[{"data":1,"prerenderedAt":1019},["ShallowReactive",2],{"article-alternates":3,"article-\u002Fen\u002Fai\u002Fllm-citation-measurement-new-seo-metric":12},{"i18nKey":4,"paths":5},"ai-002-2026-05",{"de":6,"en":7,"es":8,"fr":9,"it":10,"ru":11},"\u002Fde\u002Fai\u002Fllm-zitierungsmetriken-seo","\u002Fen\u002Fai\u002Fllm-citation-measurement-new-seo-metrics","\u002Fes\u002Fai\u002Fmedicion-de-citas-llm","\u002Ffr\u002Fai\u002Fmesure-des-citations-llm","\u002Fit\u002Fai\u002Fmisurazione-citazioni-llm","\u002Fru\u002Fai\u002Fllm-citation-measurement-new-seo-metrics",{"_path":13,"_dir":14,"_draft":15,"_partial":15,"_locale":16,"title":17,"description":18,"publishedAt":19,"modifiedAt":19,"category":14,"i18nKey":4,"tags":20,"readingTime":26,"author":27,"body":28,"_type":1013,"_id":1014,"_source":1015,"_file":1016,"_stem":1017,"_extension":1018},"\u002Fen\u002Fai\u002Fllm-citation-measurement-new-seo-metric","ai",false,"","LLM Citation Measurement — Your New SEO Metric Set","How do you measure your brand's citation rate on Perplexity, ChatGPT, Gemini? Citation tracking is the new generation SEO metric framework.","2026-05-30",[21,22,23,24,25],"llm-citation","geo","seo-metrics","ai-attribution","brand-visibility",7,"Roibase",{"type":29,"children":30,"toc":1004},"root",[31,39,46,51,56,61,67,72,83,110,128,582,587,593,598,608,629,639,649,659,764,769,775,787,797,837,847,857,863,868,886,896,906,933,939,951,963,975,987,993,998],{"type":32,"tag":33,"props":34,"children":35},"element","p",{},[36],{"type":37,"value":38},"text","Your organic traffic is declining, CTR is stagnant, but ChatGPT cites your brand 4,000 times daily. You don't know this because it doesn't show up in Google Analytics. LLM citation tracking — this is the new SEO metric framework of the generative AI era. Large language models like Perplexity, ChatGPT, and Gemini are now search's new interface. Users get answers directly without visiting your site. But if the model cites you as a source, your brand becomes part of that answer. If you don't measure this citation rate, you're essentially flying blind on visibility.",{"type":32,"tag":40,"props":41,"children":43},"h2",{"id":42},"what-citation-is-and-why-it-matters-now",[44],{"type":37,"value":45},"What Citation Is and Why It Matters Now",{"type":32,"tag":33,"props":47,"children":48},{},[49],{"type":37,"value":50},"LLM citation occurs when a language model generates an answer while crediting your brand, content, or website as a source. In classical SEO you counted backlinks; now the question is \"how many times did the model mention me.\" When ChatGPT answers a technical question with \"Roibase's server-side measurement architecture,\" that's a citation. When Perplexity shows an inline source link to your content, that builds your brand equity.",{"type":32,"tag":33,"props":52,"children":53},{},[54],{"type":37,"value":55},"Why is it critical? Because search behavior is shifting. Statcounter Q1 2026 data shows direct questions to AI chat tools hit 18% (up from 6% in Q1 2024). Google's AI Overviews now appear in 40% of search results. Instead of clicking through 10 blue links to answer \"how do I,\" users read one paragraph summary. Being included as a source in that summary may be more valuable than traffic — it's a trust signal.",{"type":32,"tag":33,"props":57,"children":58},{},[59],{"type":37,"value":60},"Classical SEO metrics (impressions, CTR, position) don't apply in the LLM context. A user asks ChatGPT \"best CDP for headless commerce,\" the model recommends three brands. Did yours get mentioned? In which prompts? Without this data, your visibility analysis is incomplete.",{"type":32,"tag":40,"props":62,"children":64},{"id":63},"how-to-set-up-citation-tracking",[65],{"type":37,"value":66},"How to Set Up Citation Tracking",{"type":32,"tag":33,"props":68,"children":69},{},[70],{"type":37,"value":71},"Measuring LLM citations requires API-based probing. Manual testing doesn't scale — you can't manually check whether your brand appears across 50 keyword variations on three different models. Automation is essential. Here's the layered approach:",{"type":32,"tag":33,"props":73,"children":74},{},[75,81],{"type":32,"tag":76,"props":77,"children":78},"strong",{},[79],{"type":37,"value":80},"Layer 1: Build your keyword pool.",{"type":37,"value":82}," Take keywords you already rank for from Google Search Console. But reformat them for LLM queries. Instead of \"roibase first party data,\" ask \"how do I set up first-party data architecture?