Duplicate Content and Canonical Tags, Explained: How Google Picks One URL

What counts as duplicate content?

Duplicate content is the same, or nearly the same, main content reachable at more than one URL. To a search engine, example.com/services and www.example.com/services/ are two different addresses, even if your website treats them as one page. When it finds several addresses with matching content, Google groups them and picks one to index and show. That chosen URL is the canonical, and the process is called canonicalization.

Google’s own list of causes is mundane. It names same-language regional variants, separate mobile URLs, HTTP and HTTPS versions, sorting and filtering, and “accidental variants” such as a demo site left open to crawlers. None involve anyone copying anything.

Three situations need different fixes:

• Technical duplicates: one page, many URLs (www and non-www, tracking parameters, trailing slashes, printer views). Fix with redirects and canonical tags.
• Near-duplicate pages: separate pages you created that say almost the same thing, such as city pages with only the name swapped. Fix by differentiating or merging them.
• Cross-site copies: your content on someone else’s domain, by agreement (syndication) or not (scraping). Fix with partner agreements or takedown requests.

Two of your articles competing for one search is a different problem, covered in our guide to keyword cannibalization.

Is there a duplicate content penalty?

No. Google’s canonicalization guide says some duplicate content on a site “is normal and it’s not a violation of Google’s spam policies.” Its September 2008 post “Demystifying the ‘duplicate content penalty’” said plainly that no such penalty exists. Google simply tries to filter out duplicates so users see less repetition. Microsoft’s Bing team wrote in December 2025 that duplicate content “doesn’t trigger search penalties on its own.”

The exception is intent. The 2008 post said duplication is not grounds for action unless it appears “to be deceptive and manipulate search engine results.” Two spam policies now draw that line:

• Scraping: republishing other sites’ content without adding value, or lightly rewriting it with synonyms.
• Doorway abuse: including “multiple domain names or pages targeted at specific regions or cities that funnel users to one page.”

So a store with every product reachable under three category paths is not what these policies target. A network of 200 near-identical city pages built to catch “plumber in [town]” searches could be.

What duplicate content actually costs you

No penalty does not mean no cost. Google gives four reasons to choose a canonical yourself: control which URL appears in results, consolidate signals such as links, simplify tracking, and avoid wasting crawling on duplicates. The canonical is also the page Google crawls most regularly.

In practice, that means a search result showing your printer-friendly page, backlinks split between HTTP and HTTPS, one page’s clicks spread across five URLs, or a partner’s copy of your article ranking above yours.

Where does duplicate content come from on a business website?

Mostly from the platform, not the people. In audits we find the same handful of causes again and again.

Location templates are the most common near-duplicate problem we see on local and multi-location sites: a regional HVAC company with 25 city pages, each one sentence different. Google will typically group them and pick one, and at scale they drift toward the doorway pattern above. Our guide to multi-location SEO covers what makes each location page earn its place.

Filters are the biggest technical source on e-commerce sites, because every combination of color, size and price creates a new URL. Google’s faceted navigation guidance recommends disallowing those URLs in robots.txt and calls canonical tags “generally less effective in the long term” for this job. Our e-commerce SEO guide has example rules.

How does Google choose the canonical URL?

Google clusters pages that are the same or very similar, then picks the one it judges most complete and useful. Its guide names “a handful of factors”: HTTP or HTTPS, redirects, sitemap inclusion and rel=“canonical” annotations. Your declared preference is one input among several. In Google’s words, it is “a hint, not a rule.”

Google’s guide on specifying canonicals ranks the methods. Permanent redirects are “a strong signal that the target of the redirect should become canonical.” A rel=“canonical” link, in the HTML or as an HTTP header, is also a strong signal. Listing a URL in your XML sitemap is “a weak signal.” Using several methods that agree increases the chance your preferred URL is chosen; methods that disagree make Google guess.

Several other preferences sit underneath. Google prefers HTTPS over HTTP as canonical unless something undermines it, such as an invalid certificate or an HTTPS page that redirects back to HTTP. In groups of language versions, it prefers URLs that reference each other reciprocally in hreflang annotations. And it may still show a duplicate when that suits the searcher better, such as a separate mobile URL for someone on a phone.

Why Google sometimes ignores your canonical

Google’s troubleshooting guide says that even when you designate a canonical, Google might choose a different one “for various reasons, such as the quality of the content or technical signals.” The causes it lists: a CMS or plugin generating incorrect canonicals or redirects, a misconfigured server, a hack that injects a cross-domain canonical, regional variants without hreflang, syndicated copies and, in rare situations, a copycat site.

The pattern we see most is the first one. A theme or SEO plugin setting points every page’s canonical at the homepage, or at the staging domain after a launch, and the site quietly tells Google that its service pages are copies. Because a canonical is only a hint, Google may overrule an obviously wrong one, but you shouldn’t rely on that.

