, \u201cwrite clearly for humans\u201d is solid baseline guidance.<\/p> The problem is what happens when that baseline guidance becomes a blanket dismissal of how the machine layer works today, even if it\u2019s unintentional. Because we are not in a \u201c10 blue links\u201d world anymore.<\/p> We are in a world where answer surfaces are expanding, search journeys are compressing, and the unit of competition is shifting from \u201cthe page\u201d to \u201cthe selected portion of the page,\u201d assembled into an answer the user never clicks past.<\/p> And that is where \u201cGEO is just SEO\u201d starts to break in my mind.<\/p> Executives love comforting statements: \u201cGoogle still dominates search. Traditional SEO still drives the most traffic. Therefore this LLM-stuff is overblown.<\/em>\u201d<\/p> The first half is true, but the conclusion is where companies get hurt.<\/p> The biggest risk here is asking the wrong question. \u201cWhere does traffic come from today?<\/em>\u201d is a dashboard question, and it\u2019s backward-looking. It tells you what has been true.<\/p> The more important questions are forward-looking:<\/p> If you want the behavior trendline in plain terms, start here, with the\u00a02024 SparkToro study, then take a look at what Danny Goodwin wrote in\u00a02024, and as a follow-up in 2025\u00a0(spoiler \u2013 zero click instances increased Y-o-Y). And while some sources are a couple of years old, you can easily find newer data showing the trend growing.<\/p> I\u2019m not using these sources to claim \u201cthe sky is falling.\u201d I\u2019m using them to reinforce a simple operational reality: If the click declines, \u201cranking\u201d is no longer the end goal. Being selected into the answer becomes the end goal.<\/p> That requires additional thinking beyond classic SEO. Not instead of it. On top of it.<\/p> If you want to understand why the public messaging is conservative, you have to look at the platform\u2019s strategic direction.<\/p> Google, for example, has been expanding AI answer surfaces, and it\u2019s not subtle. Both\u00a0AI Overviews\u00a0and\u00a0AI Mode\u00a0saw announcements of large expansions during 2025.<\/p> Again, notice what this implies at the operating level. When AI Overviews and AI Mode expand, you\u2019re not just dealing with \u201cranking signals.\u201d You\u2019re dealing with an experience layer that can answer, summarize, recommend, and route a user without a click.<\/p> Then comes the part everyone pretends not to see until it\u2019s unavoidable: Monetization follows attention.<\/p> This is no longer hypothetical. Search Engine Journal covered Google\u2019s official rollout of\u00a0ads in AI Overviews, which matters because it signals this answer layer is being treated as a durable interface surface, not a temporary experiment.<\/p> Google\u2019s own Ads\u00a0documentation reinforces the same point: This isn\u2019t just \u201csomething people noticed,\u201d it\u2019s a supported placement pattern with real operational guidance behind it. And Google noted\u00a0mid-last-year that AI Overviews monetize at a similar rate to traditional search, which is a quiet signal that this isn\u2019t a side feature.<\/p> You do not need to be cynical to read this clearly. If the answer surface becomes the primary surface, the ad surface will evolve there too. That\u2019s not a scandal so much as just the reality of where the model is evolving to.<\/p> Now connect the dots back to \u201cdon\u2019t focus on chunking\u201d-style guidance.<\/p> A platform that is actively expanding answer surfaces has multiple legitimate reasons to discourage the market from \u201cengineering for the answer layer,\u201d including quality control, spam prevention, and ecosystem stability.<\/p> Businesses, however, do not have the luxury of optimizing for ecosystem stability. Businesses must optimize for business outcomes. Their own outcomes.<\/p> That\u2019s the tension.<\/p> This isn\u2019t about blaming anyone. It\u2019s about understanding misaligned objectives, so you don\u2019t make decisions that feel safe but cost you later.<\/p> I\u2019m on record that traditional search is still an important driver, and that optimizing in this new world is additive, not an overnight replacement story. But \u201cadditive\u201d still changes the workflow.