Should you opt for subdirectories or subdomains for a multilingual site?

Why is this distinction between subdirectories and subdomains so important?

Google treats subdomains as almost independent entities. Each subdomain (de.example.com, en.example.com) operates as a distinct site in the eyes of crawl and ranking algorithms. Ranking signals — backlinks, domain authority, user signals — do not automatically transfer from one subdomain to another.

Subdirectories (example.com/de/, /fr/), on the other hand, share the same root domain. All positive signals accumulate on the same main domain. A backlink to example.com/fr/article indirectly benefits example.com/en/another-page through the shared domain structure.

This distinction becomes critical for sites with uneven language version sizes. If your English version generates 80% of the traffic and backlinks, the Spanish version may suffer from a lack of signals if it is isolated on es.example.com. In subdirectories, it benefits from the overall authority of the domain.

What does "consolidating SEO signals" really mean?

Domain authority is not an official Google score, but the consolidation of trust signals is very real. When all content resides under the same root domain, every backlink acquisition, every quality signal strengthens the domain as a whole.

Take a practical example: you launch an international PR campaign that generates 50 quality backlinks to English articles. With subdirectories, these signals benefit the entire site — including your Polish and Italian versions that received no direct backlinks. With subdomains, only en.example.com benefits.

The crawl budget also allocates differently. Google assigns a budget per domain. With multiple subdomains, you fragment this budget — it.example.com with 200 pages may receive fewer crawls than necessary if Google treats it as a weak distinct site. In subdirectories, the overall domain crawl budget distributes more effectively.

Why does analytical tracking become more complex with subdomains?

Classic analytics tools (Google Analytics 4, Matomo) consider each subdomain by default as an external traffic source. Moving from fr.example.com to de.example.com appears as an external visit, not as an internal navigation. This fragments your user journeys and complicates the analysis of multilingual conversions.

Setting up cross-domain tracking between subdomains is technically possible, but it adds a layer of technical complexity and risk of errors. Every tag manager update or configuration change can break the tracking. In subdirectories, everything works natively as a single site.

Google Search Console also segments data by domain. With subdomains, you must manage multiple separate properties, manually cross-reference data, losing the overall view of global performance. In subdirectories, a single property suffices with folder filters.

• Subdomains fragment domain authority, forcing each language version to build its own ranking signals
• Subdirectories consolidate all SEO signals under one root domain, benefitting less popular versions
• Analytical tracking becomes natively simpler in subdirectories, without complex cross-domain configurations
• The crawl budget distributes more effectively when all languages reside under the same domain
• This recommendation specifically targets sites with 3 to 5 language versions — logic changes for massive deployments

Does this recommendation really apply to all multilingual contexts?

No. Mueller clearly states "few versions (3-5 languages)", and this is not insignificant. For a site with 30 or 50 language versions, the economic and technical logic changes radically. Managing example.com/ar/, /bn/, /cs/, /da/, /el/, /fa/... up to 50 directories can create major operational complexity: heavy server configurations, management of geolocated redirects, risks of routing errors.

Major international players (Amazon, Wikipedia, Airbnb) often use country domains (ccTLD) or subdomains to compartmentalize their versions technically and organizationally. When you have dedicated teams by market with regional hosting, fragmentation becomes an asset rather than a handicap. [To verify]: Mueller doesn't provide any specific threshold where the balance tips, nor quantitative data on the actual impact of this fragmentation on real sites.

Another overlooked case: sites with radically different content by market. If your German version offers a distinct product catalog, a different editorial architecture, or even a unique business model, a subdomain can simplify technical management and team autonomy. The "signal fragmentation" becomes secondary compared to organizational gains.

Does data from the field truly confirm a measurable advantage of subdirectories?

To be honest: rigorous comparative studies on this subject are rare. Most "evidence" circulating in the industry is anecdotal — "I migrated from subdomains to subdirectories and my traffic increased by 15%". These cases systematically ignore other variables: simultaneous technical redesigns, content improvements, algorithm changes during the period.

What we observe on real migrations: the main gain often comes from correcting existing technical errors (poor hreflang implementation, cannibalization between versions, badly managed duplicate content) rather than the structure itself. A site with well-configured subdomains, clean hreflang, and a coherent content strategy can significantly outperform a poorly organized subdirectory site.

