Keyword cannibalization
When overlapping pages on your own site compete for one search intent, causing the wrong URL or a rotating set of URLs to rank.
See it
What it is
Keyword cannibalization is harmful overlap between your own pages for the same search intent. The engine cannot find a clear best answer, so it may rank the wrong URL, alternate between two URLs, or split links and relevance signals across both. It often appears after years of publishing near-identical guides, landing pages, and category pages.
Reach for the diagnosis when one query maps to several URLs and performance is unstable or the weaker page keeps winning. Choose one primary page, merge genuinely duplicate material into it, redirect retired URLs, and point internal links at the winner. If both pages deserve to exist, separate their intent clearly in the title, heading, copy, and anchor text instead of forcing both onto one keyword.
Gotcha: two pages ranking for the same phrase is not automatically a problem. A product page and a how-to guide can satisfy different needs, and owning two results can be useful. Fix evidence of conflict, not a spreadsheet match. Canonical tags are for duplicate or near-duplicate pages; they are not a substitute for deciding which distinct intent each page serves.
Ask AI for it
Find and fix keyword cannibalization using a Google Search Console Performance report export grouped by query and page. Flag queries where multiple URLs earn meaningful impressions, then distinguish healthy multi-page coverage from conflict by comparing intent, clicks, average position, ranking swaps, backlinks, and conversions. For each real conflict, name one primary URL and choose exactly one action for the others: merge and 301 redirect, retarget to a distinct intent, or apply rel="canonical" only when the content is genuinely duplicate. Rewrite titles, H1s, and internal anchor text to support the decision, and return a URL-by-URL migration table with no redirect chains.