Canonical vs Noindex: When to Use Each (Decision Table and Examples)

Canonical vs noindex comes down to one question: should this content stay in Google under another URL, or leave Google altogether? A rel="canonical" consolidates a duplicate into a preferred URL that stays indexed, while noindex removes the page from search results entirely.

They are not two strengths of the same control, and treating them as interchangeable is how pages you meant to keep disappear. This article covers what each tag does according to Google’s own documentation, a decision table for the situations that come up on real sites, what actually happens when you combine them (including with robots.txt and sitemaps), and how to check which instruction Google followed.

Key takeaways

• Canonical consolidates, noindex removes. A canonical helps Google consolidate the signals for duplicate URLs, such as links, into a single preferred URL. A crawled noindex makes Google drop the page entirely from Search results, regardless of whether other sites link to it.
• A canonical is a hint; noindex is obeyed once it is read. Google states that indicating a canonical preference is “a hint, not a rule”.
• Google does not recommend noindex as a canonicalisation tool. Its guidance says not to use noindex to prevent selection of a canonical page within a single site, because it completely blocks the page from Search.
• Most combinations are either pointless or contradictory. A self-referencing canonical on a noindexed page is harmless. A canonical to another URL plus noindex is a mixed signal, and robots.txt blocking hides both tags from Google.
• Some common cases need neither. Paginated pages, filter pages with real search demand and retired URLs are handled by a self-canonical, normal indexing or a redirect.

The difference in one table

Before the situations, the two controls side by side. Everything in this table comes from Google’s documentation on canonicalisation, consolidating duplicate URLs and blocking indexing.

	rel="canonical"	noindex
Question it answers	Which URL should represent this content?	Should this page be in Google at all?
What happens to the content	Stays indexed, under the canonical URL	Removed from search results
Link signals	Consolidated into the preferred URL	Not consolidated anywhere
Strength	A hint Google can overrule	Followed once Google crawls and extracts it
Where it lives	<link> in the <head>, or a Link HTTP header	Robots <meta> tag, or X-Robots-Tag HTTP header
Needs the page to be crawlable	Yes	Yes

If you need the syntax, the header form for PDFs, self-referencing canonicals or how frameworks output the tag, those now live in the canonical tags implementation guide. This page is about the choice.

What rel=“canonical” does

Canonicalisation is Google’s process of selecting the representative URL from a set of duplicate pages. Google clusters pages whose primary content looks the same, picks the one it considers most complete and useful, and crawls that canonical most regularly while crawling the duplicates less often.

A rel="canonical" annotation is your vote in that process. Two properties matter for the comparison.

First, nothing leaves the index. The duplicate’s signals are folded into the canonical, and the content keeps ranking under the canonical URL. A search result can still point at a duplicate when it suits the searcher better: Google gives the example that a mobile user will probably be sent to the mobile page even if the desktop page is canonical.

Second, it is a hint. Google ranks the ways of stating a preference as redirects (strong), rel="canonical" annotations (strong) and sitemap inclusion (weak), and says they stack when combined. It can still pick a different canonical for content-quality or technical reasons, which is exactly what the Search Console status Alternate page with proper canonical tag is reporting when it goes right.

What noindex does

noindex is a robots rule set with a <meta> tag or an HTTP response header. When Googlebot crawls the page and extracts the rule, Google drops that page entirely from Search results. There is no preferred URL and nothing is consolidated. How the tag is written, combined with other rules and audited is covered in the meta robots tag guide.

<meta name="robots" content="noindex">

For files that cannot carry a <meta> tag, such as PDFs, images and video, the same rule travels as a header. Google documents X-Robots-Tag: noindex for non-HTML resources, with Apache and NGINX examples for applying it to whole file types.

X-Robots-Tag: noindex

Three operating details decide whether it works:

• Google has to crawl the page to see it. If a URL is disallowed in robots.txt, its indexing rules are not found and are ignored. Putting noindex inside robots.txt is not supported by Google.
• It can be slow. Google says that, depending on the page’s importance, it may take months for Googlebot to revisit a page and notice a newly added rule. For urgent removals it points to its separate removals process.
• Negative rules add up. If you set nofollow for all robots and noindex for googlebot, Google uses the sum of the negative rules and treats the page as noindex, nofollow. none is equivalent to noindex, nofollow, which is worth knowing when a plugin setting reads “none”.

Decision table: canonical, noindex, both or neither

Match the tag to the intent. The right-hand column is the one to act on; the reasoning column is there so you can defend the choice to whoever owns the template.

