Language Targeting vs. Country Targeting

TL;DR — Language targeting serves people by the language they read. Country targeting serves people by the country they’re in — where price, currency, shipping, and legal text actually change. They’re two separate questions, not one. First decide whether your page’s content really differs by country or just by language; only then pick your hreflang codes. And remember: hreflangHreflang is an annotation (in HTML, HTTP headers, or XML sitemaps) that tells search engines which language and optional region a page targets, and which alternate versions exist. It only works when every page in the cluster references all the others. always needs a language code first — you can’t target a country by itself.

Two different questions

When you start expanding a site to more than one market, the very first fork in the road trips almost everyone up. People treat “language” and “country” as the same thing. They’re not.

• Language targeting is about what language someone reads. A Spanish page for Spanish speakers — anywhere. This is a multilingual site: the same content offered in more than one language.
• Country targeting is about where someone is and what’s different for them locally — the price, the currency, the shipping options, the legal fine print. This is a multi-regional site: content aimed at users in specific countries.

Here are two examples that make it click:

• Canada is one country with two official languages, English and French. That’s a language problem inside a single country.
• The US and the UK are two countries that mostly share a language. But prices, spelling, and shipping differ. That’s a country problem, even though the language is basically the same.

Same language, different country. Same country, different language. The two axes are genuinely independent, and lots of big sites need both at once.

Make the content decision first

Before you touch a single tag, ask one question about the page itself:

Does the actual content on this page change by country — price, currency, shipping, legal text, availability — or does only the language change?

If the only difference is language, you have a language-targeting job. If the price and shipping genuinely differ between, say, the US and the UK, you have a country-targeting job — even if both pages are in English.

Get this backwards and it’s expensive to fix, because the answer decides your URL structure, and URL structureURL structure is how the parts of a web address — scheme, domain, path, query string, and fragment — are organized and formatted. It mostly affects crawling, usability, and how engines understand a page, not rankings directly. is hard to undo later.

Then pick the tag

The technical tool for telling Google which version is which is hreflangHreflang is an annotation (in HTML, HTTP headers, or XML sitemaps) that tells search engines which language and optional region a page targets, and which alternate versions exist. It only works when every page in the cluster references all the others.. The important beginner rule:

• You can target a language on its own — es means “Spanish, everywhere.”
• You can narrow it to a country — es-MX means “Spanish, for people in Mexico.”
• You cannot target a country on its own. There is always a language code first. A code like mx by itself isn’t valid. Evidence for this claim Google's supported hreflang values begin with a language code and can optionally add a region code; a region by itself is not a supported value. Scope: Google Search-supported hreflang language and region codes. Confidence: high · Verified: 2026-07-14 Google: Supported language and region codes

So even when your reason is country-driven (Canadian pricing in English), the tag still leads with the language: en-CA, not ca.

The thing most people get wrong

Google can’t tell your target country from the language of your page. A French page could be for France, Belgium, Canada, or Switzerland. If you care about country, you have to say so explicitly — with a region code (fr-CA, not just fr) or a country-specific domain. As Google’s Gary Illyes put it, the language is not a tell for what country you’re targeting.

Want the deeper version — the ISO code rules, the Belgium example, when not to split by country, and how Bing differs? Switch to the Advanced tab.

Framework: choose the minimum viable targeting model

Use the page itself as the deciding evidence. Pricing, shipping, regulation, inventory, and market-specific intent justify country variants; a flag in the navigation does not.

Framework: content first, annotation second

1. Define the audience difference. Is the need linguistic, commercial by country, or both?
2. Prove the page difference. List what changes in visible content or transaction terms.
3. Choose the smallest URL set. Avoid country variants that would remain identical.
4. Assign codes. Use a required language code and add a region only when the page truly serves that market.
5. Connect equivalents. Publish reciprocal hreflangHreflang is an annotation (in HTML, HTTP headers, or XML sitemaps) that tells search engines which language and optional region a page targets, and which alternate versions exist. It only works when every page in the cluster references all the others. across the complete cluster.
6. Revisit the split. Consolidate variants whose content and purpose no longer differ.

This order prevents a technical label from creating pages the content strategy cannot support.

TL;DR — Language and country are two independent axes. Language targeting = multilingual (same content, different language); country targeting = multi-regional (content that differs by country: currency, price, shipping, legal). Make the content decision first — does the page’s actual content differ by country, or only by language? — then the technical decision: which hreflang codes to annotate the URL with. hreflangHreflang is an annotation (in HTML, HTTP headers, or XML sitemaps) that tells search engines which language and optional region a page targets, and which alternate versions exist. It only works when every page in the cluster references all the others. codes are ISO 639-1Two-letter language codes used in hreflang (en, es, de). language + optional ISO 3166-1 alpha-2Two-letter country codes used in hreflang region targeting (us, gb, au). region; the language is mandatory and primary, the region optional and secondary, and a region-only code is invalid. Evidence for this claim Google's supported hreflang values begin with a language code and can optionally add a region code; a region by itself is not a supported value. Scope: Google Search-supported hreflang language and region codes. Confidence: high · Verified: 2026-07-14 Google: Supported language and region codes Language never implies country (Illyes) and the HTML lang attribute isn’t a reliable signal either. Don’t over-split into per-country variants when the content is identical in a language — my 374,756-domain study foundA 302 (\"Found\") is a temporary redirect: it forwards users to a new URL while telling search engines the original URL should stay in the index. It's a weak canonicalization signal, not the zero-equity dead end of SEO folklore. 67% of hreflangHreflang is an annotation (in HTML, HTTP headers, or XML sitemaps) that tells search engines which language and optional region a page targets, and which alternate versions exist. It only works when every page in the cluster references all the others. domains have at least one issue, and more variants means more surface area to break.

