{"id":14257,"date":"2026-11-12T16:41:14","date_gmt":"2026-11-12T15:41:14","guid":{"rendered":"https:\/\/inspace.io\/?post_type=insight&#038;p=14257"},"modified":"2026-11-12T16:41:15","modified_gmt":"2026-11-12T15:41:15","slug":"why-geo-doesnt-exist-without-good-seo","status":"publish","type":"insight","link":"https:\/\/inspace.io\/en-nl\/insights\/why-geo-doesnt-exist-without-good-seo","title":{"rendered":"SEO vs. GEO: why GEO doesn\u2019t exist without good SEO"},"content":{"rendered":"<p>Every week another agency announces a GEO service. They\u2019re adding new dashboards, budget lines and job titles, and they all promise the same thing: visibility in AI answers.<\/p>\n<p>What I keep coming back to is a question nobody on those sales calls answers. <strong>Where do you think the AI gets its answers from?<\/strong><\/p>\n<p>Not from a separate index built just for chatbots. Generative engines pull from the same web that SEO has been making crawlable, structured and credible for twenty years. So when someone sells you GEO as a new discipline, they\u2019re selling you a reflection and asking you to ignore the object.<\/p>\n<p>You can\u2019t polish a reflection. You can only change what\u2019s standing in front of the mirror.<\/p>\n<p>So my thesis is simple: <strong>GEO is not a discipline. GEO is the answer AI generates from the downstream output of your good SEO.<\/strong> That means crawlable architecture, source credibility, entity-building and third-party trust. The AI answer is where that work shows up, not where it happens.<\/p>\n<h2>The traffic debate is real, but it\u2019s the wrong fight<\/h2>\n<p>There are two camps here, and both have a point.<\/p>\n<p><a href=\"https:\/\/www.mariehaynes.com\/geo-is-a-distraction\/\" target=\"_blank\" rel=\"noopener\">Marie Haynes<\/a> looked at 12 months of GA4 data across 100 websites and more than 145 million sessions. ChatGPT drove 0.62% of the sessions Google Organic did, and adding every other LLM only lifted that to 0.73%. ChatGPT visitors did convert better, 13.8% against 3.02%. But volume wins. Google Organic still drove 27 times more conversions than all LLMs combined. And ChatGPT referral growth stalled around August 2025, then started declining.<\/p>\n<p>Mike King argues the other side: <a href=\"https:\/\/ipullrank.com\/ai-search-is-a-branding-channel\" target=\"_blank\" rel=\"noopener\">judging AI search by clicks is like judging a Super Bowl ad by its click-through rate<\/a>.<\/p>\n<p>In fact, I like Mike\u2019s argument so much that I\u2019ll borrow his analogies, because they put AI search right where it belongs:<\/p>\n<blockquote>\n<p><strong>\u201cAn AI answer is earned media composed by a machine.\u201d<\/strong><\/p>\n<\/blockquote>\n<p>and<\/p>\n<blockquote>\n<p><strong>\u201cNobody has ever attributed a billboard.<\/strong> Nobody attributes a television spot, a sponsored podcast, a stadium naming deal or a celebrity holding a product on a red carpet.\u201d<\/p>\n<\/blockquote>\n<figure style=\"margin:40px 0\"><img src=\"https:\/\/inspace.io\/wp-content\/uploads\/2026\/11\/insight-seo-vs-geo-billboard-1024x704.jpg\" width=\"1024\" height=\"704\" alt=\"Billboard on a highway that reads \u2018This billboard generates clicks\u2019, with the word clicks crossed out\" loading=\"lazy\" decoding=\"async\" style=\"width:100%;height:auto;border-radius:14px;border:1px solid rgba(11,11,12,.10)\" title=\"\"><figcaption>\u201cNobody has ever attributed a billboard.