[{"data":1,"prerenderedAt":1017},["ShallowReactive",2],{"article-alternates":3,"article-\u002Fes\u002Fai\u002Fmedicion-de-citas-en-llm-nueva-metrica-seo":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":1011,"_id":1012,"_source":1013,"_file":1014,"_stem":1015,"_extension":1016},"\u002Fes\u002Fai\u002Fmedicion-de-citas-en-llm-nueva-metrica-seo","ai",false,"","Medición de Citas en LLM — Tu Nuevo Conjunto de Métricas SEO","¿Cómo medir la tasa de citas de tu marca en Perplexity, ChatGPT y Gemini? El seguimiento de citas es el nuevo conjunto de métricas SEO de generación.","2026-05-30",[21,22,23,24,25],"llm-citation","geo","seo-metrics","ai-attribution","brand-visibility",9,"Roibase",{"type":29,"children":30,"toc":1002},"root",[31,39,46,51,56,61,67,72,83,110,128,582,587,593,598,608,629,639,649,659,764,769,775,786,796,835,845,855,861,866,884,894,904,931,937,949,961,973,985,991,996],{"type":32,"tag":33,"props":34,"children":35},"element","p",{},[36],{"type":37,"value":38},"text","Tu tráfico orgánico está bajando, el CTR está estancado, pero ChatGPT menciona tu marca 4.000 veces al día. No lo sabes porque no aparece en Google Analytics. El seguimiento de citas en LLM es el nuevo conjunto de métricas SEO de la era de la IA generativa. Perplexity, ChatGPT, Gemini y otros modelos de lenguaje grande son ahora la nueva interfaz de búsqueda. El usuario accede directamente a la respuesta, podría no llegar a tu sitio. Pero si el modelo te cita como fuente, tu marca se convierte en parte de esa respuesta. Si no mides esta tasa de citas, estás perdiendo visibilidad sin saberlo.",{"type":32,"tag":40,"props":41,"children":43},"h2",{"id":42},"qué-es-citation-y-por-qué-es-crítico-ahora",[44],{"type":37,"value":45},"Qué es Citation y Por Qué es Crítico Ahora",{"type":32,"tag":33,"props":47,"children":48},{},[49],{"type":37,"value":50},"Una cita de LLM es cuando un modelo de lenguaje menciona tu marca, contenido o dominio como fuente al generar una respuesta. En SEO clásico contabas backlinks; ahora la pregunta es \"¿cuántas veces me mencionó el modelo?\". Si ChatGPT responde una pregunta técnica con \"la arquitectura de medición server-side de Roibase\", esa es una cita. Si Perplexity muestra una fuente inline, ese link alimenta tu brand equity.",{"type":32,"tag":33,"props":52,"children":53},{},[54],{"type":37,"value":55},"¿Por qué es crítico? Porque el comportamiento de búsqueda está cambiando. Los datos de Statcounter Q1 2026 muestran que hacer preguntas directas a herramientas de chat de IA alcanzó el 18% (fue 6% en Q1 2024). Las búsquedas generadas por IA de Google ahora están activas en el 40% de los resultados. El usuario no mira 10 enlaces azules para una pregunta \"cómo hacer algo\", busca un párrafo de respuesta. Ser citado como fuente en esa respuesta podría valer más que el tráfico tradicional — porque genera una señal de confianza.",{"type":32,"tag":33,"props":57,"children":58},{},[59],{"type":37,"value":60},"Las métricas de SEO clásicas (impresiones, CTR, posición) no son válidas en el entorno de LLM. Un usuario pregunta a ChatGPT \"¿cuál es el mejor CDP para comercio headless?