[{"data":1,"prerenderedAt":1772},["ShallowReactive",2],{"article-alternates":3,"article-\u002Fen\u002Fai\u002Fllm-citation-measurement":13},{"i18nKey":4,"paths":5},"ai-002-2026-07",{"de":6,"en":7,"es":8,"fr":9,"it":10,"ru":11,"tr":12},"\u002Fde\u002Fai\u002Fllm-zitierungsmetriken","\u002Fen\u002Fai\u002Fllm-citation-measurement-new-seo-metrics","\u002Fes\u002Fai\u002Fmedicion-de-citacion-llm","\u002Ffr\u002Fai\u002Fmesure-des-citations-llm","\u002Fit\u002Fai\u002Fmisurazione-citazioni-llm","\u002Fru\u002Fai\u002Fllm-citation-measurement","\u002Ftr\u002Fai\u002Fllm-citation-olcumu-yeni-seo-metrik-setiniz",{"_path":14,"_dir":15,"_draft":16,"_partial":16,"_locale":17,"title":18,"description":19,"publishedAt":20,"modifiedAt":20,"category":15,"i18nKey":4,"tags":21,"readingTime":27,"author":28,"body":29,"_type":1766,"_id":1767,"_source":1768,"_file":1769,"_stem":1770,"_extension":1771},"\u002Fen\u002Fai\u002Fllm-citation-measurement","ai",false,"","LLM Citation Measurement — Your New SEO Metric Set","How do you measure your brand's citation rate in Perplexity, ChatGPT, and Gemini? A guide to setting up critical metrics for GEO.","2026-07-25",[22,23,24,25,26],"llm-citation","geo","seo-metrics","ai-search","attribution",8,"Roibase",{"type":30,"children":31,"toc":1753},"root",[32,48,55,67,72,84,90,95,109,155,191,233,254,260,265,272,436,444,463,468,474,486,604,614,620,625,770,775,781,794,799,820,833,877,909,978,990,996,1016,1024,1040,1048,1073,1078,1099,1266,1275,1281,1293,1298,1315,1335,1694,1707,1727,1733,1747],{"type":33,"tag":34,"props":35,"children":36},"element","p",{},[37,40,46],{"type":38,"value":39},"text","Organic traffic is declining, direct traffic in Google Analytics is rising, but you can't tell which queries are now being answered by ChatGPT and bypassing your brand site entirely. By mid-2026, LLMs captured 23% of search traffic (SimilarWeb Q2 2026 data). Rather than fighting to reclaim that traffic, you need to start measuring how often LLMs ",{"type":33,"tag":41,"props":42,"children":43},"strong",{},[44],{"type":38,"value":45},"show you as a source",{"type":38,"value":47},". Add a new layer to your SEO metrics: citation rate, source prominence, retrieval frequency.",{"type":33,"tag":49,"props":50,"children":52},"h2",{"id":51},"what-is-llm-citation-and-why-measure-it-now",[53],{"type":38,"value":54},"What Is LLM Citation and Why Measure It Now",{"type":33,"tag":34,"props":56,"children":57},{},[58,60,65],{"type":38,"value":59},"LLM citation is the percentage of times a generative model references your brand or content ",{"type":33,"tag":41,"props":61,"children":62},{},[63],{"type":38,"value":64},"as a source",{"type":38,"value":66}," when answering a user query. If ChatGPT writes \"Source: roibase.com.tr,\" Perplexity provides an inline link, or Gemini lists your site in a footnote—you've earned a citation.",{"type":33,"tag":34,"props":68,"children":69},{},[70],{"type":38,"value":71},"In classical SEO, you had \"ranking\"—being in position 3 on Google. In the LLM era, you have \"citation prominence\"—when a model shows 4 sources, what's your share? Are you the first source or buried in a \"related sources\" list? This difference can shift click-through rate by 300% (Perplexity Labs internal data, Q1 2026).",{"type":33,"tag":34,"props":73,"children":74},{},[75,77,82],{"type":38,"value":76},"If you don't start measuring now, you can't establish a baseline. Six months later, you won't be able to answer \"Did our GEO work?\" First step: ",{"type":33,"tag":41,"props":78,"children":79},{},[80],{"type":38,"value":81},"build a synthetic query set",{"type":38,"value":83}," and query LLMs on a regular cadence.",{"type":33,"tag":49,"props":85,"children":87},{"id":86},"setting-up-the-measurement-architecture-synthetic-query-pipeline",[88],{"type":38,"value":89},"Setting Up the Measurement Architecture: Synthetic Query Pipeline",{"type":33,"tag":34,"props":91,"children":92},{},[93],{"type":38,"value":94},"Manual testing won't cut it for LLM citation measurement. You need to run the same 50–100 queries against Perplexity \u002F ChatGPT \u002F Gemini daily, parse the source references in the responses, and track them over time. Do this with a three-layer pipeline:",{"type":33,"tag":34,"props":96,"children":97},{},[98,103,107],{"type":33,"tag":41,"props":99,"children":100},{},[101],{"type":38,"value":102},"Layer 1: Query Set Design",{"type":33,"tag":104,"props":105,"children":106},"br",{},[],{"type":38,"value":108},"\nPull from GSC the queries that got impressions in the last 90 days, ranked 1–20, with CTR below 5%. These are \"we rank but aren't clicked\"—LLMs are already answering them. Select 50–100 queries. Mix informational and transactional; skip pure brand queries. Examples: \"server-side GTM cookie duration,\" \"BigQuery cost optimization.\"",{"type":33,"tag":34,"props":110,"children":111},{},[112,117,120,122,129,131,137,139,145,147,153],{"type":33,"tag":41,"props":113,"children":114},{},[115],{"type":38,"value":116},"Layer 2: Automated Querying",{"type":33,"tag":104,"props":118,"children":119},{},[],{"type":38,"value":121},"\nUse an n8n workflow to hit each LLM's API once daily. Perplexity API with ",{"type":33,"tag":123,"props":124,"children":126},"code",{"className":125},[],[127],{"type":38,"value":128},"model: sonar-pro",{"type":38,"value":130},", ChatGPT in ",{"type":33,"tag":123,"props":132,"children":134},{"className":133},[],[135],{"type":38,"value":136},"browsing: true",{"type":38,"value":138}," mode, Gemini with ",{"type":33,"tag":123,"props":140,"children":142},{"className":141},[],[143],{"type":38,"value":144},"grounding: web",{"type":38,"value":146},". Save the response as JSON—both body and the ",{"type":33,"tag":123,"props":148,"children":150},{"className":149},[],[151],{"type":38,"value":152},"sources",{"type":38,"value":154}," array. Important: manage rate limits (Perplexity free tier is 5 req\u002Fmin, ChatGPT Plus is 40 req\u002F3 