How keyword cannibalisation happens
It is rarely deliberate. A site publishes a service page, then a blog article on the same subject, then a location page, then a case study, and over a few years several pages end up answering the same question. Search engines have to choose one to show for that query. They may swap between them week to week, or pick the one you would not have chosen — often an old article rather than the page that sells something.
The trigger is shared search intent, not a shared keyword. Two pages can both mention the same phrase without competing, provided one answers a how-to question and the other sells the service. Two pages that both try to answer the same question, in the same way, for the same reader, are competing whatever words they use.
Why keyword cannibalisation matters
Split effort produces split results. Links, internal links and engagement that could have built one strong page are shared between weaker ones, so none of them reaches the position a single consolidated page would have held. You also lose control of the landing page, which is the part that shows up in revenue: an informational article ranking in place of a service page brings readers rather than enquiries.
It gets worse as a site grows. A large site with many overlapping pages spends crawl attention on near-duplicates, dilutes its topical clarity, and becomes harder to change, because nobody is certain which page is the one that matters.
Common mistakes with keyword cannibalisation
The biggest is treating every repeated word as cannibalisation. A guide, a service page and a city page can legitimately share vocabulary; what matters is whether they serve the same intent. Deleting perfectly good pages on the strength of a keyword overlap report destroys traffic that was never in conflict.
The second is diagnosing it from a rank tracker alone. Fluctuating positions have many causes. The reliable evidence is in Search Console: filter by query, look at the pages receiving impressions for it, and see whether they alternate over time. The third is fixing it by rewriting the weaker page to target something slightly different, which usually just moves the conflict to a new query.
What to do about it
Start by writing down, for each competing page, the one search it is supposed to win. If any two answers match, one of them is wrong. Then pick a resolution:
- Merge — combine the pages into the strongest one and redirect the others to it.
- Differentiate — keep both, but change one so it genuinely answers a different question.
- Redirect — retire a page that adds nothing and send it to the page you want to rank.
- Re-point — leave the pages alone but redirect internal links and anchors to the page you want to win.
Merging beats deleting: keep the URL that already has links and history, move the useful sections into it, and redirect the rest permanently. Afterwards, prevent the recurrence by assigning one intent per page before anything new is written, which is what a proper keyword research and mapping exercise is for. On a site with years of accumulated posts, expect to find this during a content audit rather than by looking for it.