<\/p> AI assistants are becoming measurable referrers. Not dominant, not decisive on their own, but meaningful enough to track as an early indicator. Two examples that capture this trend.<\/p> TechCrunch\u00a0noted that while it\u2019s not enough to offset the loss of traffic from search declines, news sites are seeing growth in ChatGPT referrals. And\u00a0Digiday\u00a0has data showing traffic from ChatGPT doubled from 2024 to 2025.<\/p> Why do I include these?<\/p> Because this is how platform shifts look in the early stages. They start small, then they become normal, then they become default. If you wait for the \u201cbig numbers,\u201d you\u2019re late building competence and in taking action. (Remember \u201cdirectories\u201d? Yeah, Search ate their lunch.)<\/p> And competence, in this new environment, is not\u00a0\u201chow do I rank a page<\/em>.\u201d\u00a0It\u2019s\u00a0\u201chow do I get selected, cited, and trusted when the interface is an LLM.<\/em>\u201d<\/p> This is where the \u201cGEO is just SEO\u201d framing stops being a helpful simplification and starts becoming operationally dangerous.<\/p> Let\u2019s talk about the research. The paper I\u2019m referencing here is\u00a0publicly available,\u00a0and I\u2019m going to summarize it in plain English, because most practitioners do not have time to parse academic structure during the week.<\/p> At a high level, the (\u201cE-GEO: A Testbed for Generative Engine Optimization in E-Commerce\u201d) paper tests whether common human-written rewrite heuristics actually improve performance in an LLM-mediated product selection environment, then compares that to a more systematic optimization approach. It uses ecommerce as the proving ground, which is smart for one reason: Outcomes can be measured in ways that map to money. Product rank and selection are economically meaningful.<\/p> This is important because the GEO conversation often gets stuck in \u201cvibes.\u201d In contrast, this work is trying to quantify outcomes.<\/p> Here\u2019s the key punchline, simplified:<\/p> A lot of common \u201crewrite advice\u201d does not help in this environment. Some of it can be neutral. Some of it can be negative. But when they apply a meta-optimization process, prompts improve consistently, and the optimized patterns converge on repeatable features.<\/em><\/p> That convergence is the part that should make every practitioner sit up. Because convergence suggests there are\u00a0stable signals<\/em>\u00a0the system responds to. Not mystical. Not magical. Not purely random.<\/p> Stable signals.<\/p> And this is where I come back to my earlier point: If GEO were truly \u201cjust SEO,\u201d then you would expect classic human rewrite heuristics to translate cleanly. You would expect the winning playbook to be familiar.<\/p> This paper suggests the reality is messier. Not because SEO stopped mattering, but because the unit of success changed.<\/p> I want to be careful here, as I\u2019m not telling you to treat this paper like doctrine. You should not accept it on face value and suddenly adopt this as gospel. You should treat it as a public experiment that deserves replication.<\/p> Now, the most valuable output isn\u2019t the exact numbers in their environment, but rather, it\u2019s the shape of the solution the optimizer keeps converging on. (The name of their system\/process is optimizer<\/em>.)<\/p> The optimized patterns repeatedly emphasize clarity, explicitness, and decision-support structure. They reduce ambiguity. They surface constraints. They define what the product is and is not. They make comparisons easier. They encode \u201cselection-ready\u201d information in a form that is easier for retrieval and ranking layers to use.<\/strong><\/em><\/p> That is a different goal than classic marketing copy, which often leans on narrative, brand feel, and emotional persuasion.<\/p> Those things still have a place. But if you want to be selected by an LLM acting as an intermediary, the content needs to do a second job: become machine-usable decision support.<\/p> That\u2019s not \u201canti-human.\u201d It\u2019s pro-clarity, and it\u2019s the kind of detail that will come to define what \u201cgood content\u201d means in the future, I think.