hours).",{"type":33,"tag":34,"props":156,"children":157},{},[158,163,166,168,173,175,181,183,189],{"type":33,"tag":41,"props":159,"children":160},{},[161],{"type":38,"value":162},"Layer 3: Citation Parser",{"type":33,"tag":104,"props":164,"children":165},{},[],{"type":38,"value":167},"\nIn the response JSON, check the ",{"type":33,"tag":123,"props":169,"children":171},{"className":170},[],[172],{"type":38,"value":152},{"type":38,"value":174}," key; if it's an array, scan it and match domains (",{"type":33,"tag":123,"props":176,"children":178},{"className":177},[],[179],{"type":38,"value":180},"roibase.com.tr",{"type":38,"value":182}," or subdomains). If no formal sources array, search the body for inline links (",{"type":33,"tag":123,"props":184,"children":186},{"className":185},[],[187],{"type":38,"value":188},"[roibase](...)",{"type":38,"value":190},") or plain URLs (regex). For each query, record three metrics:",{"type":33,"tag":192,"props":193,"children":194},"ol",{},[195,206,223],{"type":33,"tag":196,"props":197,"children":198},"li",{},[199,204],{"type":33,"tag":41,"props":200,"children":201},{},[202],{"type":38,"value":203},"Citation exists:",{"type":38,"value":205}," boolean (0\u002F1)",{"type":33,"tag":196,"props":207,"children":208},{},[209,214,216,221],{"type":33,"tag":41,"props":210,"children":211},{},[212],{"type":38,"value":213},"Rank:",{"type":38,"value":215}," position in the ",{"type":33,"tag":123,"props":217,"children":219},{"className":218},[],[220],{"type":38,"value":152},{"type":38,"value":222}," array (1–5, or null if absent)",{"type":33,"tag":196,"props":224,"children":225},{},[226,231],{"type":33,"tag":41,"props":227,"children":228},{},[229],{"type":38,"value":230},"Prominence:",{"type":38,"value":232}," inline in body or footnote only (inline = 2, footnote = 1, none = 0)",{"type":33,"tag":34,"props":234,"children":235},{},[236,238,244,246,252],{"type":38,"value":237},"Write this data to BigQuery in an ",{"type":33,"tag":123,"props":239,"children":241},{"className":240},[],[242],{"type":38,"value":243},"llm_citations",{"type":38,"value":245}," table—schema: ",{"type":33,"tag":123,"props":247,"children":249},{"className":248},[],[250],{"type":38,"value":251},"query_id, llm_provider, date, cited, rank, prominence",{"type":38,"value":253},".",{"type":33,"tag":49,"props":255,"children":257},{"id":256},"calculating-and-benchmarking-citation-rate",[258],{"type":38,"value":259},"Calculating and Benchmarking Citation Rate",{"type":33,"tag":34,"props":261,"children":262},{},[263],{"type":38,"value":264},"Run 50 queries once per day for 30 days across 3 LLMs: 50 query × 3 LLM × 30 days = 4,500 rows. Now compute your metrics:",{"type":33,"tag":266,"props":267,"children":269},"h3",{"id":268},"_1-overall-citation-rate",[270],{"type":38,"value":271},"1. Overall Citation Rate",{"type":33,"tag":273,"props":274,"children":278},"pre",{"className":275,"code":276,"language":277,"meta":17,"style":17},"language-sql shiki shiki-themes github-dark","SELECT \n  llm_provider,\n  COUNTIF(cited = 1) \u002F COUNT(*) AS citation_rate\nFROM `project.dataset.llm_citations`\nWHERE date >= DATE_SUB(CURRENT_DATE(), INTERVAL 30 DAY)\nGROUP BY llm_provider;\n","sql",[279],{"type":33,"tag":123,"props":280,"children":281},{"__ignoreMap":17},[282,300,309,368,383,422],{"type":33,"tag":283,"props":284,"children":287},"span",{"class":285,"line":286},"line",1,[288,294],{"type":33,"tag":283,"props":289,"children":291},{"style":290},"--shiki-default:#F97583",[292],{"type":38,"value":293},"SELECT",{"type":33,"tag":283,"props":295,"children":297},{"style":296},"--shiki-default:#E1E4E8",[298],{"type":38,"value":299}," \n",{"type":33,"tag":283,"props":301,"children":303},{"class":285,"line":302},2,[304],{"type":33,"tag":283,"props":305,"children":306},{"style":296},[307],{"type":38,"value":308},"  llm_provider,\n",{"type":33,"tag":283,"props":310,"children":312},{"class":285,"line":311},3,[313,318,323,329,334,339,344,349,354,358,363],{"type":33,"tag":283,"props":314,"children":315},{"style":296},[316],{"type":38,"value":317},"  COUNTIF(cited ",{"type":33,"tag":283,"props":319,"children":320},{"style":290},[321],{"type":38,"value":322},"=",{"type":33,"tag":283,"props":324,"children":326},{"style":325},"--shiki-default:#79B8FF",[327],{"type":38,"value":328}," 1",{"type":33,"tag":283,"props":330,"children":331},{"style":296},[332],{"type":38,"value":333},") ",{"type":33,"tag":283,"props":335,"children":336},{"style":290},[337],{"type":38,"value":338},"\u002F",{"type":33,"tag":283,"props":340,"children":341},{"style":325},[342],{"type":38,"value":343}," COUNT",{"type":33,"tag":283,"props":345,"children":346},{"style":296},[347],{"type":38,"value":348},"(",{"type":33,"tag":283,"props":350,"children":351},{"style":290},[352],{"type":38,"value":353},"*",{"type":33,"tag":283,"props":355,"children":356},{"style":296},[357],{"type":38,"value":333},{"type":33,"tag":283,"props":359,"children":360},{"style":290},[361],{"type":38,"value":362},"AS",{"type":33,"tag":283,"props":364,"children":365},{"style":296},[366],{"type":38,"value":367}," citation_rate\n",{"type":33,"tag":283,"props":369,"children":371},{"class":285,"line":370},4,[372,377],{"type":33,"tag":283,"props":373,"children":374},{"style":290},[375],{"type":38,"value":376},"FROM",{"type":33,"tag":283,"props":378,"children":380},{"style":379},"--shiki-default:#9ECBFF",[381],{"type":38,"value":382}," `project.dataset.llm_citations`\n",{"type":33,"tag":283,"props":384,"children":386},{"class":285,"line":385},5,[387,392,397,402,407,412,417],{"type":33,"tag":283,"props":388,"children":389},{"style":290},[390],{"type":38,"value":391},"WHERE",{"type":33,"tag":283,"props":393,"children":394},{"style":290},[395],{"type":38,"value":396}," date",{"type":33,"tag":283,"props":398,"children":399},{"style":290},[400],{"type":38,"value":401}," >=",{"type":33,"tag":283,"props":403,"children":404},{"style":296},[405],{"type":38,"value":406}," DATE_SUB(CURRENT_DATE(), INTERVAL ",{"type":33,"tag":283,"props":408,"children":409},{"style":325},[410],{"type":38,"value":411},"30",{"type":33,"tag":283,"props":413,"children":414},{"style":290},[415],{"type":38,"value":416}," DAY",{"type":33,"tag":283,"props":418,"children":419},{"style":296},[420],{"type":38,"value":421},")\n",{"type":33,"tag":283,"props":423,"children":425},{"class":285,"line":424},6,[426,431],{"type":33,"tag":283,"props":427,"children":428},{"style":290},[429],{"type":38,"value":430},"GROUP BY",{"type":33,"tag":283,"props":432,"children":433},{"style":296},[434],{"type":38,"value":435}," llm_provider;\n",{"type":33,"tag":34,"props":437,"children":438},{},[439],{"type":33,"tag":41,"props":440,"children":441},{},[442],{"type":38,"value":443},"2026 Q2 B2B SaaS Benchmark:",{"type":33,"tag":445,"props":446,"children":447},"ul",{},[448,453,458],{"type":33,"tag":196,"props":449,"children":450},{},[451],{"type":38,"value":452},"Perplexity: 18–24%",{"type":33,"tag":196,"props":454,"children":455},{},[456],{"type":38,"value":457},"ChatGPT browsing: 12–16%",{"type":33,"tag":196,"props":459,"children":460},{},[461],{"type":38,"value":462},"Gemini grounding: 8–14%",{"type":33,"tag":34,"props":464,"children":465},{},[466],{"type":38,"value":467},"If you're below 12% on Perplexity, you have a GEO problem—your content isn't structured for retrieval.",{"type":33,"tag":266,"props":469,"children":471},{"id":470},"_2-primary-source-rate",[472],{"type":38,"value":473},"2. Primary Source Rate",{"type":33,"tag":34,"props":475,"children":476},{},[477,479,484],{"type":38,"value":478},"When you are cited, how often are you the ",{"type":33,"tag":41,"props":480,"children":481},{},[482],{"type":38,"value":483},"first source",{"type":38,"value":485},"?",{"type":33,"tag":273,"props":487,"children":489},{"className":275,"code":488,"language":277,"meta":17,"style":17},"SELECT \n  llm_provider,\n  COUNTIF(rank = 1) \u002F COUNTIF(cited = 1) AS primary_rate\nFROM `project.dataset.llm_citations`\nWHERE cited = 1\nGROUP BY llm_provider;\n",[490],{"type":33,"tag":123,"props":491,"children":492},{"__ignoreMap":17},[493,504,511,561,572,593],{"type":33,"tag":283,"props":494,"children":495},{"class":285,"line":286},[496,500],{"type":33,"tag":283,"props":497,"children":498},{"style":290},[499],{"type":38,"value":293},{"type":33,"tag":283,"props":501,"children":502},{"style":296},[503],{"type":38,"value":299},{"type":33,"tag":283,"props":505,"children":506},{"class":285,"line":302},[507],{"type":33,"tag":283,"props":508,"children":509},{"style":296},[510],{"type":38,"value":308},{"type":33,"tag":283,"props":512,"children":513},{"class":285,"line":311},[514,519,523,527,531,535,540,544,548,552,556],{"type":33,"tag":283,"props":515,"children":516},{"style":296},[517],{"type":38,"value":518},"  COUNTIF(rank ",{"type":33,"tag":283,"props":520,"children":521},{"style":290},[522],{"type":38,"value":322},{"type":33,"tag":283,"props":524,"children":525},{"style":325},[526],{"type":38,"value":328},{"type":33,"tag":283,"props":528,"children":529},{"style":296},[530],{"type":38,"value":333},{"type":33,"tag":283,"props":532,"children":533},{"style":290},[534],{"type":38,"value":338},{"type":33,"tag":283,"props":536,"children":537},{"style":296},[538],{"type":38,"value":539}," COUNTIF(cited ",{"type":33,"tag":283,"props":541,"children":542},{"style":290},[543],{"type":38,"value":322},{"type":33,"tag":283,"props":545,"children":546},{"style":325},[547],{"type":38,"value":328},{"type":33,"tag":283,"props":549,"children":550},{"style":296},[551],{"type":38,"value":333},{"type":33,"tag":283,"props":553,"children":554},{"style":290},[555],{"type":38,"value":362},{"type":33,"tag":283,"props":557,"children":558},{"style":296},[559],{"type":38,"value":560}," primary_rate\n",{"type":33,"tag":283,"props":562,"children":563},{"class":285,"line":370},[564,568],{"type":33,"tag":283,"props":565,"children":566},{"style":290},[567],{"type":38,"value":376},{"type":33,"tag":283,"props":569,"children":570},{"style":379},[571],{"type":38,"value":382},{"type":33,"tag":283,"props":573,"children":574},{"class":285,"line":385},[575,579,584,588],{"type":33,"tag":283,"props":576,"children":577},{"style":290},[578],{"type":38,"value":391},{"type":33,"tag":283,"props":580,"children":581},{"style":296},[582],{"type":38,"value":583}," cited ",{"type":33,"tag":283,"props":585,"children":586},{"style":290},[587],{"type":38,"value":322},{"type":33,"tag":283,"props":589,"children":590},{"style":325},[591],{"type":38,"value":592}," 1\n",{"type":33,"tag":283,"props":594,"children":595},{"class":285,"line":424},[596,600],{"type":33,"tag":283,"props":597,"children":598},{"style":290},[599],{"type":38,"value":430},{"type":33,"tag":283,"props":601,"children":602},{"style":296},[603],{"type":38,"value":435},{"type":33,"tag":34,"props":605,"children":606},{},[607,612],{"type":33,"tag":41,"props":608,"children":609},{},[610],{"type":38,"value":611},"Target:",{"type":38,"value":613}," 40%+ (if cited, you should be the first source 4 out of 10 times). Below 20% means weak relevance signal—likely low embedding similarity during retrieval.",{"type":33,"tag":266,"props":615,"children":617},{"id":616},"_3-query-level-volatility",[618],{"type":38,"value":619},"3. Query-Level Volatility",{"type":33,"tag":34,"props":621,"children":622},{},[623],{"type":38,"value":624},"For each query, calculate citation variance over 30 days—some queries show you every day (low volatility), others inconsistently (high volatility). High volatility means the LLM updates its index frequently or competitors are outpacing you.",{"type":33,"tag":273,"props":626,"children":628},{"className":275,"code":627,"language":277,"meta":17,"style":17},"SELECT \n  query_id,\n  STDDEV(cited) AS citation_volatility\nFROM `project.dataset.llm_citations`\nWHERE llm_provider = 'perplexity'\nGROUP BY query_id\nHAVING COUNT(*) >= 20\nORDER BY citation_volatility