Keyword Cannibalization
A long-standing traditional-SEO failure mode: multiple pages on one site target the same query, so they compete against each other and split ranking signals (links, relevance, click-through) instead of concentrating them on a single strong page. The classic fix is consolidation — merge or canonicalize the redundant pages into one.
The LLM-era generalization
publishing-volume-hurts-seo argues the concept broadens under AI retrieval from keyword overlap to semantic / vector competition: near-duplicate pages now cannibalize each other in embedding space, so “your pages are not just competing for rankings — they are competing for embeddings.” The remedy is the same (consolidate), but the diagnosis widens — two pages need not share a keyword to cannibalize, only to occupy the same semantic region. This is one of the spoke’s clean traditional → AI-era bridges: an established SEO problem re-expressed in the vocabulary of authority-density and GEO.
Related
publishing-volume-hurts-seo · authority-density · search-marketing