Three domain structures exist for serving international content, and the choice between them compounds in cost or value for years. Each trades geographic signal strength against authority consolidation and operational complexity. Get the fit right at launch and the structure quietly supports every market you add; get it wrong and you pay for it later as a migration project that competes with every other priority.
A brand expanding into multiple countries needs to serve country-specific content at country-specific URLs. The three options Google has supported for well over a decade are country-code top-level domains (example.de, example.fr), subdomains (de.example.com, fr.example.com), and subdirectories (example.com/de/, example.com/fr/). Each appears in real implementations across major brands, and Google’s own multi-regional documentation presents all three as valid, with no preference among them.
The three structures compared #
The trade-off is structural rather than incidental: a stronger geographic signal in the URL means weaker authority consolidation, and the reverse.
| Structure | Example | Geographic signal | Authority consolidation | Operational complexity |
|---|---|---|---|---|
| ccTLD | example.de | Strongest (TLD maps to country) | Weakest (each TLD is separate) | Highest (separate domains) |
| Subdomain | de.example.com | Moderate (needs hreflang) | Moderate (treated semi-separately) | Moderate |
| Subdirectory | example.com/de/ | Weakest in the URL (needs hreflang) | Strongest (one domain) | Lowest |
A ccTLD tells Google “this is for Germany” with no extra configuration, but each ccTLD is a wholly separate domain from Google’s perspective. A subdirectory sits under one strong domain and shares its authority, but it needs an explicit signal to tell Google which country each section serves. Google’s guidance is worth internalizing here: it does not perform locale-aware crawling. It will not fetch a page from a German IP to discover a German variant. The regional relationship has to be declared, not inferred.
ccTLDs: geographic signal built into the domain #
A ccTLD such as example.de carries an automatic geographic association. Google treats .de as a strong signal that the site targets Germany, .fr for France, .co.uk for the United Kingdom, .jp for Japan. That signal is the main reason to choose the structure.
ccTLDs tend to fit when the brand genuinely operates in-country: a company with a German office, German staff, and a German legal entity benefits from the localized brand association a ccTLD reinforces. They also fit when the brand can fund marketing in each country separately, because each ccTLD needs its own content, SEO, and link building, and that investment does not transfer across TLDs. And they fit when country-specific products or pricing differ substantially, since separate domains reduce confusion for users and search engines.
The cost lands on the other side. A new example.de starts from zero domain authority even when example.com is a major global site; the links, brand mentions, and crawl frequency all begin fresh, and building them tends to take a long time. Operational overhead multiplies too, since each domain carries its own hosting, certificates, monitoring, and analytics, and a fix made on one domain does not reach the others. ccTLDs make sense for brands that operate as genuinely separate entities per country and have the resources to run each domain as a serious property.
Subdomains: the middle path #
Subdomains (de.example.com) sit between the other two. Google has described subdomains as treated somewhere between fully separate sites and fully integrated sections of the parent domain, with the exact handling depending on factors it has not fully detailed.
They fit when technical or organizational reasons force separation: different country teams owning their own subdomains, separate CMS instances per country, or infrastructure that will not run under a single domain. A subdomain like uk.example.com also inherits some brand signal from example.com while presenting a country-specific identity. The weakness is that authority consolidation stays incomplete: some signal carries from parent to subdomain, but less than would carry within one domain, so new subdomains take time to build visibility. As with subdirectories, hreflang becomes mandatory because the country target is not in the URL. Subdomains are often a compromise; the cases where they are clearly best are narrower than the cases where they get used.
Subdirectories: maximum authority consolidation #
A subdirectory structure (example.com/de/, example.com/fr/, example.com/ja/) keeps all international content under one domain, so the authority of the primary domain benefits every country section.
They fit when the brand operates as one global entity with localized content, which describes most SaaS companies, enterprise software vendors, and large media properties, and when country content is translated rather than rewritten so its relationship to the primary version is clear. They also fit when one team can run one site: single CMS, single deployment, single hosting, single SEO team, which translates directly into faster execution as new markets are added.
The trade-offs are real. The geographic signal has to be made explicit through hreflang, because it is not in the URL. Infrastructure problems compound across markets, since a performance regression in the core slows every country section at once, and a single manual action against example.com affects every subdirectory beneath it. Even so, subdirectories have become the dominant choice for global brands, because the authority benefit is substantial and operational simplicity matters more as expansion accelerates.
Hreflang and Search Console: the explicit signals #
For subdomains and subdirectories, hreflang is not optional. Hreflang declares the relationship between language and country versions of a page: this is the German version, that is the French version, this other is the English version for the United Kingdom. It matters most for these two structures precisely because the URL carries no geographic signal, so the tags become the explicit input Google uses to decide which version to show in which market.
