Should you really implement device-based redirection between mobile and desktop?

Why is this recommendation concerning separate mobile/desktop URLs important?

Google refers to an architecture that was very common between 2010 and 2015: a main site (www.example.com) alongside a distinct mobile version (m.example.com or mobile.example.com). This approach requires a technical connection between the two versions for Google to understand that it’s the same content.

Mueller's recommendation focuses on server-side redirections based on the User-Agent. A mobile visitor accessing www.example.com/page should be automatically redirected to m.example.com/page. Conversely, a desktop user landing on the mobile URL should be returned to the desktop version. This bidirectionality guarantees that each user reaches the correct version.

What’s the difference with alternate/canonical annotations?

There are two methods to connect distinct mobile/desktop URLs. The first involves alternate and canonical tags: the desktop page contains <link rel="alternate" media="only screen and (max-width: 640px)" href="m.example.com/page">, and the mobile page contains <link rel="canonical" href="www.example.com/page">.

The second method—which is the one Mueller advocates here—relies on 302 redirections based on the User-Agent. With this approach, there’s no need for tags: the server detects the type of device and automatically redirects. Google recommends this method because it immediately enhances user experience without waiting for the browser to parse the HTML and find the alternate tag.

Is this architecture still relevant today?

No. Let’s be clear: this recommendation applies to a minority of sites that still maintain separate mobile/desktop URLs. Since the introduction of responsive design and especially the widespread mobile-first indexing in 2021, Google indexes the mobile version of all sites—whether they are responsive, adaptive, or have separate URLs.

Most sites created after 2016 utilize a single URL with a responsive design. This is easier to maintain, less costly, and eliminates risks of content inconsistency between versions. Separate URLs persist mostly on older e-commerce sites, legacy media, or platforms developed before the general adoption of responsive design.

• User-Agent redirections ensure that every visitor reaches the correct version without friction
• This architecture requires double maintenance: two URLs, two templates, two content versions
• Responsive design (single URL) has become the standard as it is simpler and aligned with mobile-first indexing
• Google continues to support separate URLs but no longer recommends them for new projects
• Bidirectional redirection is critical: forgetting the desktop→mobile or mobile→desktop direction creates UX and indexing issues

Is this recommendation still consistent with real-world practices?

Yes, but only for sites that have already a separate mobile/desktop architecture. In these specific cases, Mueller’s logic holds: better to have well-configured User-Agent redirections than poorly implemented or missing alternate/canonical tags. Redirections offer an immediate user experience without depending on HTML parsing.

But be cautious—and this is where the advice can become misleading for some readers: Google does not say that you should adopt this architecture today. It’s a maintenance directive, not a design recommendation. If you’re launching a new site or redesigning an old one, moving to a single responsive URL will always be the best option in terms of development cost, SEO consistency, and alignment with mobile-first indexing.

What concrete risks do User-Agent redirections pose?

The main danger is imperfect device detection. User-Agents can be modified, outdated, and some browsers send ambiguous UAs (tablets masquerading as desktops, for example). If your redirection logic is too rigid or based on a fixed UA list, you risk redirecting users to the wrong version.

The second risk is a broken bidirectionality. Many sites correctly redirect desktop→mobile but forget mobile→desktop. The result: a desktop user clicking on a shared link pointing to m.example.com gets stuck on the mobile version. Google sometimes crawls these inconsistencies and may index the wrong version, creating duplicates or conflicting signals. [To check]: Google claims that redirections improve recognition, but we lack public data on the indexing error rates between sites with User-Agent redirections and those with properly implemented alternate/canonical tags.

In which cases does this rule not apply?

If your site uses a single URL with responsive design, this recommendation does not concern you. The same applies if you are using dynamic serving (same URL, different HTML depending on the device, with a Vary: User-Agent header)—in this case, no redirection, just a specific HTTP header.

If you are on a modern framework (Next.js, Nuxt, etc.) with server-side rendering or static site generation, the question doesn't even arise: you serve the same URL to all devices, with adaptive CSS. User-Agent redirections are a relic from an era when responsive design didn’t exist yet.

What should you concretely do if your site uses separate URLs?

First step: audit your current redirections. Test manually with both a mobile and a desktop (or change your User-Agent in DevTools) to verify that the redirection works both ways. Use a tool like Screaming Frog in ‘Mobile’ mode and then in ‘Desktop’ mode to crawl both versions and identify inconsistencies.

Next, ensure that your redirections are properly set to 302 (temporary), not 301. Google recommends 302 for context-based redirections (device, language, geo) since the “canonical” URL remains unchanged—it’s just a contextual adaptation. A 301 would signal a permanent redirection, which could mislead Google about which version to index.

What errors should you avoid during implementation?

Classic mistake: redirecting all mobile URLs to the desktop homepage (or vice versa) instead of mapping URL by URL. If a mobile user lands on m.example.com/product/shoes and you redirect them to www.example.com instead of www.example.com/product/shoes, you break the experience and lose qualified traffic.

Another pitfall: combining User-Agent redirections AND alternate/canonical tags. Google states that redirections are sufficient—adding tags can create conflicting signals. If you’re redirecting, no need for tags. If you’re using tags, no need to redirect. Choose one method, not both. Finally, watch out for redirection chains: if m.example.com/page redirects to www.example.com/page, which then redirects to www.example.com/page/ (with final slash), you create unnecessary latency and risk losing crawl budget.

How can you check if your site complies with this recommendation?

Use Google Search Console: in the ‘Settings > Crawling’ section, check that Google properly detects both versions (mobile and desktop) and that there are no redirection errors. Also, consult the ‘Coverage’ report to identify any potential duplicates or pages incorrectly indexed on the wrong version.

Test with the URL inspection tool in GSC: inspect a desktop URL, then its mobile equivalent. Google should clearly identify each version and not report any canonical conflict. If you see warnings like ‘submitted URL not selected as canonical’, it means your redirections or tags are not consistent.

• Ensure that each mobile URL redirects to its desktop equivalent (and vice versa), not to the homepage
• Use 302 redirections, not 301, for context-based redirections based on the device
• Do not combine User-Agent redirections and alternate/canonical tags—choose one method
• Test with Screaming Frog in Mobile then Desktop mode to detect mapping inconsistencies
• Check Google Search Console to confirm Google is indexing the correct versions without redirection errors
• Avoid redirection chains (m.example.com → www.example.com → www.example.com/) that slow down crawling