Two axes, not one spectrum

The single most useful distinction in this whole topic is that language targeting and country targeting are different things. They are orthogonal axes, not two ends of one slider:

• Language targeting is for multilingual sites — the same content offered in more than one language.
• Country targeting is for multi-regional sites — content aimed at users in different countries, sometimes in the same language.

Google’s own docs draw exactly this line: a multilingual website is one that offers content in more than one language, a multi-regional website is one that explicitly targets users in different countries, and — this is the part people miss — a site can be both at once. Google’s example: a site with different versions for the USA and Canada, and both French and English versions of the Canadian content.

The content decision vs. the technical decision

Nearly every article on this topic jumps straight to hreflang syntax. That skips the harder, prior question. There are really two decisions here, in order:

1. The content decision — who is this page actually written for? Ask whether the page’s substance differs by country: price, currency, shipping terms, legal disclosures, product availability, tax handling. If the only thing that changes between markets is the language of the words, you have a language job. If the dollars-vs-pounds, the returns policy, or the regulated claims change, you have a country job — even when both pages are in English.

2. The technical decision — which hreflang codes do you annotate the URL with? Only after you’ve answered the content question do you pick codes: language-only, language+region, or a mix. The technical decision follows the content decision; it doesn’t lead it.

Getting the order backwards is one of the most expensive early mistakes in international SEOInternational SEO is the practice of optimizing a site so search engines understand which countries and/or languages it targets, and serve the right version to each user. It spans URL structure, hreflang, and on-page localization., because the content decision drives your URL structureURL structure is how the parts of a web address — scheme, domain, path, query string, and fragment — are organized and formatted. It mostly affects crawling, usability, and how engines understand a page, not rankings directly. — ccTLDCountry-code top-level domain like .de or .co.uk — a strong geotargeting signal. vs. subdomain vs. subdirectory, one page per language vs. one page per country — and URL structure is painful to unwind once it’s live and indexedStoring a crawled page in the search index so it can appear in results. Crawled is not the same as indexed — Google selects what to keep, and indexing isn't guaranteed..

How hreflang expresses the difference

hreflang codes are built from two independent parts. Per Google’s docs, the first code is the language in ISO 639-1 format, followed by an optional second code that is the region in ISO 3166-1 alpha-2 format. That structure maps directly onto the two axes:

• Language-only codes — es, de, fr. These target the language regardless of country. Google’s docs describe de as German-language content, independent of region. es is every Spanish speaker on earth.
• Language+region codes — es-MX, en-GB, de-ES. These narrow the language to a specific country’s speakers. Google’s examples: en-GB is English content for users in the UK; de-ES is German content for users in Spain (yes, German for people in Spain — the two codes really are independent).
• Region-only codes — invalid. You cannot specify a country by itself. Google is explicit: the first code stands for the language, and Google doesn’t automatically derive the language from a country code. Specifying the region alone is not valid.

That last rule is the technical shadow of the strategic one. You can’t “target a country” independent of some language, because the tag always leads with language — even when the business reason is entirely country-driven (en-CA for Canadian English pricing). And only codes that appear in the ISO 639-1 and ISO 3166-1 alpha-2 standards are supported — invented regional codes like es-419 (Latin America) aren’t in those standards, so they don’t work.

The Belgium example — one country, three languages

The cleanest illustration lives in Google’s own hreflang examples table. Belgium is a single country with three languages, and each gets its own valid code:

• de-be — German for users in Belgium
• nl-be — Dutch for users in Belgium
• fr-be — French for users in Belgium

And the trap Google flags: be on its own is bad, because be is the language code for Belarusian — not a country code for Belgium. One country, three languages, and a code that looks like “Belgium” but actually means a completely different language. That’s the whole language-vs-country problem in four lines.

TIP Turn the targeting decision into valid hreflang syntax

The generator validates language and optional region codes and produces reciprocal annotations. It cannot decide whether the content genuinely warrants a country-specific page—that business decision comes first.

Once the page set is defined, enter each locale in my Hreflang Generator to catch invalid codes and export a consistent cluster. Hreflang Generator Free

1. Decide whether each page differs by language, country, or both.
2. Use a language code first and add a supported region only when the content is market-specific.
3. Generate the annotations, deploy them on every member, and verify live reciprocity separately.

Tap to enlarge

Real-world scenarios that make it concrete

• Canada — one country, two languages. English and French Canadians see different language content at the same prices under the same laws. This is a language job inside a single country: en-CA and fr-CA.
• US / UK — two countries, one shared language. Different currency, spelling, shipping, and sometimes legal text, but effectively one language. This is a country job despite the shared language: en-US and en-GB.
• A global SaaS with identical pricing everywhere. If the product and price are the same worldwide and only the interface language changes, language-only targeting is enough. You probably don’t need per-country pages at all — one page per language does the job.
• A retailer with local prices, currency, and shipping per country. Here the content genuinely differs by country, so language+country variants (and often country-specific URL structureURL structure is how the parts of a web address — scheme, domain, path, query string, and fragment — are organized and formatted. It mostly affects crawling, usability, and how engines understand a page, not rankings directly.) are justified.