\u201d<\/figcaption><\/figure>\n<p><strong>AI search is a brand channel<\/strong> for trust and positioning at the moment people compare options. With hundreds of millions to billions of users across ChatGPT, AI Overviews, AI Mode and Gemini, presence matters even without the click.<\/p>\n<p>At first this looks like a clicks-vs-brand argument. I don\u2019t think it is. <strong>It\u2019s a measurement argument.<\/strong> Both camps are telling you the same thing: <strong>the way we measure AI visibility right now is broken<\/strong>.<\/p>\n<h2>Prompt tracking measures the wrong thing<\/h2>\n<p>There! Finally got that off my chest. Sigh\u2026 Most GEO dashboards take rank-tracking logic and apply it to a system that doesn\u2019t rank. That\u2019s the core mistake.<\/p>\n<p>In classic search, a keyword returns roughly the same results for everyone. In generative search, the prompt is only part of the input. ChatGPT also uses memory, context, platform behaviour and location, then rewrites the actual search query before answering.<\/p>\n<p>Take one prompt: <em>\u201cWho is the best basketball player ever?\u201d<\/em><\/p>\n<ul>\n<li>User A values championships and legacy. They get Jordan.<\/li>\n<li>User B values longevity and all-round stats. They get LeBron.<\/li>\n<li>And I value mentality and being part of my basketball journey. So I would say Kobe Bryant.<\/li>\n<\/ul>\n<p>All three answers are correct, and none of them is a ranking. A prompt monitor only replays the text string. It doesn\u2019t replay the person.<\/p>\n<p>So vendors measure the probability of getting cited for a prompt:<\/p>\n<blockquote>\n<p>P(citation | prompt)<\/p>\n<\/blockquote>\n<p>But the real thing is the probability of getting cited for the intent and context of the user, on a platform, at a moment in time:<\/p>\n<blockquote>\n<p>P(citation | intent, context, platform, time)<\/p>\n<\/blockquote>\n<p>That\u2019s a different equation, so it needs a different instrument.<\/p>\n<p>And it gets worse. Google\u2019s <strong>query fan-out<\/strong> splits one question into many synthetic sub-queries to build an answer. Those sub-queries are retrieval operations, not human demand. That\u2019s exactly where some agency reporting goes wrong. Generate hundreds of machine sub-queries, show up in more of them, and you get a slide that says \u201cAI visibility up from 18% to 64%.\u201d <strong>You\u2019ve measured how the system behaves, not who you actually reached.<\/strong><\/p>\n<p>The research points the same way. <a href=\"https:\/\/arxiv.org\/abs\/2610.15033\" target=\"_blank\" rel=\"noopener\">A preprint by Mahe Chen and Nick Koudas analysed about 158,000 ChatGPT answers and 468,000 citations<\/a>. Search retrieval explained roughly 32% of the variance in citation counts. On-page \u201cGEO\u201d features explained 2.8%.<\/p>\n<p><strong>So the issue isn\u2019t formatting. It\u2019s retrieval.<\/strong> AI cites you because search found you and you fit the context, not because you dressed the page up for a robot.<\/p>\n<p>Even the platforms say this out loud. Bing Webmaster Tools states that AI citations are not rankings and not a signal of authority, and that the data is sampled.