\", el modelo sugiere 3 marcas. ¿Fue mencionada la tuya? ¿En cuántas consultas similares aparece tu nombre? Sin estos datos, tu análisis de visibilidad está incompleto.",{"type":32,"tag":40,"props":62,"children":64},{"id":63},"cómo-configurar-el-seguimiento-de-citas",[65],{"type":37,"value":66},"Cómo Configurar el Seguimiento de Citas",{"type":32,"tag":33,"props":68,"children":69},{},[70],{"type":37,"value":71},"Medir citas de LLM requiere un enfoque basado en API con sondeos automatizados. Las pruebas manuales no escalan — no puedes revisar manualmente si tu marca se menciona en 50 combinaciones de palabras clave en 3 modelos diferentes. Necesitas automatización. Aquí están las capas:",{"type":32,"tag":33,"props":73,"children":74},{},[75,81],{"type":32,"tag":76,"props":77,"children":78},"strong",{},[79],{"type":37,"value":80},"Capa 1: Construir tu pool de palabras clave.",{"type":37,"value":82}," Toma las palabras clave que ya obtienes de Google Search Console. Pero conviértelas al formato para preguntas de LLM. En lugar de \"roibase first party data\", convierte a \"¿cómo construir una arquitectura de datos first-party?\". Porque los usuarios hacen preguntas a los modelos, no queries de motor de búsqueda. Si tienes 100 palabras clave, conviértelas en 100 preguntas.",{"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},"Capa 2: Configurar sondeos de API.",{"type":37,"value":92}," Envías cada pregunta a las APIs de ChatGPT, Claude y Gemini. Obtienes la respuesta y la escaneas — ¿aparece el nombre de tu marca, tu URL de sitio, el nombre de tu producto\u002Fservicio? Usa regex o similitud de embedding. Para Perplexity, que proporciona citas inline, verifica si tu dominio está en el array ",{"type":32,"tag":94,"props":95,"children":97},"code",{"className":96},[],[98],{"type":37,"value":99},"sources",{"type":37,"value":101},". Para ChatGPT, aunque no genera referencias en forma de nota al pie, si la búsqueda web está habilitada, revisa ",{"type":32,"tag":94,"props":103,"children":105},{"className":104},[],[106],{"type":37,"value":107},"search_results",{"type":37,"value":109}," en los metadatos para ver si tu sitio está incluido.",{"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},"Capa 3: Agregación de logs.",{"type":37,"value":119}," Escribe cada resultado de sondeo en una base de datos de series temporales (InfluxDB, TimescaleDB, o BigQuery). Esquema: ",{"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},". Sin estos datos no puedes ver tendencias.",{"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","# Ejemplo de sondeo simplificado (API de ChatGPT)\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# Uso\nresult = check_citation(\n    \"¿Cómo construir una arquitectura de datos first-party?\",\n    [\"Roibase\", \"roibase.com.tr\"]\n)\nprint(result)\n","python",[135],{"type":32,"tag":94,"props":136,"children":137},{"__ignoreMap":16},[138,150,166,176,196,215,240,288,297,325,334,358,382,444,452,483,491,500,518,531,559,568],{"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},"# Ejemplo de sondeo simplificado (API de ChatGPT)\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":243},{"class":141,"line":242},7,[244,249,253,258,263,268,273,278,283],{"type":32,"tag":139,"props":245,"children":246},{"style":221},[247],{"type":37,"value":248},"        messages",{"type":32,"tag":139,"props":250,"children":251},{"style":156},[252],{"type":37,"value":209},{"type":32,"tag":139,"props":254,"children":255},{"style":162},[256],{"type":37,"value":257},"[{",{"type":32,"tag":139,"props":259,"children":260},{"style":231},[261],{"type":37,"value":262},"\"role\"",{"type":32,"tag":139,"props":264,"children":265},{"style":162},[266],{"type":37,"value":267},": ",{"type":32,"tag":139,"props":269,"children":270},{"style":231},[271],{"type":37,"value":272},"\"user\"",{"type":32,"tag":139,"props":274,"children":275},{"style":162},[276],{"type":37,"value":277},", ",{"type":32,"tag":139,"props":279,"children":280},{"style":231},[281],{"type":37,"value":282},"\"content\"",{"type":32,"tag":139,"props":284,"children":285},{"style":162},[286],{"type":37,"value":287},": keyword_question}]\n",{"type":32,"tag":139,"props":289,"children":291},{"class":141,"line":290},8,[292],{"type":32,"tag":139,"props":293,"children":294},{"style":162},[295],{"type":37,"value":296},"    )\n",{"type":32,"tag":139,"props":298,"children":299},{"class":141,"line":26},[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":267},{"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":277},{"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":267},{"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},"# Uso\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},"    \"¿Cómo construir una arquitectura de datos first-party?\"",{"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":277},{"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},"En una configuración real necesitas procesamiento por lotes — en lugar de enviar 500 palabras clave en línea, usa colas asincrónicas. Añade gestión de límite de velocidad, lógica de reintentos, seguimiento de costos. Cada llamada a API cuesta 0.01–0.03$ (según el modelo y tokens), lo que suma alrededor de 150$ mensuales en costos de sondeo (500 palabras clave × 3 modelos × 10 pruebas\u002Fmes).",{"type":32,"tag":40,"props":588,"children":590},{"id":589},"define-tu-conjunto-de-métricas",[591],{"type":37,"value":592},"Define Tu Conjunto de Métricas",{"type":32,"tag":33,"props":594,"children":595},{},[596],{"type":37,"value":597},"¿Qué números rastrearás en el seguimiento de citas? En lugar de \"posición\" y \"CTR\" de tu dashboard de SEO clásico, vienen estos:",{"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}," El porcentaje de palabras clave probadas en las que tu marca fue citada, del total de palabras clave. Probaste 100 palabras clave, tu marca apareció en 18 → 18% de tasa de citas. Es como \"share of voice\" pero en el mundo de LLM.",{"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},"Share por Modelo:",{"type":37,"value":617}," Podrías tener 22% de tasa de citas en ChatGPT, 14% en Claude, 9% en Gemini. Las diferencias por modelo son normales — los datos de entrenamiento, mecanismos de recuperación y ajuste de prompts difieren. Saber en cuál modelo eres fuerte orienta tu estrategia de ",{"type":32,"tag":619,"props":620,"children":624},"a",{"href":621,"rel":622},"https:\u002F\u002Fwww.roibase.com.tr\u002Fes\u002Fgeo",[623],"nofollow",[625],{"type":37,"value":626},"Optimización para Motores Generativos",{"type":37,"value":628},".",{"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}," ¿En qué lugar aparece tu marca en la respuesta del modelo? ¿En los primeros 3 sugeridos o en la sección \"otras opciones\"? La posición importa — los usuarios se enfoque generalmente en las primeras 2–3 fuentes.",{"type":32,"tag":33,"props":640,"children":641},{},[642,647],{"type":32,"tag":76,"props":643,"children":644},{},[645],{"type":37,"value":646},"Citation Context Quality Score:",{"type":37,"value":648}," Tu marca fue citada, pero ¿en qué contexto? \"Agencias como Roibase\" versus \"la solución server-side GTM de Roibase\" llevan diferentes valores de equity. Analiza el snippet semánticamente (polaridad positiva\u002Fneutral\u002Fnegativa + grado de especificidad).",{"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}," ¿Cuál es la tasa de citas de competidores en las mismas palabras clave? Si \"CDP para first-party data\" genera menciones para Segment, mParticle y Roibase, compartes un mercado de 3. ¿Está creciendo tu parte con el