\" — because users ask chatbots questions, not search engines. Convert your 100 keywords into 100 natural questions.",{"type":32,"tag":33,"props":84,"children":85},{},[86,91,93,100,102,108],{"type":32,"tag":76,"props":87,"children":88},{},[89],{"type":37,"value":90},"Layer 2: API probe structure.",{"type":37,"value":92}," Send each question to ChatGPT API, Claude API, and Gemini API. Capture the response. Use regex or embedding similarity to check if your brand name, site URL, or product name appears in the answer. Perplexity API provides inline citations — check if your domain appears in the ",{"type":32,"tag":94,"props":95,"children":97},"code",{"className":96},[],[98],{"type":37,"value":99},"sources",{"type":37,"value":101}," array. ChatGPT doesn't footnote sources in the response, but if web search is enabled, check the ",{"type":32,"tag":94,"props":103,"children":105},{"className":104},[],[106],{"type":37,"value":107},"search_results",{"type":37,"value":109}," metadata for your domain.",{"type":32,"tag":33,"props":111,"children":112},{},[113,118,120,126],{"type":32,"tag":76,"props":114,"children":115},{},[116],{"type":37,"value":117},"Layer 3: Log aggregation.",{"type":37,"value":119}," Write every probe result to a time-series database (InfluxDB, TimescaleDB, or BigQuery). Schema: ",{"type":32,"tag":94,"props":121,"children":123},{"className":122},[],[124],{"type":37,"value":125},"{timestamp, model, keyword, cited: boolean, citation_type, position, context_snippet}",{"type":37,"value":127},". Without this data, you can't see trends.",{"type":32,"tag":129,"props":130,"children":134},"pre",{"className":131,"code":132,"language":133,"meta":16,"style":16},"language-python shiki shiki-themes github-dark","# Simplified probe example (ChatGPT API)\nimport openai, re\n\ndef check_citation(keyword_question, brand_terms):\n    response = openai.ChatCompletion.create(\n        model=\"gpt-4o\",\n        messages=[{\"role\": \"user\", \"content\": keyword_question}]\n    )\n    answer = response.choices[0].message.content\n    \n    for term in brand_terms:\n        if re.search(term, answer, re.IGNORECASE):\n            return {\"cited\": True, \"term\": term, \"snippet\": answer[:200]}\n    \n    return {\"cited\": False}\n\n# Usage\nresult = check_citation(\n    \"How do I set up first-party data architecture?\",\n    [\"Roibase\", \"roibase.com.tr\"]\n)\nprint(result)\n","python",[135],{"type":32,"tag":94,"props":136,"children":137},{"__ignoreMap":16},[138,150,166,176,196,215,240,287,296,325,334,358,382,444,452,483,491,500,518,531,559,568],{"type":32,"tag":139,"props":140,"children":143},"span",{"class":141,"line":142},"line",1,[144],{"type":32,"tag":139,"props":145,"children":147},{"style":146},"--shiki-default:#6A737D",[148],{"type":37,"value":149},"# Simplified probe example (ChatGPT