Should you really cross canonical tags between mobile and desktop?

Why is Google still addressing the issue of separate URLs for mobile?

This statement may seem anachronistic: responsive design has dominated the web for years. Yet, configurations with dedicated mobile URLs (like m.example.com or example.com/mobile/) persist, particularly on older e-commerce platforms or high-traffic sites that have not migrated.

Google acknowledges this reality and provides a clear technical framework: bidirectional canonical tags are legitimate for signaling equivalence between versions. This allows the search engine to consolidate signals (backlinks, engagement, authority) toward a single reference URL, even if the content is served on two distinct domains or paths.

How does this cross-canonicalization system actually work?

The principle is simple: your mobile page (m.example.com/page-a) includes a rel="canonical" tag pointing to the desktop version (example.com/page-a). Simultaneously, the desktop version points to itself or... to the mobile version if you want to prioritize the latter in indexing.

Google treats these signals as consolidation indicators: it understands that the two URLs are equivalent and chooses one of them as the canonical version for indexing. This is not an absolute directive — Google might ignore it if the pages present substantial differences in content or structure. Thus, consistency between the two versions remains essential.

In what context is this practice still relevant today?

Real-world use cases are limited. Sites still in m-dot configuration (separate mobile domain) or with distinct paths for mobile/desktop are the main ones concerned. These architectures are often inherited from old technical constraints: rigid CMS, proprietary platforms, or strategic choices for differentiating experiences.

Some e-commerce giants or high-traffic media maintain this separation for reasons of server performance or advanced personalization. In these contexts, cross-canonicalization prevents index duplication and preserves consolidated PageRank. But for the majority of modern projects, this complexity is no longer necessary.

• Bidirectional canonical tags are a consolidation tool for distinct mobile/desktop URLs.
• Google uses these signals to avoid authority dilution between equivalent versions.
• This setup is a technical legacy: responsive design makes this practice obsolete for most sites.
• The consistency of content between the two versions is critical: major discrepancies can cause canonicalization to fail.
• Google reserves the right to choose another URL than the one indicated if the signals are contradictory.

Is this statement consistent with practices observed in the field?

Absolutely. Google has been emphasizing for years that canonical tags are signals, not orders. This clarification on mobile/desktop bidirectionality confirms what we observe in audits: sites with separate URLs that properly manage their canonicals actually see their signals consolidated in Search Console.

The catch? The proportion of sites that poorly implement this setup is high. Contradictory canonicals (mobile pointing to desktop, desktop pointing to a third URL), truncated content on mobile not reported, missing markup on certain pages: all configurations that neutralize the expected effect. [To be verified]: Google does not publish statistics on the failure rate of these implementations, but field feedback suggests significant fragility.

What nuances should we add to this official claim?

Google says "correct", not "optimal". The wording is revealing: this practice is accepted, not recommended. The implicit message? If you are still on this architecture, do it properly. But if you are building a new site, choose responsive.

Another nuance: bidirectionality works if and only if the pages are truly equivalent. Lightweight mobile content, missing features, truncated sections: all signals that can lead Google to ignore your canonicals. Observations show that discrepancies of more than 30-40% of content between versions are enough to cause consolidation to fail.

In what cases can this configuration cause issues?

First scenario: you have a mobile-first version that is more complete than desktop (rare, but observable in some media). If your canonical points to the desktop, you risk losing mobile engagement signals that do not find an equivalent on the desktop side.

Second scenario: loading times significantly differ between the two versions. If your mobile is fast but the desktop is slow (or vice versa), Google may favor the most performant version regardless of your canonicals. Core Web Vitals now influence the choice of the canonical version.

What should you do if your site uses separate URLs?

First imperative: audit the consistency of your canonical tags. Crawl your two versions (mobile and desktop) and check that each pair of equivalent URLs correctly points to each other. Use Screaming Frog, Oncrawl, or Sitebulb to extract all canonicals and cross-reference the data.

Second action: compare the actual content between versions. An automated diff (tools like Diffchecker or custom scripts) allows you to identify substantial gaps. If you find major discrepancies, either harmonize the content or accept that Google may arbitrarily choose which version it prefers to index.

What mistakes should you absolutely avoid in this setup?

Classic error number one: looping canonicals. Mobile points to desktop, desktop points to an intermediate URL that points back to mobile. Google gives up and indexes what it wants. Ensure that each chain of canonicals resolves in a single hop to the final URL.

Error number two: forgetting about alternate hreflang. If you have mobile/desktop URLs AND multilingual content, you must cross-reference the two markup systems. This is complex, prone to frequent errors, and another reason to migrate to responsive if possible. Hybrid configurations combining canonical + hreflang + alternate mobile are bug-prone.

How can you check that Google respects your canonicals?

Search Console is your main ally. Go to "Coverage" > "Excluded" and filter for "Other pages with appropriate canonical tag". You should find your non-canonical URLs (mobile or desktop depending on your choice) listed there. If they appear in "Indexed", it means Google chose to ignore your signals.

Complement this with a site:m.example.com query in Google. If mobile pages appear while you've canonicalized to desktop, dig deeper: either the markup is missing/broken, or Google deems the versions too divergent. Server logs can also reveal whether Googlebot crawls both versions intensively instead of focusing on the canonical.

• Crawl both versions (mobile and desktop) and extract all canonical tags to detect inconsistencies and loops.
• Compare content across versions with a diff tool to identify substantial gaps (threshold: <30% divergence).
• Check in Search Console that non-canonical URLs appear correctly in "Excluded" with the appropriate status.
• Manually test a sample of URLs using the URL Inspection tool to confirm which version Google considers canonical.
• Monitor server logs: Googlebot should focus its crawl on canonical URLs, not equally on both versions.
• If you also manage hreflang, check the cross-compatibility of the markups (canonical + alternate + hreflang) with a specialized validator.