Keyword cannibalization: how to find and fix it

Short answer: start with Search Console, not with paid tools. Open the Performance report, expand the full report and filter by a single query — then count how many of your URLs show up. If traffic for that keyword is split evenly across two or three pages, that is cannibalization.

The second signal is ranking history. A page held the top of the results, you published a new article on a close topic, and afterwards both slipped. The third is keywords stuck in the twenties to fifties for years, with different URLs rotating on them: Google cannot decide which page to show, so it shows none of them properly.

How to hunt this at scale:

• a crawler (Screaming Frog, Sitebulb, Botify, Deepcrawl) — export every title and H1, then look for duplicates and near-duplicates;
• rank trackers (Semrush, Ahrefs, Authority Labs) — keywords with more than one ranking URL;
• check canonicals, robots.txt and whatever your plugins changed during the last update — the cause is often purely technical.

One caveat matters, and Adam Riemer insists on it in the Ask An SEO column on Search Engine Journal: not every pair of similar pages competes. A how-to guide and a conversion page serve different intents at different stages of the journey — merge them and you simply lose the top of the funnel.

When it really is cannibalization, the lossless fix looks like this: keep the stronger page (by links and conversions) and either rewrite the weaker one for a different intent, point its canonical at the main version, or block it from the index with meta robots — easy to apply programmatically to whole folders and parameters. Then fix internal links: anchors and context must point to the page that matches the query intent. And the step that prevents a relapse — put covered topics on an editorial list and ship nothing new on them without SEO sign-off.