International SEO

International SEO is the practice of structuring and optimising a website to perform in multiple countries and languages. It covers how a search engine decides which regional or language version of a page to show each user, and what signals influence that decision.

Most of what follows is sourced from Google, and applies to Google. Engines differ here more than in most areas of SEO: Microsoft’s Fabrice Canel has said on record that “hreflang is indeed a far weaker signal than content-language at Bing”, with Bing inferring language and market from links and visitors as well.¹ Bing’s own guidance does now recommend hreflang, so the accurate summary is that hreflang is load-bearing for Google and secondary for Bing, rather than universal.

What is international SEO?

When a website serves users in more than one country or language, organic search creates problems that standard SEO does not address. Google needs to know which version of a page targets which audience, whether similar pages in different languages compete with one another, and which URL represents the canonical version of each piece of content.

International SEO provides those signals. It is the sum of structural decisions (how URLs are organised across markets), technical implementation (hreflang annotations that communicate language and regional targeting), and content strategy (how pages are adapted for each market rather than duplicated across them).

Core international SEO elements

These split into two areas: the technical targeting and structure, and the content and market research behind it.

Targeting and structure

• International URL Structure. The choice between ccTLDs, subdomains, and subdirectories determines how Google understands regional targeting and how link equity is distributed across your international presence.
• Hreflang and International SEO. The technical implementation reference: syntax, ISO codes, delivery methods, and common error patterns.
• Hreflang Best Practices. The strategic guide: when hreflang is needed, how to maintain it at scale, and how to audit for the errors that cause ranking failures.
• Geotargeting in Google Search Console. Google removed the International Targeting report in 2022, taking both country targeting and hreflang error reporting with it. What replaced them, and how to monitor hreflang now.

Content and market research

• Multilingual Content Strategy. Which markets to prioritise, when to translate versus localise, and how to maintain quality and consistency across languages.
• International Keyword Research. Search volume, phrasing, and intent differ across markets even for identical topics. Research must be done independently for each target language and country.
• International Link Building. Earning links from locally relevant sources in each target market, and why a global link profile is not a substitute for local authority.
• Common International SEO Mistakes. The errors that consistently undermine international campaigns: broken hreflang, IP-based redirects, machine translation without review, and misapplied geo-targeting.

Why is international SEO different?

Domestic SEO assumes a single audience, a single language, and a single competitive set. International SEO multiplies each variable by the number of markets you target. A French competitor for the French version of your product is not the same entity as a UK competitor. The content, keywords, and links that work in one market do not transfer automatically to another. Skipping localisation and market-specific research produces untargeted content and link profiles that never reach local relevance thresholds, which is a separate failure from getting the technical annotations wrong.

The structural complexity is also distinct. A single-market site has one URL per piece of content. An international site may have ten or more URLs for conceptually the same page, each targeting a different language or region. Correct hreflang tells Google which of those to serve to whom.

Two common misconceptions are worth heading off, because both lead people to expect the wrong thing from hreflang. It does not consolidate authority across versions: that is what canonicalisation does, and Google states that canonical annotations carrying hreflang “are not used for canonicalization”.² And translated pages are not a duplicate content problem to begin with, since “localized versions of a page are only considered duplicates if the main content of the page remains untranslated”.³

Where versions genuinely can undercut each other is narrower: near-identical pages in the same language for different regions, such as en-GB and en-IE differing only cosmetically. Google tends to consolidate those onto one canonical rather than serving both. The fix is either meaningful differentiation or one page for both markets, and Google’s documented approach where the content really is similar is to use rel="canonical" and hreflang together.⁴

International SEO and AI search

Google’s AI surfaces are now broadly multilingual: AI Overviews reached “more than 200 countries and territories and in more than 40 languages” by May 2025,⁵ and AI Mode expanded to “over 200 countries and territories total” by October 2025.⁶ So the question of which language and market version gets used is live in AI answers as well as in classic results.

What is not established is any mechanism connecting hreflang to those surfaces. Google has published nothing on how hreflang interacts with AI Overviews or AI Mode, and its guidance on AI features says only that “there are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary”.⁷ Claims that broken hreflang confuses AI answers are not supported by any platform statement, and are worth treating with scepticism.

The structural point that can be made is narrower. Hreflang is a serving signal, applying at the moment Google decides which URL to show a user. AI answers select and synthesise sources before that stage, which means a market-specific page can be valid, indexed and correctly annotated and still not be used. Bing is the one platform to have described an adjacent mechanism: “LLMs group near-duplicate URLs into a single cluster and then choose one page to represent the set. If the differences between pages are minimal, the model may select a version that is outdated or not the one you intended to highlight.”⁸

The practical conclusion matches the one on the clusters: keep hreflang for classic search, where it still does its job, and rely on genuine market differentiation rather than annotation to determine whether a local page earns its place in an AI answer.

What signals does Google use to target a market?

Since Search Console’s country targeting setting was removed in 2022, the question comes up constantly. Google documents four groups of signals for determining a page’s target locale:⁴

• ccTLDs, which “provide a strong signal” of country intent, with the caveat that Google treats around twenty vanity ccTLDs (including .ai, .co, .io, .me and .tv) as generic.
• hreflang statements, “whether in tags, headers, or sitemaps”.
• Server location, qualified as “not a definitive signal” given CDNs and hosting choices.
• Other signals, including “local addresses and phone numbers on the pages, the use of local language and currency, links from other local sites, or signals from your Business Profile”.

The last group is the one most often ignored and the cheapest to act on. Note also that language is determined from what a reader sees: “Google uses the visible content of your page to determine its language. We don’t use any code-level language information such as lang attributes, or the URL.”⁴

International SEO and E-E-A-T

Trust operates locally. A business that is well known in its home market arrives in a new one with none of that recognition attached, and the signals that establish credibility are market-specific: local contact details and a real address, pricing in local currency, compliance with local regulatory expectations, coverage in local publications, and reviews from customers in that market.

The E-E-A-T framework applies with the usual caveat that it is a concept from Google’s quality rater guidelines rather than a measurable ranking factor. Its practical relevance to international SEO is that experience and trustworthiness do not translate along with the copy. A page that reads as written for a market, by people who operate in it, is doing work that no amount of correct hreflang can substitute for. This matters most in regulated sectors, where local legal and professional context is part of demonstrating competence at all.

Where to start

For most sites entering new international markets, the priority order is:

1. Choose a URL structure and implement it before adding content.
2. Decide which market variants genuinely warrant their own pages, rather than creating one per market by default.
3. Implement hreflang across the variants you keep.
4. Do keyword research independently for each target market.
5. Localise content for each market rather than translating directly from the home-market version.
6. Build links from locally relevant sources in each target market.
7. Monitor performance in Search Console, segmented by country, and audit hreflang with a crawler since Search Console does not report hreflang errors.

Footnotes

1. Bing: hreflang is a weak signal, content-language is stronger — Search Engine Roundtable ↩

2. How to specify a canonical page — Google Search Central ↩

3. Tell Google about localized versions of your page — Google Search Central ↩

4. Managing multi-regional and multilingual sites — Google Search Central ↩ ↩² ↩³

5. AI Overviews expansion, May 2025 — Google ↩

6. AI Mode expands to more languages and locations — Google ↩

7. AI features and your website — Google Search Central ↩

8. Does duplicate content hurt SEO and AI search visibility? — Bing Webmaster Blog ↩