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

For equivalent pages in the same language but targeting different countries, it is recommended to use hreflang in addition to geographical targeting in Search Console, as this provides an extra signal to Google.
227:52
🎥 Source video

Extracted from a Google Search Central video

⏱ 985h14 💬 EN 📅 26/02/2021 ✂ 39 statements
Watch on YouTube (227:52) →
Other statements from this video 38
  1. 21:28 Les sitemaps suffisent-ils vraiment à déclencher un recrawl rapide de vos pages modifiées ?
  2. 21:28 Peut-on forcer Google à recrawler immédiatement après un changement de prix ?
  3. 40:33 La taille de police influence-t-elle réellement le classement Google ?
  4. 40:33 La taille de police CSS impacte-t-elle vraiment vos positions dans Google ?
  5. 70:28 Le contenu masqué derrière un bouton Read More est-il vraiment indexé par Google ?
  6. 70:28 Le contenu masqué derrière un bouton « Lire plus » est-il vraiment indexé par Google ?
  7. 98:45 Le maillage interne surpasse-t-il vraiment le sitemap pour signaler vos pages stratégiques à Google ?
  8. 98:45 Le maillage interne est-il vraiment plus décisif que le sitemap pour hiérarchiser vos pages ?
  9. 111:39 Pourquoi l'API Search Console ne remonte-t-elle pas les URLs référentes des 404 ?
  10. 144:15 Pourquoi Google continue-t-il à crawler des URLs 404 vieilles de plusieurs années ?
  11. 182:01 Faut-il vraiment s'inquiéter d'avoir 30% d'URLs en 404 sur son site ?
  12. 182:01 Un taux de 404 élevé peut-il vraiment pénaliser votre référencement ?
  13. 217:15 Comment cibler plusieurs pays avec un seul domaine sans perdre son référencement local ?
  14. 217:15 Peut-on vraiment cibler différents pays sur un même domaine sans passer par les sous-domaines ?
  15. 227:52 Faut-il vraiment utiliser hreflang quand on cible plusieurs pays avec la même langue ?
  16. 276:47 Pourquoi vos breadcrumbs en données structurées n'apparaissent-ils pas dans les SERP ?
  17. 285:28 Pourquoi vos rich results disparaissent dans les SERP classiques alors qu'ils s'affichent en recherche site: ?
  18. 293:25 Les breadcrumbs invisibles bloquent-ils vraiment vos rich results dans Google ?
  19. 325:12 Faut-il vraiment optimiser l'hydration JavaScript pour Googlebot en SSR ?
  20. 347:05 Le nombre de mots est-il vraiment inutile pour ranker sur Google ?
  21. 347:05 Le nombre de mots est-il vraiment un facteur de classement pour Google ?
  22. 400:17 Le volume de trafic de votre site impacte-t-il votre score Core Web Vitals ?
  23. 415:20 Le volume de trafic influence-t-il vraiment vos Core Web Vitals ?
  24. 420:26 Les Core Web Vitals comptent-ils vraiment dans le classement Google ?
  25. 422:01 Les Core Web Vitals peuvent-ils vraiment booster votre classement sans contenu pertinent ?
  26. 510:42 Pourquoi Google ne peut-il pas garantir l'affichage de la bonne version locale de votre site ?
  27. 529:29 Faut-il vraiment dupliquer tous les codes pays dans le hreflang pour cibler plusieurs régions ?
  28. 531:48 Pourquoi hreflang en Amérique latine impose-t-il tous les codes pays un par un ?
  29. 574:05 PageSpeed Insights mesure-t-il vraiment la performance de votre site ?
  30. 598:16 Peut-on vraiment passer du long-tail au short-tail sans changer de stratégie ?
  31. 616:26 Peut-on vraiment masquer les dates dans les résultats de recherche Google ?
  32. 635:21 Faut-il arrêter de mettre à jour les dates de publication pour améliorer son référencement ?
  33. 649:38 Google réécrit-il vraiment vos titres pour vous rendre service ?
  34. 650:37 Google réécrit vos balises title : peut-on vraiment l'en empêcher ?
  35. 688:58 Faut-il vraiment signaler les bugs SERP avec des requêtes génériques pour espérer une réponse de Google ?
  36. 870:33 Les nouveaux sites e-commerce doivent-ils d'abord prouver leur légitimité hors de Google ?
  37. 937:08 La longueur du title est-elle vraiment un facteur de classement sur Google ?
  38. 940:42 La longueur des balises title est-elle vraiment un critère de classement Google ?
📅
Official statement from (5 years ago)
TL;DR

Google recommends using hreflang even when geographical targeting is set up in Search Console for equivalent pages in the same language. This approach overlays two signals instead of favoring just one, enhancing clarity for algorithms. Specifically, this means that a .com site with /fr-fr/ and /fr-be/ versions must deploy hreflang in addition to the GSC settings, not as a substitute.