API)\n",{"type":32,"tag":139,"props":151,"children":153},{"class":141,"line":152},2,[154,160],{"type":32,"tag":139,"props":155,"children":157},{"style":156},"--shiki-default:#F97583",[158],{"type":37,"value":159},"import",{"type":32,"tag":139,"props":161,"children":163},{"style":162},"--shiki-default:#E1E4E8",[164],{"type":37,"value":165}," openai, re\n",{"type":32,"tag":139,"props":167,"children":169},{"class":141,"line":168},3,[170],{"type":32,"tag":139,"props":171,"children":173},{"emptyLinePlaceholder":172},true,[174],{"type":37,"value":175},"\n",{"type":32,"tag":139,"props":177,"children":179},{"class":141,"line":178},4,[180,185,191],{"type":32,"tag":139,"props":181,"children":182},{"style":156},[183],{"type":37,"value":184},"def",{"type":32,"tag":139,"props":186,"children":188},{"style":187},"--shiki-default:#B392F0",[189],{"type":37,"value":190}," check_citation",{"type":32,"tag":139,"props":192,"children":193},{"style":162},[194],{"type":37,"value":195},"(keyword_question, brand_terms):\n",{"type":32,"tag":139,"props":197,"children":199},{"class":141,"line":198},5,[200,205,210],{"type":32,"tag":139,"props":201,"children":202},{"style":162},[203],{"type":37,"value":204},"    response ",{"type":32,"tag":139,"props":206,"children":207},{"style":156},[208],{"type":37,"value":209},"=",{"type":32,"tag":139,"props":211,"children":212},{"style":162},[213],{"type":37,"value":214}," openai.ChatCompletion.create(\n",{"type":32,"tag":139,"props":216,"children":218},{"class":141,"line":217},6,[219,225,229,235],{"type":32,"tag":139,"props":220,"children":222},{"style":221},"--shiki-default:#FFAB70",[223],{"type":37,"value":224},"        model",{"type":32,"tag":139,"props":226,"children":227},{"style":156},[228],{"type":37,"value":209},{"type":32,"tag":139,"props":230,"children":232},{"style":231},"--shiki-default:#9ECBFF",[233],{"type":37,"value":234},"\"gpt-4o\"",{"type":32,"tag":139,"props":236,"children":237},{"style":162},[238],{"type":37,"value":239},",\n",{"type":32,"tag":139,"props":241,"children":242},{"class":141,"line":26},[243,248,252,257,262,267,272,277,282],{"type":32,"tag":139,"props":244,"children":245},{"style":221},[246],{"type":37,"value":247},"        messages",{"type":32,"tag":139,"props":249,"children":250},{"style":156},[251],{"type":37,"value":209},{"type":32,"tag":139,"props":253,"children":254},{"style":162},[255],{"type":37,"value":256},"[{",{"type":32,"tag":139,"props":258,"children":259},{"style":231},[260],{"type":37,"value":261},"\"role\"",{"type":32,"tag":139,"props":263,"children":264},{"style":162},[265],{"type":37,"value":266},": ",{"type":32,"tag":139,"props":268,"children":269},{"style":231},[270],{"type":37,"value":271},"\"user\"",{"type":32,"tag":139,"props":273,"children":274},{"style":162},[275],{"type":37,"value":276},", ",{"type":32,"tag":139,"props":278,"children":279},{"style":231},[280],{"type":37,"value":281},"\"content\"",{"type":32,"tag":139,"props":283,"children":284},{"style":162},[285],{"type":37,"value":286},": keyword_question}]\n",{"type":32,"tag":139,"props":288,"children":290},{"class":141,"line":289},8,[291],{"type":32,"tag":139,"props":292,"children":293},{"style":162},[294],{"type":37,"value":295},"    )\n",{"type":32,"tag":139,"props":297,"children":299},{"class":141,"line":298},9,[300,305,309,314,320],{"type":32,"tag":139,"props":301,"children":302},{"style":162},[303],{"type":37,"value":304},"    answer ",{"type":32,"tag":139,"props":306,"children":307},{"style":156},[308],{"type":37,"value":209},{"type":32,"tag":139,"props":310,"children":311},{"style":162},[312],{"type":37,"value":313}," response.choices[",{"type":32,"tag":139,"props":315,"children":317},{"style":316},"--shiki-default:#79B8FF",[318],{"type":37,"value":319},"0",{"type":32,"tag":139,"props":321,"children":322},{"style":162},[323],{"type":37,"value":324},"].message.content\n",{"type":32,"tag":139,"props":326,"children":328},{"class":141,"line":327},10,[329],{"type":32,"tag":139,"props":330,"children":331},{"style":162},[332],{"type":37,"value":333},"    \n",{"type":32,"tag":139,"props":335,"children":337},{"class":141,"line":336},11,[338,343,348,353],{"type":32,"tag":139,"props":339,"children":340},{"style":156},[341],{"type":37,"value":342},"    for",{"type":32,"tag":139,"props":344,"children":345},{"style":162},[346],{"type":37,"value":347}," term ",{"type":32,"tag":139,"props":349,"children":350},{"style":156},[351],{"type":37,"value":352},"in",{"type":32,"tag":139,"props":354,"children":355},{"style":162},[356],{"type":37,"value":357}," brand_terms:\n",{"type":32,"tag":139,"props":359,"children":361},{"class":141,"line":360},12,[362,367,372,377],{"type":32,"tag":139,"props":363,"children":364},{"style":156},[365],{"type":37,"value":366},"        if",{"type":32,"tag":139,"props":368,"children":369},{"style":162},[370],{"type":37,"value":371}," re.search(term, answer, re.",{"type":32,"tag":139,"props":373,"children":374},{"style":316},[375],{"type":37,"value":376},"IGNORECASE",{"type":32,"tag":139,"props":378,"children":379},{"style":162},[380],{"type":37,"value":381},"):\n",{"type":32,"tag":139,"props":383,"children":385},{"class":141,"line":384},13,[386,391,396,401,405,410,414,419,424,429,434,439],{"type":32,"tag":139,"props":387,"children":388},{"style":156},[389],{"type":37,"value":390},"            return",{"type":32,"tag":139,"props":392,"children":393},{"style":162},[394],{"type":37,"value":395}," {",{"type":32,"tag":139,"props":397,"children":398},{"style":231},[399],{"type":37,"value":400},"\"cited\"",{"type":32,"tag":139,"props":402,"children":403},{"style":162},[404],{"type":37,"value":266},{"type":32,"tag":139,"props":406,"children":407},{"style":316},[408],{"type":37,"value":409},"True",{"type":32,"tag":139,"props":411,"children":412},{"style":162},[413],{"type":37,"value":276},{"type":32,"tag":139,"props":415,"children":416},{"style":231},[417],{"type":37,"value":418},"\"term\"",{"type":32,"tag":139,"props":420,"children":421},{"style":162},[422],{"type":37,"value":423},": term, ",{"type":32,"tag":139,"props":425,"children":426},{"style":231},[427],{"type":37,"value":428},"\"snippet\"",{"type":32,"tag":139,"props":430,"children":431},{"style":162},[432],{"type":37,"value":433},": answer[:",{"type":32,"tag":139,"props":435,"children":436},{"style":316},[437],{"type":37,"value":438},"200",{"type":32,"tag":139,"props":440,"children":441},{"style":162},[442],{"type":37,"value":443},"]}\n",{"type":32,"tag":139,"props":445,"children":447},{"class":141,"line":446},14,[448],{"type":32,"tag":139,"props":449,"children":450},{"style":162},[451],{"type":37,"value":333},{"type":32,"tag":139,"props":453,"children":455},{"class":141,"line":454},15,[456,461,465,469,473,478],{"type":32,"tag":139,"props":457,"children":458},{"style":156},[459],{"type":37,"value":460},"    return",{"type":32,"tag":139,"props":462,"children":463},{"style":162},[464],{"type":37,"value":395},{"type":32,"tag":139,"props":466,"children":467},{"style":231},[468],{"type":37,"value":400},{"type":32,"tag":139,"props":470,"children":471},{"style":162},[472],{"type":37,"value":266},{"type":32,"tag":139,"props":474,"children":475},{"style":316},[476],{"type":37,"value":477},"False",{"type":32,"tag":139,"props":479,"children":480},{"style":162},[481],{"type":37,"value":482},"}\n",{"type":32,"tag":139,"props":484,"children":486},{"class":141,"line":485},16,[487],{"type":32,"tag":139,"props":488,"children":489},{"emptyLinePlaceholder":172},[490],{"type":37,"value":175},{"type":32,"tag":139,"props":492,"children":494},{"class":141,"line":493},17,[495],{"type":32,"tag":139,"props":496,"children":497},{"style":146},[498],{"type":37,"value":499},"# Usage\n",{"type":32,"tag":139,"props":501,"children":503},{"class":141,"line":502},18,[504,509,513],{"type":32,"tag":139,"props":505,"children":506},{"style":162},[507],{"type":37,"value":508},"result ",{"type":32,"tag":139,"props":510,"children":511},{"style":156},[512],{"type":37,"value":209},{"type":32,"tag":139,"props":514,"children":515},{"style":162},[516],{"type":37,"value":517}," check_citation(\n",{"type":32,"tag":139,"props":519,"children":521},{"class":141,"line":520},19,[522,527],{"type":32,"tag":139,"props":523,"children":524},{"style":231},[525],{"type":37,"value":526},"    \"How do I set up first-party data architecture?