The pattern: split by country only when the content actually differs by country. If it doesn’t, you’re manufacturing variants that add risk without adding value.

Why language can’t tell you the country

Google can’t infer your target country from the language of your text — and it says so plainly. Gary Illyes, on Google’s Search Off the Record podcast, put it as bluntly as it gets: the language is absolutely not a tell for what country you are targeting. A French page could be for France, Belgium, Canada, or Switzerland. If country matters, you have to signal it explicitly.

The HTML lang attribute doesn’t rescue you either. Illyes told a story about a CMSA content management system (CMS) is software that lets users create, manage, and publish digital content — like blog posts and pages — without writing raw code. WordPress, Drupal, and Joomla are the most common open-source CMS platforms. (Joomla, in his telling) that shipped with the lang attribute hardcoded to English on a page that was, in his words, entirely German — with no easy way to override it. The attribute is easy to get wrong by default, so search engines don’t lean on it as a targeting signal with any confidence.

Google draws a hard line between language detection and language targeting signals: per its own documentation, it determines a page’s actual language from the page’s visible content using its own algorithms — not from hreflang, the HTML lang attribute, or the URL. hreflang and lang are separate systems from detection; they tell Google which alternate version to serve, not what language a page is actually written in. That’s also why translating only the boilerplate — nav, footer, a couple of headers — while leaving the main content in one language doesn’t create a real language variant: Google is still reading mostly-untranslated content and can end up treating the pages as near-duplicatesThe same or very similar primary content reachable at more than one URL. There's no general duplicate content penalty — the real costs are possible signal dilution, the wrong URL getting chosen, and less-efficient crawling..

What does Google use to identify the intended country audience? Per its multi-regional guidance, signals like local addresses and phone numbers, local language and currency on the page, and links from other local sites — plus ccTLDs, which it calls a strong signal about the target country. What it does not use: geolocation meta tags and IP-based location analysis are not treated as reliable signals.

There’s also evidence that Google’s ranking systems treat language-match and country-match as separate, independently-weighted signals rather than one collapsed signal. Illyes described an algorithm he called LDCP — “language demotion, country promotion” — where, for example, if someone searches in German and your page is in English, you’d get a demotion (Search Engine Journal coverage). In the same conversation he suggested the ccTLD-based country boost is getting ambiguous and will fade over the years — which matters for the decision below.

Choosing your approach

A practical framework, in order:

1. Start with data, not assumption. Look at whether search volume and competitor SERP patterns actually differ meaningfully by country before you commit to a country-targeted build. Plenty of markets rank a single global page just fine.
2. Answer the content question. Does price, currency, shipping, legal text, or availability genuinely differ by country? If yes → country targeting. If only the language changes → language targeting.
3. Don’t over-split. If content is identical across countries within a language, collapsing them into per-country variants adds error surface for no benefit. Illyes’ point on the same podcast: as soon as you have to vary things, that’s where you start introducing errors. My own study backs this up quantitatively — every extra variant is another node in the reciprocal hreflangIf page A lists page B as an alternate, B must list A back — or Google ignores the pair. cluster that can break.
4. Add country targeting later if evidence justifies it. Start language-targeted; layer country targeting onto a specific market once it shows enough activity to earn the extra build cost.

Architecturally, my preference at scale is one page per language rather than one page per country/locale. It produces fewer, stronger pages, enables dynamic personalization for the country-specific bits, and sidesteps a lot of hreflang complexity. I ran international SEOInternational SEO is the practice of optimizing a site so search engines understand which countries and/or languages it targets, and serve the right version to each user. It spans URL structure, hreflang, and on-page localization. at IBM scale, and the per-country sprawl is where things quietly fall apart.

How this interacts with URL structure

Your language/country answer should drive your URL structure — not the other way around. A ccTLD (example.de) is the strongest country signal but the most expensive to run; a subdirectory (example.com/de/) is the lowest-maintenance and what I’d pick for most sites. If the honest answer is “language only, no real country differences,” a country-specific ccTLD or a dedicated per-country folder is usually overkill.

And don’t bank the decision on the ranking boost alone. Illyes has suggested, in podcast remarks, that the ccTLD-linked ranking boost (his LDCP framing) is getting ambiguous and may fade over time — but that’s a rep’s forward-looking comment, not a policy change reflected in Google’s maintained documentation, which still calls ccTLDs a strong country signal today. Treat “the boost is fading” as a trend to watch, not a settled fact to build a URL-structure decision on — the business/content argument (does the content genuinely differ by country) should carry more weight than either version of the SEO-signal argument. See the sibling piece on choosing between ccTLD, subdomain, and subdirectory for the full trade-off.

hreflang is a hint, not enforcement

Even correctly-coded language/country variants aren’t guaranteed separate indexingStoring a crawled page in the search index so it can appear in results. Crawled is not the same as indexed — Google selects what to keep, and indexing isn't guaranteed.. John Mueller has reminded people that hreflang tags are hints, not directives — if two variants are the same (his example: fr-fr and fr-be), it’s common for Google to pick one as canonical and consolidate them (Search Engine Journal). So the payoff of getting the decision right is mostly about which URL gets swapped into the SERP for the right user — not a guarantee of distinct indexed pages.

