Canonical Tag vs Noindex: Choose by the Job the Landing Page Must Do

Use a canonical tag when multiple landing-page URLs represent the same page and one should be the preferred search version. Use noindex when a landing page must remain accessible but should be removed from search results after it is recrawled.

Those are different jobs:

• Canonicalization: signal that several URLs are versions of one piece of content and identify the preferred URL.
• Noindexing: instruct a supporting search engine not to show this URL in its results.

For duplicate landing pages, canonical is usually the better fit. Google recommends rel="canonical" rather than noindex when the goal is to prevent selection of a duplicate as canonical: noindex removes the page from Google Search instead of identifying a representative URL for the duplicate set. Canonicalization can also consolidate signals such as links to duplicate URLs into the selected canonical (Google Search Central).

The decision table

Landing-page situation	Preferred control	Why
Tracking parameters produce the same content	Canonical to the clean URL	Keeps campaign URLs usable while identifying one search version
Persistent campaign URLs closely duplicate an organic page	Canonical to the organic page	The URLs are alternate versions of the same content
A paid-only or operational landing page should be excluded from organic search	noindex	Exclusion, not consolidation, is the requirement
A duplicate URL is obsolete and users no longer need it	Permanent redirect	Removes the unnecessary choice for users and crawlers
Alternate URL exists only for a temporary A/B test	Canonical to the original; use a temporary redirect if routing users	Preserves the original as the preferred search URL during the test
Pages target materially different products, locations or intents	Self-canonical and improve differentiation	They are not true duplicates and should not be collapsed merely because their templates overlap

The test is not “Are these pages similar?” It is: Should the alternate URL remain a destination, with its content represented in search by another URL? If yes, use canonical. If the URL should be excluded from search regardless of any duplicate relationship, use noindex.

What a canonical tag does—and does not do

Google clusters duplicate or highly similar pages and chooses a representative URL. A declared canonical is a strong signal, not a command: Google may select a different URL when other signals disagree (Google’s canonicalization overview).

For a duplicate campaign page, return a normal 200 OK response and put a link element in the document head with rel="canonical" and an absolute href pointing to the preferred URL. Put a self-referencing canonical on the preferred page as well.

Then align the surrounding signals:

• Use an absolute HTTPS URL in the canonical element.
• Link internally to the preferred URL rather than its duplicates.
• Include only preferred URLs in the XML sitemap.
• Keep the duplicate crawlable so Google can read its canonical.
• Do not specify competing targets through document markup, HTTP headers, redirects and sitemaps.

Canonical is not a reliable way to keep sensitive, private or legally restricted material out of search. It is a preference, not an access-control mechanism or guaranteed exclusion rule.

When noindex is the right choice

Apply noindex when the operating requirement is search exclusion. Examples include a paid-acquisition page with no organic-search purpose, an internal-results page that remains publicly reachable, or a short-lived partner landing page that should not appear in results.

For an HTML document, put a robots meta element in the document head with name="robots" and content="noindex". For HTML or non-HTML resources, send this HTTP response header instead:

X-Robots-Tag: noindex

Google must crawl the URL to discover either rule. If robots.txt blocks the page, Google cannot read its noindex; a previously known URL may remain in results. Google also does not support placing a noindex rule in robots.txt (Google’s noindex documentation).

Noindex does not prevent people from opening a URL, following a shared link or discovering the page outside search. If an offer or document must be limited to authorized visitors, enforce that requirement with authentication or another access-control mechanism.

Remove noindexed landing pages from XML sitemaps. As an operating policy, the sitemap should reinforce the URLs intended for indexing rather than contradict page-level instructions.

Do not combine canonical and noindex by default

Putting noindex on a duplicate while canonicalizing it elsewhere expresses two different desired outcomes: exclude this URL, but also treat it as a version of another URL. Even if a crawler can process both, the combination makes the intended behavior harder to reason about and verify.

Choose the control that matches the requirement:

• Need one representative from a duplicate set? Canonical.
• Need exclusion after recrawl? noindex.
• No longer need the duplicate as a destination? Redirect it.

For a retired duplicate, a server-side 301 or 308 is cleaner than leaving a crawlable page with a canonical. Google uses permanent redirects as a signal that the target should be canonical; temporary redirects generally preserve the source as the search URL (Google’s redirect guidance).

Special case: landing-page experiments

For an A/B test with alternate URLs, Google recommends canonicalizing the variants to the original URL rather than noindexing them. If the experiment redirects users from the original to a variant, use a temporary 302 rather than a permanent 301, and remove experiment URLs and routing when the test ends (Google’s website-testing guidance).

That advice applies to genuine test variants. If a “test” becomes a permanent page serving a different intent, revisit the URL strategy rather than leaving experimental controls in place indefinitely. A predeclared test plan can make that stopping decision explicit; see the content experiment design framework.

Verify the result in production

Do not stop at checking a template or CMS field. For representative URLs in each landing-page class:

1. Fetch the production response and confirm its status, redirect behavior and X-Robots-Tag headers.
2. Inspect the rendered document head for the final canonical and robots directives.
3. Confirm that robots.txt permits crawling when Google needs to see a canonical or noindex.
4. Check that internal links and sitemaps support the same preferred URL.
5. Use Search Console URL Inspection to compare the declared canonical with Google’s selected canonical. Google recommends checking its selected canonical first when troubleshooting a disagreement (Google’s canonicalization troubleshooting guide).
6. Monitor duplicate, canonical and excluded-by-noindex states after Google recrawls the pages.

If duplicate landing pages are multiplying faster than they can be reviewed, treat that as a URL-inventory problem rather than adding directives page by page. The broader content publishing and crawl budget workflow covers inventory cleanup and measuring crawl lag before increasing output.