\"",{"type":32,"tag":139,"props":528,"children":529},{"style":162},[530],{"type":37,"value":239},{"type":32,"tag":139,"props":532,"children":534},{"class":141,"line":533},20,[535,540,545,549,554],{"type":32,"tag":139,"props":536,"children":537},{"style":162},[538],{"type":37,"value":539},"    [",{"type":32,"tag":139,"props":541,"children":542},{"style":231},[543],{"type":37,"value":544},"\"Roibase\"",{"type":32,"tag":139,"props":546,"children":547},{"style":162},[548],{"type":37,"value":276},{"type":32,"tag":139,"props":550,"children":551},{"style":231},[552],{"type":37,"value":553},"\"roibase.com.tr\"",{"type":32,"tag":139,"props":555,"children":556},{"style":162},[557],{"type":37,"value":558},"]\n",{"type":32,"tag":139,"props":560,"children":562},{"class":141,"line":561},21,[563],{"type":32,"tag":139,"props":564,"children":565},{"style":162},[566],{"type":37,"value":567},")\n",{"type":32,"tag":139,"props":569,"children":571},{"class":141,"line":570},22,[572,577],{"type":32,"tag":139,"props":573,"children":574},{"style":316},[575],{"type":37,"value":576},"print",{"type":32,"tag":139,"props":578,"children":579},{"style":162},[580],{"type":37,"value":581},"(result)\n",{"type":32,"tag":33,"props":583,"children":584},{},[585],{"type":37,"value":586},"Real implementations need batch processing — use async queues instead of serial requests. Add rate limit management, retry logic, and cost tracking. Each API call costs 0.01–0.03$ (depending on model and tokens); expect ~$150\u002Fmonth for probes (500 keywords × 3 models × 10 tests\u002Fmonth).",{"type":32,"tag":40,"props":588,"children":590},{"id":589},"define-your-metric-set",[591],{"type":37,"value":592},"Define Your Metric Set",{"type":32,"tag":33,"props":594,"children":595},{},[596],{"type":37,"value":597},"What numbers do you track in citation monitoring? Instead of \"position\" and \"CTR\" from your classical SEO dashboard, focus on these:",{"type":32,"tag":33,"props":599,"children":600},{},[601,606],{"type":32,"tag":76,"props":602,"children":603},{},[604],{"type":37,"value":605},"Citation Rate:",{"type":37,"value":607}," The percentage of tested keywords where your brand gets cited. Test 100 keywords, get mentioned in 18 → 18% citation rate. Think \"share of voice,\" but in the LLM space.",{"type":32,"tag":33,"props":609,"children":610},{},[611,616,618,627],{"type":32,"tag":76,"props":612,"children":613},{},[614],{"type":37,"value":615},"Model-Specific Share:",{"type":37,"value":617}," You might see 22% citation rate on ChatGPT, 14% on Claude, 9% on Gemini. Models differ because of training data, retrieval mechanisms, and prompt tuning. Knowing which model favors you shapes your ",{"type":32,"tag":619,"props":620,"children":624},"a",{"href":621,"rel":622},"https:\u002F\u002Fwww.roibase.com.tr\u002Fen\u002Fgeo",[623],"nofollow",[625],{"type":37,"value":626},"Generative Engine Optimization",{"type":37,"value":628}," strategy.",{"type":32,"tag":33,"props":630,"children":631},{},[632,637],{"type":32,"tag":76,"props":633,"children":634},{},[635],{"type":37,"value":636},"Citation Position:",{"type":37,"value":638}," Where does your brand appear in the model's answer? In the \"top 3 recommendations\" or in \"other options\"? Position matters — users focus on the first 2–3 sources.",{"type":32,"tag":33,"props":640,"children":641},{},[642,647],{"type":32,"tag":76,"props":643,"children":644},{},[645],{"type":37,"value":646},"Context Quality Score:",{"type":37,"value":648}," Your brand was mentioned, but how? \"Agencies like Roibase\" versus \"Roibase's server-side GTM solution\" carry different equity. Run semantic analysis on snippets (positive\u002Fneutral\u002Fnegative + specificity grade).",{"type":32,"tag":33,"props":650,"children":651},{},[652,657],{"type":32,"tag":76,"props":653,"children":654},{},[655],{"type":37,"value":656},"Competitive Displacement:",{"type":37,"value":658}," What share do competitors hold for the same keywords? If \"first-party data CDP\" gets answers citing Segment, mParticle, and Roibase, you're in a three-way split. Is your share growing over