Some of the most stubborn ranking problems are not caused by competitors, penalties or technical errors. Both pages involved are healthy, indexed and well written. The problem is that they are on the same site and aim at the same search phrase — and the search engine keeps changing its mind about which one to show. This is called keyword cannibalisation.
This article explains how to recognise it, why it happens and how to fix it without losing the work already put into either page.
What cannibalisation is
Cannibalisation happens when two or more pages of the same site compete for the same query with the same intent. Instead of one strong page, the site offers the search engine two weaker candidates. The search engine picks one, then another, and neither settles into a stable position.
Two pages ranking for the same phrase is not always a problem. If a site holds positions one and two for a query, that is a success. Cannibalisation is the situation where the pages take turns, or where both sit lower than one combined page would.
Symptoms
- The ranking URL keeps switching. For the same query, the result shows one page this week and another next week.
- Positions jump without explanation. A query moves between, say, the second and the fifth page of results with no changes on the site.
- The wrong page ranks. An old blog post appears where the service page should be, and visitors land on content that does not lead to an enquiry.
- Neither page reaches the top. Both stay in the middle of the results despite good content.
The first symptom is the most reliable. A ranking that changes address is a clear signal that the search engine sees two candidates for the same intent.
Why it happens
- Content written over time without a plan. A blog post from two years ago and a new guide cover the same topic because nobody checked what already existed.
- Category and product pages overlap. A category page and its main product page target the same product name.
- Location or variant pages that differ only in a word. "Plumber in North London" and "Plumber in London North" are the same page to a search engine.
- Technical duplicates. The same page available at several addresses — with and without parameters, with and without a trailing slash.
- Weak internal linking. When the site itself links to both pages with the same anchor text, it gives no signal about which one is the main page.
How to fix it
Start by deciding, for each affected query, which page should rank. Choose the page that best matches the intent and is most useful for the business — often the one closer to a conversion. Then choose one of the following approaches.
1. Merge and redirect
If the pages cover the same topic, combine the useful content into the chosen page and permanently redirect (301) the other address to it. This is usually the strongest fix: links and history from both pages are consolidated into one.
2. Differentiate
If both pages have a reason to exist, make their topics clearly different. A general guide and a pricing page can both live on a site if one explains the subject and the other is about cost. Adjust titles, headings and content so each answers a distinct question.
3. Fix internal links
Link to the chosen page with the target phrase from other relevant pages, and stop using that phrase as anchor text for the competing page. Internal links are one of the clearest ways a site tells search engines which page is the main one for a topic.
4. Use canonical tags for true duplicates
When the pages are technically the same content at different addresses, a canonical tag on the duplicates pointing to the main version tells search engines which address to keep. Canonical tags are a hint, not a command, and they are not a substitute for merging pages that are genuinely different.
What not to do
- Do not delete a page without redirecting it — its links and visitors are lost.
- Do not add
noindexto one page as a quick fix if it has useful links; merging is usually better. - Do not rewrite both pages at once and hope; change one thing and observe.
Preventing it
The simplest prevention is a map of topics and target pages. Before writing new content, check which page is already assigned to the topic. Grouping keywords by intent and assigning one page per group turns a growing site into a planned one.
Finding cannibalisation with SEO Melody
Many tools detect cannibalisation only from Search Console data, which requires connecting your Google account and enough traffic to produce it. SEO Melody takes a different route.
Page Conflicts detects two of your own pages competing for the same phrase from search results data collected while tracking your positions. It works without Search Console and shows which pages are involved for each phrase.
Positions records events for every tracked query, including the moment the ranking URL changes. A query whose URL keeps changing is the most direct evidence of cannibalisation, and the event history shows how often it happens.
For prevention, Clusters groups your keywords into topics and assigns a target page to each group, and Internal Links suggests which pages should link to the chosen page.
Summary
Cannibalisation is a quiet problem: no errors, no warnings, just rankings that will not settle. Watch for ranking URLs that switch, decide which page should own each phrase, and consolidate or differentiate the rest. One clear page per intent is almost always stronger than two competing ones.