Situation	Use	Why
Tracking parameters (?utm_source=, ?gclid=) on a real page	Canonical to the clean URL	Same content, and you want any links to the tagged URL credited to the clean one.
HTTP vs HTTPS, www vs non-www	Neither as the main fix: a 301 redirect, with a canonical as backup	Redirects are the strongest canonical signal, and there is no reason for the duplicate to resolve at all.
Colour or size variants of one product with near-identical copy	Canonical to the main product URL	A duplicate cluster where one URL should rank.
Sort-order variants (?order=price)	noindex	Google’s ecommerce guidance is to block unwanted sort and filter variations from being indexed with noindex, or discourage crawling with robots.txt.
Filter combinations nobody searches for	noindex (robots.txt only if crawl volume is the real problem)	No preferred version to consolidate into; it should not be in results.
Filter pages with real search demand (for example /dresses/green/)	Neither: index it with a self-referencing canonical	It is a landing page, not a duplicate. Give it unique copy.
Paginated pages 2 to n	Neither: each page canonicalises to itself	Google says not to use the first page of a paginated sequence as the canonical. See pagination SEO.
Internal site search results	noindex	Endless low-value URLs with nothing to consolidate into.
Thank-you, order confirmation, account pages	noindex	Should never be a search landing page.
Thin tag or author archives	noindex, or merge them	Not duplicates of one page, so a canonical has no sensible target.
Your article republished on a partner’s site	noindex on the partner’s copy	Google says the canonical element is not recommended for avoiding duplication by syndication partners, and the most effective solution is for partners to block indexing.
Same-language regional pages (US and UK)	Self-canonical on each, plus hreflang	Google recommends using both canonicalisation and hreflang for regional variants. See hreflang and canonical tags.
A PDF that duplicates an HTML page	Canonical via the Link HTTP header	Google supports a rel="canonical" HTTP header for non-HTML documents.
A PDF that should not appear in search	noindex via X-Robots-Tag	Removal, not consolidation.
A page you are retiring that has a replacement	Neither: 301 redirect	Google says to use redirects when you want to get rid of existing duplicate pages. See SEO redirects.
A staging or demo copy of the site	Neither on its own: put it behind authentication, then noindex as a backup	Google lists a demo site accidentally left open to crawlers as a common source of duplicates. A tag that gets copied to production is the bigger risk.

Two patterns fall out of the table. Canonical is for content that deserves to rank somewhere, and you are choosing which URL. noindex is for content that should not rank anywhere. When neither description fits, the answer is usually a redirect or a normal, indexable page.

What happens when you combine them

This is where most of the confusion sits, because the combinations behave differently. Here they are one at a time.

Combination	What Google sees	Verdict
noindex + self-referencing canonical	A page that names itself as canonical and asks not to be indexed	Harmless. Fine to leave.
noindex + canonical to another URL	”This is a duplicate of B” and “drop me” on the same page	Contradictory. Pick one.
Page A canonical to B, and B is noindex	A points its signals at a page that is being removed	Broken. Fix the canonical target.
noindex + robots.txt Disallow	Nothing: the page is never fetched, so the rule is never read	The noindex does not work.
Canonical on a duplicate + robots.txt Disallow on that duplicate	Nothing on the duplicate: the canonical is never read	Do not use robots.txt for canonicalisation.
Noindexed URL listed in the XML sitemap	A page suggested as canonical and also asked to be dropped	Mixed signal. Take it out of the sitemap.

noindex with a self-referencing canonical

Many templates print a self-referencing canonical on every page, so any page you noindex ends up carrying both. Conductor’s write-up of this case concludes that a self-referencing canonical plus a robots noindex is not harmful: the canonical says there is only one version of this page, and the noindex says not to index it. There is nothing to consolidate, so there is nothing to conflict. Do not spend a sprint removing it.

noindex with a canonical to another URL

This is the combination people mean when they ask about using both. It asks Google to treat the page as a duplicate of B and fold its signals into B, while also asking Google to remove it.

Google’s public statements on it have not been consistent. In a 2021 office-hours answer reported by Search Engine Journal, John Mueller said that with a noindex the page is “definitely not shown”, with a rel canonical it is “more likely not shown”, and you “can also do both” so that “maybe some of the signals” can be forwarded when external links point at the page. Another Mueller comment, quoted by Conductor, points the other way: you shouldn’t mix them because they are “very contradictory pieces of information”, and Google will “generally pick the rel=canonical and use that over the noindex”.

The documentation is the tie-breaker, and it is clear: don’t use noindex to prevent selection of a canonical page within a single site; rel="canonical" is the preferred solution. So decide the intent. If the page should pass its signals on, use the canonical alone. If it should vanish and its links do not matter, use noindex alone. If it should vanish and its links matter, a 301 redirect usually does both jobs better than either tag.