time?",{"type":32,"tag":660,"props":661,"children":662},"table",{},[663,687],{"type":32,"tag":664,"props":665,"children":666},"thead",{},[667],{"type":32,"tag":668,"props":669,"children":670},"tr",{},[671,677,682],{"type":32,"tag":672,"props":673,"children":674},"th",{},[675],{"type":37,"value":676},"Metric",{"type":32,"tag":672,"props":678,"children":679},{},[680],{"type":37,"value":681},"Definition",{"type":32,"tag":672,"props":683,"children":684},{},[685],{"type":37,"value":686},"Target",{"type":32,"tag":688,"props":689,"children":690},"tbody",{},[691,710,728,746],{"type":32,"tag":668,"props":692,"children":693},{},[694,700,705],{"type":32,"tag":695,"props":696,"children":697},"td",{},[698],{"type":37,"value":699},"Citation Rate",{"type":32,"tag":695,"props":701,"children":702},{},[703],{"type":37,"value":704},"% of keywords earning citations",{"type":32,"tag":695,"props":706,"children":707},{},[708],{"type":37,"value":709},">15% (vs. category leader)",{"type":32,"tag":668,"props":711,"children":712},{},[713,718,723],{"type":32,"tag":695,"props":714,"children":715},{},[716],{"type":37,"value":717},"First-Position Rate",{"type":32,"tag":695,"props":719,"children":720},{},[721],{"type":37,"value":722},"% of mentions in top position",{"type":32,"tag":695,"props":724,"children":725},{},[726],{"type":37,"value":727},">5%",{"type":32,"tag":668,"props":729,"children":730},{},[731,736,741],{"type":32,"tag":695,"props":732,"children":733},{},[734],{"type":37,"value":735},"Context Positivity",{"type":32,"tag":695,"props":737,"children":738},{},[739],{"type":37,"value":740},"% of positive context snippets",{"type":32,"tag":695,"props":742,"children":743},{},[744],{"type":37,"value":745},">80%",{"type":32,"tag":668,"props":747,"children":748},{},[749,754,759],{"type":32,"tag":695,"props":750,"children":751},{},[752],{"type":37,"value":753},"Competitive Share",{"type":32,"tag":695,"props":755,"children":756},{},[757],{"type":37,"value":758},"Citation share vs. rivals",{"type":32,"tag":695,"props":760,"children":761},{},[762],{"type":37,"value":763},"Top 3",{"type":32,"tag":33,"props":765,"children":766},{},[767],{"type":37,"value":768},"Add these to your weekly dashboard. Trend graph: X-axis is time, Y-axis is citation rate. You should see citation rate rise 2–4 weeks after publishing new content (models have indexing lag).",{"type":32,"tag":40,"props":770,"children":772},{"id":771},"optimize-your-content-strategy-around-citations",[773],{"type":37,"value":774},"Optimize Your Content Strategy Around Citations",{"type":32,"tag":33,"props":776,"children":777},{},[778,780,785],{"type":37,"value":779},"If citation rate drops, what do you do? The classical SEO approach of \"build more backlinks\" doesn't work. LLMs don't count backlinks directly. Instead, focus on: ",{"type":32,"tag":76,"props":781,"children":782},{},[783],{"type":37,"value":784},"content depth, structured data, and authoritative signals",{"type":37,"value":786},".",{"type":32,"tag":33,"props":788,"children":789},{},[790,795],{"type":32,"tag":76,"props":791,"children":792},{},[793],{"type":37,"value":794},"Depth:",{"type":37,"value":796}," LLMs don't skip shallow content, but they respond to \"is this source thorough?\" signals. Write 2,000-word technical guides instead of 500-word blogs. Add code samples, tables, step-by-step instructions. The model's retrieval processes \"this source is actionable.\"",{"type":32,"tag":33,"props":798,"children":799},{},[800,805,807,813,814,820,822,828,830,835],{"type":32,"tag":76,"props":801,"children":802},{},[803],{"type":37,"value":804},"Structured Data:",{"type":37,"value":806}," Schema.org markup helps LLMs parse your content. Add ",{"type":32,"tag":94,"props":808,"children":810},{"className":809},[],[811],{"type":37,"value":812},"Article",{"type":37,"value":276},{"type":32,"tag":94,"props":815,"children":817},{"className":816},[],[818],{"type":37,"value":819},"HowTo",{"type":37,"value":821},", and ",{"type":32,"tag":94,"props":823,"children":825},{"className":824},[],[826],{"type":37,"value":827},"FAQPage",{"type":37,"value":829}," schemas. ",{"type":32,"tag":94,"props":831,"children":833},{"className":832},[],[834],{"type":37,"value":827},{"type":37,"value":836}," especially — models extract Q&A pairs directly.",{"type":32,"tag":33,"props":838,"children":839},{},[840,845],{"type":32,"tag":76,"props":841,"children":842},{},[843],{"type":37,"value":844},"Authoritativeness:",{"type":37,"value":846}," Author bios, organization info, publication date matter. Models can tell \"this was written in 2023 — it's outdated.