What you need to understand

Why does Google emphasize this dual approach?

The geographical targeting in Search Console allows you to inform Google that a section of your site (such as a subdirectory) targets a specific country. It's a global, declarative setting applied at the domain or subdirectory level.

Hreflang, on the other hand, operates page by page and specifies relationships of linguistic and geographical equivalence. It tells Google: "this French page for France has a French equivalent for Belgium." The two mechanisms do not operate at the same level of granularity.

Google doesn't say that one replaces the other. It says they complement each other. GSC targeting gives a macro indication, while hreflang refines it at the micro level. Together, they reduce ambiguity for algorithms, especially in cases where the same language serves multiple countries (French, English, Spanish, German, etc.).

In what scenarios does this recommendation truly apply?

This directive primarily targets multilingual and multi-country sites with identical or very similar content in the same language. A typical example: an e-commerce site that operates in France, Belgium, and Switzerland with content in French, sometimes strictly identical.

If your site serves only one country with a single language, hreflang is irrelevant. If you have clearly different language versions (US English vs Mexican Spanish), GSC targeting alone may suffice in some cases — but hreflang remains preferable to avoid any confusion.

The crucial nuance: Google refers to equivalent pages. If the content differs substantially (different products, currencies, legal mentions specific to each country), hreflang remains relevant but the notion of equivalence becomes blurred. You signal variations, not clones.

What happens if one only implements one or the other?

In practice, many sites only use hreflang and do very well. Geographical targeting in GSC is optional and can sometimes even be counterproductive if misconfigured (for example, targeting France on a .com while also having subfolders for other countries).

Conversely, relying solely on GSC targeting without hreflang exposes users to display errors: a Belgian user might come across the French version, a Swiss user the Belgian version, etc. Google does its best to guess, but without explicit page-by-page signals, the results are less reliable.

Mueller clearly states: "additional signal." This means that hreflang reinforces GSC targeting but does not replace it — and vice versa. One without the other works, but together they maximize accuracy.

  • Hreflang operates at the page level, establishing relationships of linguistic and geographical equivalence.
  • GSC targeting operates at the domain/subdirectory level, indicating a global geographical intent.
  • The two complement each other, not substituting for one another.
  • For a single-country, single-language site, hreflang is unnecessary.
  • For a multi-country, same language site, both are recommended.

SEO Expert opinion

Is this recommendation consistent with observed field practices?

Yes and no. In practice, hreflang alone works very well for the majority of well-configured international sites. Geographical targeting in GSC is often ignored or misconfigured, especially on generic domains (.com, .org) that serve multiple countries.

Mueller's recommendation primarily reflects a defensive posture from Google: by multiplying signals, we reduce outlier cases where the algorithm has to guess. This makes sense from Google's perspective, but for practitioners, it adds a layer of complexity — especially when GSC does not allow targeting multiple countries on the same root domain.

Specifically, if you have a .com with /fr-fr/, /fr-be/, /fr-ch/, you cannot use GSC targeting helpfully. Hreflang then becomes the main signal, or even the only one. Mueller's recommendation is better suited for sites with ccTLDs (.fr, .be, .ch) or dedicated subdomains (fr.example.com, be.example.com).

What limits and grey areas still exist?

Google does not specify how much weight each signal receives. Is hreflang 10 times more important than GSC targeting? 2 times? We don’t know. It remains a black box. [To be verified]

Another point: what to do if hreflang and GSC targeting contradict each other? For instance, a targeted subfolder "France" in GSC but with hreflang tags "fr-be." Google does not document this case. Field experience suggests that hreflang prevails, but this is not officially confirmed.

Finally, Mueller speaks of "equivalent pages in the same language." But what exactly is equivalence? An identical page word for word? A page with 80% common content? A page on the same subject but with different prices? Google remains vague, and this is problematic for e-commerce sites with partly overlapping catalogs.

Should you really apply this recommendation strictly?

It depends on your architecture. If you are on separate ccTLDs (.fr, .be, .ch) with French content, then yes: configure GSC targeting country by country AND implement hreflang. This is the ideal setup, and this is what Mueller describes.

If you are on a .com with subfolders, GSC targeting becomes unusable at the domain level — you can only target one country. In this case, hreflang alone is sufficient, and this is the standard practice observed across thousands of successful international sites.

Warning: never configure geographical targeting in GSC at the root domain level if that domain serves multiple countries. You would send a contradictory signal with hreflang, and the results would be unpredictable. It is better to not configure anything in GSC rather than misconfigure it.

Practical impact and recommendations

What should you actually do based on your architecture?

