What Is Keyword Cannibalisation in SEO?
Two of your own pages competing for the same search, so the site divides its effort between pages that should have been one. We have hit this repeatedly in our own work, which means the method below is one we have actually used rather than one we recommend.
Two Of Your Own Pages Competing
Somebody searches for something you answer. Two of your pages both address it, neither is clearly the answer, then your site ends up competing against itself for a search it should own outright.
What actually happens. Division.
Whatever standing the subject has earned is split between two addresses rather than accumulating on one. Links, references and attention all divide, so neither page reaches what a single page would have.
Why the word is dramatic. It exaggerates slightly.
Nothing is being eaten and nothing is being punished. It is a dilution problem. The melodrama of the term makes businesses treat it more urgently than it usually deserves.
What it is not. A penalty.
No sanction is applied for having two pages on a subject. Anybody presenting this as something a site gets marked down for is overstating it.
Why it still matters. The cost is real.
A search you could have won is being won by somebody else, because your effort arrived in two places. That is an opportunity cost rather than a punishment. It is also genuine.
Where it accumulates. Sites that publish.
A site of ten pages rarely has it. A site publishing regularly for years almost certainly does somewhere, which is block two.
It Is A Symptom Of Not Deciding
It almost always comes from publishing without checking what already exists. Nobody creates competing pages deliberately. It accumulates quietly on sites that publish regularly.
How it arrives. Reasonably, each time.
Somebody has a good idea for a page and writes it. The idea is genuinely good and the page is genuinely useful. Nobody established whether the site already covered that ground.
Why nobody checks. There is nothing to check against.
Most sites have no record of which page owns which subject. Checking would mean reading everything, which nobody does, so the decision is made by omission.
Why it worsens with contributors. Different people, different times.
A site written by several people over several years accumulates overlap faster, because each contributor knows their own work rather than the whole. Agencies and staff changes both accelerate it.
What makes it invisible. Every page looks fine.
Nothing about an individual page indicates the problem. It exists only in the relationship between pages, which is why nobody encounters it while looking at any one of them.
Why that framing is useful. It points at the fix.
If the cause is not deciding, the permanent answer is deciding, which is block eight. Everything between here and there is repair work.
How To Tell You Have It
The signal is several pages appearing weakly for the same search and swapping places, rather than one appearing strongly. That pattern is distinctive once you know to look for it.
The swapping. The clearest indicator.
One page appears for a search this week and a different one next week. Something is choosing between them repeatedly, which means neither is clearly the answer.
The weakness. The second half.
Both pages appearing, neither well. A single strong page would be preferable to two mediocre ones. The arithmetic of that is the whole problem.
What it is not. Two pages both doing well.
If two of your pages both appear strongly for related searches, nothing is wrong. Presence is not evidence of competition, which block four develops.
Where to look. Your own search reporting.
The question is which pages appear for a given search over time. Our Google Search Console material covers the reports and how to operate them.
The manual version. Search your own site.
Restricting a search to your own domain shows what you have on a subject. Crude, immediate and frequently revealing on a site nobody has audited.
It Is Not Always A Problem
Pages covering the same subject from genuinely different angles are fine. Every site has overlap. The test is whether they compete for the same search rather than whether they share a topic.
The distinction. Topic against search.
Several pages can discuss headings without competing, if one defines them, one covers structuring them and one reviews them on an existing page. Same topic, three different searches.
What legitimate overlap looks like. Different questions.
A page answering what something is, plus another answering how to fix it when it goes wrong. Related, adjacent and serving people at different points.
Why the distinction matters. Merging is destructive.
Consolidating pages that were serving different searches produces one page serving neither properly. That is a worse outcome than the overlap it was meant to fix.
How to test it. Name each page's search.
If you can name a distinct search for each, they are not competing. If you name the same one twice, they are. That takes a minute rather than an analysis.
The realistic position. Some overlap always.
A site covering a subject thoroughly will always have adjacent pages. Chasing overlap to zero produces a thin site, which is the opposite of the objective.
Fixing It
Consolidate into one better page, differentiate so each targets a distinct search, else remove the weaker one. The order is not arbitrary: it runs from most preserved to least.
