How Keyword cannibalisation works
It tends to appear as sites grow: a blog post and a service page both chase "payroll outsourcing", or a new article quietly repeats an old one. Search engines usually show one or two URLs per site for a query, and when signals are split they may alternate between pages or pick the weaker one, which shows up in Search Console as one query with several of your URLs taking turns.
The common misconception is that any two pages sharing a keyword are cannibalising. They are not if the intent differs: a guide on how payroll outsourcing works and a page selling the service can both rank, because they meet different needs. Google has also said there is no penalty here; the cost is lost potential, not punishment. Real cannibalisation is overlap in intent, not just in vocabulary.
The fix depends on why the overlap exists. Merging and redirecting suits duplicated content, sharpening each page suits pages that should serve distinct intents, and a canonical tag suits true variants such as sorted versions of one list.
How to improve it
- 01Filter Google Search Console by a priority query and check whether several of your URLs receive impressions for it over time.
- 02For each overlapping pair, decide which page best serves the intent and should be the one that ranks.
- 03Merge useful material from the weaker page into the stronger one, then 301 redirect the old URL.
- 04Where both pages should exist, refocus their titles, headings and copy on clearly different intents.
- 05Point internal links with descriptive anchor text at the chosen page.
- 06Before publishing anything new, check whether an existing page already covers the query.
A fictional IT support company has three articles on managed IT costs written over several years, plus a pricing page, and all four rotate for the same query. It folds the best material from the articles into the pricing page, redirects two of them, and turns the third into a guide on choosing a provider that links to pricing.