<\/p>\n<h3>How much each method actually proves<\/h3>\n<div style=\"overflow-x:auto;-webkit-overflow-scrolling:touch\">\n<table>\n<thead>\n<tr>\n<th>Method<\/th>\n<th>What it proves<\/th>\n<th>How far to trust it<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Fixed prompt monitoring<\/td>\n<td>You appeared in one sampled synthetic run<\/td>\n<td>Low: high variance, no context, open to cherry-picked prompts<\/td>\n<\/tr>\n<tr>\n<td>Context-stratified tests<\/td>\n<td>How often you appear across defined personas<\/td>\n<td>Moderate: good for diagnosis, says nothing about real reach<\/td>\n<\/tr>\n<tr>\n<td>Platform-native data (GSC, Bing WMT)<\/td>\n<td>Observed exposure from real users<\/td>\n<td>Stronger, but sampled, and citations \u2260 ranking<\/td>\n<\/tr>\n<tr>\n<td>AI referral traffic + conversions<\/td>\n<td>Real sessions and revenue<\/td>\n<td>Strong for tracked visits, blind to zero-click influence<\/td>\n<\/tr>\n<tr>\n<td>Controlled before\/after tests<\/td>\n<td>Possible incremental impact<\/td>\n<td>Depends on design; noisy around updates and seasonality<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<h2>Visibility isn\u2019t one thing. It has six states.<\/h2>\n<p>\u201cTo be or not to be.\u201d \u201cTo be visible or to be invisible.\u201d As good as that sounds, that framing needs to go.<\/p>\n<p>What you\u2019re really looking at is a chain:<\/p>\n<figure style=\"margin:40px 0\"><img src=\"https:\/\/inspace.io\/wp-content\/uploads\/2026\/11\/insight-seo-vs-geo-chain-1024x538.jpg\" width=\"1024\" height=\"538\" alt=\"Chain of six states: accessible, retrieved, cited, mentioned, recommended, selected\" loading=\"lazy\" decoding=\"async\" style=\"width:100%;height:auto;border-radius:14px;border:1px solid rgba(11,11,12,.10)\" title=\"\"><figcaption>The six states of AI visibility<\/figcaption><\/figure>\n<p><strong>Accessible \u2192 Retrieved \u2192 Cited \u2192 Mentioned \u2192 Recommended \u2192 Selected<\/strong><\/p>\n<ul>\n<li><strong>Accessible:<\/strong> crawlers can reach and parse you.<\/li>\n<li><strong>Retrieved:<\/strong> the engine fetched your page during a search or fan-out.<\/li>\n<li><strong>Cited:<\/strong> your URL shows up as a source.<\/li>\n<li><strong>Mentioned:<\/strong> your name shows up, with or without a link.<\/li>\n<li><strong>Recommended:<\/strong> the AI frames you as the right fit for this person.<\/li>\n<li><strong>Selected:<\/strong> someone (or an agent acting for them) actually clicks, books or buys.<\/li>\n<\/ul>\n<p>Most GEO reporting stops at \u201ccited\u201d and calls it a win. But being cited isn\u2019t being recommended.<\/p>\n<p>Picture someone asking: <em>\u201cWhich mountain bike is best for a 125 kg rider on asphalt and gravel?\u201d<\/em> Retailer A wrote a buying guide with weight-limit data, so it gets cited. Retailer B sells a bike built for exactly that load and terrain, so it gets recommended.<\/p>\n<p><strong>A got the citation. B got the customer.<\/strong><\/p>\n<h3>Correct exclusion is a win<\/h3>\n<p>This is where most dashboards look at the wrong thing.<\/p>\n<p>Take a university of applied sciences that only offers full-time, on-campus marketing programs. Three people ask:<\/p>\n<ul>\n<li>A school leaver who wants a four-year, full-time degree<\/li>\n<li>A professional working 40 hours who needs evening classes<\/li>\n<li>A career changer looking for an accelerated certificate<\/li>\n<\/ul>\n<p>If the AI recommends this school to the second or third person, that\u2019s not visibility. It\u2019s a bad lead, a bad experience and a dent in trust. If the AI leaves the school out for those two, the dashboard calls it \u201clost visibility.\u201d It isn\u2019t. <strong>It\u2019s the system working.<\/strong><\/p>\n<p>So the metric worth tracking isn\u2019t how often you\u2019re included. It\u2019s whether you\u2019re included in the right places:<\/p>\n<blockquote>\n<p><strong>Contextual Recommendation Accuracy<\/strong> = (appropriate recommendations + correct exclusions) \u00f7 total evaluated scenarios<\/p>\n<\/blockquote>\n<p>That changes the question from \u201care we in the answer?