Keyword Cannibalisation and How to Fix It

Keyword cannibalisation happens when two or more pages on the same website target the same search intent, so they compete with each other for one query instead of supporting each other. Google usually shows only one of them, its choice can change from week to week and neither page gathers the links and relevance a single stronger page would. It is most common on websites that publish regularly over many years, which is why it matters so much in SEO for B2B companies with large content libraries.

The fix is rarely complicated once the problem is confirmed, but confirming it is where most of the work sits. The sections below cover what cannibalisation is and is not, how to find it with free tools, how to decide which page should win and how to resolve a clash without throwing away rankings the website already has.

What Keyword Cannibalisation Means

Keyword cannibalisation means several pages on one website are trying to answer the same question for the same searcher. Search engines have to choose which of those pages to rank, while the pages split the links and relevance that would otherwise belong to one.

The term is often used loosely, so it helps to be precise about what counts. The Ahrefs guide to keyword cannibalisation makes the point that there is only a real problem when multiple pages target the same keyword and that overlap hurts the website’s organic performance. Two pages that both mention a phrase are not automatically competing, because most pages rank for many different queries at once.

The distinction matters because the usual cure for real cannibalisation, which is merging pages, would damage a website if it were applied to pages that serve different purposes. Before changing anything, the job is to show that two pages serve the same intent and that the overlap is costing traffic, clicks or enquiries.

Why Keyword Cannibalisation Hurts Rankings

Keyword cannibalisation hurts rankings because it splits the evidence Google uses to judge a page across two or more URLs. Links, internal anchor text and topical relevance that could have built one strong page are shared between several weaker ones, so none of them reaches the position the combined page might.

The damage tends to show up in the same handful of ways. Some of them are visible in ranking reports, while others only appear when enquiries and sales are compared against search traffic.

• Links from other websites are divided between pages rather than concentrated on one
• Internal links point at different URLs for the same topic, which blurs which page matters most
• Google alternates between the pages, so the ranking position and the URL shown keep changing
• The page that ranks is sometimes the weaker answer, such as an old blog post shown instead of the service page
• Visitors land on a page that was never meant to convert them, so enquiries fall even when traffic holds

The conversion problem is the one most often missed. A query might bring a similar number of visits whichever page ranks, but a visitor who lands on an old news story is far less likely to get in touch than one who lands on the service page built for that query.

When Overlap Between Pages Is Not a Problem

Overlap between pages is not a problem when each page serves a different search intent, even if they share keywords. A guide explaining how something works and a service page selling it can both use the same phrase, because Google treats a research query and a buying query differently and will usually rank each page for different searches.

Intent is the test that separates harmless overlap from real competition. If a searcher would be equally satisfied landing on either page, the two pages are competing. If one page answers a question and the other offers to do the work, they are serving different stages of the same journey. The intent behind a query is covered in more depth in our guide to user intent in SEO.

Two pages from the same website ranking side by side for one query can also be a good outcome rather than a symptom. If the two pages appear on page one and together win more clicks than one page would, there is nothing to fix. The time to act is when one page holds the other back, which usually shows as a ranking URL that keeps switching or a position that stalls below where the content deserves to be.

What Causes Keyword Cannibalisation

Most keyword cannibalisation is caused by publishing over several years without a record of which page owns which query. Each piece made sense when it was written, but a new blog post on a familiar topic, a refreshed guide published at a new URL or a service page that repeats a popular article slowly builds a set of pages that answer the same thing.

Technical causes are easier to fix than editorial ones because they follow a pattern. Google’s guidance on managing faceted navigation URLs explains how filter and sort parameters can create a very large number of URLs that crawlers visit before working out they are not useful. Editorial causes need a decision about each page, which is why they take longer to resolve.

Internal linking is the other technical cause that is easy to overlook. When the navigation, related posts and links in body copy all point at different pages for the same topic, Google gets mixed signals about which page the website itself considers most important.

How to Find Keyword Cannibalisation

The quickest way to find keyword cannibalisation is to look for queries in Google Search Console where more than one page on the website earns impressions. Free tools are enough for most websites, with paid rank trackers mostly speeding up the same checks.

The checks below work best in order, starting with Google’s own data and finishing with a list of every page and the query it targets. Each one narrows the list of suspected clashes before any page is changed.

Search Console is the most reliable starting point because it shows Google’s own data rather than an estimate. The Performance report documentation explains how the Queries and Pages dimensions group the data, which is what makes one query spread across several URLs visible. Look at clicks and average position for each page as well as impressions, since a page with impressions but almost no clicks is often the one being shown in the wrong place.

A site: search is broader and less precise. Google’s documentation on the site: search operator notes that an indexed URL can appear for related site: queries but that this is not guaranteed, so treat the results as a list of candidates rather than proof. The same Ahrefs guide suggests adding &filter=0 to the end of a Google results URL, which turns off host clustering and shows whether two pages from the same website sit close together in the rankings.

On larger websites a crawler such as Screaming Frog speeds up the inventory step by exporting every title and main heading at once. Pages with matching or near identical titles are the first place to look, since they are usually targeting the same query whether anyone intended it or not.

How to Confirm a Cannibalisation Problem

A cannibalisation problem is confirmed when two pages serve the same intent, each earns impressions for the same query and the overlap is holding performance back. All three conditions need to be true before any page is merged, redirected or rewritten.

Start with the search results themselves. Search the query and look at what Google ranks from other websites, because the type of page in the top results shows which intent Google has settled on. If the top results are all guides, a service page is unlikely to win that query no matter how well it is optimised, so the guide is the natural page to keep.

Then compare the two pages against each other. Read them side by side and ask whether a visitor would get the same answer from both. Check whether the ranking URL has switched over recent months and whether the combined clicks are lower than the stronger page managed on its own before the second page existed.

