Keyword cannibalisation is several pages on your own site competing for the same query, so none ranks as well as one strong page would. The mechanism is less dramatic than the name suggests, and the name causes a lot of unnecessary surgery.
What actually happens
Google picks one page per site for most queries anyway, so your pages are not fighting each other into oblivion. The harm is quieter: internal links, external links and topical signals get split across 3 mediocre pages instead of concentrating into one good one, and Google sometimes picks the page you didn’t want.
The telltale pattern in Search Console is a query where the ranking URL keeps swapping between 2 or 3 of your own pages, with mediocre positions throughout and no version ever settling.
How to diagnose it properly
- Open the Search Console performance report and filter to a specific query.
- Switch to the Pages tab. If several of your URLs appear against that one query, note them.
- Compare position over time. Instability plus mediocrity is the signature. A stable winner with other pages picking up scraps is usually fine.
- Read the competing pages. Ask honestly whether they serve different intents or just repeat each other.
Do this for your commercially important queries rather than running a sitewide cannibalisation audit, which mostly generates work.
Not every overlap is a problem
A guide and a product page ranking for related terms are serving different intents, and that’s healthy coverage rather than cannibalisation. A category page and a blog post about the category can coexist. The problem is 3 blog posts each trying to be the definitive answer to the same question.
Genuine cannibalisation has a specific shape: same intent, same query, no meaningful difference in what the pages offer.
The 4 fixes, in order of how often they’re right
| Fix | When | Effort |
|---|---|---|
| Fix internal links | Anchors point inconsistently at competing pages | Low, and resolves a lot of cases alone |
| Consolidate and 301 | Pages genuinely duplicate each other | Medium, and usually the right answer |
| Differentiate | Each page has a real separate purpose | High: requires rewriting both |
| Noindex the weakest | The page exists for users but shouldn’t compete | Low |
Internal links first is deliberate. Inconsistent anchor text pointing at 2 different pages for the same concept is the most common underlying cause, and it’s the cheapest to fix.
Common mistakes
- Over-diagnosing it. Plenty of sites are told they have a cannibalisation problem when they have a quality problem, and merging 2 thin pages produces 1 slightly longer thin page.
- Consolidating different intents because the pages share a word.
- Redirecting without checking traffic. The page you judged weaker may be the one earning conversions.
- Treating it as a one-off audit. It reappears every time someone publishes without checking what already exists.
Related
Internal linking covers the most common underlying cause. Canonical tags handle the duplicate-URL version of this problem, which is a different thing with a similar smell.