How Bing differs

Bing’s whole mechanism is different. It weights the content-language signal (the HTTP header or meta tag, plus <html lang>) above hreflang, and its principal program manager Fabrice Canel has said hreflang is a far weaker signal than content-language at Bing (Search Engine Roundtable). The practical Bing advice is directly on-point for this topic: don’t manufacture country-specific URLs just to tag them if the content isn’t actually different — Bing leans instead on content-language, <html lang>, inbound links, and visitor geography. Worth flagging: Bing doesn’t maintain one current reference page laying this out the way Google’s hreflang docs do — this reflects rep statements and blog coverage gathered over several years rather than a single up-to-date technical contract, so treat the specific weighting as directional guidance rather than a fixed spec.

One more thing worth knowing: neither engine gives you a manual “set my country” override anymore. Google’s GSCA free Google service that reports how a site performs in Google Search and surfaces problems with how Google crawls, indexes, and serves it. It's first-party data straight from Google — but you don't need it to appear in results. International Targeting report was deprecated in 2022 (deemed to have little value for the ecosystem), and Bing removed its GeoGenerative Engine Optimization — visibility inside AI answer engines. Targeting feature in 2020. Once you’ve made the country decision, it has to be expressed entirely through structural signals — URL/ccTLD, hreflang region code, on-page localizationLocalization is adapting content for a specific target market — not just translating the words, but adjusting currency, formats, idioms, cultural references, local search terms, and trust signals so the experience feels native. — with no dashboard toggle to fall back on.

Implementation traps that undo a correct decision

Even with the right content and code decision, a few common implementation mistakes quietly break it:

• Serving locale-adapted content from one URL instead of separate URLs. Adapting content by cookie, browser language, or IP on a single URL — rather than publishing a separate URL per language/country and connecting them with hreflang — can leave variants undiscovered. GooglebotGooglebot is Google's web crawler — the software that fetches pages so Google can index and rank them. It comes in two variants, Googlebot Smartphone (primary, under mobile-first indexing) and Googlebot Desktop, and runs an evergreen Chromium renderer. commonly crawls from US IP addresses and doesn’t send an Accept-Language header (though geo-distributed crawlingCrawling is how search engines use automated bots (like Googlebot and Bingbot) to discover URLs and download pages. A page has to be crawlable to be indexed, but crawling on its own isn't a ranking factor. exists too), so a design that depends on a crawlerA crawler — also called a spider or bot — is an automated program that fetches web pages, extracts their links, and queues new URLs to visit. Search engines use crawlers to discover and download content for their index. exercising every location or language branch will miss coverage. Google’s own recommendation is locale-specific URLs plus hreflang, not URL-level adaptation.
• Auto-redirecting users based on inferred language or location. Google advises against automatically redirecting visitors to a “detected” language version — it can trap people on the wrong variant and get in the way of Google crawling the others. Give users an explicit link to switch instead, and treat locale targeting as directional: even correct signals aren’t a guarantee every visitor lands on the intended version.
• Same-language regional duplicates with no canonical. Near-identical en-US and en-GB pages still need a canonical, not just hreflang — Google recommends picking a preferred version and pointing the canonical at it (same language, or the closest substitute) so regional duplicates don’t compete against each other. hreflang alone doesn’t guarantee every regional URL stays separately indexed.
• Translating only the boilerplate. Covered above — nav/footer translated, main content not, and the page reads as a near-duplicate rather than a real language variant.

Where this sits

This is the prerequisite decision that everything else in international SEO follows from. Once you’ve answered it, the implementation topics take over: hreflang mechanics, the ccTLD vs. subdomain vs. subdirectory choice, reciprocal hreflang, the x-defaultx-default is the reserved hreflang value that points to a fallback URL — the page Google shows when a user's locale doesn't match any of your other hreflang tags. It is optional and does not mean \"English.\" fallback, and how duplicate contentThe same or very similar primary content reachable at more than one URL. There's no general duplicate content penalty — the real costs are possible signal dilution, the wrong URL getting chosen, and less-efficient crawling. interacts with same-language variants. Those are covered in the sibling pieces in this cluster.

AI summary

A condensed take on the Advanced version:

• Two independent axes, not one spectrum. Language targeting = multilingual (same content, different language). Country targeting = multi-regional (content that differs by country: currency, price, shipping, legal text). A page can need one, the other, or both.
• Content decision first, technical decision second. Ask whether the page’s actual content differs by country or only by language. Only then pick hreflang codes. Getting the order backwards is expensive because it drives URL structureURL structure is how the parts of a web address — scheme, domain, path, query string, and fragment — are organized and formatted. It mostly affects crawling, usability, and how engines understand a page, not rankings directly..
• hreflangHreflang is an annotation (in HTML, HTTP headers, or XML sitemaps) that tells search engines which language and optional region a page targets, and which alternate versions exist. It only works when every page in the cluster references all the others. code rules: ISO 639-1Two-letter language codes used in hreflang (en, es, de). language (mandatory, primary) + optional ISO 3166-1 alpha-2 region (secondary). es = the language everywhere; es-MX = narrowed to Mexico; a region-only code is invalid — the tag always leads with a language.
• Language ≠ country. Illyes: the language is absolutely not a tell for what country you are targeting. A French page could be for France, Belgium, Canada, or Switzerland. The HTML lang attribute isn’t a reliable signal either.
• What Google uses for country: local address/phone, local currency, local links, ccTLDsCountry-code top-level domain like .de or .co.uk — a strong geotargeting signal.. What it does not use: geolocation meta tagsMeta tags are HTML elements in a page's head that pass metadata about the page to search engines and browsers. For SEO only a few matter — the title element, the meta description, and the robots meta tag — while meta keywords and most others are ignored. and IP-based analysis.
• Don’t over-split. Split by country only when content genuinely differs. Illyes: as soon as you vary things, you introduce errors. My 374,756-domain study foundA 302 (\"Found\") is a temporary redirect: it forwards users to a new URL while telling search engines the original URL should stay in the index. It's a weak canonicalization signal, not the zero-equity dead end of SEO folklore. 67% of hreflangHreflang is an annotation (in HTML, HTTP headers, or XML sitemaps) that tells search engines which language and optional region a page targets, and which alternate versions exist. It only works when every page in the cluster references all the others. domains have at least one issue.
• hreflang is a hint — same-language variants (fr-fr/fr-be) can get consolidated.
• Bing differs: weights content-language over hreflang; don’t manufacture country-specific URLs just to tag them. No manual country override survives on either engine (GSCA free Google service that reports how a site performs in Google Search and surfaces problems with how Google crawls, indexes, and serves it. It's first-party data straight from Google — but you don't need it to appear in results. report gone 2022, Bing GeoGenerative Engine Optimization (GEO) is the practice of optimizing content and brand presence so AI-powered search engines and assistants — Google AI Overviews, ChatGPT, Perplexity — cite, recommend, or mention you when generating answers. Google's position is that it's still SEO. Targeting gone 2020).
• Implementation traps that undo a good decision: serving locale-adapted content from one URL instead of separate URLs + hreflang, auto-redirecting users on inferred language, same-language regional duplicates with no canonical, and translating only the boilerplate while leaving main content untranslated.
• My default at scale: one page per language, not one page per country.

Official documentation

Primary-source documentation from the search engines.

Google

• Managing multi-regional and multilingual sites — the definitional split (multilingual vs. multi-regional), the “both at once” case, and the locale signals Google uses (local address/phone, currency, links, ccTLDCountry-code top-level domain like .de or .co.uk — a strong geotargeting signal.).
• Localized versions of your pages — the hreflang code format (ISO 639-1Two-letter language codes used in hreflang (en, es, de). + optional ISO 3166-1 alpha-2Two-letter country codes used in hreflang region targeting (us, gb, au).), the “region alone is not valid” rule, and the supported-codes list, plus the Belgium examples.
• Locale-adaptive pages — why geolocation meta tagsMeta tags are HTML elements in a page's head that pass metadata about the page to search engines and browsers. For SEO only a few matter — the title element, the meta description, and the robots meta tag — while meta keywords and most others are ignored. and IP-based location aren’t used as reliable signals.
• International Targeting report deprecation (2022) — the GSCA free Google service that reports how a site performs in Google Search and surfaces problems with how Google crawls, indexes, and serves it. It's first-party data straight from Google — but you don't need it to appear in results. country-targeting toggle is gone.

Bing / Microsoft

• Going International (Bing Webmaster Blog, 2009) — Bing’s foundational international guidance, ccTLD as a primary country signal.
• Bing Webmaster Guidelines — the current guidelines hub.

Quotes from the source

On-the-record statements from Google, plus rep commentary. Google’s docs support verbatim deep links; the rep quotes from podcasts and social posts are relayed through published coverage and are paraphrased here rather than quoted, since I couldn’t confirm the exact transcript wording against a primary source.

Google docs — the definitional split

• “A multilingual website is any website that offers content in more than one language. […] A multi-regional website is one that explicitly targets users in different countries.” — Google Search Central. Jump to source

Google docs — you can’t target a country alone

• “Specifying the region alone is not valid. The first code stands for the language and Google doesn’t automatically derive the language from a country code.” — Google Search Central, Localized versions of your pages. Jump to source

Google docs — the code format

• “The first code of the hreflang attributeHreflang is an annotation (in HTML, HTTP headers, or XML sitemaps) that tells search engines which language and optional region a page targets, and which alternate versions exist. It only works when every page in the cluster references all the others. is the language code (in ISO 639-1Two-letter language codes used in hreflang (en, es, de). format) followed by an optional second code that represents the region code (in ISO 3166-1 Alpha 2 format) of an alternate URL.” — Google Search Central, Localized versions of your pages. Jump to source

Gary Illyes, Google (Search Off the Record, July 2024 — relayed via International Web Mastery’s write-up; paraphrased, confirm transcript wording before quoting)

• The language is not a tell for what country you’re targeting — you have to signal the target country separately, because a page’s language alone can map to many countries.
• On the HTML lang attribute: a CMSA content management system (CMS) is software that lets users create, manage, and publish digital content — like blog posts and pages — without writing raw code. WordPress, Drupal, and Joomla are the most common open-source CMS platforms. shipped with lang set to English on a page that was actually all German, with no way to override it — so the attribute isn’t a reliable targeting signal.
• On over-splitting: as soon as you have to vary things, that’s where you start introducing errors.

