Keyword cannibalization happens when two or more pages on your site compete for the same search query with the same intent, so none of them gets the position one strong page would. The classic case is a startup blog with “How to write a pitch deck”, “Pitch deck guide” and “Pitch deck tips”, all answering the same question in slightly different words.
What is keyword cannibalization?
Google picks which of your URLs to show for a query. When several of your pages fit equally well, Google may rotate them, show a weaker one, or show two of them where one would do. Links and anchor text also spread across the near-duplicates instead of landing on one page.
Not every overlap is a problem. Google’s John Mueller has said that different pages ranking for the same query is not a problem in itself, and a query can legitimately show a product page and a guide. The overlap hurts only when it costs you something:
- the wrong page ranks (a blog post instead of your pricing page for a buying query),
- positions swap between two URLs week after week,
- both pages sit at positions 8 to 20 where a merged page could reach the top 5,
- the pages say the same thing and each one earns half the links.
Why does it matter for SEO?
A single strong page concentrates relevance, links and click history in one URL. Two thin ones divide them. Duplicate coverage also wastes your crawl and editing time: you maintain two pages where one would do, and visitors who land on the weaker one get a worse answer. Startups with small blogs hit this fast, because they publish many posts around one subject.
How do you find keyword cannibalization, step by step?
- Open Search Console and go to Performance. Set the last 3 months and sort by impressions.
- Check your target queries one at a time. Click a query to filter by it, then open the Pages tab. If two or more URLs show impressions, note each page’s clicks and average position.
- Compare with intent. Open each URL. If both answer the same question for the same reader (see search intent), that is a cannibalization candidate. If one is a product page and the other a tutorial, leave them.
- Spot the flip-flop. Add the Date dimension for the query and look at which page gets the impressions over time. Alternating pages is the strongest sign. A rank tracker with ranking URL history shows it faster (see rank tracking).
- Run a title and H1 audit. Crawl the site and sort by title and H1. Near-identical titles point to overlapping pages before any data does.
- Try
site:searches. Searchsite:yoursite.com "pitch deck template"to see every page that mentions the phrase. Treat it as a hint, not a measurement. - Pick the winner. Choose the page with the most links, the best content and the right URL. If it is not the one currently ranking, still pick it when it is the better page.
How do you fix it?
Pick the option that matches what you found:
| Situation | Fix |
|---|---|
| Two pages, same intent, similar content | Merge the best parts into the winner and 301 redirect the loser to it |
| Two pages, same intent, loser has no value | 301 redirect or delete with a redirect |
| Near-identical pages you must keep (print version, filtered view) | Use a canonical tag to name the main page |
| Different intents, similar titles | Rewrite titles, H1s and openings so each page targets its own query |
| One page is a hub, others are details | Link detail pages up to the hub with the target anchor text (see internal linking) |
Google says to use a permanent server-side redirect (HTTP 301 or 308) when you are sure the redirect won’t be reverted, and that it treats this as a signal that the target is the canonical URL (see redirects). After a merge:
- Move the useful sections, data and links into the winning page and make the page better than both originals.
- Redirect the old URL straight to the winner. A chain of redirects slows crawling.
- Update internal links to point to the winner directly, and drop the old URL from the sitemap.
- Request indexing of the updated winner in Search Console.
How do you measure the result?
Watch the winning URL for 4 to 8 weeks: combined clicks and impressions for the target query against the total of the two pages before, and average position. Expect a rise in position for the merged page. If total clicks fall for the query, check that the redirect works and the merged page really covers what the old one did.
What mistakes should you avoid?
- Treating every shared query as an error. Check intent first. A homepage and a blog post can both appear for a brand name.
- Deleting instead of redirecting. A deleted page loses its links and leaves a 404.
- Redirecting to an unrelated page. Redirect to the closest match, or Google may treat it as a soft 404.
- Using
noindexon one page and keeping its internal links. It removes the page from search but does not point its signals at the winner the way a redirect does. - Fixing it with a canonical tag when the pages differ. Canonicals are for duplicates, and Google treats them as a hint.
- Planning new posts without checking existing ones. Search your own site first. Keep one target query per page in a simple sheet.
Where can you learn more?
Google’s page on redirects covers permanent redirects. The free digital marketing module of the Fe/male Switch startup school teaches how to plan one page per topic. Next steps: map every page to one target query with keyword research data and prune what is left over with content pruning.