<\/p> What follows is not me inventing a process out of thin air. This is me reverse-engineering what their optimization process converged toward, and turning it into a repeatable template you can apply to product descriptions and other decision-heavy content.<\/p> Treat it as a starting point, then test it. Revise it, create your own version, whatever.<\/p> Step 1:<\/strong> State the product\u2019s purpose in one sentence, with explicit context. Example pattern: Step 2:<\/strong> Declare the selection criteria you satisfy, plainly. Include what the buyer cares about most in that category. If the category is knives, it\u2019s steel type, edge retention, maintenance, balance, handle material. If it\u2019s software, it\u2019s integration, security posture, learning curve, time-to-value.<\/p> Make it explicit.<\/p> Step 3:\u00a0<\/strong>Surface constraints and qualifiers early, not buried. Examples of qualifiers that matter: Step 4:<\/strong> State what it is, and what it is not. Pattern: Step 5:<\/strong> Convert benefits into testable claims. Do not fabricate. Do not inflate. This is not about hype. It\u2019s about clarity.<\/p> Step 6:\u00a0<\/strong>Provide structured comparison hooks. Give the model clean hooks: Step 7:\u00a0<\/strong>Add evidence anchors that improve trust. This is not about adding fluff. It\u2019s about making your claims attributable and your product legible.<\/p> Step 8:<\/strong> Close with a decision shortcut. Pattern: That\u2019s the template*.<\/p> Notice what it does. It turns a product description into structured decision support, which is not how most product copy is written today. And it is an example of why \u201cGEO is just SEO\u201d fails as a blanket statement.<\/p> SEO fundamentals help you get crawled, indexed, and discovered. This helps you get selected when discovery is mediated by an LLM.<\/p> Different layer. Different job.<\/p> Saying GEO = SEO and SEO = GEO is an oversimplification that will become normalized and lead to people missing the fact that the details matter. The differences, even small ones, matter. And they can have impacts and repercussions.<\/p> *A much deeper-dive pdf version of this process is available for my Substack subscribers for free via\u00a0my resources page.<\/em><\/p> Here\u2019s the part I want to be explicit about. This paper is interesting because it\u2019s measurable, and because it suggests the system responds to repeatable features.<\/p> But you should treat it as a starting point, not a law of physics. Results like this are sensitive to context: industry, brand authority, page type, and even the model and retrieval stack sitting between the user and your content.<\/p> That\u2019s why replication matters. The only way we learn what holds, what breaks, and what variables actually matter is by running controlled tests in our own environments and publishing what we find. If you work in SEO, content, product marketing, or growth, here is the invitation.<\/p> Read the paper here.<\/p> Then run a controlled test on a small, meaningful slice of your site.<\/p> Keep it practical:<\/p> And if you can, track whether these pages are being surfaced, cited, paraphrased, or selected in the AI answer interfaces your customers are increasingly using.<\/p> <\/blockquote> You are not trying to win a science fair. You are trying to reduce uncertainty with a controlled test. If your results disagree with the paper, that\u2019s not failure. That\u2019s signal.<\/p> Publish what you find, even if it\u2019s messy. Even if it\u2019s partial. Even if the conclusion is \u201cit depends.\u201d Because that is exactly how a new discipline becomes real. Not through repeating platform talking points. Not through tribal arguments. Through measurement.<\/p> Platform guidance is one input, not your operating system. Your operating system is your measurement program. SEO is still necessary. If you can\u2019t get crawled, you can\u2019t get chosen.<\/p> But GEO, meaning optimizing for selection inside LLM-mediated discovery, is an additional competence layer. Not a replacement. A layer. If you decide to ignore that layer because a platform said \u201cdon\u2019t optimize,\u201d you\u2019re outsourcing your business risk to someone else\u2019s incentive structure.<\/p> And that\u2019s not a strategy. The strategy is simple: learn the layer by testing the layer.<\/p> We need more people doing exactly that.