Consolidate. Keeps both pages' value.
Merge the content into one page and send the old address to it. Whatever each page had earned accumulates on one, which is the best available outcome.
Differentiate. Keeps both pages.
Change what each targets so they stop competing. You retain two assets and neither gains the other's standing, which is right where both subjects genuinely deserve a page.
Remove. Keeps neither.
Delete the weaker page and redirect it. Appropriate only where it genuinely has nothing worth merging, which is less often than people assume.
Why that is the order. Preservation.
Consolidation concentrates two efforts into one. Differentiation converts a conflict into two assets. Removal discards something. Prefer the option that loses least.
What decides between the first two. Whether both subjects deserve a page.
If the second page exists because somebody wrote it rather than because the subject needs it, consolidate. If both subjects are genuinely distinct and were merely aimed badly, differentiate.
Consolidation Is Usually Right
Merging preserves the value of both pages and produces something better than either. It is also the option most often skipped, because deleting one page is quicker than combining two.
What merging involves. More than copying.
Taking what is genuinely useful from both and writing one page that covers the subject properly. A page assembled by pasting two together reads as two pages stuck end to end.
Which address to keep. Usually the stronger one.
Keep the page already performing better and bring the other's content to it. Where they are equal, keep the address that describes the merged subject more accurately.
Why the address handling matters. It is where value transfers.
The old address must permanently redirect to the surviving one. Without that, whatever the removed page earned is discarded, which defeats the purpose of merging.
What else needs updating. Internal links.
Anything on your site pointing at the old page should point at the new one directly. Our site migrations material covers doing that properly.
The result worth expecting. One page that wins.
A merged page is usually more complete than either original, which is the second benefit after concentration. Both effects work in the same direction.
Retitling Is The Cheaper Fix
Changing what a page targets, rather than removing it, frequently resolves the problem without any address change at all. That makes it the cheapest option available and it is routinely overlooked.
What it involves. A change of aim.
Retitle the page, adjust its opening and headings, then point it at a search nothing else on the site owns. The content largely survives and the conflict disappears.
Why it is cheap. Nothing moves.
No redirects, no internal links to update and no risk associated with changing addresses. An afternoon of editing rather than a migration.
Our own worked example. Two guides converged.
Building this programme, two of our own pages ended up near-synonymous: one on structuring a blog post and one on optimising one. Same search, two pages, neither clearly the answer.
What we did about it. Changed one page's job.
One kept the writing guidance. The other became a review pass, checking a finished post against on-page fundamentals. Different searches, both pages retained, no address changed.
Why that was better than merging. Both jobs were real.
Writing a post and reviewing one are genuinely different tasks. Merging would have produced a longer page serving both badly, which block four warns against.
When it will not work. No second subject.
If there is no distinct search for the second page to own, retitling just relabels the conflict. That is when consolidation is the answer instead.
Prevent It At The Planning Stage
A record showing which page owns which subject stops this happening. Everything above is repair. This is the only thing that prevents the repair being needed again.
What the record contains. Very little.
Each page, the search it owns and a line on who it is for. Not a document anybody enjoys maintaining, though short enough that keeping it current is realistic.
How it gets used. One check before publishing.
Before writing anything, look at whether the search is already owned. That single habit prevents almost every instance, taking moments once the record exists.
What it also reveals. The gaps.
A map showing what you cover shows what you do not. That is more valuable than the prevention, because it turns publishing from opportunistic into deliberate.
Why most sites lack one. Nobody starts with it.
Sites accumulate rather than being planned, so the map would have to be built retrospectively. That is a genuine piece of work and it is done once.
Where this connects. Topical coverage.
Our advanced material covers deciding what a site should cover as a whole, which is the fuller version of this. The audit page covers where checking for this sits. The full series is on the on page SEO guide.
Name the search
each page
owns.
If you can name a distinct one for each, they are not competing and merging them would produce a page serving neither properly. If you name the same search twice, you have found it. That takes a minute rather than an analysis.
How we resolve it:
Chasing overlap to zero produces a thin site, which is the opposite of the objective.
Every guide.
One practice.
What a page says about itself, headings, keywords, search intent, images, page types, user behaviour, credibility and how to review an existing site.