For ccTLDs, hreflang is helpful but less critical for basic country targeting, since the TLD already signals the country. It still clarifies language relationships where a single country runs more than one language, as a Belgian site might with French and Dutch versions. The implementation patterns are the same across all three structures: annotations in the HTML head, equivalent tags in HTTP headers, or entries in XML sitemaps. Common failures such as missing return tags, wrong language codes, and broken page mappings cause hreflang to be ignored, which pushes Google back to guessing.
One point worth correcting, because older guides still repeat it: Search Console country targeting no longer exists. Google deprecated the International Targeting report on September 22, 2022, having decided that manual country targeting for a generic TLD added little value. The report’s hreflang error reporting moved into the Enhancements section, and geographic relevance is now derived from signals Google collects directly. Those signals, per Google’s documentation, are the ccTLD, hreflang, server location, and other on-page cues such as local addresses, phone numbers, language, and currency. Google explicitly ignores locational meta tags like geo.position. If a strategy still assumes a Search Console country-targeting toggle, that assumption is outdated.
Mixed structures and migrations #
Some brands run mixed structures: ccTLDs in major markets, subdirectories for emerging ones. Google handles the mix, but complexity compounds, because hreflang has to span the structures correctly, each needs its own Search Console property, and analytics has to bridge the splits.
Migrations between structures are substantial projects rather than tuning exercises. Consolidating from ccTLDs to subdirectories requires large volumes of 301 redirects per country, careful content mapping, and a recovery period during which rankings stay unstable while Google reprocesses the new structure. Documented consolidations, such as NFON’s move from separate country domains onto a single gTLD, show both sides: NFON reported organic visibility recovering and then growing within a few months in its earliest markets, though results varied by market and its UK domain struggled against the international English version. Expect months of volatility, not weeks. Moving the other way, from subdirectories to ccTLDs, surrenders the authority advantage and restarts each domain from low visibility. Migration is rarely worth the cost unless the current structure is causing real business problems; it is a project of last resort, not an optimization.
What Google says about ranking impact #
Google representatives have said repeatedly that none of the three structures is preferred or penalized, and its multi-regional documentation states the same. Each structure signals different things, and Google combines those signals with hreflang, content quality, and its other ranking factors to decide what to show in each market. In practice, a well-implemented subdirectory can outrank a poorly-implemented ccTLD in the same country, because authority and content carry more weight than the structure label; a subdirectory with no hreflang and mixed languages on one page will struggle even on a strong domain. The structure decision sits upstream of ranking but does not determine it. The execution does.
A decision framework #
A short set of questions usually points to the answer:
- Does the brand operate as one entity with localized content, or as separate entities per country? One entity leans subdirectory; separate entities lean ccTLD.
- Can the team maintain multiple domain properties properly? If not, a subdirectory is operationally easier.
- Do country-specific products differ substantially? Substantially different offerings favor ccTLDs; translated equivalents favor subdirectories.
- Is there a strong existing domain with real authority? If so, subdirectories preserve it; if starting fresh, the question matters less.
- Are there legal, regulatory, or organizational reasons for separation? If so, a ccTLD or subdomain may be required.
Regulation can constrain the choice even when SEO would point elsewhere. Data-residency rules in some jurisdictions push toward ccTLDs paired with regional infrastructure, and country-specific advertising rules for pharmaceuticals, gambling, or financial products segment more naturally along ccTLD boundaries. Heavily regulated markets often end up on ccTLDs for compliance and operational separation rather than for any SEO advantage.
FAQ #
Does Google rank one structure higher than the others?
No. Google’s documentation and its representatives are consistent that none of the three is preferred or penalized. Authority, content quality, and correct hreflang implementation drive ranking far more than the structure choice.
Can I still set a country target in Search Console?
No. The International Targeting report was deprecated on September 22, 2022. Geographic relevance now comes from the ccTLD, hreflang, server location, and on-page signals such as local address, language, and currency.
Should an established brand pick subdirectories by default?
For many global brands operating as one entity with localized content, subdirectories are the reasonable default because they consolidate authority and keep operations simple. The clear exceptions are brands that genuinely run as separate country entities with separate products, teams, or legal structures.
Pick the structure you would choose if SEO did not exist, then confirm SEO will not break it. The structure that fits how the business actually operates is usually the one the SEO work can support, and it holds up for the years these decisions have to last. Choosing against business fit because an article claimed ccTLDs rank better in some market is how brands end up in post-migration recovery, undoing a decision that only looked clean on paper.