Is it true that hreflang can indicate duplicate content in multilingual subdomains?

Why does Google mention duplicate content for language versions?

Mueller's statement creates a conceptual confusion. Normally, identical content across multiple URLs constitutes classic duplicate content. Google must then decide which version to index and serve.

However, when discussing differing language versions, the content is actually not identical: a text in French and its translation in English are two distinct pieces of content. The term "identical subdomains in different languages" is ambiguous. Mueller likely refers to cases where the structure and template are identical, but the textual content differs by language.

How does hreflang fit into this issue?

The hreflang tags serve to indicate to Google that a page exists in several linguistic or regional variants. They do not strictly resolve a duplicate content issue. They convey: "This page in French and this page in English are regional equivalents, not redundant duplicates."

Without hreflang, Google may indeed treat two nearly identical pages (same structure, same images, only the text changes) as duplicate content. It will then canonicalize one at the expense of the other, sabotaging your multilingual strategy. Hreflang prevents this wild canonicalization by explicitly documenting the relationship between versions.

What happens if the content is truly identical in multiple languages?

If you publish the exact same text on fr.example.com and en.example.com, hreflang will not save you. Google will detect the duplicate and choose a canonical version. Hreflang only works if the content actually differs by language.

Some multilingual sites fall into this trap: they duplicate English content across multiple subdomains (UK, US, AU) without any real adaptation. In this case, hreflang is not the solution: content must be differentiated or explicitly canonicalized to a single version.

• Hreflang is not an anti-duplicate shield: it documents legitimate linguistic variants.
• Structurally identical pages that are translated are not considered duplicate content.
• If the content is truly identical in multiple languages, explicitly canonicalize to a reference version.
• Google can arbitrarily canonicalize if hreflang is missing or poorly implemented.
• The confusion arises from the term "identical": Mueller speaks of identical structure, not identical text.

Is this statement consistent with real-world observations?

Yes, but it lacks precision. In practice, it is observed that Google handles multilingual sites differently based on the quality of their hreflang implementation. A site with well-configured hreflang sees all its language versions indexed and served correctly. Without hreflang, Google often arbitrarily chooses one version, usually the English one.

The issue is that Mueller uses the term "duplicate content" somewhat loosely. Technically, translated content is not duplicate. What hreflang resolves is the ambiguity for Google: without a clear signal, the search engine does not know if two similar pages are redundant duplicates or intentional regional variants.

When is hreflang not enough?

Hreflang does not compensate for a poorly designed multilingual architecture. If your URLs do not follow conventions (subdomains, subdirectories, or consistent parameters), hreflang will be ignored or misinterpreted. [To be verified] Google has never published statistics on hreflang error rates, but Search Console reports indicate that it is one of the most frequently broken annotations.

Another edge case: sites that serve nearly identical content across various regional versions (en-US, en-GB, en-AU) with only a few words modified. Hreflang will declare them as variants, but Google may still treat them as thin content or duplicate if differentiation is insufficient. Here, the real solution is to produce truly differentiated content by market or to canonicalize to a unique version.

What common mistakes invalidate hreflang?

The most frequent issue: non-reciprocal hreflang links. If the French page points to the English one in hreflang, but the English page does not point back to the French one, Google ignores the annotation. Another classic mistake: declaring URLs that return 404, 301, or do not themselves contain hreflang.

We also see sites mixing methods (HTTP headers, XML sitemaps, HTML tags) without consistency. Google prefers a single, clean implementation. Finally, using incorrect language codes ("en-uk" instead of "en-gb") ruins everything. These errors turn hreflang into noise, and Google then reverts to its default behavior: arbitrary canonicalization.

What practical steps should you take for a multilingual site?

Start by auditing your architecture: subdomains (fr.site.com), subdirectories (site.com/fr/), or distinct domains (site.fr)? Each option has implications for hreflang. Subdomains and subdirectories are the easiest to manage.

Then, implement hreflang consistently on all equivalent pages. Each page must point to all its linguistic variants, including itself (self-reference). Preferably use HTML tags in the or a dedicated XML sitemap if you have thousands of pages. Ensure that each URL declared in hreflang is accessible, indexable, and contains hreflang itself.

How can you check if hreflang is functioning correctly?

Google Search Console displays hreflang errors in the "Coverage" and "Enhancements" sections. Track missing returns, orphan URLs, and invalid language codes. Use tools like Screaming Frog or Sitebulb to crawl your site and validate the reciprocity of annotations.

Also, test manually: search for a specific term on Google.fr and Google.com. You should see the appropriate language version based on location and language of the engine. If you see the wrong version, it means hreflang is not being considered or Google has selected a different canonical.

What mistakes should you absolutely avoid?

Never declare hreflang pointing to canonically defined pages elsewhere. If your French page has a canonical tag pointing to the English one, hreflang will be ignored: it’s contradictory. Canonical and hreflang are complementary, but canonical always takes precedence.

Avoid duplicating identical content across multiple language variants hoping that hreflang will mask the issue. Google will detect the duplicate and will canonicalize anyway. Hreflang is not a permission to duplicate. Finally, do not mix levels: if you are using subdomains for languages, don’t switch some languages to subdirectories. Architectural consistency is essential.

• Audit the multilingual architecture (subdomains, subdirectories, distinct domains).
• Implement hreflang on all equivalent pages with strict reciprocity.
• Validate language codes (ISO 639-1 for language, ISO 3166-1 Alpha 2 for country).
• Check in Search Console for hreflang errors and unintended canonicalizations.
• Test manually on different local versions of Google.
• Never mix canonical and hreflang in contradictory ways.