Technical SEO
Keyword Cannibalization Checker
List the pages on your own site — URL, title, meta description, H1 and target keyword — and find every pair that overlaps enough to compete with each other in search results, ranked by risk.
Format: URL | Title | Meta description | H1 | Target keyword (one page per line, separated by " | "). Only URL and Title are required — meta description, H1 and target keyword are optional but sharpen the result.
Matching is stopword-filtered word overlap (Jaccard), so it works for any Latin-script language. Languages written without spaces between words (e.g. Chinese, Thai, Khmer) get a coarser signal — treat the score as directional for those.
What this tool checks
Keyword cannibalization happens when two or more of your own pages target the same search intent — the same keyword, near-identical title, or overlapping H1 and meta description. Search engines then have to pick one to rank, often splitting authority and links between them and hurting both. This tool compares every pair of pages you list and flags the ones likely to be cannibalizing each other, with the specific fields that overlap.
How the score is calculated
Titles, H1s, meta descriptions and target keywords are each tokenized and stripped of common stopwords, then compared pairwise with the Jaccard index (shared words divided by total unique words). The four field scores are combined into one weighted cannibalization score per pair, with an exact target-keyword or exact-title match forcing the pair straight to critical.
Title & H1 overlap
Titles carry the most ranking weight, so title overlap is weighted heaviest, followed by H1. Two pages with near-identical titles are telling both users and search engines the same story.
Target keyword match
If you list a target keyword per page, an exact match (after stripping "a/an/the") between two different URLs is the strongest possible cannibalization signal and is flagged as critical regardless of the other scores.
Meta description overlap
Meta descriptions do not directly affect rankings, but heavy overlap usually means the pages were written to answer the same query — a secondary confirming signal.
Combined risk score
Each pair gets one 0-100% score, blending whichever fields you supplied, then a severity: critical (exact keyword or title match), high (≥55%), or moderate (≥30%). Pairs below that are not shown — negligible overlap is normal on a real site.
Frequently asked questions
What should I do about a cannibalization pair the tool flags?
Pick the page that ranks better or has more backlinks and expand it to fully cover the intent, then either 301-redirect the weaker page to it, add a canonical tag, or rewrite the weaker page around a genuinely different angle or sub-keyword and cross-link the two.
Is some overlap between pages always a problem?
No. A pillar page and a supporting article can legitimately share some vocabulary. The tool is tuned to only flag pairs with substantial overlap (30%+, or an exact keyword/title match) — that threshold is a signal worth reviewing, not an automatic verdict.
Where do I get the title, meta description and H1 for each page?
View each page's source (or use this site's "View Page Source Online" and "Get HTML Meta Header" tools) and copy the <title>, meta description and <h1> text. For a handful of pages this only takes a minute per page.
Does this tool send my data anywhere?
The list you paste is sent once, over HTTPS, to our server to run the pairwise comparison (needed to keep the matching logic server-side and consistent), and is not stored beyond the request.