Citation Decay

Citation Decay happens silently and without warning in traditional analytics tools. Content that is not maintained, updated or confirmed by other sources will gradually fall out of AI models' preferred citation sources – either because competitors produce better content, because prompt patterns shift, or because the information is perceived as outdated. In 3LA this is tracked via time-series data in brand_probe_runs with created_at timestamps, available as day or week buckets over a 30-day window. Aggregation is supported both per tenant and per individual prompt. A drop in mention rate exceeding 20% over three consecutive weeks should trigger a content review of pages linked to the relevant prompts. Citation Decay is an early symptom – the underlying cause may be Model Drift, competitor activity or content staleness.