Keyword cannibalization finder (lite)
Reads a Search Console export with both query and page columns and lists the queries where more than one of your URLs gets a real share of impressions.
When to use
- "Are my pages cannibalizing each other?", "two pages rank for the same keyword", "why does Google swap which page ranks?"
- The user has a GSC performance export (or a Looker Studio / API export) at query + page level.
How to run
- Export from GSC: Performance → add both Query and Page dimensions (the API, Looker Studio or a GSC export tool; the UI alone exports one dimension at a time). Columns needed: query, page, clicks, impressions, position. Names are auto-detected.
- Run:
python3 scripts/cannibalization.py gsc.csv python3 scripts/cannibalization.py gsc.csv --min-impressions 50 --min-share 15 --exclude "brand name" - Present the table. For each HIGH row, recommend one action: merge and 301 the weaker page, de-optimise its title/H1 for the query, or differentiate intent. Ask the user before suggesting a redirect of a page that earns clicks.
Severity
- HIGH: the second page gets ≥30% of the query's impressions and ranks within 10 positions of the first.
- MEDIUM: the second page gets ≥20%.
- LOW: anything else above the thresholds (often harmless, such as a homepage plus a product page).
Limits
It does not fetch the pages or judge intent overlap, and it does not choose a fix for each page. Some splits are fine (different intents served by different pages), so treat LOW as informational.
Example
best running shoes → primary /best-running-shoes (6.2) and /blog/running-shoes-guide (14.8, 31% of impressions) → HIGH.
Want a fix per page?
- Paid skill that gives a merge, canonical or differentiate action for each competing page: Fix Keyword Cannibalization Using Google Search Console on Agensi
- Hosted version: Keyword Cannibalization Finder on Apify
- Free MCP server with a
find_keyword_cannibalizationtool: cygnix-seo-mcp