\u201d to \u201care we in the right answers, and out of the wrong ones?\u201d<\/p>\n<h2>Why SEO still owns the retrieval layer<\/h2>\n<p>Generative engines don\u2019t read the web from scratch. They sit on top of search infrastructure and RAG. Google\u2019s own documentation says its AI features run on standard Search systems.<\/p>\n<p>Which makes most \u201cGEO tactics\u201d a distraction:<\/p>\n<ul>\n<li><strong>llms.txt:<\/strong> sold as essential, but no major search engine requires it for generative visibility. Normal crawling still decides what gets in.<\/li>\n<li><strong>\u201cAI schema\u201d:<\/strong> doesn\u2019t exist as a requirement. Standard schema.org already carries the entity information.<\/li>\n<li><strong>Micro-chunking for vectors:<\/strong> chopping pages into disconnected snippets gives you no retrieval advantage over a well-structured page, and it makes the page worse to read.<\/li>\n<li><strong>Writing for machines:<\/strong> flattening your voice into robotic fact-lists works against systems built to reward people who are satisfied with what they read.<\/li>\n<\/ul>\n<div style=\"overflow-x:auto;-webkit-overflow-scrolling:touch\">\n<table>\n<thead>\n<tr>\n<th>Core SEO<\/th>\n<th>The \u201cGEO\u201d version<\/th>\n<th>What actually happens<\/th>\n<\/tr>\n<\/thead>\n<tbody>\n<tr>\n<td>Crawlability &amp; indexation<\/td>\n<td>llms.txt, AI-specific crawling<\/td>\n<td>Standard indexes feed RAG<\/td>\n<\/tr>\n<tr>\n<td>Structured data<\/td>\n<td>Proprietary \u201cAI schema\u201d<\/td>\n<td>Schema.org already covers it<\/td>\n<\/tr>\n<tr>\n<td>Topical authority<\/td>\n<td>Micro-chunking<\/td>\n<td>Comprehensive expert pages get retrieved<\/td>\n<\/tr>\n<tr>\n<td>Digital PR &amp; mentions<\/td>\n<td>Prompt-stuffing, LLM seeding<\/td>\n<td>Models synthesise consensus from third parties<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n<\/div>\n<p><strong>The tactic is the symptom. The infrastructure is the cause.<\/strong><\/p>\n<h3>Off-page is where recommendations get decided<\/h3>\n<p>PromptWatch data on cited asset types puts listicles and roundups at 18.7% of citations, landing pages at 15.2% and product detail pages at 14.0%.<\/p>\n<p>But here\u2019s the catch, and it matters. When a brand publishes its own \u201cTop 10\u201d and puts itself at #1, AI systems seem to see through it. <a href=\"https:\/\/promptwatch.com\/glossary\/self-promotional-listicles\" target=\"_blank\" rel=\"noopener\">Lily Ray found that in Google\u2019s AI Overviews, these self-promotional listicles were often cited as a source while the AI still recommended a different brand.<\/a> The system takes the list and leaves out the biased host.<\/p>\n<p>So it isn\u2019t enough to publish the list. You need to be on someone else\u2019s list. That means digital PR, editorial coverage, review platforms, Reddit and community threads: the unowned places where models check what people actually think.<\/p>\n<h2>The next layer: search as an agent manager<\/h2>\n<p>If you pull this wider, the shift gets bigger than citations.<\/p>\n<p>Google\u2019s leadership, Sundar Pichai and Demis Hassabis, describe search moving from a link index to an <strong>agent manager<\/strong>: something that coordinates tasks, runs workflows and completes transactions for the user.<\/p>\n<p>That changes the whole thing. It\u2019s no longer \u201cuser searches \u2192 clicks \u2192 buys.\u201d It\u2019s \u201cuser states intent and constraints \u2192 agent compares, negotiates and transacts.