\" Fresh content gets a boost. Update old articles with new publication dates.",{"type":32,"tag":33,"props":848,"children":849},{},[850,855],{"type":32,"tag":76,"props":851,"children":852},{},[853],{"type":37,"value":854},"Tradeoff:",{"type":37,"value":856}," Optimizing for citations doesn't mean abandoning traffic — it's a priority shift. For example: the generic keyword \"Shopify plugins\" drives traffic but low LLM citations (models generate their own lists). The specific phrase \"server-side Shopify checkout tracking\" drives less traffic but higher citations (fewer sources exist, yours is deep). Balance both — allocate 60% effort to traffic keywords, 40% to citation-favorable ones.",{"type":32,"tag":40,"props":858,"children":860},{"id":859},"connect-citation-data-to-your-attribution-pipeline",[861],{"type":37,"value":862},"Connect Citation Data to Your Attribution Pipeline",{"type":32,"tag":33,"props":864,"children":865},{},[866],{"type":37,"value":867},"Don't isolate citation tracking. Integrate it with classical marketing attribution. A user sees your brand on ChatGPT, searches for you on Google two days later, and converts. If you don't connect these events, you miss LLM's contribution.",{"type":32,"tag":33,"props":869,"children":870},{},[871,876,878,884],{"type":32,"tag":76,"props":872,"children":873},{},[874],{"type":37,"value":875},"UTM tagging:",{"type":37,"value":877}," If Perplexity provides an inline link, tag it (",{"type":32,"tag":94,"props":879,"children":881},{"className":880},[],[882],{"type":37,"value":883},"utm_source=perplexity&utm_medium=citation",{"type":37,"value":885},"). You'll see Perplexity traffic in Google Analytics. But ChatGPT only mentions your brand name—no direct attribution link.",{"type":32,"tag":33,"props":887,"children":888},{},[889,894],{"type":32,"tag":76,"props":890,"children":891},{},[892],{"type":37,"value":893},"Brand search lift:",{"type":37,"value":895}," Does brand search volume grow when citation rate rises? Monitor branded keywords in Google Trends or Search Console. If your ChatGPT citation rate hits 25% and stays there for three months, expect +15% brand search growth. Not direct attribution, but a strong signal.",{"type":32,"tag":33,"props":897,"children":898},{},[899,904],{"type":32,"tag":76,"props":900,"children":901},{},[902],{"type":37,"value":903},"Survey attribution:",{"type":37,"value":905}," Add \"AI chatbot (ChatGPT, Perplexity, etc.)\" as an option in your \"how did you hear about us?\" survey. Small sample but directional.",{"type":32,"tag":33,"props":907,"children":908},{},[909,914,916,922,924,931],{"type":32,"tag":76,"props":910,"children":911},{},[912],{"type":37,"value":913},"First-party event tracking:",{"type":37,"value":915}," When a user lands on your site with no referrer but the landing page targets AI-related keywords (e.g., ",{"type":32,"tag":94,"props":917,"children":919},{"className":918},[],[920],{"type":37,"value":921},"\u002Fblog\u002Fllm-citation",{"type":37,"value":923},"), that's an indirect signal. Combine these signals in your CDP using ",{"type":32,"tag":619,"props":925,"children":928},{"href":926,"rel":927},"https:\u002F\u002Fwww.roibase.com.tr\u002Fen\u002Ffirstparty",[623],[929],{"type":37,"value":930},"First-Party Data & Measurement Architecture",{"type":37,"value":932}," to build an \"AI exposure\" segment in your customer