Duplicate pages are several different URLs on the same site that serve identical or nearly identical content. Google can’t tell which copy to show in search results, splits link equity and behavioral signals between them, and in the end no page ranks as high as a single strong one could. Keyword cannibalization is a related problem: several of your pages compete for the same search query at once. Below we look at how to find both problems at scale and fix them with the canonical tag, 301 redirects and noindex.
Key points in 30 seconds
- Duplicates are the same content on different URLs; cannibalization is different pages targeting the same query.
- Both problems dilute link equity and stop a single strong page from reaching the top.
- Common sources: URL parameters, pagination, catalog filters, http/https, trailing slashes and www.
- They are found at scale with Search Console, a site crawler and the site: operator.
- They are fixed with three tools: rel=canonical, 301 redirects and noindex, each suited to its own situation.
What duplicate pages are and where they come from
A duplicate is a clone of an existing page available at a separate address. Duplicates usually appear not because someone deliberately copied content but because of how the CMS works: the same product opens at several addresses, sorting parameters and UTM tags “stick” to URLs, and the site responds on both http and https, with and without a trailing slash. For a person it’s one page; for a search engine it’s several different ones.
It helps to break duplicate types down by source, so you can see right away how to fix each one.
| Duplicate type | Example | How to fix |
| Technical (protocol, slash, www) | http:// and https://, /page and /page/ | 301 redirect to the canonical URL |
| URL parameters | ?sort=price?utm=?sessionid | rel=canonical to the clean URL |
| Pagination | /blog, /blog/page/2 | Self-referencing canonical + unique title, not a canonical to page one |
| Catalog filters | ?color=red&size=42 | Canonical or noindex according to your indexing rules |
| Full copies | print version, duplicate product page | 301 redirect or noindex |
| Partial duplicates | similar category descriptions | Rewrite or merge the pages |
Full duplicates are more dangerous than partial ones: they compete directly with each other for the same spot in search results. But partial duplicates, such as nearly identical category or product descriptions, also dilute relevance and make it harder for Google to pick out the main page.
What keyword cannibalization is and why two pages are worse than one
Cannibalization is when two or more pages on a site are optimized for the same query and compete with each other in search results. A classic example: a “how to choose” guide and a category page fighting over the same commercial keyword; or two similar blog posts that are essentially about the same thing.
Why it hurts. Google picks one page for a query, and its choice can “jump”: one page ranks today, another tomorrow, and the position is unstable. The links you earn get split between competing pages instead of strengthening one. CTR drops because the SERP alternates between a weaker title and a less relevant URL. Two average pages almost always lose to one strong page: by merging them, you combine link equity, behavioral signals and content into one powerful signal.
How to find duplicates and cannibalization at scale
Checking pages by hand is pointless: you need tools that show the problem across the whole site at once.
Google Search Console. The Indexing → Pages report shows which URLs Google considered duplicates: the statuses “Duplicate, Google chose different canonical than user” and “Duplicate without user-selected canonical”. This is the first place to look; how to respond to these reports is described in the article on technical SEO audits.
Site crawler. Screaming Frog, Netpeak Spider or a similar tool crawls the site like a bot and exports all URLs with their titles, descriptions and content hashes. Sort by identical titles and descriptions, and duplicates surface in batches.
The site: operator. A quick manual check: a query like site:yoursite.com “text fragment” shows how many pages contain the same piece of content.
Finding cannibalization. In GSC (the Performance report → filter by a specific query), look at how many different URLs get impressions for it and how they alternate in rankings. If 2–3 pages rank for one keyword and the position keeps jumping, that’s cannibalization.
How to fix it: canonical, 301 redirect and noindex
Fixing duplicates doesn’t mean “delete everything extra” but choosing the right tool for the specific situation. There are essentially three.
| Tool | When to use it | What happens |
| rel=canonical | The duplicate needs to stay accessible (filters, sorting, parameters) | Equity is consolidated on the specified URL, both pages remain available to users |
| 301 redirect | The duplicate isn’t needed by people or search (http, extra slash, old URL) | The pages are merged, equity is passed on, the duplicate disappears from search results |
| noindex | The page is needed by people but not by search (filters, site search results) | It is removed from the index but still crawled and passes equity through its links |
| Merging | Cannibalization: two weak pages targeting one query | The content is combined into one page, and the second gets a 301 to the first |
The main rule: use a 301 redirect when the duplicate isn’t needed by people or search; use canonical when the page must remain accessible to users but only one version should be indexed; use noindex when the page is useful to people but doesn’t belong in search results. The syntax for canonicals, redirects and noindex is covered in detail in the article on technical SEO. For cannibalization, merging usually works best: combine two articles into one stronger piece and set a 301 from the second to the first.
Pagination and filters: where duplicates appear most often
Online stores generate the most duplicates: filter and sorting pages multiply nearly identical URLs by the hundreds. The general logic: filter combinations that are useful to users and have search demand (for example, “Nike sneakers”) can stay indexed with a self-referencing canonical, while endless technical combinations should be handled with canonical or noindex. Today it is not recommended to canonicalize pagination (page/2, page/3) to the first page; each page should reference itself and have a unique title. More on this in the breakdown of SEO for online stores.
If there are many duplicates and you’re not sure what to close and how, that is the first sign you need a technical SEO audit: it will show a full map of duplicates and cannibalization along with fix priorities, so you don’t break working pages while trying to clean things up.
Analyzing duplicates and cannibalization is part of a comprehensive technical audit: we find all duplicates, build a map of redirects and canonicals and hand over a prioritized fix plan. To gauge the scale of the problem on your site, order an SEO audit, and we’ll calculate the cost for your project. Prices are approximate and depend on the niche, volume and competition; we’ll calculate the exact cost for your project.