\u201d<\/p>\n<p>Four layers to watch:<\/p>\n<ol>\n<li><strong>Google\u2019s Universal Cart.<\/strong> Multi-retailer checkout inside Google Wallet, tracking stock, price changes and promotions. The move here: treat your Merchant Center feed as a product, not a chore. Real-time accuracy, clean pricing markup, nothing stale.<\/li>\n<li><strong>Universal Commerce Protocol (UCP).<\/strong> An open standard co-developed by Google and Shopify that lets agents discover products, compare them and check out while you stay the merchant of record. The move: add agentic endpoints like <code>\/.well-known\/ucp<\/code> to your technical audits.<\/li>\n<li><strong>WebMCP.<\/strong> Exposes a site\u2019s interactive tools (calculators, schedulers, troubleshooters) directly to agents, so they call your functions instead of scraping your text. The move: list your decision tools and make them callable.<\/li>\n<li><strong>Smart glasses and multimodal search.<\/strong> Search moves off the screen and into the room. <em>\u201cLook at this appliance, diagnose the noise, order the part.\u201d<\/em> The move: map real-world problem \u2192 solution scenarios to structured data and visual assets.<\/li>\n<\/ol>\n<p>And this is the part most organizations will trip over. Technical SEO, feed management, development and PR usually sit in separate rooms. In an agentic world they\u2019re one pipeline. The agent can\u2019t check out your product if dev hasn\u2019t exposed the endpoint, or the feed shows the wrong price, or PR never built the trust that got you shortlisted.<\/p>\n<p><strong>This isn\u2019t a channel problem anymore. It\u2019s an org-chart problem.<\/strong><\/p>\n<h2>Who should actually invest, and who should fix the basement first<\/h2>\n<p>Not every business needs to chase this now.<\/p>\n<p><strong>Invest now if:<\/strong><\/p>\n<ul>\n<li>Your buyer journey is complex and high-consideration: education, B2B software, finance, healthcare, technical equipment.<\/li>\n<li>A wrong AI answer about your specs creates real operational, financial or legal damage.<\/li>\n<li>Your SEO foundation, site architecture and conversion tracking already work.<\/li>\n<\/ul>\n<p><strong>Fix the foundation first if:<\/strong><\/p>\n<ul>\n<li>You still have unresolved technical SEO, indexation or mobile issues.<\/li>\n<li>You\u2019re not tracking conversions. In Haynes\u2019 dataset, 45 of the 100 sites weren\u2019t, which means they couldn\u2019t even measure the thing they want to improve.<\/li>\n<li>You sell low-consideration impulse products that nobody researches through an AI.<\/li>\n<\/ul>\n<h3>The sequence<\/h3>\n<figure style=\"margin:40px 0\"><img src=\"https:\/\/inspace.io\/wp-content\/uploads\/2026\/11\/insight-seo-vs-geo-sequence-1024x704.jpg\" width=\"1024\" height=\"704\" alt=\"Five steps: fix core SEO and data, build a bottom-of-funnel asset engine, move link-building into digital PR and UGC, repurpose high-converting queries across channels, pilot agentic endpoints\" loading=\"lazy\" decoding=\"async\" style=\"width:100%;height:auto;border-radius:14px;border:1px solid rgba(11,11,12,.10)\" title=\"\"><figcaption>The sequence: five steps, foundation first<\/figcaption><\/figure>\n<ol>\n<li><strong>Fix core SEO and data.<\/strong> Crawl errors, entity schema, GA4 conversions. Without measurement, everything after this is guesswork.<\/li>\n<li><strong>Build a bottom-of-funnel asset engine.<\/strong> Structured landing pages and product pages. Together they made up almost 30% of citations in the PromptWatch data.<\/li>\n<li><strong>Move link-building into digital PR and UGC.