journey.",{"type":32,"tag":40,"props":934,"children":936},{"id":935},"risks-and-blind-spots",[937],{"type":37,"value":938},"Risks and Blind Spots",{"type":32,"tag":33,"props":940,"children":941},{},[942,944,949],{"type":37,"value":943},"What are the limits of LLM citation tracking? First: ",{"type":32,"tag":76,"props":945,"children":946},{},[947],{"type":37,"value":948},"sampling bias",{"type":37,"value":950},". You test 500 keywords, but real users ask 50,000 different questions. Your test set might not be representative. Solution: pull keyword pools from Search Console and convert them into prompts — this proxies real demand.",{"type":32,"tag":33,"props":952,"children":953},{},[954,956,961],{"type":37,"value":955},"Second: ",{"type":32,"tag":76,"props":957,"children":958},{},[959],{"type":37,"value":960},"model update churn",{"type":37,"value":962},". ChatGPT cites you today; two weeks later a model update drops your citation rate from 18% to 9%. Like algorithm updates, you can't control it. Your only defense: multi-model diversification. Don't rely on ChatGPT alone — earn citations across Claude, Gemini, and Perplexity.",{"type":32,"tag":33,"props":964,"children":965},{},[966,968,973],{"type":37,"value":967},"Third: ",{"type":32,"tag":76,"props":969,"children":970},{},[971],{"type":37,"value":972},"cost",{"type":37,"value":974},". 500 keywords × 3 models × 4 weeks = 6,000 API calls monthly. At $0.02 per call, that's $120\u002Fmonth. Startups can handle it, but enterprise scales to 5,000 keywords and costs hit $1,200\u002Fmonth. If budget constraints exist, tier your keywords — Tier 1 (high-value, weekly tests), Tier 2 (medium-value, monthly tests).",{"type":32,"tag":33,"props":976,"children":977},{},[978,980,985],{"type":37,"value":979},"Fourth: ",{"type":32,"tag":76,"props":981,"children":982},{},[983],{"type":37,"value":984},"false positives",{"type":37,"value":986},". You search for \"Roibase\" and the model says \"agencies like Roibase.\" Is that a citation? Technically yes, but brand equity is zero. Context quality scores solve this — don't just count mentions; add sentiment + specificity scoring.",{"type":32,"tag":40,"props":988,"children":990},{"id":989},"what-to-do-now",[991],{"type":37,"value":992},"What to Do Now",{"type":32,"tag":33,"props":994,"children":995},{},[996],{"type":37,"value":997},"Citation tracking isn't mainstream yet, but by 2027 it'll be standard. Start early and you establish a baseline—when competitors start, you're already seeing trends. First step: take 50 critical keywords, convert them into prompts, test manually on ChatGPT and Perplexity. How often does your brand appear? In what context? Two hours of work, full baseline. Next: automate with an API probe using n8n or Python, pull weekly reports. If citation rate is low, increase content depth and structured data. If it's high, wire it into your attribution pipeline and measure brand lift. LLM citation is SEO's new frontier — ranking position 1 on Google is secondary; getting mentioned in ChatGPT's answer is the goal.",{"type":32,"tag":999,"props":1000,"children":1001},"style",{},[1002],{"type":37,"value":1003},"html .default .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}html .shiki span {color: var(--shiki-default);background: var(--shiki-default-bg);font-style: var(--shiki-default-font-style);font-weight: var(--shiki-default-font-weight);text-decoration: var(--shiki-default-text-decoration);}",{"title":16,"searchDepth":168,"depth":168,"links":1005},[1006,1007,1008,1009,1010,1011,1012],{"id":42,"depth":152,"text":45},{"id":63,"depth":152,"text":66},{"id":589,"depth":152,"text":592},{"id":771,"depth":152,"text":774},{"id":859,"depth":152,"text":862},{"id":935,"depth":152,"text":938},{"id":989,"depth":152,"text":992},"markdown","content:en:ai:llm-citation-measurement-new-seo-metric.md","content","en\u002Fai\u002Fllm-citation-measurement-new-seo-metric.md","en\u002Fai\u002Fllm-citation-measurement-new-seo-metric","md",1782079495825]