John Mueller, Google (via Search Engine Journal; paraphrased)

• hreflangHreflang is an annotation (in HTML, HTTP headers, or XML sitemaps) that tells search engines which language and optional region a page targets, and which alternate versions exist. It only works when every page in the cluster references all the others. tags are hints, not directives; near-identical same-language variants (like fr-fr and fr-be) commonly get consolidated to one canonical.

Fabrice Canel, Microsoft Bing (via Search Engine Roundtable; paraphrased)

• At Bing, hreflang is a far weaker signal than content-language.

Which should I target — language, country, or both?

Work top to bottom. The first question is about your content, not your tags.

Step 1 — Does this page’s actual content differ by country? Check the real substance: price, currency, shipping options, legal/regulatory text, tax handling, product availability.

• No — only the language changes. → You have a language-targeting job. Go to Step 2.
• Yes — dollars vs. pounds, different shipping, different legal text. → You have a country-targeting job (possibly plus language). Go to Step 3.

Step 2 — Language targeting path

• One language, one audience worldwide → a single global page may be enough.
• Multiple languages, same content and pricing → one page per language, language-only codes (en, es, de). Don’t add region codes you don’t need.
• Multiple languages inside one country (Canada, Belgium, Switzerland) → language+region codes that share the country (en-CA/fr-CA; de-be/nl-be/fr-be).

Step 3 — Country targeting path

• Content genuinely differs per country → language+region codes (en-US/en-GB), and consider country-specific URL structureURL structure is how the parts of a web address — scheme, domain, path, query string, and fragment — are organized and formatted. It mostly affects crawling, usability, and how engines understand a page, not rankings directly. if the market justifies it.
• Remember: you still can’t use a region code alone. gb is invalid; it must be en-GB.
• Same language across countries with identical content → don’t split. Consolidate to one language variant and revisit later if a market grows.

Step 4 — Sanity checks before you build

• Is there data (search volume, competitor SERPs) showing the country split is worth it, or are you assuming?
• Will each new variant have real, maintained differences — or are you manufacturing near-duplicatesThe same or very similar primary content reachable at more than one URL. There's no general duplicate content penalty — the real costs are possible signal dilution, the wrong URL getting chosen, and less-efficient crawling. that Google will just consolidate?
• Did you let the content answer drive the URL structureURL structure is how the parts of a web address — scheme, domain, path, query string, and fragment — are organized and formatted. It mostly affects crawling, usability, and how engines understand a page, not rankings directly., and not the reverse?

Default when unsure: start language-only, one page per language. Add country targeting later, per market, once evidence justifies the extra build and maintenance.

Common mistakes

Concrete ways the language/country decision goes wrong — and what to do instead.

1. Using a country code by itself (gb, de, mx). Why it’s wrong: hreflangHreflang is an annotation (in HTML, HTTP headers, or XML sitemaps) that tells search engines which language and optional region a page targets, and which alternate versions exist. It only works when every page in the cluster references all the others. always leads with a language. A region-only value is invalid, and Google doesn’t derive the language from a country code — so the tag is ignored. Do instead: always put a language code first (en-GB, de-DE, es-MX). If your reason is country-driven, the language is still primary: en-CA, not ca.

2. Assuming Google infers your target country from the page’s language. Why it’s wrong: a French page could be for France, Belgium, Canada, or Switzerland. Illyes: the language is not a tell for the target country. Do instead: if country matters, signal it explicitly — a region code (fr-CA) or a ccTLDCountry-code top-level domain like .de or .co.uk — a strong geotargeting signal. — plus on-page cues (local currency, address, phone).

3. Trusting the HTML lang attribute as a targeting signal. Why it’s wrong: it’s easy to get wrong by default — Illyes described a CMSA content management system (CMS) is software that lets users create, manage, and publish digital content — like blog posts and pages — without writing raw code. WordPress, Drupal, and Joomla are the most common open-source CMS platforms. that shipped lang="en" on an all-German page. Engines don’t lean on it for targeting. Do instead: set lang correctly for accessibilityWeb accessibility means designing sites so people with disabilities can use them, per the W3C's WCAG guidelines. It overlaps with SEO in specific, checkable ways — alt text, heading structure, descriptive link text, captions, and page speed all serve both audiences — but Google has said accessibility itself is not a ranking factor, and most WCAG success criteria (keyboard focus order, ARIA live regions, form labels) have no SEO effect at all., but drive targeting through hreflangHreflang is an annotation (in HTML, HTTP headers, or XML sitemaps) that tells search engines which language and optional region a page targets, and which alternate versions exist. It only works when every page in the cluster references all the others., URL structureURL structure is how the parts of a web address — scheme, domain, path, query string, and fragment — are organized and formatted. It mostly affects crawling, usability, and how engines understand a page, not rankings directly., and localized on-page signals — not the attribute alone.

