Definition
The term describes an outcome, not a rule the search engine enforces. Google has no feature by that name and documents no penalty for it. What happens is mechanical: for most queries the results show one page per site, so when a site offers several candidates of similar relevance the engine's choice between them is unstable. The page shown changes from week to week, neither page accumulates the full weight of the site's internal and external links, and the listing's title and snippet keep changing, which costs clicks.
The intent test
Three conditions together mark genuine cannibalisation:
- Same intent. The pages would satisfy the same searcher with the same need, judged from the results page rather than from the words on the pages.
- Alternating URLs. The site's ranking page for the query changes between periods, or several pages share its impressions.
- Split performance. Clicks and impressions are divided between the pages, and the best position is worse than the site's strength in the topic would suggest.
If only the first holds, there is a risk to watch. If the pages serve different intents, there is no problem to fix, whatever the keyword overlap.
In practice
Search Console's performance report is where the signs appear. Filtered to a single query, it lists every page on the site that received impressions for it; several pages with meaningful impressions, or a top page that changes between date ranges, is the signature. Exported page-and-query data grouped by query shows the same pattern across many queries at once. Rank tracking adds the time dimension, and a site crawl finds candidates by similar titles and headings, though similar titles alone prove nothing.
Prevention sits upstream of all of this. A keyword map assigns one owner URL to each cluster of queries, and keyword clustering decides which queries belong together because one page can satisfy them. The SEO page publishing checklist turns that into a single check before launch: the primary keyword maps to this page only.
When two pages for one keyword are fine
Overlap is normal on any site with depth. A product category page and a buying guide both mention the product type; a tool and the article that explains the metric it calculates share most of their keywords; a hub page and its supporting pages share the subject's vocabulary by design. In a well-built topic cluster the hub covers the subject broadly and each supporting page owns one narrower need, so their keyword overlap is deliberate. The problem appears only when a supporting page restates the hub, or two supporting pages answer the same narrow question.
Fixes in brief
Each competing group needs one of four decisions:
- Merge the useful content into the stronger page.
- Redirect the weaker page to it, so its links and history follow.
- Differentiate the pages onto distinct intents, rewriting one so it clearly serves another need.
- Re-link, so internal anchors for the query point at one owner only.
The keyword cannibalisation guide works through choosing between these, and the internal linking guide explains how identical anchors aimed at different pages reinforce the confusion.
A buying guide and a category page trading places
Tisanecrest Tea Merchants tracks "loose leaf green tea" and finds its ranking page alternating every week or two between the green tea category page and a buying guide, at positions between 7 and 14. Search Console splits the query's 3,800 monthly impressions roughly evenly between them, with 70 clicks in total. Both pages aim to help someone choose and buy a green tea, so they share an intent. The shop moves the guide's brewing and grade advice into the category page, redirects the guide to it and points its internal links at the category page. Over the next three months the category page holds position 4, and the query brings 210 clicks a month.
Related terms
Search IntentThe goal behind a search query: to learn, to find a site, to compare or to act. is the test that separates competing pages from complementary ones. An Internal LinkA link from one page to another page on the same website. with the query as its anchor tells the engine which page the site considers the owner, so inconsistent anchors make cannibalisation worse. A 301 RedirectA permanent server redirect telling browsers and search engines a URL has moved for good. is how a merged page hands its signals to the survivor. The Canonical TagTells search engines the preferred URL. handles a different case, exact or near-exact duplicates, and is not the right tool for two distinct pages that happen to compete.
Common confusions
- Shared keywords are not cannibalisation. Shared intent is.
- It is not a penalty. Nothing is imposed; the site simply wastes its own signals.
- Duplicates are a separate problem. Identical content is a canonicalisation issue with a technical fix.
- Not every case is worth fixing. Two old posts sharing a rare query may cost more to merge than they lose.
- Similar titles are a clue, not proof. Query data decides.