A canonical pointing at a noindexed page

The spread-out version of the conflict: page A canonicalises to B, and B carries noindex. A says “B is the version that counts”, and B asks to be removed. Because the canonical is a hint and the noindex is followed once read, you have aimed A’s consolidation at a page that will not be in the index.

This rarely happens on purpose. It is usually a CMS template or plugin pointing canonicals at the wrong URL, which Google lists among the most common canonicalisation issues. The fix is to make every canonical target an indexable, 200-status URL, and to make sure no page that other pages canonicalise to carries noindex. A crawl that records each URL’s canonical target and robots tags will surface these in one pass.

Either tag plus a robots.txt block

robots.txt controls crawling, not indexing. If the URL is disallowed, Google never fetches the HTML, so it never reads a canonical or a noindex on it. Google says it ignores indexing and serving rules on pages disallowed in robots.txt, and separately warns not to use robots.txt for canonicalisation, because Google may still index disallowed URLs without their content.

The practical rule: if you want a noindex or a canonical to work, the page has to be crawlable while Google reads it. Block crawling afterwards only if crawl volume is the problem you are solving. The robots.txt guide for developers covers that sequence.

Sitemaps

Google treats every URL listed in a sitemap as a suggested canonical. Listing a noindexed URL, or a duplicate that canonicalises elsewhere, sends a weak signal in the opposite direction to your tag. Google also says not to specify different canonical URLs for the same page using different techniques. Generate the sitemap from the same rule that sets your canonicals and robots tags, so the three cannot drift apart.

How to check which instruction Google followed

You cannot see a canonical or noindex conflict on the rendered page. Read the source, then check what Google decided.

1. URL Inspection in Search Console. It shows which page Google considers canonical and lets you see the HTML Googlebot received, which is how you confirm a noindex is actually there for Googlebot and not just in your CMS preview.
2. The Page Indexing report. Google says you can use it to monitor the pages from which Googlebot extracted a noindex rule. An unexpected URL in that list is usually a template or plugin setting. Pages Google has clustered as duplicates show up under statuses such as the alternate-page one, and pages it chose not to index for other reasons land in Crawled, currently not indexed.
3. Budget for the lag. Once you fix the content or signals, Google may hold pages in a duplicate cluster for up to two weeks before re-evaluating, and a newly added noindex waits for the next crawl. Request Indexing is subject to quotas, so save it for the URLs that matter.

Where this fits

Canonical and noindex are two of the indexation controls, alongside redirects, robots.txt and the sitemap. They collide most during a site migration, where templates re-stamp canonicals and a staging noindex can ride into production. Checking that every indexable page self-canonicalises, that no canonical points at a noindexed or redirected URL, and that the sitemap agrees with both is one section of a technical SEO audit.

Frequently asked questions

Can you use canonical and noindex together on the same page?

With a self-referencing canonical, yes, and it does no harm. With a canonical that points to a different URL, you should not: Google’s documentation recommends against using noindex to influence canonical selection, and the two tags ask for opposite outcomes. Decide whether the page should pass its signals on (canonical) or disappear (noindex), and use one.

Does a canonical tag remove a page from Google?

No. A canonical tells Google which URL should represent a set of duplicates and consolidates their signals into that URL. The content stays indexed under the canonical URL, and Google can still overrule your choice. To remove a page from results, use noindex.

Should I use canonical or noindex for duplicate content?

Use a canonical when the content should rank under one preferred URL, such as parameter, protocol or product-variant duplicates. Use noindex when there is no preferred version to consolidate into and the page should not appear at all, such as internal search results or sort-order variants. If the duplicate URL does not need to exist, a 301 redirect beats both.

Should paginated pages use canonical or noindex?

Neither, in the usual case. Google says to give each page in a paginated sequence its own canonical URL rather than pointing them at page 1, because pages 2 onwards list items page 1 does not. Noindexing them risks the products or articles that only appear deeper in the series.

Why is my noindexed page still showing in Google?

Usually because Google has not read the rule yet. It has to crawl the page to see a noindex, and for a low-priority page that revisit can take months. The other common cause is a robots.txt Disallow on the same URL, which stops Google from ever seeing the tag. Remove the block, check the HTML Googlebot received in URL Inspection, and request a recrawl.

Is noindex a directive or a hint?

Google follows noindex once it crawls the page and extracts the rule, so it behaves as a directive. A canonical is explicitly a hint, not a rule, and Google can choose a different canonical. That difference in strength is one reason the two should not be stacked to say the same thing.

Next step

Canonical picks which duplicate URL Google indexes, noindex removes a page, and the decision table above covers the cases where the answer is neither. For how to write, test and deploy the canonical itself, continue with the canonical tags implementation guide. If you would rather have the crawl-and-indexing layer designed and monitored than reconciled by hand after each release, that is part of GEO and technical SEO consulting.