DESC;\n",[629],{"type":33,"tag":123,"props":630,"children":631},{"__ignoreMap":17},[632,643,651,668,679,700,712,747],{"type":33,"tag":283,"props":633,"children":634},{"class":285,"line":286},[635,639],{"type":33,"tag":283,"props":636,"children":637},{"style":290},[638],{"type":38,"value":293},{"type":33,"tag":283,"props":640,"children":641},{"style":296},[642],{"type":38,"value":299},{"type":33,"tag":283,"props":644,"children":645},{"class":285,"line":302},[646],{"type":33,"tag":283,"props":647,"children":648},{"style":296},[649],{"type":38,"value":650},"  query_id,\n",{"type":33,"tag":283,"props":652,"children":653},{"class":285,"line":311},[654,659,663],{"type":33,"tag":283,"props":655,"children":656},{"style":296},[657],{"type":38,"value":658},"  STDDEV(cited) ",{"type":33,"tag":283,"props":660,"children":661},{"style":290},[662],{"type":38,"value":362},{"type":33,"tag":283,"props":664,"children":665},{"style":296},[666],{"type":38,"value":667}," citation_volatility\n",{"type":33,"tag":283,"props":669,"children":670},{"class":285,"line":370},[671,675],{"type":33,"tag":283,"props":672,"children":673},{"style":290},[674],{"type":38,"value":376},{"type":33,"tag":283,"props":676,"children":677},{"style":379},[678],{"type":38,"value":382},{"type":33,"tag":283,"props":680,"children":681},{"class":285,"line":385},[682,686,691,695],{"type":33,"tag":283,"props":683,"children":684},{"style":290},[685],{"type":38,"value":391},{"type":33,"tag":283,"props":687,"children":688},{"style":296},[689],{"type":38,"value":690}," llm_provider ",{"type":33,"tag":283,"props":692,"children":693},{"style":290},[694],{"type":38,"value":322},{"type":33,"tag":283,"props":696,"children":697},{"style":379},[698],{"type":38,"value":699}," 'perplexity'\n",{"type":33,"tag":283,"props":701,"children":702},{"class":285,"line":424},[703,707],{"type":33,"tag":283,"props":704,"children":705},{"style":290},[706],{"type":38,"value":430},{"type":33,"tag":283,"props":708,"children":709},{"style":296},[710],{"type":38,"value":711}," query_id\n",{"type":33,"tag":283,"props":713,"children":715},{"class":285,"line":714},7,[716,721,725,729,733,737,742],{"type":33,"tag":283,"props":717,"children":718},{"style":290},[719],{"type":38,"value":720},"HAVING",{"type":33,"tag":283,"props":722,"children":723},{"style":325},[724],{"type":38,"value":343},{"type":33,"tag":283,"props":726,"children":727},{"style":296},[728],{"type":38,"value":348},{"type":33,"tag":283,"props":730,"children":731},{"style":290},[732],{"type":38,"value":353},{"type":33,"tag":283,"props":734,"children":735},{"style":296},[736],{"type":38,"value":333},{"type":33,"tag":283,"props":738,"children":739},{"style":290},[740],{"type":38,"value":741},">=",{"type":33,"tag":283,"props":743,"children":744},{"style":325},[745],{"type":38,"value":746}," 20\n",{"type":33,"tag":283,"props":748,"children":749},{"class":285,"line":27},[750,755,760,765],{"type":33,"tag":283,"props":751,"children":752},{"style":290},[753],{"type":38,"value":754},"ORDER BY",{"type":33,"tag":283,"props":756,"children":757},{"style":296},[758],{"type":38,"value":759}," citation_volatility ",{"type":33,"tag":283,"props":761,"children":762},{"style":290},[763],{"type":38,"value":764},"DESC",{"type":33,"tag":283,"props":766,"children":767},{"style":296},[768],{"type":38,"value":769},";\n",{"type":33,"tag":34,"props":771,"children":772},{},[773],{"type":38,"value":774},"Volatility > 0.4? Check manually—likely a freshness issue (your content is 6 months old; the LLM favors newer pieces).",{"type":33,"tag":49,"props":776,"children":778},{"id":777},"attribution-tradeoff-direct-traffic-or-llm-referral",[779],{"type":38,"value":780},"Attribution Tradeoff: Direct Traffic or LLM Referral?",{"type":33,"tag":34,"props":782,"children":783},{},[784,786,792],{"type":38,"value":785},"One side effect of earning LLM citations: direct traffic rises in Google Analytics, but you can't tell it came from an LLM. ChatGPT web-interface clicks show as ",{"type":33,"tag":123,"props":787,"children":789},{"className":788},[],[790],{"type":38,"value":791},"(direct) \u002F (none)",{"type":38,"value":793},"—no referrer header.",{"type":33,"tag":34,"props":795,"children":796},{},[797],{"type":38,"value":798},"Solve this with two methods:",{"type":33,"tag":34,"props":800,"children":801},{},[802,807,810,812,818],{"type":33,"tag":41,"props":803,"children":804},{},[805],{"type":38,"value":806},"Method 1: UTM Injection (in LLM API)",{"type":33,"tag":104,"props":808,"children":809},{},[],{"type":38,"value":811},"\nIf you submit content to LLM APIs (e.g., Perplexity Publisher API), add ",{"type":33,"tag":123,"props":813,"children":815},{"className":814},[],[816],{"type":38,"value":817},"?utm_source=perplexity&utm_medium=llm&utm_campaign=citation",{"type":38,"value":819}," to your URLs. GA4 will then show the source. This works only for API-integrated LLMs—you can't inject UTMs into ChatGPT web crawl.",{"type":33,"tag":34,"props":821,"children":822},{},[823,828,831],{"type":33,"tag":41,"props":824,"children":825},{},[826],{"type":38,"value":827},"Method 2: Server-Side Fingerprinting",{"type":33,"tag":104,"props":829,"children":830},{},[],{"type":38,"value":832},"\nLLM bots use distinct user-agent patterns:",{"type":33,"tag":445,"props":834,"children":835},{},[836,847,866],{"type":33,"tag":196,"props":837,"children":838},{},[839,841],{"type":38,"value":840},"Perplexity: ",{"type":33,"tag":123,"props":842,"children":844},{"className":843},[],[845],{"type":38,"value":846},"PerplexityBot",{"type":33,"tag":196,"props":848,"children":849},{},[850,852,858,860],{"type":38,"value":851},"ChatGPT: ",{"type":33,"tag":123,"props":853,"children":855},{"className":854},[],[856],{"type":38,"value":857},"ChatGPT-User",{"type":38,"value":859}," or ",{"type":33,"tag":123,"props":861,"children":863},{"className":862},[],[864],{"type":38,"value":865},"GPTBot",{"type":33,"tag":196,"props":867,"children":868},{},[869,871],{"type":38,"value":870},"Gemini: ",{"type":33,"tag":123,"props":872,"children":874},{"className":873},[],[875],{"type":38,"value":876},"Google-Extended",{"type":33,"tag":34,"props":878,"children":879},{},[880,882,891,893,899,901,907],{"type":38,"value":881},"Filter server logs for these user-agents and send them to GA4 as server-side events via ",{"type":33,"tag":883,"props":884,"children":888},"a",{"href":885,"rel":886},"https:\u002F\u002Fwww.roibase.com.tr\u002Fen\u002Ffirstparty",[887],"nofollow",[889],{"type":38,"value":890},"First-Party Data & Measurement Architecture",{"type":38,"value":892},". Event name: ",{"type":33,"tag":123,"props":894,"children":896},{"className":895},[],[897],{"type":38,"value":898},"llm_visit",{"type":38,"value":900},", parameter: ",{"type":33,"tag":123,"props":902,"children":904},{"className":903},[],[905],{"type":38,"value":906},"llm_provider",{"type":38,"value":908},". Now you can isolate LLM traffic within your \"direct\" bucket.",{"type":33,"tag":910,"props":911,"children":912},"table",{},[913,937],{"type":33,"tag":914,"props":915,"children":916},"thead",{},[917],{"type":33,"tag":918,"props":919,"children":920},"tr",{},[921,927,932],{"type":33,"tag":922,"props":923,"children":924},"th",{},[925],{"type":38,"value":926},"Method",{"type":33,"tag":922,"props":928,"children":929},{},[930],{"type":38,"value":931},"Advantage",{"type":33,"tag":922,"props":933,"children":934},{},[935],{"type":38,"value":936},"Limitation",{"type":33,"tag":938,"props":939,"children":940},"tbody",{},[941,960],{"type":33,"tag":918,"props":942,"children":943},{},[944,950,955],{"type":33,"tag":945,"props":946,"children":947},"td",{},[948],{"type":38,"value":949},"UTM Injection",{"type":33,"tag":945,"props":951,"children":952},{},[953],{"type":38,"value":954},"Auto-sources in GA4",{"type":33,"tag":945,"props":956,"children":957},{},[958],{"type":38,"value":959},"API-only",{"type":33,"tag":918,"props":961,"children":962},{},[963,968,973],{"type":33,"tag":945,"props":964,"children":965},{},[966],{"type":38,"value":967},"Server-Side Fingerprint",{"type":33,"tag":945,"props":969,"children":970},{},[971],{"type":38,"value":972},"Works for all LLMs",{"type":33,"tag":945,"props":974,"children":975},{},[976],{"type":38,"value":977},"Requires log parsing",{"type":33,"tag":34,"props":979,"children":980},{},[981,983,988],{"type":38,"value":982},"Either way, your goal: ",{"type":33,"tag":41,"props":984,"children":985},{},[986],{"type":38,"value":987},"correlate citation rate with referral traffic",{"type":38,"value":989},". If citations rise 20% but LLM referrals don't, users see you cited but aren't clicking—prominence or snippet quality is the issue.",{"type":33,"tag":49,"props":991,"children":993},{"id":992},"citation-prominence-inline-vs-footnote",[994],{"type":38,"value":995},"Citation Prominence: Inline vs Footnote",{"type":33,"tag":34,"props":997,"children":998},{},[999,1001,1006,1008,1014],{"type":38,"value":1000},"The LLM cited you, but ",{"type":33,"tag":41,"props":1002,"children":1003},{},[1004],{"type":38,"value":1005},"how",{"type":38,"value":1007},"? Did Perplexity embed you as an inline link in the answer (numbered ",{"type":33,"tag":123,"props":1009,"children":1011},{"className":1010},[],[1012],{"type":38,"value":1013},"[1]",{"type":38,"value":1015},"), or list you at the end in \"Related sources\"? This difference drives 4x difference in click-through rate (Roibase internal A\u002FB, n=2,300 queries).",{"type":33,"tag":34,"props":1017,"children":1018},{},[1019],{"type":33,"tag":41,"props":1020,"children":1021},{},[1022],{"type":38,"value":1023},"Inline citation example:",{"type":33,"tag":1025,"props":1026,"children":1027},"blockquote",{},[1028],{"type":33,"tag":34,"props":1029,"children":1030},{},[1031,1033,1038],{"type":38,"value":1032},"\"Server-side GTM cookie duration can be extended to 730 days ",{"type":33,"tag":883,"props":1034,"children":1036},{"href":1035},"roibase.com.tr\u002F...",[1037],{"type":38,"value":1013},{"type":38,"value":1039},".\"",{"type":33,"tag":34,"props":1041,"children":1042},{},[1043],{"type":33,"tag":41,"props":1044,"children":1045},{},[1046],{"type":38,"value":1047},"Footnote citation example:",{"type":33,"tag":1025,"props":1049,"children":1050},{},[1051,1061],{"type":33,"tag":34,"props":1052,"children":1053},{},[1054,1056,1059],{"type":38,"value":1055},"\"...multiple approaches exist.",{"type":33,"tag":104,"props":1057,"children":1058},{},[],{"type":38,"value":1060},"\nSources:",{"type":33,"tag":192,"props":1062,"children":1063},{},[1064,1068],{"type":33,"tag":196,"props":1065,"children":1066},{},[1067],{"type":38,"value":1035},{"type":33,"tag":196,"props":1069,"children":1070},{},[1071],{"type":38,"value":1072},"competitor.com\u002F...\"",{"type":33,"tag":34,"props":1074,"children":1075},{},[1076],{"type":38,"value":1077},"Inline citations get clicked as users read—context is present. Footnotes get clicked only if the user seeks more detail—conversion intent is lower.",{"type":33,"tag":34,"props":1079,"children":1080},{},[1081,1086,1089,1091,1097],{"type":33,"tag":41,"props":1082,"children":1083},{},[1084],{"type":38,"value":1085},"Calculate prominence score:",{"type":33,"tag":104,"props":1087,"children":1088},{},[],{"type":38,"value":1090},"\nFor each citation, log ",{"type":33,"tag":123,"props":1092,"children":1094},{"className":1093},[],[1095],{"type":38,"value":1096},"position_type",{"type":38,"value":1098}," (inline \u002F footnote \u002F sidebar). Average over 30 days:",{"type":33,"tag":273,"props":1100,"children":1102},{"className":275,"code":1101,"language":277,"meta":17,"style":17},"SELECT \n  AVG(CASE \n    WHEN position_type = 'inline' THEN 3\n    WHEN position_type = 'footnote' THEN 1\n    ELSE 0\n  END) AS avg_prominence_score\nFROM `project.dataset.llm_citations`\nWHERE cited = 