4. Over-splitting into per-country variants when the content is identical. Why it’s wrong: every extra variant is another node in the reciprocal hreflangIf page A lists page B as an alternate, B must list A back — or Google ignores the pair. cluster that can break — and near-identical same-language variants (fr-fr/fr-be) often get consolidated anyway. Illyes: as soon as you vary things, you introduce errors. Do instead: split by country only when the content genuinely differs. Otherwise, one language variant.

5. Relying on a Search ConsoleA free Google service that reports how a site performs in Google Search and surfaces problems with how Google crawls, indexes, and serves it. It's first-party data straight from Google — but you don't need it to appear in results. or Bing Webmaster ToolsMicrosoft's free portal for monitoring and improving how a site appears in Bing search — the peer to Google Search Console, plus IndexNow instant indexing, richer backlink data, and keyword volumes. Because Bing's index also feeds Microsoft Copilot, it doubles as a window into AI-search visibility. country setting. Why it’s wrong: Google deprecated the GSC International Targeting report in 2022, and Bing removed GeoGenerative Engine Optimization (GEO) is the practice of optimizing content and brand presence so AI-powered search engines and assistants — Google AI Overviews, ChatGPT, Perplexity — cite, recommend, or mention you when generating answers. Google's position is that it's still SEO. Targeting in 2020. There’s no manual override to fall back on. Do instead: express the country decision structurally — URL/ccTLD, hreflang region code, on-page localizationLocalization is adapting content for a specific target market — not just translating the words, but adjusting currency, formats, idioms, cultural references, local search terms, and trust signals so the experience feels native..

6. Deciding country targeting on a ccTLD ranking “boost.” Why it’s wrong: Illyes says the underlying country-boost mechanism is getting ambiguous and will fade. Building for an eroding ranking edge is a bad basis for an expensive, hard-to-undo URL decision. Do instead: decide on whether the content genuinely differs by country. Let the business case, not a fading signal, drive the split.

7. Letting URL structureURL structure is how the parts of a web address — scheme, domain, path, query string, and fragment — are organized and formatted. It mostly affects crawling, usability, and how engines understand a page, not rankings directly. drive the content decision (backwards order). Why it’s wrong: teams pick ccTLDs or per-country folders first, then scramble to fill them with content that doesn’t actually differ. Do instead: answer “does the content differ by country?” first; let that answer choose the URL structure.

8. Serving locale-adapted content from a single URL instead of separate URLs + hreflang. Why it’s wrong: cookie-, browser-language-, or IP-based adaptation on one URL can leave variants undiscovered — GooglebotGooglebot is Google's web crawler — the software that fetches pages so Google can index and rank them. It comes in two variants, Googlebot Smartphone (primary, under mobile-first indexing) and Googlebot Desktop, and runs an evergreen Chromium renderer. commonly crawls from US IPs and doesn’t send an Accept-Language header. Do instead: publish a separate URL per language/country and connect them with hreflang, per Google’s own recommendation.

9. Auto-redirecting users to a “detected” language version. Why it’s wrong: Google advises against automatic redirectsA redirect sends browsers and crawlers from a requested URL to a different one. An HTTP redirect specifically is a 3xx status code paired with a Location header; meta refresh and JavaScript redirects achieve a similar navigation without being a 3xx response themselves. Permanent redirects (301/308) are Google's signal the target should be canonical; temporary ones (302/303/307) aren't. based on inferred language — they can trap users on the wrong variant and interfere with Google crawlingCrawling is how search engines use automated bots (like Googlebot and Bingbot) to discover URLs and download pages. A page has to be crawlable to be indexed, but crawling on its own isn't a ranking factor. the others. Do instead: give users an explicit link to switch versions; treat locale signals as directional, not a guarantee every visitor lands on the intended page.

10. Publishing same-language regional duplicates with no canonical. Why it’s wrong: hreflang alone doesn’t guarantee every regional URL stays separately indexed — near-identical en-US/en-GB pages can compete against each other. Do instead: set a canonical to a preferred version (same language, or closest substitute) alongside hreflang, per Google’s recommendation.

11. Translating only the boilerplate. Why it’s wrong: if the nav/footer are translated but the main content isn’t, Google is still reading mostly one-language content — it doesn’t read as a genuine language variant, so the pages can look like near-duplicates. Do instead: translate the content that actually matters, not just the chrome around it.

Worked examples

Four scenarios showing how the content decision drives the codes.

1. Global SaaS, identical pricing everywhere (language only) A developer tool priced the same worldwide, UI available in English, Spanish, and German. Nothing about the content differs by country — only the language.

• Decision: language targeting. One page per language.
• Codes: en, es, de (language-only). No region codes.
• Why: adding en-US/en-GB/en-AU here would create near-duplicatesThe same or very similar primary content reachable at more than one URL. There's no general duplicate content penalty — the real costs are possible signal dilution, the wrong URL getting chosen, and less-efficient crawling. with no real differences — more to maintain, more to break, and Google would likely consolidate them.

2. Canada — one country, two languages (language inside a country) A Canadian retailer, same prices and shipping, serving English and French Canadians.

• Decision: language targeting, inside one country.
• Codes: en-CA and fr-CA.
• Why: the country is fixed; the axis that varies is language. The shared -CA keeps both aimed at Canada while the language code does the real work.

3. US vs. UK — two countries, one language (country despite shared language) An ecommerce brand selling in the US and UK: different currency (USD vs. GBP), different shipping and returns, minor spelling differences, sometimes different legal copy.

