How to Find and Fix Keyword Cannibalization in Search Console

A step-by-step method to find pages competing for the same queries using Search Console and a spreadsheet, and a decision table for merging, refocusing or redirecting them.

Keyword cannibalisation happens when two or more pages on the same site compete for the same queries, so search engines alternate between them and neither ranks as well as one strong page would. To find it, export query and page data from Search Console together, list the queries where several of your URLs get impressions, and check whether their positions swap over time. To fix it, merge pages with the same intent into one and redirect the rest, or make each page clearly answer a different question.

When two ranking pages are not a problem

Not every query with two URLs is cannibalisation. A home page and a service page both appearing for a brand search is normal. So is a guide and a product page for a broad query with mixed intent. It becomes a problem when pages with the same intent share the impressions, positions move up and down as Google switches between them, and the page you want to rank is not the one that does.

Step 1: get query and page data together

The Performance report in Search Console shows queries and pages separately. To see them together:

For a quick check of one query, click it in the Queries tab and then open the Pages tab to see every URL that received impressions for it. For the whole site, connect Search Console to Looker Studio with the “URL Impression” table, build a table with Query and Landing Page as dimensions, and export it to Google Sheets.

Step 2: flag queries with several URLs

In the sheet (columns: A query, B page, C clicks, D impressions, E position), add a column counting how many pages share each query:

=COUNTIF(A:A, A2)

Filter to queries where the count is 2 or more and total impressions are meaningful, for example 100 or more over three months. Then add each page’s share of the query’s impressions:

=D2 / SUMIF(A:A, A2, D:D)

Queries where two pages each hold a large share, such as 30% or more, are the strongest candidates.

Step 3: confirm with the timeline

For each candidate, filter the Performance report by the query and compare the pages over time. Signs of real cannibalisation:

  • The ranking URL changes from week to week
  • Average position jumps up and down with the switches
  • Neither page reaches the top three
  • The page that ranks is not the one you would choose

Step 4: decide what to do

Situation Action
Two pages answer the same question Merge the best of both into the stronger page and 301 redirect the other
Pages overlap but serve different intents Sharpen each: different titles, headings and focus, and link between them
A blog post outranks the service page for a buying query Refocus the post on the informational question and link it prominently to the service page
Near-duplicate pages created by filters or parameters Canonical tags to the main version, or prevent the duplicates from being created
An old page with no value Redirect to the relevant page, or remove it

Step 5: consolidate the signals

  • Update internal links to point to the chosen page, with anchor text that describes it.
  • Remove redirected URLs from the XML sitemap.
  • Keep redirects in place permanently.
  • Record what you changed and when, so you can compare results after the pages are recrawled.

Why it matters for AI search too

When an AI assistant searches for a sub-question, two weak pages from your site each have a smaller chance of being the best passage than one strong, focused page. Clear ownership of each question, as described in our guide to topical authority and topic maps, prevents cannibalisation in the first place.

Frequently asked questions

How long does it take to see results after merging?

Usually a few weeks, once Google recrawls both URLs and processes the redirect. Watch the chosen page’s position and impressions for the affected queries.

Should I use noindex instead of redirecting?

Only for pages that must stay available to users but should not appear in search. For duplicates, a redirect passes signals to the chosen page; noindex does not.

Can cannibalisation happen with category and product pages?

Yes, especially on shops where category pages repeat product descriptions. Give category pages their own introductory content and keep product details on the product pages.

Sources

  1. Google Search Central, “Consolidate duplicate URLs” and “Redirects and Google Search”
  2. Search Console Help, “Performance report”
  3. Looker Studio Help, “Search Console connector”