1;\n",[1103],{"type":33,"tag":123,"props":1104,"children":1105},{"__ignoreMap":17},[1106,1117,1138,1170,1198,1211,1232,1243],{"type":33,"tag":283,"props":1107,"children":1108},{"class":285,"line":286},[1109,1113],{"type":33,"tag":283,"props":1110,"children":1111},{"style":290},[1112],{"type":38,"value":293},{"type":33,"tag":283,"props":1114,"children":1115},{"style":296},[1116],{"type":38,"value":299},{"type":33,"tag":283,"props":1118,"children":1119},{"class":285,"line":302},[1120,1125,1129,1134],{"type":33,"tag":283,"props":1121,"children":1122},{"style":325},[1123],{"type":38,"value":1124},"  AVG",{"type":33,"tag":283,"props":1126,"children":1127},{"style":296},[1128],{"type":38,"value":348},{"type":33,"tag":283,"props":1130,"children":1131},{"style":290},[1132],{"type":38,"value":1133},"CASE",{"type":33,"tag":283,"props":1135,"children":1136},{"style":296},[1137],{"type":38,"value":299},{"type":33,"tag":283,"props":1139,"children":1140},{"class":285,"line":311},[1141,1146,1151,1155,1160,1165],{"type":33,"tag":283,"props":1142,"children":1143},{"style":290},[1144],{"type":38,"value":1145},"    WHEN",{"type":33,"tag":283,"props":1147,"children":1148},{"style":296},[1149],{"type":38,"value":1150}," position_type ",{"type":33,"tag":283,"props":1152,"children":1153},{"style":290},[1154],{"type":38,"value":322},{"type":33,"tag":283,"props":1156,"children":1157},{"style":379},[1158],{"type":38,"value":1159}," 'inline'",{"type":33,"tag":283,"props":1161,"children":1162},{"style":290},[1163],{"type":38,"value":1164}," THEN",{"type":33,"tag":283,"props":1166,"children":1167},{"style":325},[1168],{"type":38,"value":1169}," 3\n",{"type":33,"tag":283,"props":1171,"children":1172},{"class":285,"line":370},[1173,1177,1181,1185,1190,1194],{"type":33,"tag":283,"props":1174,"children":1175},{"style":290},[1176],{"type":38,"value":1145},{"type":33,"tag":283,"props":1178,"children":1179},{"style":296},[1180],{"type":38,"value":1150},{"type":33,"tag":283,"props":1182,"children":1183},{"style":290},[1184],{"type":38,"value":322},{"type":33,"tag":283,"props":1186,"children":1187},{"style":379},[1188],{"type":38,"value":1189}," 'footnote'",{"type":33,"tag":283,"props":1191,"children":1192},{"style":290},[1193],{"type":38,"value":1164},{"type":33,"tag":283,"props":1195,"children":1196},{"style":325},[1197],{"type":38,"value":592},{"type":33,"tag":283,"props":1199,"children":1200},{"class":285,"line":385},[1201,1206],{"type":33,"tag":283,"props":1202,"children":1203},{"style":290},[1204],{"type":38,"value":1205},"    ELSE",{"type":33,"tag":283,"props":1207,"children":1208},{"style":325},[1209],{"type":38,"value":1210}," 0\n",{"type":33,"tag":283,"props":1212,"children":1213},{"class":285,"line":424},[1214,1219,1223,1227],{"type":33,"tag":283,"props":1215,"children":1216},{"style":290},[1217],{"type":38,"value":1218},"  END",{"type":33,"tag":283,"props":1220,"children":1221},{"style":296},[1222],{"type":38,"value":333},{"type":33,"tag":283,"props":1224,"children":1225},{"style":290},[1226],{"type":38,"value":362},{"type":33,"tag":283,"props":1228,"children":1229},{"style":296},[1230],{"type":38,"value":1231}," avg_prominence_score\n",{"type":33,"tag":283,"props":1233,"children":1234},{"class":285,"line":714},[1235,1239],{"type":33,"tag":283,"props":1236,"children":1237},{"style":290},[1238],{"type":38,"value":376},{"type":33,"tag":283,"props":1240,"children":1241},{"style":379},[1242],{"type":38,"value":382},{"type":33,"tag":283,"props":1244,"children":1245},{"class":285,"line":27},[1246,1250,1254,1258,1262],{"type":33,"tag":283,"props":1247,"children":1248},{"style":290},[1249],{"type":38,"value":391},{"type":33,"tag":283,"props":1251,"children":1252},{"style":296},[1253],{"type":38,"value":583},{"type":33,"tag":283,"props":1255,"children":1256},{"style":290},[1257],{"type":38,"value":322},{"type":33,"tag":283,"props":1259,"children":1260},{"style":325},[1261],{"type":38,"value":328},{"type":33,"tag":283,"props":1263,"children":1264},{"style":296},[1265],{"type":38,"value":769},{"type":33,"tag":34,"props":1267,"children":1268},{},[1269,1273],{"type":33,"tag":41,"props":1270,"children":1271},{},[1272],{"type":38,"value":611},{"type":38,"value":1274}," 2.0+ (if cited, more than half should be inline). Below 1.5 means the LLM treats you as supplementary, not primary. Fix: structure your content for LLM quotation—single-sentence definitions, fact boxes, code snippets.",{"type":33,"tag":49,"props":1276,"children":1278},{"id":1277},"competitive-analysis-query-level-source-overlap",[1279],{"type":38,"value":1280},"Competitive Analysis: Query-Level Source Overlap",{"type":33,"tag":34,"props":1282,"children":1283},{},[1284,1286,1291],{"type":38,"value":1285},"Which queries do you miss citations on, while competitors don't? Parse ",{"type":33,"tag":41,"props":1287,"children":1288},{},[1289],{"type":38,"value":1290},"all sources",{"type":38,"value":1292}," each LLM shows per query, not just yourself.",{"type":33,"tag":34,"props":1294,"children":1295},{},[1296],{"type":38,"value":1297},"Example: On \"BigQuery cost optimization,\" Perplexity shows:",{"type":33,"tag":192,"props":1299,"children":1300},{},[1301,1306,1310],{"type":33,"tag":196,"props":1302,"children":1303},{},[1304],{"type":38,"value":1305},"competitor-a.com",{"type":33,"tag":196,"props":1307,"children":1308},{},[1309],{"type":38,"value":180},{"type":33,"tag":196,"props":1311,"children":1312},{},[1313],{"type":38,"value":1314},"competitor-b.com",{"type":33,"tag":34,"props":1316,"children":1317},{},[1318,1320,1326,1327,1333],{"type":38,"value":1319},"Write all sources to ",{"type":33,"tag":123,"props":1321,"children":1323},{"className":1322},[],[1324],{"type":38,"value":1325},"llm_all_sources",{"type":38,"value":245},{"type":33,"tag":123,"props":1328,"children":1330},{"className":1329},[],[1331],{"type":38,"value":1332},"query_id, llm_provider, date, source_domain, rank",{"type":38,"value":1334},". Now build an overlap matrix:",{"type":33,"tag":273,"props":1336,"children":1338},{"className":275,"code":1337,"language":277,"meta":17,"style":17},"SELECT \n  a.source_domain AS source_1,\n  b.source_domain AS source_2,\n  COUNT(DISTINCT a.query_id) AS