tiempo?",{"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},"Métrica",{"type":32,"tag":672,"props":678,"children":679},{},[680],{"type":37,"value":681},"Definición",{"type":32,"tag":672,"props":683,"children":684},{},[685],{"type":37,"value":686},"Valor Objetivo",{"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},"% de palabras clave mencionadas",{"type":32,"tag":695,"props":706,"children":707},{},[708],{"type":37,"value":709},">15% (según líder de categoría)",{"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},"% mencionado en primer lugar",{"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},"% de snippets de contexto positivo",{"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},"Citas vs. competidores",{"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},"Incluye estas métricas en tu dashboard semanal. Gráfico de tendencia: eje X es tiempo, eje Y es tasa de citas. Después de publicar contenido, deberías ver un aumento en la tasa de citas dentro de 2–4 semanas (hay lag de indexación en los modelos).",{"type":32,"tag":40,"props":770,"children":772},{"id":771},"optimiza-tu-estrategia-de-contenido-alrededor-de-citation",[773],{"type":37,"value":774},"Optimiza Tu Estrategia de Contenido Alrededor de Citation",{"type":32,"tag":33,"props":776,"children":777},{},[778,780,785],{"type":37,"value":779},"Si tu tasa de citas es baja, ¿qué haces? El enfoque clásico de SEO \"obtén más backlinks\" no funciona aquí. Los LLM no cuentan backlinks (al menos no directamente). En su lugar: ",{"type":32,"tag":76,"props":781,"children":782},{},[783],{"type":37,"value":784},"profundidad del contenido, datos estructurados, señales de autoridad",{"type":37,"value":628},{"type":32,"tag":33,"props":787,"children":788},{},[789,794],{"type":32,"tag":76,"props":790,"children":791},{},[792],{"type":37,"value":793},"Profundidad:",{"type":37,"value":795}," Los LLM no ignoran contenido superficial, pero responden a \"¿esta fuente es exhaustiva?\". En lugar de un blog de 500 palabras, escribe una guía técnica de 2.000 palabras. Incluye ejemplos de código, tablas, instrucciones paso a paso. El modelo recibe la señal \"este contenido es accionable\".",{"type":32,"tag":33,"props":797,"children":798},{},[799,804,806,812,813,819,820,826,828,833],{"type":32,"tag":76,"props":800,"children":801},{},[802],{"type":37,"value":803},"Datos Estructurados:",{"type":37,"value":805}," El markup de Schema.org facilita que los LLM analicen tu contenido. Añade ",{"type":32,"tag":94,"props":807,"children":809},{"className":808},[],[810],{"type":37,"value":811},"Article",{"type":37,"value":277},{"type":32,"tag":94,"props":814,"children":816},{"className":815},[],[817],{"type":37,"value":818},"HowTo",{"type":37,"value":277},{"type":32,"tag":94,"props":821,"children":823},{"className":822},[],[824],{"type":37,"value":825},"FAQPage",{"type":37,"value":827},". Especialmente ",{"type":32,"tag":94,"props":829,"children":831},{"className":830},[],[832],{"type":37,"value":825},{"type":37,"value":834}," — el modelo puede extraer pares pregunta-respuesta directamente.",{"type":32,"tag":33,"props":836,"children":837},{},[838,843],{"type":32,"tag":76,"props":839,"children":840},{},[841],{"type":37,"value":842},"Autoridad:",{"type":37,"value":844}," Bio del autor, información de la organización, fecha de publicación. El modelo puede detectar \"esto se escribió en 2023, está desactualizado\". Existe sesgo de contenido fresco. Actualiza contenido antiguo con fechas de revisión.",{"type":32,"tag":33,"props":846,"children":847},{},[848,853],{"type":32,"tag":76,"props":849,"children":850},{},[851],{"type":37,"value":852},"Trade-off:",{"type":37,"value":854}," Optimizar