<\/p> More Resources:<\/strong><\/p> This post was originally published on Duane Forrester Decodes.<\/em><\/p> Featured Image: Rawpixel.com\/Shutterstock<\/em><\/p> <\/div> SEO#Platforms #Dont #Optimize #Smart #Teams #Run #Experiments #sejournal #DuaneForrester1769094605<\/p> ","protected":false},"excerpt":{"rendered":" A quick note up front, so we start on the right foot. The research I\u2019m about to reference is not mine. I did not run these experiments. I\u2019m not affiliated with the authors. I\u2019m not here to \u201cendorse\u201d a camp, pick a side, or crown a winner. What I am going to endorse, loudly and […]<\/p>\n","protected":false},"author":1,"featured_media":1975,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[16],"tags":[458,387,3820,183,4713,537,80,4411,252],"class_list":["post-1974","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-accessibility","tag-dont","tag-duaneforrester","tag-experiments","tag-optimize","tag-platforms","tag-run","tag-sejournal","tag-smart","tag-teams"],"acf":[],"_links":{"self":[{"href":"http:\/\/longzhuplatform.com\/index.php?rest_route=\/wp\/v2\/posts\/1974","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/longzhuplatform.com\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/longzhuplatform.com\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/longzhuplatform.com\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/longzhuplatform.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=1974"}],"version-history":[{"count":0,"href":"http:\/\/longzhuplatform.com\/index.php?rest_route=\/wp\/v2\/posts\/1974\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"http:\/\/longzhuplatform.com\/index.php?rest_route=\/wp\/v2\/media\/1975"}],"wp:attachment":[{"href":"http:\/\/longzhuplatform.com\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=1974"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/longzhuplatform.com\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=1974"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/longzhuplatform.com\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=1974"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}![\"When](\"https:\/\/cdn.searchenginejournal.com\/wp-content\/uploads\/2026\/01\/run-experiments-43.jpg\" "\\"When")
The Narrative And The Incentives<\/h2>
The Wrong Question: \u201cIs Google Still The Biggest Traffic Driver?\u201d<\/h2>
The Platform Footprint Is Changing, And The Business Model Is Following<\/h2>
Discovery Is Fragmenting Beyond Google, And Early Signals Matter<\/h2>
Now, The Receipts: A Paper That Tests GEO Tactics And Shows Measurable Differences<\/h2>
What The Optimizer Keeps Finding, And Why That Matters<\/h2>
The Universal LLM-Optimization Rewrite Recipe, Framed As A Reusable Template<\/h2>
Not \u201cpremium quality.\u201d Not \u201cbest in class.\u201d Purpose.<\/p>
This is a designed for [specific use case] in [specific constraints], for people who need [core outcome].<\/p>
This is where you stop writing like a brochure and start writing like a spec sheet with a human voice.<\/p>
Most marketing copy hides the \u201cbuts\u201d until the end. Machines do not reward that ambiguity.<\/p>
Not ideal for [X]. Works best when [Y]. Requires [Z]. Compatible with [A], not [B]. This matters if you [C].<\/p>
This is one of the simplest ways to reduce ambiguity for both the user and the model.<\/p>
This is for [audience]. It is not for [audience].
This is optimized for [scenario]. It is not intended for [scenario].<\/p>
Instead of \u201cdurable,\u201d say what durable means in practice. Instead of \u201cfast,\u201d define what \u201cfast\u201d looks like in a workflow.<\/p>
LLMs often behave like comparison engines because users ask comparative questions.<\/p>
Compared to [common alternative], this offers [difference] because [reason].
If you\u2019re choosing between [A] and [B], pick this when [condition].<\/p>
This can be certifications, materials, warranty terms, return policies, documented specs, and other verifiable signals.<\/p>
Make the \u201cif you are X, do Y\u201d moment explicit.<\/p>
Choose this if you need [top 2\u20133 criteria]. If your priority is [other criteria], consider [alternative type].<\/p> What To Do Next: Read The Paper, Then Replicate It In Your Environment<\/h2>
One Final Level-Set, For The Executives Reading This<\/h2>