Pages that pass all three tests are a real cannibalisation problem. Pages that fail one of them are usually better left alone or given clearer internal links rather than merged.

How to Decide Which Page Should Win

The page that should win is usually the one with the strongest existing signals and the closest match to the intent Google rewards for the query. Commit to one winner per query before touching anything, since every fix that follows depends on that decision.

Older pages tend to have more history, but age alone is not a reason to keep a page. Sometimes the newer page is simply the better answer, with clearer structure and more current information. In that case the right move is to fold the older page and its links into the newer one.

When the signals point in different directions, give the most weight to intent and conversions. Links can be moved to another URL with a redirect, but a page in the wrong format will struggle to rank for the query regardless of how many links it has.

Record the decision for each query in a spreadsheet with the winning URL, the losing URLs and the planned fix. That record becomes the start of a keyword map, which is what stops the same clash appearing again later.

How to Fix Keyword Cannibalisation

Keyword cannibalisation is fixed by merging pages, giving each page a distinct purpose, correcting internal links or, for pages that must stay live, telling Google which version to index. Most clean ups use a mixture of these across different queries, chosen page by page rather than applied in bulk.

The order of preference matters. Merging and redirecting keeps the most value, rewriting keeps each page useful and technical signals such as canonical tags and noindex are reserved for pages that cannot be combined. Where almost identical URLs rather than competing articles are the cause, Search Console usually reports them as duplicates without a canonical chosen by you.

Merge the Pages and Redirect the Weaker URL

Merging works best when two pages overlap heavily. Move anything useful from the weaker page into the winner, update the winner so it is the complete answer, then redirect the weaker URL to it. Google’s guidance on redirects and Google Search recommends a permanent server side redirect whenever a page’s URL needs to change in search results, using a 301 or 308 status code.

Priority Pixels handles consolidation the same way in its content optimisation work, keeping the strongest URL, redirecting the weaker one and updating the internal links that pointed at it. The internal link step is the one most often skipped, which leaves visitors and crawlers passing through a redirect on every click.

Give Each Page a Different Job

Rewriting is the better choice when both pages have value but are aimed at the same query. Keep one page on the original query and refocus the other on a related but distinct intent, such as turning a second overview into a detailed guide to one part of the topic. Change the title, main heading and opening section of the refocused page so the new purpose is obvious to readers and to Google.

Ahrefs warns against trying to remove a single keyword from a page, because stripping out internal links or mentions to push one query away tends to affect its rankings for other queries too. Refocusing works because it gives the page a new target, not because it hides the old one.

Correct the Internal Links

Internal links tell Google which page the website itself considers the main resource on a topic. Point every relevant link in body copy, menu item and related posts block at the winning page, using anchor text that describes its subject. Google’s documentation on crawlable links and anchor text describes good anchor text as descriptive, reasonably concise and relevant to both the page it sits on and the page it links to.

The way a clear structure of internal links supports rankings is covered in more detail in our guide to internal linking and website architecture. Fixing links on their own is rarely enough for a heavy overlap, but it often resolves mild cases where the content is already distinct.

Use Canonical Tags and Noindex Carefully

Canonical tags and noindex suit pages that have to stay live for visitors but should not compete in search. Google’s guide to consolidating duplicate URLs describes redirects and rel=”canonical” annotations as strong signals for which URL should become canonical. The same guide advises against using noindex to choose a canonical page within one website, because that blocks the page from Search completely.

A canonical tag is the right tool for near duplicates such as product variants or print versions of a page, not for two articles written separately on the same subject. Noindex fits thin category or tag archives that visitors use but that add nothing as a search result, since Google’s documentation on blocking indexing with noindex confirms a noindexed page is dropped from Google Search results entirely.

Deleting a page is the last resort rather than the first. A page with no links, no traffic and nothing worth keeping can be removed, but anything that has earned links from other websites should be redirected so that value passes to the page that remains.

Whichever fix is chosen, make one change per query at a time where possible. Changing titles, links and redirects for several overlapping pages at once makes it hard to tell which change caused any movement that follows.

What to Expect After Fixing Keyword Cannibalisation

Changes after a fix show up once Google has recrawled the affected pages and processed any redirects, so the timing varies with how often the website is crawled. Important URLs can be submitted through the URL Inspection tool in Search Console, which can prompt a fresh crawl but does not change how the page ranks.

Track the query in Search Console using the same filters as the diagnosis. The signs of success are one URL taking the impressions that were previously split, a steadier average position and clicks moving to the page that was meant to win. Note the date of each change, so that movement can be tied to the work rather than to a Google update or seasonal demand.

Expect some movement while Google works out the new structure. A short period of fluctuation after a merge is not a reason to reverse the change, as long as the redirect works and the winning page is indexed.

How to Prevent Keyword Cannibalisation

Keyword cannibalisation is prevented by giving every page one primary query and checking new content against that list before it is written. A keyword map does this job, recording the query, intent and URL for each page so that anyone planning content can see what is already covered.

Topic clusters build the same discipline into the structure of the website. One pillar page owns the broad term, supporting pages own the specific questions beneath it and internal links make the hierarchy obvious, as set out in our guide to building topical authority with topic clusters.

Titles are the cheapest early warning of a future clash. Google’s guidance on title links asks for distinct text in the title element of each page. Two pages with near identical titles are often the first sign that they target the same query, so check new titles against existing ones before anything is published.

Updating an existing page is usually better than publishing a new one on the same subject. If a guide has gone out of date, refresh it at the same URL rather than writing a replacement, because the original keeps the links and history it has built up. Older content that is losing ground is covered in our guide to content decay.

A regular audit catches the overlap that planning misses. A Priority Pixels content audit checks for cannibalisation alongside duplication, decay and gap analysis, with every page given a next action.