How do you add a canonical tag correctly?

Put one <link rel="canonical"> element in the <head> of each page, pointing to the full, absolute URL of the version you want indexed. Google says the element “is only accepted if it appears in the <head> section,” so a tag that a script or page builder drops into the body is wasted.

On the page you want indexed, the canonical points to itself. On a duplicate, such as the same page reached with a tracking parameter, it points to the main version:

<!-- On https://www.example.com/water-heater-repair/?utm_source=gbp -->
<head>
  <title>Water Heater Repair in Columbus | Example Plumbing</title>
  <link rel="canonical" href="https://www.example.com/water-heater-repair/" />
</head>

For files that have no <head>, such as a PDF version of a guide, Google supports the same instruction as an HTTP response header, which your developer or host sets on the server:

Link: <https://www.example.com/guides/furnace-checklist/>; rel="canonical"

When the duplicate doesn’t need to exist at all, redirect it instead. On an Apache server, this common .htaccess pattern sends HTTP and non-www requests to one HTTPS www address in a single 301 redirect. Many hosts offer it as a setting:

RewriteEngine On
RewriteCond %{HTTPS} off [OR]
RewriteCond %{HTTP_HOST} !^www\. [NC]
RewriteRule ^(.*)$ https://www.example.com/$1 [L,R=301]

The rules that trip people up

• Use absolute URLs. Google recommends https://www.example.com/page/, not /page/.
• One method, one answer. Choose the HTML element or the HTTP header. Google supports both together but calls it more error-prone, and says not to give different canonicals through different methods.
• No fragments. Don’t point a canonical at a URL with #section in it.
• Point to a live, indexable page. In audits we regularly find canonicals aimed at URLs that redirect, return 404 or carry noindex. Each sends Google a contradiction.
• Make internal links agree. Google’s best practices include linking internally to the canonical URL. A menu linking to /services while every canonical says /services/ is a mixed message.
• Keep it in the source HTML. For JavaScript-built sites, Google prefers the canonical in the HTML; if a script must inject it, make sure no script changes it afterwards.

How do you check which URL Google chose?

Use the URL Inspection tool in Google Search Console. Its Page indexing section shows the user-declared canonical (what your tag says) and the Google-selected canonical (what Google actually chose). If they match, you’re fine. If they don’t, Google indexed a different URL from the one you asked for.

One catch: the live test can’t tell you this. Google’s help page says the live test “cannot predict whether or not the tested version will be considered canonical,” because that choice is made at indexing time. You see Google’s choice only for the version already indexed.

Reading the Page indexing report

For a site-wide view, open Indexing > Pages and look at the reasons pages aren’t indexed. Three of them involve canonicals, and only one is usually a problem:

• Alternate page with proper canonical tag. The page points to a canonical that is indexed. Google says there is nothing you need to do; this is the system working.
• Duplicate without user-selected canonical. Google found a duplicate with no canonical declared and picked one itself. Google calls this working as intended. Inspect a few: if Google picked the wrong version, declare your preferred canonical explicitly; if the pages aren’t really duplicates, make their content more distinct.
• Duplicate, Google chose different canonical than user. You declared a canonical and Google disagreed. This is the one to investigate.

For that last status, compare the two URLs in URL Inspection. The Page indexing help page notes that Google won’t choose a declared canonical that isn’t similar to the page. So either your tag points to the wrong page, or the two pages differ more than you think. Our guide to the technical SEO checklist puts this check alongside the other indexing tests.

Which fix should you use for each type of duplicate?

Choose by asking whether visitors still need the duplicate URL. If they don’t, redirect it. If they do, keep it live and canonicalize it. If there are too many URLs to manage one by one, stop crawlers reaching them. And if two pages only look alike because one is thin, the fix is content, not code.

1. 301 redirect: the duplicate can goOld URLs after a redesign, HTTP and non-www versions, a retired page replaced by a better one. Google recommends redirects “only when deprecating a duplicate page,” and says server-side redirects take effect fastest.
2. rel=“canonical”: the duplicate must stay reachableTracking-parameter URLs, printer views, a product listed under several categories, near-identical variants. Visitors keep their page; search engines get your preference.
3. robots.txt: there are thousands of themFaceted filter combinations and internal search results. This saves crawling rather than consolidating, so keep one clean, crawlable version of each listing page.
4. Differentiate or merge: the pages are “different” on paper onlyCity pages and service pages with swapped names. Add what is genuinely local or specific, or combine them into one stronger page and redirect the rest.
5. Noindex: the page is useful but shouldn’t be in searchThin tag archives or thank-you pages. Not for choosing between two real pages, since noindex removes rather than consolidates.

