Should you really keep canonical and alternate tags after switching to mobile-first?

Why does Google insist on maintaining existing tags?

The mobile-first indexing has created confusion among many SEO practitioners. Some thought that by switching to this mode, the historical tags linking desktop and mobile versions lost their utility.

However, Google continues to crawl both versions in most cases. The canonical and alternate annotations still serve to clarify the relationship between these versions, even if indexing prioritizes mobile. Removing these tags would confuse Googlebot.

What does this mean in practice for a site with a separate configuration?

A site with distinct URLs for desktop (www.example.com) and mobile (m.example.com) relies on a cross-signaling system: the desktop page points to its mobile equivalent via rel="alternate", and the mobile page refers back to the desktop via rel="canonical".

With mobile-first, Googlebot primarily crawls m.example.com. But it still needs to understand that this mobile version is not an isolated page — that it indeed corresponds to www.example.com. Without these tags, there is a risk of duplication or loss of signal.

Does this statement also apply to responsive sites?

No. A responsive site only has one URL per content, so there is no need for alternate or canonical tags between desktop/mobile versions. This guideline from Mueller applies only to separate configurations (m.example.com) or sites using dynamic serving with content variations based on the user-agent.

If your site is responsive, you are not affected by this statement. You can ignore this debate and focus on other aspects of mobile-first.

• Separate configuration (distinct URLs): absolutely maintain canonical and alternate tags
• Dynamic serving (same URL, different content): the Vary: User-Agent directive remains essential
• Responsive: no action required, one HTML for all devices
• Switching to mobile-first does not change the necessity for these technical annotations
• Removing these tags may lead to indexing errors or duplication

Is this guideline consistent with real-world observations?

Overall, yes. We are indeed seeing indexing issues on sites that have removed their canonical/alternate tags thinking they were doing the right thing after transitioning to mobile-first. Google continued to crawl both versions and found itself lost without these signals.

What is less clear is why Google maintains such a systematic double crawl. If indexing is truly mobile-first, why this persistent reliance on desktop annotations? [To be verified] to what extent Googlebot continues to actively crawl the desktop versions of separate sites — official data is lacking.

What nuances should be added to this rule?

Mueller's statement is correct, but it obscures a reality: having separate URLs for mobile and desktop is an inherited architecture, technically heavy and prone to recurring errors. If you are still in that situation, the real question is not 'should I keep my tags?', it's 'when will I switch to responsive?'

Maintaining two synchronized versions, with impeccable cross-markup, is an operational nightmare. Configuration errors are common: canonical pointing to the wrong page, missing alternate, non-equivalent content between versions. All this creates noise and dilutes the SEO signal.

In what cases might this directive not apply?

If you have a site with an extremely simplified mobile version, almost a subset of content compared to desktop, it may be tempting to canonicalize mobile to itself and abandon the desktop version as the reference. However, Google might view this as reverse cloaking.

Another edge case: sites that have migrated to responsive but keep m.example.com as a permanent 301 redirect. Here, the alternate/canonical tags indeed become obsolete once the redirect is in place. But be cautious not to remove the tags before the migration is completed and stable in the index.

What should you do concretely on a site with a separate configuration?

First, check the integrity of your existing tags. Every desktop page should have a rel="alternate" pointing to its mobile equivalent, and every mobile page should have a rel="canonical" pointing to the desktop. No exceptions, no orphan pages.

Next, ensure the content consistency between both versions. If the mobile page displays significantly different or slimmed-down content compared to the desktop, Google may ignore your tags and index both versions separately, creating cannibalization.

What mistakes should be absolutely avoided?

Never remove canonical/alternate tags thinking that mobile-first renders them useless. This is the most common mistake since the deployment of mobile-first. Google continues to rely on these signals to consolidate the indexing of both versions.

Another trap: creating chains of redirects or loops between mobile and desktop. If m.example.com redirects to www.example.com, which redirects back to m.example.com based on the user-agent, you are creating a labyrinth for Googlebot. Ensure your redirects are clear and unidirectional.

How can you verify that your configuration is compliant?

Use the URL inspection tool in Search Console. Test a desktop page and check that Google correctly detects the alternate mobile version. Then test the mobile page and confirm that the canonical points to the desktop.

Crawl your site with Screaming Frog or a similar tool in mobile user-agent mode, then in desktop mode. Compare the two exports to identify inconsistencies: pages without alternate, missing canonical, mismatched content. Correct methodically.

• Verify that each desktop page has a rel="alternate" to its mobile version
• Check that each mobile page has a rel="canonical" to the desktop
• Compare the content of both versions to avoid significant discrepancies
• Test URL inspection in Search Console on both versions
• Crawl the site in mobile and desktop to detect inconsistencies
• Never remove tags without prior validation in the index