What does Google say about SEO? /
Quick SEO Quiz

Test your SEO knowledge in 3 questions

Less than 30 seconds. Find out how much you really know about Google search.

🕒 ~30s 🎯 3 questions 📚 SEO Google

Official statement

Hreflang annotations are only relevant between canonical and indexable pages. If pages are noindex, they are not canonical, so hreflang links are not needed on those pages. Hreflang should only be used on URLs that are indeed canonical or at least crawlable and indexable.
25:10
🎥 Source video

Extracted from a Google Search Central video

⏱ 55:38 💬 EN 📅 07/05/2021 ✂ 15 statements
Watch on YouTube (25:10) →
Other statements from this video 14
  1. 1:33 La longueur des URL affecte-t-elle vraiment votre classement Google ?
  2. 1:33 Les points dans les URLs sont-ils vraiment sans danger pour le SEO ?
  3. 2:07 Les URLs courtes sont-elles vraiment privilégiées par Google pour la canonicalisation ?
  4. 5:02 Faut-il vraiment attendre 3 mois après une migration 301 pour récupérer son trafic ?
  5. 7:57 Les iframes tuent-elles vraiment l'indexation de votre contenu ?
  6. 11:04 Un redesign de site peut-il vraiment casser votre ranking Google ?
  7. 19:59 Pourquoi Google continue-t-il à crawler des URLs redirigées en 301 depuis plus d'un an ?
  8. 22:04 Fusionner deux sites : pourquoi le trafic combiné n'est jamais garanti ?
  9. 37:54 Pourquoi Google ne traite-t-il pas toutes les erreurs 404 de la même manière dans Search Console ?
  10. 40:01 Le maillage interne accélère-t-il vraiment l'indexation de vos nouvelles pages ?
  11. 43:06 Les content clusters sont-ils réellement reconnus par Google ?
  12. 44:41 Le breadcrumb suffit-il vraiment comme seul linking interne ?
  13. 46:15 La homepage a-t-elle vraiment plus de poids SEO que les autres pages ?
  14. 49:52 Le duplicate content pénalise-t-il vraiment votre référencement ?
📅
Official statement from (4 years ago)
TL;DR

Hreflang annotations only apply to canonical, indexable pages. If a page is noindex, it is not considered canonical by Google, so hreflang is unnecessary there. In practical terms: don't waste time implementing hreflang on URLs you block from indexing—focus your efforts only on the canonical versions intended to appear in the SERPs.

What you need to understand

Why does Google make this connection between hreflang and canonization?<\/h3>

Hreflang indicates the language or regional variants of the same page. Google considers that only canonical pages can be legitimate variants within a multilingual or multi-regional set. Since a noindex page is never considered canonical, it cannot be part of a valid hreflang cluster.<\/p>

The logic is clear: if you prohibit the indexing of a URL, you are explicitly signaling to Google that it should not appear in the results. So why associate it with hreflang signals that are specifically designed to guide display in the SERPs based on the user's language or region?<\/p>

In what scenarios can we still find hreflang on noindex pages?<\/h3>

This situation often arises during poorly prepared migrations or misconfigured automated implementations. A CMS may automatically generate hreflang annotations on all versions of a page, including those temporarily blocked from indexing. Alternatively, a team may noindex underperforming regional variants without cleaning up the hreflang code.<\/p>

Another common case: staging or development environments in noindex that retain hreflang code pointing to production. These hybrid configurations create noise and disrupt crawl budget without providing any SEO benefit.<\/p>

What happens if we leave hreflang on noindex pages?<\/h3>

Google simply ignores those annotations. They do not generate blocking errors but represent a waste of resources—wasted crawl time, superfluous code, unnecessary technical complexity. In some cases, they can create confusion in Search Console if hreflang clusters become incomplete or inconsistent.<\/p>

More concerning: if your noindex pages point via hreflang to indexable pages, but not vice versa, you create one-way hreflang chains that break the reciprocity principle. Google may decide not to take the entire cluster into account.<\/p>

  • Hreflang only works on canonical and indexable URLs<\/strong><\/li>
  • A noindex page is never considered canonical by Google<\/li>
  • Implementing hreflang on noindex = waste of crawl budget and unnecessary complexity<\/li>
  • Hreflang annotations must be reciprocal<\/strong> between all variants of a cluster<\/li>
  • Regularly clean up your implementations to remove orphan hreflang or those on blocked pages<\/li><\/ul>

SEO Expert opinion

Does this rule contradict observed practices in the field?<\/h3>

No, it confirms what is empirically observed: Google never uses noindex pages in its hreflang clusters. I have audited hundreds of multilingual sites—in all cases where hreflang was present on blocked pages, Search Console reported no usage<\/strong> of these annotations. They simply appeared as ignored.<\/p>

What can sometimes be surprising is the persistence of this error. Many agencies or junior SEOs think they need to "maintain code consistency" and leave hreflang everywhere, including in pre-production environments or temporarily de-indexed pages. This is a dogmatic approach<\/strong> that does not reflect Google's technical reality.<\/p>

Are there edge cases where this logic could falter?<\/h3>

A marginal scenario: you have a temporary<\/strong> noindex page (say, for a past event you plan to reactivate next year). If you remove the hreflang now and forget to put it back when reactivated, you create a gap in your cluster. But let's be honest—this is a governance problem<\/strong>, not a technical SEO issue.<\/p>

