{
  "kind": "answer",
  "studySlug": "citation-decay",
  "slug": "how-should-teams-interpret-short-lived-citation-spikes",
  "title": "How should teams interpret short-lived citation spikes?",
  "description": "Interpret them cautiously. A spike can be real and still fail to persist because most citations are one-and-done and the median URL lifespan is 0 days.",
  "lastUpdated": "2026-03-30",
  "lastTested": "2026-03-30",
  "sourceStudyUrl": "/trakkr-research/citation-decay",
  "sourceStudyTitle": "The Half-Life of AI Citations",
  "claimIds": [
    "citation-decay:one-and-done",
    "citation-decay:median-life",
    "citation-decay:still-active"
  ],
  "relatedSlugs": [
    "answer:how-long-do-ai-citations-last",
    "answer:what-does-a-thirty-one-day-brand-half-life-actually-mean",
    "fact:brand-level-ai-presence-halves-in-about-thirty-one-days",
    "tracker:visibility-stability-tracker"
  ],
  "methodologySummary": "Built from 857,138 reports, 108,650 citations, 10,991 brands, and 8 tracked models across a 10-month observation window.",
  "limitations": [
    "Decay describes observed persistence, not the full causal mechanism behind why a citation disappears.",
    "Brand- and domain-level averages can hide very different retention patterns by query class or model.",
    "Some of the stickiest or most volatile domains come from niche query pockets and should be read as examples, not universal leaders."
  ],
  "keywords": [
    "citation decay",
    "AI half life",
    "AI citation persistence",
    "AI visibility churn",
    "citation spikes AI",
    "interpreting volatility"
  ],
  "schemaHints": {
    "pageType": "Article",
    "includeDataset": true
  },
  "question": "How should teams interpret short-lived citation spikes?",
  "directAnswer": "Mostly with caution. A spike can be real and still fail to persist because 72.8% of citations are one-and-done and the median URL lifespan is 0 days.",
  "answerSummary": "The right question is not whether a spike occurred, but whether the visibility survived repeated observation over time.",
  "keyFacts": [
    {
      "label": "One-and-done citations",
      "value": "72.8%",
      "detail": "Citations that appear once and vanish.",
      "claimId": "citation-decay:one-and-done"
    },
    {
      "label": "Median citation lifespan",
      "value": "0 days",
      "detail": "Most citations disappear immediately.",
      "claimId": "citation-decay:median-life"
    },
    {
      "label": "Still active citations",
      "value": "6.8%",
      "detail": "Small share of URLs still active at the end of the window.",
      "claimId": "citation-decay:still-active"
    }
  ],
  "evidenceTable": [
    {
      "label": "One-and-done citations",
      "value": "72.8%",
      "note": "Citations that appear once and vanish."
    },
    {
      "label": "Median citation lifespan",
      "value": "0 days",
      "note": "Most citations disappear immediately."
    },
    {
      "label": "Still active citations",
      "value": "6.8%",
      "note": "Small share of URLs still active at the end of the window."
    }
  ],
  "whyItMatters": "This turns study findings into an operating rule teams can use to decide what to publish, refresh, or measure next, avoiding overinvestment in transient visibility.",
  "whatToDo": [
    "Measure visibility as a moving system rather than a one-time citation snapshot.",
    "Refresh and monitor citation-driving pages on the cadence your models actually decay.",
    "Separate durable wins from temporary spikes to prevent overreacting to short-lived mentions."
  ],
  "faqs": [
    {
      "question": "What percentage of citations disappear immediately?",
      "answer": "According to the study, 72.8% of citations are one-and-done, resulting in a median citation lifespan of 0 days."
    },
    {
      "question": "How many citations remain active over the long term?",
      "answer": "Only 6.8% of citations are still active at the end of the observation window."
    }
  ]
}