• Decision: country targeting, even though the language is effectively the same.
• Codes: en-US and en-GB (note: en-GB, never the invalid en-UK).
• Why: the content genuinely differs by country, so the region code earns its place.

4. Belgium — one country, three languages (Google’s own example) Content for Belgium in German, Dutch, and French.

• Decision: language targeting across three languages inside one country.
• Codes: de-be, nl-be, fr-be.
• Watch out: be on its own is invalid and misleading — be is the language code for Belarusian, not a country code for Belgium.

The through-line: in every case the content answer came first, and the codes just encoded it. Language-only when only language differs; language+region when the country’s specifics genuinely differ; and always a language code first.

Language vs. country — cheat sheet

The two axes

hreflang code rules

• Language code = ISO 639-1Two-letter language codes used in hreflang (en, es, de). (mandatory, comes first).
• Region code = ISO 3166-1 alpha-2Two-letter country codes used in hreflang region targeting (us, gb, au). (optional, comes second).
• The tag always leads with a language — even for country-driven pages (en-CA).

Decision order

1. Does the content differ by country (price/currency/shipping/legal)? → yes = country; no = language.
2. Pick codes to match: language-only, or language+region.
3. Let that answer choose your URL structureURL structure is how the parts of a web address — scheme, domain, path, query string, and fragment — are organized and formatted. It mostly affects crawling, usability, and how engines understand a page, not rankings directly. — not the reverse.

What signals country to Google (when the answer is “country”): ccTLDCountry-code top-level domain like .de or .co.uk — a strong geotargeting signal., region code in hreflangHreflang is an annotation (in HTML, HTTP headers, or XML sitemaps) that tells search engines which language and optional region a page targets, and which alternate versions exist. It only works when every page in the cluster references all the others., local currency/address/phone, links from local sites. What does NOT: the page’s language, the HTML lang attribute alone, geolocation meta tags, IP-based analysis, or a GSCA free Google service that reports how a site performs in Google Search and surfaces problems with how Google crawls, indexes, and serves it. It's first-party data straight from Google — but you don't need it to appear in results./Bing country setting (both deprecated).

Bing: weights content-language over hreflangHreflang is an annotation (in HTML, HTTP headers, or XML sitemaps) that tells search engines which language and optional region a page targets, and which alternate versions exist. It only works when every page in the cluster references all the others.; don’t manufacture country URLs just to tag them.

Test yourself: Language vs. Country Targeting

Five quick questions on the two axes and how hreflangHreflang is an annotation (in HTML, HTTP headers, or XML sitemaps) that tells search engines which language and optional region a page targets, and which alternate versions exist. It only works when every page in the cluster references all the others. expresses them. Pick an answer for each, then check.

Resources worth your time

My related writing

• Over 67% of Domains Using Hreflang Have Issues (Study of 374,756 Domains) — my Ahrefs study quantifying how often sites get hreflangHreflang is an annotation (in HTML, HTTP headers, or XML sitemaps) that tells search engines which language and optional region a page targets, and which alternate versions exist. It only works when every page in the cluster references all the others. wrong, including using a country code where a language code belongs.
• Hreflang: The Easy Guide for Beginners — the Ahrefs beginner guide I contributed to; the plain-English version of when a region code is actually needed.

My speaking

• Hreflang Study and Interesting Issues — brightonSEO 2023 — the conference version of the 374,756-domain study, including Google’s most-specific matching order (language+country → language → x-defaultx-default is the reserved hreflang value that points to a fallback URL — the page Google shows when a user's locale doesn't match any of your other hreflang tags. It is optional and does not mean \"English.\") and the most common language/country code mistakes. (Talk listing · video)
• International SEO: The Weird Technical Parts — Pubcon Vegas 2019 — the deeper technical nuances, including hreflangHreflang is an annotation (in HTML, HTTP headers, or XML sitemaps) that tells search engines which language and optional region a page targets, and which alternate versions exist. It only works when every page in the cluster references all the others. depending on the indexedStoring a crawled page in the search index so it can appear in results. Crawled is not the same as indexed — Google selects what to keep, and indexing isn't guaranteed. URL, not the canonical.
• You’re Going To Screw Up International SEO — Pubcon Vegas 2017 — implementation chaos, generally.

From around the industry

• Managing multi-regional and multilingual sites (Google) — the official definitional split and the “both at once” case.
• Localized versions of your pages (Google) — the hreflang code format, the region-alone-is-invalid rule, and the Belgium examples.
• Google Hints at Lowering SEO Value of ccTLDs (Search Engine Journal) — coverage of Illyes’ LDCP comments and the fading ccTLDCountry-code top-level domain like .de or .co.uk — a strong geotargeting signal. boost.
• Google Reminds That hreflang Tags Are Hints, Not Directives (Search Engine Journal) — Mueller on same-language variants getting consolidated.
• Bing: hreflang Is a Far Weaker Signal Than content-language (Search Engine Roundtable) — Fabrice Canel on Bing’s mechanism.
• Insights from Google’s Search Off the Record podcast — internationalization and hreflang (International Web Mastery) — the write-up relaying the Illyes and Splitt comments referenced above.
• International SEO strategy: country vs language targeting (Click Consult) — a competing overview of the same decision.

Changelog