Another nuance: some third-party SEO tools (SEMrush, Screaming Frog, etc.) may report "hreflang errors" if annotations are present on noindex pages. These alerts do not necessarily indicate a blocking bug on Google's side, but they clutter your dashboards and complicate the detection of real issues. [To be verified]<\/strong> whether your tool stack tolerates this configuration well before deciding to clean it up or not.<\/p>

What is the best practice to avoid this trap?<\/h3>

Automate the logic: if a page is noindex (or canonicalized to another URL), simply do not generate hreflang tags<\/strong> on that page. From an implementation perspective, this means adding a condition in your CMS or template generator: "If robots = noindex OR canonical ≠ self, then do not inject hreflang".<\/p>

For complex sites with multiple environments, set up firewall rules or server configuration<\/strong> to prevent hreflang injection outside of indexable production. This avoids accidental leaks during deployments or partial migrations.<\/p>

Practical impact and recommendations

How to quickly audit your site to detect this issue?<\/h3>

Run a complete crawl with Screaming Frog<\/strong> or OnCrawl filtering for noindex URLs. Then export the pages that contain hreflang tags. If you find matches, you have identified superfluous code to clean up. Also, check pages with canonical pointing elsewhere than self—they should not carry hreflang either.<\/p>

In Search Console, go to the “International Coverage”<\/strong> section. If you see errors like “hreflang missing on return pages” or “incomplete clusters,” cross-reference with your list of noindex pages. Often, these errors come from blocked pages that break the cluster's reciprocity.<\/p>

What corrective actions should be implemented immediately?<\/h3>

Manually or via a script, remove hreflang annotations from all noindex or non-canonical pages<\/strong>. If you are using a CMS plugin (Yoast, RankMath, WPML), check the settings to disable automatic hreflang generation on these pages. Some plugins have dedicated options, others require custom code.<\/p>

For sites with hreflang in HTTP headers or XML sitemaps, clean up those sources as well. A noindex page should never appear<\/strong> in an XML sitemap—and if it does, it certainly should not carry hreflang annotations in the associated HTTP headers.<\/p>

How to avoid this issue in future implementations?<\/h3>

Integrate this rule into your technical guidelines<\/strong> and pre-launch checklists. Train developers never to generate hreflang if a page does not simultaneously meet these conditions: indexable (no noindex), canonical (canonical self or absent), and crawlable (no disallow robots.txt).<\/p>

For migrations or redesigns, include an hreflang audit in your QA phase. Automate it with end-to-end tests (Cypress, Selenium) that check that no noindex page<\/strong> carries a hreflang tag before going live. This saves hours of post-launch debugging.<\/p>

  • Crawl the site to identify noindex pages with hreflang<\/li>
  • Remove hreflang annotations from all noindex or non-canonical URLs<\/li>
  • Check CMS/plugin settings to disable automatic generation on these pages<\/li>
  • Clean the HTTP headers and XML sitemaps of any hreflang reference to blocked pages<\/li>
  • Add automated tests to detect this problem before each deployment<\/li>
  • Train dev and editorial teams on this rule to avoid regressions<\/li><\/ul>
    Hreflang only applies to pages intended to appear in search results. Any noindex or non-canonical page should be exempt from it. Clean your existing implementations and automate this logic for the future. These multilingual optimizations can be complex to orchestrate on international architectures—if your team lacks bandwidth or technical expertise, working with a specialized SEO agency ensures a clean, sustainable implementation without the risk of regression during future site updates.<\/div>

❓ Frequently Asked Questions

Peut-on mettre du hreflang sur une page avec une balise canonical vers une autre URL ?
Non. Si une page pointe vers une autre URL via canonical, elle n'est pas considérée comme canonique elle-même, donc le hreflang n'y a pas sa place. Seules les URLs canoniques doivent porter des annotations hreflang.
Que se passe-t-il si on laisse du hreflang sur des pages en noindex sans le corriger ?
Google ignore simplement ces annotations. Elles ne causent pas d'erreur bloquante, mais elles gaspillent du crawl budget, compliquent la maintenance et peuvent générer des alertes dans la Search Console si le cluster devient incomplet.
Le hreflang fonctionne-t-il sur des pages bloquées par robots.txt mais sans noindex ?
Google ne peut pas crawler ces pages pour lire le hreflang. Même si elles ne sont pas en noindex, l'absence de crawl empêche la prise en compte des annotations. Le hreflang nécessite un crawl effectif pour être interprété.
Doit-on retirer le hreflang sur une page temporairement mise en noindex ?
Oui. Tant qu'elle est en noindex, elle n'est pas canonique, donc le hreflang est inutile. Si vous prévoyez de la réindexer, documentez-le pour ne pas oublier de réinjecter le hreflang lors de la réactivation.
Comment vérifier si mes clusters hreflang respectent bien cette règle ?
Crawlez votre site et filtrez les pages noindex ou avec canonical vers une autre URL. Si elles portent du hreflang, c'est une erreur de configuration à corriger. Vérifiez aussi la Search Console pour détecter des clusters incomplets.

💬 Comments (0)

Be the first to comment.

2000 characters remaining
🔔

Get real-time analysis of the latest Google SEO declarations

Be the first to know every time a new official Google statement drops — with full expert analysis.

No spam. Unsubscribe in one click.