co_citation_count\nFROM `project.dataset.llm_all_sources` a\nJOIN `project.dataset.llm_all_sources` b \n  ON a.query_id = b.query_id \n  AND a.llm_provider = b.llm_provider\n  AND a.date = b.date\nWHERE a.source_domain != b.source_domain\nGROUP BY source_1, source_2\nHAVING co_citation_count > 5\nORDER BY co_citation_count DESC;\n",[1339],{"type":33,"tag":123,"props":1340,"children":1341},{"__ignoreMap":17},[1342,1353,1380,1405,1449,1466,1483,1525,1562,1600,1638,1651,1674],{"type":33,"tag":283,"props":1343,"children":1344},{"class":285,"line":286},[1345,1349],{"type":33,"tag":283,"props":1346,"children":1347},{"style":290},[1348],{"type":38,"value":293},{"type":33,"tag":283,"props":1350,"children":1351},{"style":296},[1352],{"type":38,"value":299},{"type":33,"tag":283,"props":1354,"children":1355},{"class":285,"line":302},[1356,1361,1365,1370,1375],{"type":33,"tag":283,"props":1357,"children":1358},{"style":325},[1359],{"type":38,"value":1360},"  a",{"type":33,"tag":283,"props":1362,"children":1363},{"style":296},[1364],{"type":38,"value":253},{"type":33,"tag":283,"props":1366,"children":1367},{"style":325},[1368],{"type":38,"value":1369},"source_domain",{"type":33,"tag":283,"props":1371,"children":1372},{"style":290},[1373],{"type":38,"value":1374}," AS",{"type":33,"tag":283,"props":1376,"children":1377},{"style":296},[1378],{"type":38,"value":1379}," source_1,\n",{"type":33,"tag":283,"props":1381,"children":1382},{"class":285,"line":311},[1383,1388,1392,1396,1400],{"type":33,"tag":283,"props":1384,"children":1385},{"style":325},[1386],{"type":38,"value":1387},"  b",{"type":33,"tag":283,"props":1389,"children":1390},{"style":296},[1391],{"type":38,"value":253},{"type":33,"tag":283,"props":1393,"children":1394},{"style":325},[1395],{"type":38,"value":1369},{"type":33,"tag":283,"props":1397,"children":1398},{"style":290},[1399],{"type":38,"value":1374},{"type":33,"tag":283,"props":1401,"children":1402},{"style":296},[1403],{"type":38,"value":1404}," source_2,\n",{"type":33,"tag":283,"props":1406,"children":1407},{"class":285,"line":370},[1408,1413,1417,1422,1427,1431,1436,1440,1444],{"type":33,"tag":283,"props":1409,"children":1410},{"style":325},[1411],{"type":38,"value":1412},"  COUNT",{"type":33,"tag":283,"props":1414,"children":1415},{"style":296},[1416],{"type":38,"value":348},{"type":33,"tag":283,"props":1418,"children":1419},{"style":290},[1420],{"type":38,"value":1421},"DISTINCT",{"type":33,"tag":283,"props":1423,"children":1424},{"style":325},[1425],{"type":38,"value":1426}," a",{"type":33,"tag":283,"props":1428,"children":1429},{"style":296},[1430],{"type":38,"value":253},{"type":33,"tag":283,"props":1432,"children":1433},{"style":325},[1434],{"type":38,"value":1435},"query_id",{"type":33,"tag":283,"props":1437,"children":1438},{"style":296},[1439],{"type":38,"value":333},{"type":33,"tag":283,"props":1441,"children":1442},{"style":290},[1443],{"type":38,"value":362},{"type":33,"tag":283,"props":1445,"children":1446},{"style":296},[1447],{"type":38,"value":1448}," co_citation_count\n",{"type":33,"tag":283,"props":1450,"children":1451},{"class":285,"line":385},[1452,1456,1461],{"type":33,"tag":283,"props":1453,"children":1454},{"style":290},[1455],{"type":38,"value":376},{"type":33,"tag":283,"props":1457,"children":1458},{"style":379},[1459],{"type":38,"value":1460}," `project.dataset.llm_all_sources`",{"type":33,"tag":283,"props":1462,"children":1463},{"style":296},[1464],{"type":38,"value":1465}," a\n",{"type":33,"tag":283,"props":1467,"children":1468},{"class":285,"line":424},[1469,1474,1478],{"type":33,"tag":283,"props":1470,"children":1471},{"style":290},[1472],{"type":38,"value":1473},"JOIN",{"type":33,"tag":283,"props":1475,"children":1476},{"style":379},[1477],{"type":38,"value":1460},{"type":33,"tag":283,"props":1479,"children":1480},{"style":296},[1481],{"type":38,"value":1482}," b \n",{"type":33,"tag":283,"props":1484,"children":1485},{"class":285,"line":714},[1486,1491,1495,1499,1503,1508,1513,1517,1521],{"type":33,"tag":283,"props":1487,"children":1488},{"style":290},[1489],{"type":38,"value":1490},"  ON",{"type":33,"tag":283,"props":1492,"children":1493},{"style":325},[1494],{"type":38,"value":1426},{"type":33,"tag":283,"props":1496,"children":1497},{"style":296},[1498],{"type":38,"value":253},{"type":33,"tag":283,"props":1500,"children":1501},{"style":325},[1502],{"type":38,"value":1435},{"type":33,"tag":283,"props":1504,"children":1505},{"style":290},[1506],{"type":38,"value":1507}," =",{"type":33,"tag":283,"props":1509,"children":1510},{"style":325},[1511],{"type":38,"value":1512}," b",{"type":33,"tag":283,"props":1514,"children":1515},{"style":296},[1516],{"type":38,"value":253},{"type":33,"tag":283,"props":1518,"children":1519},{"style":325},[1520],{"type":38,"value":1435},{"type":33,"tag":283,"props":1522,"children":1523},{"style":296},[1524],{"type":38,"value":299},{"type":33,"tag":283,"props":1526,"children":1527},{"class":285,"line":27},[1528,1533,1537,1541,1545,1549,1553,1557],{"type":33,"tag":283,"props":1529,"children":1530},{"style":290},[1531],{"type":38,"value":1532},"  