para citation no significa sacrificar tráfico, pero las prioridades cambian. Ejemplo: la palabra clave genérica \"plugins de Shopify\" genera tráfico pero tiene baja tasa de citas (el modelo genera su propia lista). \"Seguimiento de checkout en Shopify server-side\" genera menos tráfico pero mayor tasa de citas (menos fuentes, la tuya es profunda). Equilibra — dedica 60% del esfuerzo a palabras clave de tráfico, 40% a palabras clave de citation.",{"type":32,"tag":40,"props":856,"children":858},{"id":857},"vincula-citation-data-a-tu-pipeline-de-attribution",[859],{"type":37,"value":860},"Vincula Citation Data a Tu Pipeline de Attribution",{"type":32,"tag":33,"props":862,"children":863},{},[864],{"type":37,"value":865},"No aísles el seguimiento de citas. Intégralo con tu atribución de marketing clásica. Porque un usuario ve tu marca en ChatGPT, espera 2 días, después busca en Google y te visita. Si no conectas este viaje, no ves la contribución del LLM.",{"type":32,"tag":33,"props":867,"children":868},{},[869,874,876,882],{"type":32,"tag":76,"props":870,"children":871},{},[872],{"type":37,"value":873},"Etiquetado UTM:",{"type":37,"value":875}," Si Perplexity proporciona un enlace directo, etiquétalo con UTM (",{"type":32,"tag":94,"props":877,"children":879},{"className":878},[],[880],{"type":37,"value":881},"utm_source=perplexity&utm_medium=citation",{"type":37,"value":883},"). En Google Analytics verás tráfico proveniente de \"perplexity\". Pero ChatGPT no proporciona enlaces, solo menciona la marca — no hay atribución directa allí.",{"type":32,"tag":33,"props":885,"children":886},{},[887,892],{"type":32,"tag":76,"props":888,"children":889},{},[890],{"type":37,"value":891},"Lift de Búsqueda de Marca:",{"type":37,"value":893}," ¿Aumenta el volumen de búsqueda de marca cuando sube tu tasa de citation? Monitorea tus palabras clave de marca en Google Trends o Search Console. Si tu tasa de citation en ChatGPT alcanza 25% durante 3 meses, podrías ver +15% en búsquedas de marca. No es atribución perfecta, pero es una señal fuerte.",{"type":32,"tag":33,"props":895,"children":896},{},[897,902],{"type":32,"tag":76,"props":898,"children":899},{},[900],{"type":37,"value":901},"Atribución por Encuesta:",{"type":37,"value":903}," Añade a tu encuesta de usuario la opción \"¿Dónde nos descubriste?\" con \"Chatbot de IA (ChatGPT, Perplexity, etc.)\". Pequeña muestra, pero datos direccionales.",{"type":32,"tag":33,"props":905,"children":906},{},[907,912,914,920,922,929],{"type":32,"tag":76,"props":908,"children":909},{},[910],{"type":37,"value":911},"Eventos First-Party Tracking:",{"type":37,"value":913}," Cuando un usuario llega a tu sitio sin referrer pero la página de destino es de IA (ej. ",{"type":32,"tag":94,"props":915,"children":917},{"className":916},[],[918],{"type":37,"value":919},"\u002Fblog\u002Fllm-citation",{"type":37,"value":921},"), eso es una señal indirecta. Con ",{"type":32,"tag":619,"props":923,"children":926},{"href":924,"rel":925},"https:\u002F\u002Fwww.roibase.com.tr\u002Fes\u002Ffirstparty",[623],[927],{"type":37,"value":928},"arquitectura First-Party Data & Medición",{"type":37,"value":930},", puedes consolidar estas señales en tu CDP y crear un segmento \"exposición a IA\" en el customer journey.",{"type":32,"tag":40,"props":932,"children":934},{"id":933},"riesgos-y-puntos-ciegos",[935],{"type":37,"value":936},"Riesgos y Puntos Ciegos",{"type":32,"tag":33,"props":938,"children":939},{},[940,942,947],{"type":37,"value":941},"¿Cuáles son los límites del seguimiento de citas en LLM? Primero: ",{"type":32,"tag":76,"props":943,"children":944},{},[945],{"type":37,"value":946},"sesgo de muestreo",{"type":37,"value":948},". Pruebas 500 palabras clave, pero los usuarios reales hacen 50.000 preguntas diferentes. Tu conjunto de pruebas podría no ser representativo. Solución: extrae tu pool de palabras clave de Search Console y conviértelas en plantillas de prompts — así proxies la demanda real.",{"type":32,"tag":33,"props":950,"children":951},{},[952,954,959],{"type":37,"value":953},"Segundo: ",{"type":32,"tag":76,"props":955,"children":956},{},[957],{"type":37,"value":958},"cambios de modelo",{"type":37,"value":960},". ChatGPT te cita hoy, pero dentro de 2 semanas hay una actualización del modelo — tu tasa de citation cae de 18% a 9%. Es como una actualización de algoritmo — no lo controlas. La única defensa: diversificación multi-modelo. No dependas solo de ChatGPT; obtén citas en Claude, Gemini y Perplexity.",{"type":32,"tag":33,"props":962,"children":963},{},[964,966,971],{"type":37,"value":965},"Tercero: ",{"type":32,"tag":76,"props":967,"children":968},{},[969],{"type":37,"value":970},"costo",{"type":37,"value":972},". 500 palabras clave × 3 modelos × 4 semanas = 6.000 llamadas a API. A 0.02$ por llamada, eso es 120$\u002Fmes. Para una startup es tolerable, pero en enterprise con 5.000 palabras clave, subes a 1.200$\u002Fmes. Si el presupuesto es limitado, estratifica tus palabras clave — Tier 1 (alto valor, prueba semanal), Tier 2 (valor medio, prueba mensual).",{"type":32,"tag":33,"props":974,"children":975},{},[976,978,983],{"type":37,"value":977},"Cuarto: ",{"type":32,"tag":76,"props":979,"children":980},{},[981],{"type":37,"value":982},"falsos positivos",{"type":37,"value":984},". Buscas \"Roibase\" con regex, el modelo dice \"agencias pequeñas como Roibase\". ¿Técnicamente es citation? Sí. ¿Pero con valor de equity? Cero. El Citation Context Quality Score lo resuelve — no solo cuentes menciones, añade puntuación de sentimiento + especificidad.",{"type":32,"tag":40,"props":986,"children":988},{"id":987},"qué-hacer-ahora",[989],{"type":37,"value":990},"Qué Hacer Ahora",{"type":32,"tag":33,"props":992,"children":993},{},[994],{"type":37,"value":995},"El seguimiento de citas de LLM aún no es mainstream, pero para 2027 será una métrica estándar. Si comienzas pronto, estableces baseline — cuando tus competidores comiencen, tú ya ves tendencias. Primer paso: toma 50 palabras clave críticas, conviértelas en prompts de preguntas, prueba manualmente en ChatGPT + Perplexity. ¿Cuántas veces aparece tu marca? ¿En qué contexto? Esto toma 2 horas y te da el estado actual. Siguiente paso: automatiza el sondeo de API. Usa workflow n8n o script Python, obtén reportes semanales. Si tu tasa de citation es baja, aumenta la profundidad del contenido y datos estructurados. Si es alta, vincula a tu pipeline de atribución y mide el lift de marca. LLM citation es la nueva frontera del SEO — no se trata solo de estar en posición 1 de Google, sino de ser mencionado en la respuesta de ChatGPT.",{"type":32,"tag":997,"props":998,"children":999},"style",{},[1000],{"type":37,"value":1001},"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":1003},[1004,1005,1006,1007,1008,1009,1010],{"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":857,"depth":152,"text":860},{"id":933,"depth":152,"text":936},{"id":987,"depth":152,"text":990},"markdown","content:es:ai:medicion-de-citas-en-llm-nueva-metrica-seo.md","content","es\u002Fai\u002Fmedicion-de-citas-en-llm-nueva-metrica-seo.md","es\u002Fai\u002Fmedicion-de-citas-en-llm-nueva-metrica-seo","md",1782079498290]