SEO Cannibalization

Cannibalisation is when two or more pages on your site compete for the same search query. Google sees both, splits its confidence, and neither ranks as well as one strong page could. I worked with a running shoe site that had three posts: 'Best Running Shoes', 'Best Running Shoes for Men', 'Best Running Shoes for Women'. All three got impressions for the root query 'best running shoes'. Each attracted about 150 monthly visits. After I merged them into one full walkthrough and used a [301 redirect](/301-redirect/) from the weaker URLs, that single page now pulls 700 visits. The fix seems obvious now, but at the time I thought more pages meant more coverage. It does not. The key is to check whether two pages serve the same intent. If they do, consolidate. If not, ensure titles and internal links clearly signal the difference. A counter-argument I hear is that multiple pages can dominate the SERP for different variants. In practice, Google often picks the wrong one or shows both erratically, which hurts click-through rate. The decision rule I now apply: query the core keyword in incognito. If two of your pages appear within the first 20 results, you probably have a cannibalisation problem. When the content is near-identical — like product variants with minor differences — a [canonical tag](/canonical-tag/) can help, but true [duplicate content](/duplicate-content/) scenarios are better handled by consolidation.

Cannibalisation becomes a real drag once your site passes 50 pages. I see it most often in three scenarios. First, blogs that publish frequent updates on the same topic. A client once had 'SEO Tips 2025' and 'SEO Tips for Beginners' — both targeting 'SEO tips'. The older page had backlinks, the newer one had fresher content. They split traffic and link equity. Second, eCommerce sites with product variants. A T-shirt shop listed 'Blue T-Shirt', 'Blue T-Shirt Cotton', 'Blue T-Shirt Slim Fit' — all three competing for 'blue T-shirt'. Google could not decide which to rank. The fix was not to delete pages but to add canonical tags pointing to the master product page and use internal linking to reinforce it. Third, content refreshes done poorly. I used to publish an updated version of a post and leave the old one live, thinking both would rank. They just split impressions. The number to watch is in Google Search Console: look for any query that drives impressions to more than one URL. If those URLs collectively get fewer clicks than a single URL from a competitor, you have a problem. An edge case I still debate: programmatic SEO sites with thousands of similar pages. Here, [website structure](/website-structure/) and [pagination](/pagination/) often cause cannibalisation. The decision rule for those is to use parameter handling in Search Console and consolidate thin pages. [Technical SEO](/technical-seo/) hygiene — like proper canonicalisation and sitemap management — becomes critical at scale.

I have made three serious mistakes with cannibalisation. First, I used to redirect the losing URL without updating internal links. The redirected page passed equity to the new one, but every link on the site still pointed to the old URL, creating extra hops and wasting link juice. I now run a full [SEO audit](/seo-audit/) to find and update every internal link to the preferred target. Second, I thought canonical tags were a fix for everything. They are not. If two pages contain substantially different content but target the same intent, a canonical tag tells Google to ignore one, which can waste the content you wrote. Only use canonical tags when pages are near-identical — for example, session IDs or printer-friendly versions. For real content overlap, merge and redirect. Third, I flagged every shared keyword as a crisis. I once saw two pages both targeting 'best running shoes for flat feet' and panicked. But one was a guide and the other a product category page — they served different intents. The guide informed, the category sold. Google showed the category for transactional queries and the guide for informational ones. That is not cannibalisation; it is good information architecture. The mistake was not checking the search intent behind the keyword. Now I always map each page to a specific query intent before I act. If the intents match and the pages exist, I consolidate. If not, I adjust titles and internal links to reinforce the difference.