AND",{"type":33,"tag":283,"props":1534,"children":1535},{"style":325},[1536],{"type":38,"value":1426},{"type":33,"tag":283,"props":1538,"children":1539},{"style":296},[1540],{"type":38,"value":253},{"type":33,"tag":283,"props":1542,"children":1543},{"style":325},[1544],{"type":38,"value":906},{"type":33,"tag":283,"props":1546,"children":1547},{"style":290},[1548],{"type":38,"value":1507},{"type":33,"tag":283,"props":1550,"children":1551},{"style":325},[1552],{"type":38,"value":1512},{"type":33,"tag":283,"props":1554,"children":1555},{"style":296},[1556],{"type":38,"value":253},{"type":33,"tag":283,"props":1558,"children":1559},{"style":325},[1560],{"type":38,"value":1561},"llm_provider\n",{"type":33,"tag":283,"props":1563,"children":1565},{"class":285,"line":1564},9,[1566,1570,1574,1578,1583,1587,1591,1595],{"type":33,"tag":283,"props":1567,"children":1568},{"style":290},[1569],{"type":38,"value":1532},{"type":33,"tag":283,"props":1571,"children":1572},{"style":325},[1573],{"type":38,"value":1426},{"type":33,"tag":283,"props":1575,"children":1576},{"style":296},[1577],{"type":38,"value":253},{"type":33,"tag":283,"props":1579,"children":1580},{"style":325},[1581],{"type":38,"value":1582},"date",{"type":33,"tag":283,"props":1584,"children":1585},{"style":290},[1586],{"type":38,"value":1507},{"type":33,"tag":283,"props":1588,"children":1589},{"style":325},[1590],{"type":38,"value":1512},{"type":33,"tag":283,"props":1592,"children":1593},{"style":296},[1594],{"type":38,"value":253},{"type":33,"tag":283,"props":1596,"children":1597},{"style":325},[1598],{"type":38,"value":1599},"date\n",{"type":33,"tag":283,"props":1601,"children":1603},{"class":285,"line":1602},10,[1604,1608,1612,1616,1620,1625,1629,1633],{"type":33,"tag":283,"props":1605,"children":1606},{"style":290},[1607],{"type":38,"value":391},{"type":33,"tag":283,"props":1609,"children":1610},{"style":325},[1611],{"type":38,"value":1426},{"type":33,"tag":283,"props":1613,"children":1614},{"style":296},[1615],{"type":38,"value":253},{"type":33,"tag":283,"props":1617,"children":1618},{"style":325},[1619],{"type":38,"value":1369},{"type":33,"tag":283,"props":1621,"children":1622},{"style":290},[1623],{"type":38,"value":1624}," !=",{"type":33,"tag":283,"props":1626,"children":1627},{"style":325},[1628],{"type":38,"value":1512},{"type":33,"tag":283,"props":1630,"children":1631},{"style":296},[1632],{"type":38,"value":253},{"type":33,"tag":283,"props":1634,"children":1635},{"style":325},[1636],{"type":38,"value":1637},"source_domain\n",{"type":33,"tag":283,"props":1639,"children":1641},{"class":285,"line":1640},11,[1642,1646],{"type":33,"tag":283,"props":1643,"children":1644},{"style":290},[1645],{"type":38,"value":430},{"type":33,"tag":283,"props":1647,"children":1648},{"style":296},[1649],{"type":38,"value":1650}," source_1, source_2\n",{"type":33,"tag":283,"props":1652,"children":1654},{"class":285,"line":1653},12,[1655,1659,1664,1669],{"type":33,"tag":283,"props":1656,"children":1657},{"style":290},[1658],{"type":38,"value":720},{"type":33,"tag":283,"props":1660,"children":1661},{"style":296},[1662],{"type":38,"value":1663}," co_citation_count ",{"type":33,"tag":283,"props":1665,"children":1666},{"style":290},[1667],{"type":38,"value":1668},">",{"type":33,"tag":283,"props":1670,"children":1671},{"style":325},[1672],{"type":38,"value":1673}," 5\n",{"type":33,"tag":283,"props":1675,"children":1677},{"class":285,"line":1676},13,[1678,1682,1686,1690],{"type":33,"tag":283,"props":1679,"children":1680},{"style":290},[1681],{"type":38,"value":754},{"type":33,"tag":283,"props":1683,"children":1684},{"style":296},[1685],{"type":38,"value":1663},{"type":33,"tag":283,"props":1687,"children":1688},{"style":290},[1689],{"type":38,"value":764},{"type":33,"tag":283,"props":1691,"children":1692},{"style":296},[1693],{"type":38,"value":769},{"type":33,"tag":34,"props":1695,"children":1696},{},[1697,1699,1705],{"type":38,"value":1698},"This shows: \"Co-cited with competitor-a 47 times.\" Divide ",{"type":33,"tag":123,"props":1700,"children":1702},{"className":1701},[],[1703],{"type":38,"value":1704},"co_citation_count",{"type":38,"value":1706}," by competitor-a's standalone citations—this is your \"overlap ratio.\" >60% means direct competition; \u003C30% means different niches.",{"type":33,"tag":34,"props":1708,"children":1709},{},[1710,1715,1718,1720,1725],{"type":33,"tag":41,"props":1711,"children":1712},{},[1713],{"type":38,"value":1714},"Turn into action:",{"type":33,"tag":104,"props":1716,"children":1717},{},[],{"type":38,"value":1719},"\nHigh overlap but you're not cited while competitor-a is? Close the content gap. Read their page—what facts do they include? What format (table \u002F list \u002F code)? Present the same facts more ",{"type":33,"tag":41,"props":1721,"children":1722},{},[1723],{"type":38,"value":1724},"structurally",{"type":38,"value":1726}," (JSON-LD, tables, bullet lists)—LLM retrieval favors these.",{"type":33,"tag":49,"props":1728,"children":1730},{"id":1729},"what-to-measure-starting-now",[1731],{"type":38,"value":1732},"What to Measure Starting Now",{"type":33,"tag":34,"props":1734,"children":1735},{},[1736,1738,1745],{"type":38,"value":1737},"Start by designing a synthetic query set—pull low-CTR, high-impression queries from GSC. Set up daily querying via n8n, write responses to BigQuery. Establish a baseline over 30 days: citation rate, primary source rate, prominence score. Then measure the impact of your ",{"type":33,"tag":883,"props":1739,"children":1742},{"href":1740,"rel":1741},"https:\u002F\u002Fwww.roibase.com.tr\u002Fen\u002Fgeo",[887],[1743],{"type":38,"value":1744},"Generative Engine Optimization",{"type":38,"value":1746}," work—which content changes lifted citations? Which hurt? You're cited but traffic doesn't follow? That's a prominence problem—optimize for inline references. Run competitive analysis to spot co-citation patterns and patch content gaps. Add these metrics to your SEO dashboard. By year-end 2026, you'll track \"organic + LLM visibility\" instead of organic traffic alone.",{"type":33,"tag":1748,"props":1749,"children":1750},"style",{},[1751],{"type":38,"value":1752},"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":17,"searchDepth":311,"depth":311,"links":1754},[1755,1756,1757,1762,1763,1764,1765],{"id":51,"depth":302,"text":54},{"id":86,"depth":302,"text":89},{"id":256,"depth":302,"text":259,"children":1758},[1759,1760,1761],{"id":268,"depth":311,"text":271},{"id":470,"depth":311,"text":473},{"id":616,"depth":311,"text":619},{"id":777,"depth":302,"text":780},{"id":992,"depth":302,"text":995},{"id":1277,"depth":302,"text":1280},{"id":1729,"depth":302,"text":1732},"markdown","content:en:ai:llm-citation-measurement.md","content","en\u002Fai\u002Fllm-citation-measurement.md","en\u002Fai\u002Fllm-citation-measurement","md",1785967475856]