Write several posts on one subject and something odd starts happening. A piece you know is good will not climb, and sitting in its place is another of your own posts on the same subject, ranking indifferently. Check again next week and the two have swapped.
This is cannibalisation: two of your pages competing for one query while the search engine keeps changing its mind about which to show. The cost of that indecision is that neither page ranks as well as it could.
Where it comes from — usually diligence
Cannibalisation is not a symptom of writing too little. It shows up precisely on sites that publish consistently, by three familiar routes.
First, similar posts written far apart in time: you forget last year’s article and approach the subject again. Second, a blurred line between pillar and cluster, where the broad page drifts into the detail its clusters were meant to own. Third, a service page and an article saying the same thing.
Finding it — two methods are enough
The first method is your performance data. In Search Console, look at which pages appear for each query; you will find queries where two or more of your URLs surface. Pay particular attention to queries where both rank mediocrely — one page at position one and another at forty is usually not a problem at all.
The second method is a site search. Type site:yourdomain.com keyword and see which of your pages come back for it. It takes a minute and needs no tooling, and doing it before writing anything new stops the overlap happening in the first place.
Three fixes
Once you have the overlapping pair, you have three options. Which one applies turns on a single question: are the two posts answering different questions?
Merging is right more often than not. One deep article beats two shallow ones, and the act of combining usually improves both. Keep the URL that already earns traffic and 301 the one that disappears.
Differentiating applies when the two really do answer different questions. What needs rewriting then is not the body but the title and the opening — that is where both readers and search engines decide what a page is about.
Redirecting is for a page that has aged out or lost its role. Deleting alone throws away the links and traffic pointing at that address, so decide where it goes before you remove it.
After the fix, wait
Rankings wobble for a while after a merge or a redirect while the change is being reflected. This is the moment impatience does damage: reverting makes things worse than either state. Note the date you changed it and check the same queries a few weeks later.
More on redirect handling and how indexing recovers lives in the SEO archive, and the step-by-step way we run a restructure safely is published on our process page.
Next part
With relationships between posts sorted out, the next step is stating those relationships explicitly. The following part is internal linking — what anchor text and link direction each do.