For a site on multiple ccTLDs (.fr, .be, .ch), the steps are clear: check in Search Console that each property is targeted to the correct country (France for .fr, Belgium for .be, etc.). Then, implement hreflang on all equivalent pages to connect the versions together.

For a site on a generic domain with subfolders (/fr-fr/, /fr-be/, /fr-ch/), do not touch the geographical targeting in GSC — leave it on "Not targeted." Focus on a rigorous hreflang implementation, preferably via XML sitemap for easier maintenance.

If you use subdomains (fr.example.com, be.example.com), you can create a GSC property for each subdomain and configure country targeting individually. Complete with hreflang to reinforce the signal. This is the architecture that allows you to apply Mueller's recommendation 100%.

What mistakes should you absolutely avoid?

Never configure GSC targeting that contradicts your hreflang tags. If your subfolder /fr-be/ is targeted "France" in GSC but your hreflang tags indicate "fr-be," Google receives two contradictory signals and may ignore both.

Avoid incomplete hreflang chains. If you indicate that /fr-fr/ has an equivalent /fr-be/, ensure that /fr-be/ reciprocates. Hreflang operates in a bidirectional loop. A single missing relationship can break the entire structure.

Do not multiply hreflang implementations (HTML + sitemap + HTTP header). Choose one, preferably XML sitemap, and stick to it. Mixing multiple methods increases the risk of errors and makes debugging nightmarish.

How can you check that everything is in order?

Use the Hreflang report in Search Console (under "Coverage" or "Pages" depending on the version). Google reports errors: missing tags, inaccessible URLs, invalid language codes. Systematically correct these errors.

Test with a third-party tool like Screaming Frog, Oncrawl, or Sitebulb. Set up a simulated multi-country crawl to verify that each page correctly points to its equivalents. Check that the hreflang annotations are consistent, complete, and that the URLs are canonical.

Finally, monitor impressions by country in GSC. If you see that your /fr-be/ version generates a lot of impressions in France, or vice versa, it’s a sign that Google has misunderstood your structure. Audit hreflang and GSC targeting immediately.

  • Check that your architecture (ccTLD, subfolders, subdomains) allows for GSC targeting to be applied without contradiction.
  • Implement hreflang via XML sitemap for all equivalent pages, with bidirectional annotations.
  • Only configure GSC geographical targeting if it makes sense (ccTLD, subdomains), otherwise leave it as "Not targeted."
  • Regularly audit the hreflang report in GSC and correct the reported errors.
  • Test with a crawler to validate the consistency of your annotations.
  • Monitor impressions by country in GSC to detect targeting inconsistencies.
Mueller's recommendation fully applies to sites on ccTLDs or dedicated subdomains. For generic domains with subfolders, hreflang alone remains the standard and effective approach. The key: never create contradictions between the two signals. These multi-country configurations can quickly become complex, especially at scale — if you manage dozens of languages and markets, partnering with a specialized SEO agency can save you precious time and avoid costly mistakes.

❓ Frequently Asked Questions

Hreflang est-il obligatoire si j'ai déjà configuré le ciblage géographique dans Search Console ?
Non, hreflang n'est jamais strictement obligatoire, mais Google recommande de l'utiliser en complément du ciblage GSC pour renforcer le signal. Sur un domaine avec sous-dossiers, hreflang devient de facto indispensable car GSC ne peut cibler qu'un pays au niveau domaine.
Peut-on utiliser hreflang sans configurer le ciblage géographique dans GSC ?
Oui, et c'est même la pratique courante sur les domaines .com avec sous-dossiers. Hreflang seul fonctionne très bien si correctement implémenté. Le ciblage GSC n'est utile que sur ccTLD ou sous-domaines.
Que se passe-t-il si hreflang et ciblage GSC se contredisent ?
Google n'a jamais documenté ce cas précisément. L'expérience terrain suggère que hreflang prime, mais mieux vaut éviter toute contradiction pour ne pas brouiller les signaux.
Dois-je utiliser hreflang entre des pages en français pour France et Belgique si le contenu est strictement identique ?
Oui, même avec un contenu identique, hreflang permet de dire à Google quelle version afficher selon la localisation de l'utilisateur. Sans cela, Google devine, avec des résultats parfois aléatoires.
Le ciblage géographique dans GSC a-t-il un impact sur le classement ?
Pas directement. C'est un signal d'intention géographique, pas un facteur de ranking. Il aide Google à comprendre pour quel pays une section du site est pensée, mais ne booste pas mécaniquement les positions dans ce pays.
🏷 Related Topics
Domain Age & History AI & SEO Search Console International SEO

🎥 From the same video 38

Other SEO insights extracted from this same Google Search Central video · duration 985h14 · published on 26/02/2021

🎥 Watch the full video on YouTube →

Related statements

💬 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.