<\/strong> Get onto third-party roundups, editorial pieces and review hubs. Look after real sentiment where models read it.<\/li>\n<li><strong>Repurpose high-converting queries across channels.<\/strong> Short video, LinkedIn, X. Not for reach theatre, but to build entity density across the web.<\/li>\n<li><strong>Pilot agentic endpoints.<\/strong> Run small quarterly experiments: Merchant Center for Universal Cart eligibility, WebMCP on one interactive tool.<\/li>\n<\/ol>\n<h2>So what actually changes<\/h2>\n<p>Nobody tricks an LLM into recommending them. Not with llms.txt, not with \u201cAI schema,\u201d not with chunked robot-copy.<\/p>\n<p>Compare AI visibility with a billboard: it\u2019s more branding than performance, and it doesn\u2019t primarily generate traffic.<\/p>\n<p>What changes is the question you ask. Not <em>\u201chow do we show up in AI?\u201d<\/em> but <em>\u201cis our brand true, findable and trusted enough that any system, human or machine, picks us for the right person and leaves us out for the wrong one?\u201d<\/em><\/p>\n<p>That\u2019s not a new discipline. It\u2019s SEO, finally measured by the right outcome. We just call it GEO.<\/p>\n<h3>Sources<\/h3>\n<ul>\n<li><a href=\"https:\/\/searchengineland.com\/chatgpt-sends-less-than-googles-traffic-497705\" target=\"_blank\" rel=\"noopener\">ChatGPT sends less than 1% of Google\u2019s traffic (Search Engine Land)<\/a><\/li>\n<li><a href=\"https:\/\/martech.org\/chatgpt-sends-less-than-1-of-googles-traffic\/\" target=\"_blank\" rel=\"noopener\">ChatGPT sends less than 1% of Google\u2019s traffic (MarTech)<\/a><\/li>\n<li><a href=\"https:\/\/www.info-lux.com\/trafic-chatgpt-contre-google-etude-100-sites\/intelligence-artificielle\/\" target=\"_blank\" rel=\"noopener\">Haynes study coverage, incl. 45\/100 sites not tracking conversions (info-lux)<\/a><\/li>\n<li><a href=\"https:\/\/www.shopify.com\/be-fr\/ucp\" target=\"_blank\" rel=\"noopener\">Universal Commerce Protocol (Shopify)<\/a><\/li>\n<li><a href=\"https:\/\/ahrefs.com\/blog\/self-promotional-content-ai-seo-experiment\/\" target=\"_blank\" rel=\"noopener\">Self-Promotional Content Works, Until It Backfires (Ahrefs)<\/a><\/li>\n<\/ul>\n","protected":false},"excerpt":{"rendered":"<p>Every week another agency launches a GEO service. But AI answers come from the same web that SEO has built for twenty years. GEO is the output of good SEO, not a new discipline.<\/p>\n","protected":false},"featured_media":14247,"template":"","insight_category":[549,548],"class_list":["post-14257","insight","type-insight","status-publish","has-post-thumbnail","hentry","insight_category-ai-search","insight_category-geo"],"acf":[],"_links":{"self":[{"href":"https:\/\/inspace.io\/en-nl\/wp-json\/wp\/v2\/insight\/14257","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/inspace.io\/en-nl\/wp-json\/wp\/v2\/insight"}],"about":[{"href":"https:\/\/inspace.io\/en-nl\/wp-json\/wp\/v2\/types\/insight"}],"version-history":[{"count":1,"href":"https:\/\/inspace.io\/en-nl\/wp-json\/wp\/v2\/insight\/14257\/revisions"}],"predecessor-version":[{"id":14258,"href":"https:\/\/inspace.io\/en-nl\/wp-json\/wp\/v2\/insight\/14257\/revisions\/14258"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/inspace.io\/en-nl\/wp-json\/wp\/v2\/media\/14247"}],"wp:attachment":[{"href":"https:\/\/inspace.io\/en-nl\/wp-json\/wp\/v2\/media?parent=14257"}],"wp:term":[{"taxonomy":"insight_category","embeddable":true,"href":"https:\/\/inspace.io\/en-nl\/wp-json\/wp\/v2\/insight_category?post=14257"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}