What cannibalisation actually is
Keyword cannibalisation is when two or more pages on your site compete for the same query, and the competition makes both perform worse than one page would.
That last clause is the whole article. Two pages appearing for one term is not cannibalisation. It is frequently a good outcome — two results on page one takes space a competitor would otherwise have.
Cannibalisation is when the competition costs you something: Google keeps swapping which page it shows, neither page consolidates authority, and the ranking sits lower than one strong page would achieve.
The term gets applied to any overlap at all, which leads to pages being merged or deleted that were doing no harm.
The symptoms that are real
Four signals, and you want more than one before acting.
Google keeps changing which URL it ranks. The clearest sign. One week page A appears for the query, the next week page B. Each swap resets whatever position history had accumulated. In Search Console, filter to the query and look at pages over time — flip-flopping is visible immediately.
Both pages rank, and both rank poorly. Positions 14 and 19 for the same query, neither improving. Two pages splitting the signals one page would have held.
The wrong page ranks. Your thin blog post outranks your commercial page for a buying query. Traffic arrives and does not convert, because the page it lands on was not built to.
Rankings fell when you published something. A new article went up and an established page dropped for its main term shortly after. That is the most direct evidence available.
The symptoms that are not
Two pages appearing for one query. Often ideal. A guide and a service page both ranking for a term with mixed intent serves both kinds of searcher.
Several pages mentioning the same keyword. Mentioning is not targeting. Cannibalisation is about which page is built to *answer* a query, not which pages contain a phrase.
A drop after an algorithm update. Rankings move for many reasons. Before restructuring content, check whether the timing matches an update rather than anything you did.
Overlap between a category and a product page. These serve different intents — browse and buy — and normally coexist.
How to diagnose it properly
Three steps, all in Search Console, and none takes long.
1 · Find queries with multiple ranking URLs. Performance report → filter by query → look at the Pages tab for that query. More than one URL appearing is your candidate list, not your problem list.
2 · Check whether the ranking URL is unstable. For each candidate, compare two date ranges. If the same URL ranks in both, there is no swapping and probably no problem. If the URL changes between periods, that is the flip-flopping that costs you.
3 · Decide whether the pages actually serve the same intent. Open both. Ask what a searcher typing that query wants, and whether both pages try to give it to them. Two pages answering the same question compete. A guide and a product page usually do not.
The step people skip is 3, and it is the one that prevents needless merges.
What to do when it is real
Four options, in order of how much you should prefer them.
1 · Differentiate the pages. Often the cheapest fix. Re-target one page at a genuinely different angle — a different stage of the funnel, a different sub-question. Retitle it, rewrite the introduction, and re-point its internal links. Both pages survive and stop colliding.
2 · Consolidate into one page. When two pages genuinely say the same thing, merge the better material into the stronger URL and 301 the weaker one to it. This is the strongest fix because it stops splitting signals entirely. It is also irreversible in practice, so be sure first.
3 · Fix the internal links. Sometimes the pages are fine and your own linking is the problem — half your internal links point at page A with one anchor, half at page B with the same anchor. Deciding which page owns the term and linking consistently can resolve it without touching the content.
4 · Canonicalise one to the other. Use when both pages must stay reachable for people but only one should compete in search. Remember a canonical is a hint — if your internal links contradict it, it will be ignored.
Deleting pages is rarely the right answer. A page with traffic, links or conversions is an asset even when it overlaps something else. Merge or differentiate before deleting.
Preventing it
Cannibalisation is a planning failure that shows up months later, and the prevention is unexciting: decide which page owns which query, and write it down.
A keyword map — one primary term per page, recorded — makes the check trivial before publishing rather than forensic afterwards. Without one, the same term gets targeted repeatedly as different people write different articles, and nobody notices until rankings wobble.
The discipline is: before publishing, search the map for the primary keyword. If it is already owned, either pick a different angle or improve the page that owns it.
Frequently asked questions
Is it bad if two pages rank for the same keyword?
Not necessarily. Two results take space a competitor would otherwise hold. It becomes a problem when Google keeps swapping which page it shows, or when both rank poorly because signals are split.
How do I find keyword cannibalisation?
In Search Console, filter the Performance report by a query and check which pages appear for it. Then compare two date ranges: if the ranking URL changes between them, that instability is the real symptom. One stable URL means no problem.
Should I delete cannibalising pages?
Rarely. A page with traffic, backlinks or conversions is an asset. Differentiate the two pages, or merge the weaker into the stronger with a 301. Deleting discards accumulated value that merging would keep.
Does cannibalisation cause a Google penalty?
No. There is no penalty for it. The cost is mechanical — ranking signals split across URLs instead of accumulating on one, so the best page ranks lower than it otherwise would.
Find out which pages are actually competing
Cannibalisation is over-diagnosed and under-verified. The expensive version is not the problem itself — it is merging or deleting pages that were never in conflict, and losing what they earned.
If you want the overlaps identified properly, with a decision per page rather than a blanket merge, the Content & Information Architecture Audit maps what each page targets and where the genuine conflicts are — $1,497, one-off, no retainer.