The point on which Mueller is objectively right: simplified tracking is a fact, not an opinion. The reduction of analytical complexity is measurable and real. But regarding pure SEO impact? Field feedback shows that the structure matters less than execution — correct hreflang signals, non-duplicated content, positive user signals.

What real risks does this recommendation underestimate?

The risk of cross-language cannibalization is rarely mentioned. In subdirectories, all your language versions live under the same roof. If your hreflang tags have even slight issues or if Google doubts a page's target language, you can create situations where example.com/en/shoes and example.com/fr/chaussures cannibalize each other in some countries' SERPs.

With subdomains, this natural separation acts as an implicit safety barrier. fr.example.com and en.example.com are less likely to compete for the same positions. This is not an argument to systematically use subdomains, but it's a nuance that Mueller overlooks.

Another point: migrating from subdomains to subdirectories is not trivial. It involves massive 301 redirects, a transition period during which Google must crawl and reindex the entire site, and risks of temporary ranking losses. For a site generating 500K€/month in organic revenue, the risk/benefit deserves tighter analysis than a simple "it's generally preferable".

How should I concretely decide between subdirectories and subdomains for my site?

Start by mapping your actual situation, not your ideal situation. How many active language versions do you have today? How many planned for the next 24 months? If you're at 3 languages with a plan to reach 8, you’re in the zone where subdirectories remain manageable. Beyond 15-20 versions, reassess.

Evaluate the SEO maturity disparity among your versions. If your English version generates 95% of traffic and the others are embryonic, consolidating signals through subdirectories will give them a real boost. Conversely, if you have balanced versions with autonomous local teams, the fragmentation of subdomains may be acceptable.

Another criterion rarely mentioned: your technical stack and hosting capabilities. Serving 8 language versions from the same origin server via subdirectories may present latency issues for geographically dispersed users. Regional hosting subdomains may offer better performance — and Core Web Vitals now weigh heavily in rankings.

What implementation errors must be absolutely avoided?

The number one error: choosing subdirectories and then botching the hreflang implementation. If your hreflang tags are inconsistent, incomplete, or missing, you lose the primary benefit of this architecture. Google won’t know which version to serve to which user, resulting in cross-language duplicate content.

Error two: duplicating translated content automatically without added value. Having example.com/en/, /fr/, /de/ with unedited Google Translate content is worse than having just one well-done version. User signals (bounce rate, time on page) will be catastrophic and contaminate the entire domain in subdirectories.

Error three: neglecting server geolocation and CDNs. A site in subdirectories hosted solely in Europe with Asian users will suffer from latency, regardless of your URL structure's beauty. A geo-distributed CDN becomes essential to compensate for the physical content centralization.

Should I migrate an existing site from subdomains to subdirectories?

Only if you have identified concrete and measurable SEO issues that this migration would solve. "Google recommends it" is not a sufficient reason to risk redesigning the architecture of an established site. Ask yourself the right questions: are your secondary versions stagnating despite quality content? Is your analytics tracking broken to the point of hindering optimization?

If you migrate, do so gradually and with a rollback plan. Start with a secondary language version, measure the impact over 3 months, then expand. Document each 301 redirect, monitor crawl logs, meticulously track rankings and traffic by version.

Keep in mind that this migration will mobilize significant technical and SEO resources: pre-migration audit, redirect plan, updating all internal links, revising sitemaps and hreflang, post-migration monitoring. This type of project can quickly become complex and may require specialized support. If you lack internal expertise or bandwidth, enlisting an experienced SEO agency for multilingual migrations can help you avoid costly mistakes and speed up return on investment.

• Audit the current architecture and map all existing and planned language versions
• Check the current hreflang implementation with tools like Screaming Frog or OnCrawl
• Assess the SEO performance disparity between versions (traffic, backlinks, rankings) using segmented Search Console data
• Test loading times by geography to identify CDN or regional hosting needs
• Document all cross-language internal links and prepare a redirect plan if migration is considered
• Set up clean cross-language analytics tracking (GA4 with folder filters or cross-domain configuration per chosen architecture)