Then make your other signals agree. List only canonical URLs in your XML sitemap, link internally to canonical URLs and keep HTTP URLs out of sitemaps and hreflang annotations, as Google advises. Our guide to XML sitemaps and robots.txt covers both files in detail, and regional versions in the same language need hreflang as well as canonicals, which our international SEO guide explains.

Expect a delay. Google has to recrawl the affected URLs, and its troubleshooting guide says it “might hold pages in a duplicate cluster for up to two weeks” even after you fix the content. Pages split out faster when the differences are clear and significant. Use Request Indexing in URL Inspection for your most important URLs only, as there is a daily limit.

What about syndicated content and AI answers?

Syndication is where Google and Microsoft disagree, and where a duplicate can cost you the citation. If a trade publication republishes your article in full, search engines see two versions on two domains, and the bigger site’s copy can win.

Google and Bing disagree on the fix

Google’s troubleshooting guide says the canonical link element is “not recommended for those who want to avoid duplication by syndication partners, because the pages are often very different.” It calls the most effective solution having partners block indexing of your content, for example with a noindex tag on their copy. Microsoft’s December 2025 Bing post recommends a different tool: ask partners to add a canonical tag pointing to your original “when agreements allow,” and syndicate excerpts with a clear link back rather than full articles where possible.

Our view: negotiate for a noindex on the partner’s copy first, because it satisfies Google’s advice and leaves no doubt. If the partner refuses, ask for a canonical to your URL, which Bing explicitly supports. At a minimum, publish on your own site first and insist on a prominent link to the original. And for content copied without permission, Google points to contacting the host or filing a DMCA request.

How duplicates affect AI answers

Google and Microsoft have both published guidance here. Google says a page must be “indexed and eligible to be shown in Google Search with a snippet” to be a supporting link in AI Overviews or AI Mode. Its Search Console documentation adds that Google doesn’t serve duplicate pages. Our reading: within a duplicate cluster, only the URL Google selected as canonical is realistically in contention for that link. If Google picked a partner’s copy, the partner’s URL is the one that can be cited.

Microsoft is more direct. Its Bing post says large language models “group near-duplicate URLs into a single cluster and then choose one page to represent the set,” and may pick a version that is outdated or not the one you intended. It recommends IndexNow, a protocol that notifies participating search engines when URLs are added, updated or deleted, so consolidation is picked up faster.

None of this is a special AI technique. Google’s guide to optimizing for generative AI search, updated July 10, 2026, says that work is “still SEO,” and lists reducing duplicate content under building a clear technical structure. Clean canonicalization is ordinary SEO hygiene that now also affects which of your URLs an assistant can find and link.

How can you check your own site in 30 minutes?

Work through these steps with Search Console open. No paid tools are needed for a first pass.

1. Type four versions of your domainhttp and https, with and without www. All four should land on one HTTPS address in a single redirect.
2. View the source of five key pagesSearch for “canonical”. Each page you want indexed should have one tag, in the head, with its own absolute URL.
3. Add a parameter and recheckLoad a service page with ?utm_source=test on the end. The canonical should still show the clean URL.
4. Inspect your 20 most valuable URLsIn URL Inspection, compare user-declared and Google-selected canonicals. Note every mismatch.
5. Read the Page indexing reasonsOpen “Duplicate, Google chose different canonical than user” and look for patterns, such as all city pages or all products in one category.
6. Search a sentence from your best articlePut a distinctive sentence in quotes in Google. If another domain shows up with the same text, decide whether it is a partner (ask for noindex) or a copier (contact the host).
7. Check your staging siteSearch Google for your staging or development subdomain. If it appears, put it behind a password.

If you’d rather start from a report, our free SEO + GEO report analyzes your website, including redirects, indexability and canonicals, and ranks every issue from critical to low with how to fix it. Fixing what it finds is the kind of technical work our SEO + GEO services include. In our experience, duplicates travel with weak internal linking and unindexed pages, so fix them together.

Common questions

Sources

1. Google Search Central: What is URL canonicalization
2. Google Search Central: How to specify a canonical URL with rel="canonical" and other methods
3. Google Search Central: How to fix canonicalization issues
4. Google Search Central Blog (September 12, 2008): Demystifying the "duplicate content penalty"
5. Google Search Central: Spam policies for Google web search
6. Search Console Help: Page indexing report
7. Search Console Help: URL Inspection tool
8. Google Search Central: Managing crawling of faceted navigation URLs
9. Google Search Central: Pagination, incremental page loading, and their impact on Google Search
10. Google Search Central: AI features and your website
11. Google Search Central: Optimizing for generative AI search
12. Bing Webmaster Blog (December 19, 2025): Does duplicate content hurt SEO and AI search visibility?