Why does Google overlook the canonical tags on the mobile version?

What is the technical reasoning behind this rule?

Google has always prioritized the desktop version as the canonical reference, even after the shift to Mobile-First Indexing. This statement may seem counterintuitive at first glance: why ignore signals from the mobile version when it is the one crawled first?

The answer lies in the separation of roles: crawling happens on mobile, but the canonical authority remains defined by the desktop. Canonical tags on mobile are considered redundant or potentially contradictory. Google prefers a single source of truth to avoid conflicting signals between the two versions.

How does this affect the treatment of internal links?

The point about links is crucial. When Googlebot crawls your mobile version and encounters a link, it does not treat that link as belonging to the mobile version. It automatically transposes this link to its desktop equivalent.

Specifically? A link https://example.com/m/page on your mobile site will be interpreted as pointing to https://example.com/page if that is your desktop structure. This automatic normalization aims to avoid dilution of PageRank between the mobile and desktop versions of the same resource.

What does this change for responsive sites?

For a responsive design site, this rule has no practical impact. There is only one URL, so only one canonical tag. No mobile/desktop distinction.

The issue arises only for architectures with distinct URLs: m-dot sites (m.example.com) or with suffixes like /mobile/. In these configurations, placing canonical tags on the mobile version is unnecessary. Only those on the desktop count.

• Mobile canonical tags are ignored: Google does not consider them to determine which version to index
• Desktop remains the canonical reference: even in Mobile-First, it defines which URL to index
• Mobile links are transposed: Googlebot treats them as if they came from the desktop version
• Responsive circumvents the issue: one URL = one canonical signal, no possible conflict

Is this statement consistent with what we observe in the field?

Yes, and it is indeed a recurring source of confusion. We regularly see sites with contradictory canonical tags between mobile and desktop, without generating any glaring errors in Search Console. Why? Because Google simply ignores the mobile version.

However, what sometimes gets tricky is when sites remove the canonicals from the desktop thinking that those from the mobile would suffice after the shift to Mobile-First. Classic mistake. The result: loss of canonical consolidation, duplicated pages reappearing in the index.

What nuances should we add to this rule?

Google talks here about canonical tags but does not mention other canonicalization signals. The 301/302 redirects from mobile to desktop are also processed but differently: they work, unlike the canonicals.

Another point: [To verify] for sites with dynamic serving (same URL, different content based on user-agent). Google claims to crawl primarily with a mobile user-agent, but sometimes we observe inconsistencies when the mobile content differs too much from the desktop. In this case, it's hard to know which version is truly taken into account.

In what scenarios could this rule cause problems?

The classic trap: sites that have a separate AMP version in addition to a traditional mobile version. The canonical tag for the AMP page must point to the canonical mobile version (or desktop depending on the architecture), but if the latter has its own mobile canonical ignored by Google, we end up with a shaky chain of canonicals.

Another problematic case: international sites with distinct mobile URLs by language. If the hreflang tags are only present on mobile, they risk being ignored or misinterpreted. Let's be honest: it's rare, but when it happens, it's a nightmare to debug.

What should I check on my site now?

First step: audit the consistency of canonical tags between your mobile and desktop versions if they exist separately. Open your site on mobile and desktop, inspect the source code, compare. Do the desktop canonicals properly point to the right URLs?

If your site is in m-dot or has distinct mobile URLs, ensure that each desktop page has its own canonical tag. The mobile version may have one for symmetry (some CMS generate them automatically), but don't rely on that to fix issues on the desktop side.

What mistakes should you absolutely avoid?

Never remove canonical tags from the desktop under the pretext that Google crawls mobile first. Crawling and indexing are two distinct steps. Google can crawl mobile but refer to the desktop for canonical signals.

Another common mistake: leaving mobile canonicals pointing to mobile URLs while the desktop points to different desktop URLs. This creates a logical inconsistency that Google resolves by ignoring mobile, but it often reveals a deeper architectural problem.

How can I verify that everything is in order?

Use Google Search Console: in the Coverage section, check that the indexed URLs match what you expect. If you see mobile URLs indexed while you have desktop canonicals, that's a warning sign.

Run a crawl with Screaming Frog in desktop mode and mobile mode separately. Compare the extracted canonical tags. If they diverge on important pages, dig deeper: either you fix the desktop or you accept that the mobile version may be decorative from an SEO standpoint.

• Audit the desktop canonical tags and ensure they point to the right URLs
• Compare the mobile/desktop canonicals on a sample of key pages
• Check in Search Console that the indexed URLs are consistent with your desktop canonicals
• Crawl the site in desktop and mobile mode to identify divergences
• Remove or correct contradictory canonicals between mobile and desktop