hreflang-implementation

What this datapoint measures

Correctness of hreflang tag syntax and implementation for multilingual sites. Whether the brand’s per-language pages declare their language and locale correctly, whether the cross-references between language variants are reciprocal, and whether the implementation follows the Google/W3C specification.

This datapoint is relevant only for multilingual brands. Single-language brands score at a neutral baseline because hreflang is not applicable.

What high looks like

• Every page on every language variant has hreflang tags
• hreflang tags reciprocate (page A declares hreflang to page B; page B declares hreflang to page A)
• Language and locale codes use BCP 47 (e.g., en, en-US, en-GB, ja, zh-Hant)
• x-default declared on appropriate pages indicating the fallback language
• hreflang declarations match the actual language of the linked content
• Implementation in either HTML head, HTTP headers, or XML sitemap (consistent choice across the site)

What low looks like

• hreflang tags missing on some language variants
• hreflang tags with non-reciprocal cross-references (page A links to B but B doesn’t link back)
• Incorrect language codes (e.g., jp instead of ja, cn instead of zh)
• hreflang pointing to non-existent or 404 URLs
• x-default missing from sites with clear default-language preference

What at floor looks like

A multilingual brand at floor on hreflang-implementation has substantial hreflang gaps, errors, or absence. AI systems and search engines cannot reliably determine which language variant of a page to surface for which audience, producing degraded multilingual visibility across all languages.

This pattern is common in multilingual brands that built each language variant independently without coordinating hreflang implementation across them. The remediation is auditing all language variants and establishing reciprocal hreflang implementation, typically scoped within O-4.

For single-language brands, the datapoint is neutral; the brand is not penalized for not having hreflang it doesn’t need.

What affects this datapoint

• Whether hreflang tags are implemented at all (for multilingual sites)
• Reciprocity of hreflang cross-references
• Correctness of language and locale codes
• Match between declared language and actual content language
• x-default declaration
• Implementation method consistency

OMG actions that influence this datapoint

Action	Influence
O-4 Technical Infrastructure, Performance & International Foundation	Direct, primary. The “International Foundation” component of O-4 covers hreflang.

Multilingual considerations

This datapoint is itself fundamentally multilingual. Considerations specific to Avonetiq’s primary five languages:

• English: standard en or locale-specific (en-US, en-GB) as appropriate
• Indonesian: id (not in, which is a deprecated code)
• Japanese: ja (not jp, which is a country code, not a language code)
• Korean: ko (not kr, which is a country code)
• Traditional Chinese: zh-Hant (not zh-TW, although zh-TW is also acceptable in some contexts and is more specific to Taiwan-localized content)

Brands operating in Simplified and Traditional Chinese should use zh-Hans and zh-Hant respectively to disambiguate, not just zh.

Common failure modes

• Using country codes instead of language codes (jp, kr, cn instead of ja, ko, zh)
• Non-reciprocal hreflang causing the implementation to be ignored
• hreflang pointing to URLs that 404 or 301 redirect
• Mixing hreflang implementation methods (some pages in HTML head, others in sitemap, others in HTTP headers)
• x-default missing on a brand that has a clear default-language preference, leaving fallback selection to engines
• Subdomain-based and path-based language variants implemented without consistent hreflang

Diagnostic interpretation

For multilingual brands, hreflang-implementation at floor combined with multilingual-readiness also low indicates a multilingual brand whose international foundation work has not been done. O-4 work targeting both is required.

For single-language brands, this datapoint should be at neutral baseline; if it is not, the implementation has likely incorrectly added hreflang tags that don’t apply, which itself produces signals that confuse engines.

A multilingual brand with hreflang-implementation at high but multilingual-readiness at low indicates that the hreflang structure is correct but the underlying content isn’t comparable across languages — translated or localized content has substantial depth gaps relative to the primary language. Manifest-pillar work in the lagging languages is the remedy.

---