There’s no single “best” structure Google ranks higher, and Google has said so directly: its documentation on international and multilingual sites states there is no structural preference in ranking among country-code top-level domains, subdomains, or subfolders. The choice is a tradeoff between geo-targeting clarity, authority consolidation, and operational resourcing, not a ranking decision. For most organizations scaling hreflang across many markets, subfolders offer the best combination of crawl efficiency and authority consolidation, ccTLDs offer the clearest unambiguous geographic signal at the cost of starting each market from zero, and subdomains sit in between without cleanly capturing the advantage of either extreme.
Why the three structures aren’t interchangeable in practice, even without a ranking preference
ccTLDs (example.de, example.fr) send the strongest and least ambiguous geotargeting signal available, because the country association is baked into the domain itself and requires no configuration in Search Console or elsewhere to reinforce. Google doesn’t need to infer anything; the ccTLD is a direct, unspoofable declaration of country association recognized at the DNS level. The cost is that each ccTLD is treated as a fully separate domain for authority purposes: backlinks, crawl history, and trust signals accumulated by example.com do not transfer to example.de. Each ccTLD has to independently build the link profile, crawl history, and topical authority that search engines use to trust a domain. At scale, across a dozen or more markets, this means running a dozen or more independent authority-building efforts in parallel, which is a substantial and ongoing resourcing burden, and newer ccTLDs will generally take longer to rank competitively simply because they’re new domains regardless of content quality.
Subdomains (de.example.com, fr.example.com) are technically part of the same root domain from a registration standpoint, but historically Google’s systems have sometimes evaluated subdomains with characteristics closer to separate entities than subfolders, depending on hosting setup, content independence, and how consistently the subdomains share templates, navigation, and trust signals with the main domain. This isn’t a fixed rule so much as a tendency rooted in how subdomains are frequently used in the wild (separate blogs, separate app platforms, sometimes entirely separate teams or even separate hosting providers), so Google’s systems don’t automatically assume subdomain content should inherit the root domain’s full authority the way a subfolder does. In practice this makes subdomains a genuine middle ground: better authority-sharing potential than a ccTLD, but less reliably consolidated than a subfolder, and the outcome depends heavily on implementation consistency.
Subfolders (example.com/de/, example.com/fr/) consolidate everything under one domain. All backlinks, all crawl equity, and all accumulated trust signals apply to the single root domain, and every language/region path benefits from that shared foundation immediately, without needing to independently earn it. This is generally the most crawl-efficient option too: a single domain means Google’s crawlers are already established on your infrastructure, already have a stable understanding of your site’s crawl patterns and server response behavior, and don’t need to build separate crawl budget allocations and trust from scratch the way they effectively do for new ccTLDs. The tradeoff is a less explicit geographic signal than a ccTLD provides; subfolder geotargeting relies more heavily on hreflang annotations, Search Console’s International Targeting settings (for generic top-level domains), and content signals like language and local references to communicate regional relevance, since the URL structure itself doesn’t carry country information the way a ccTLD does.
Comparison at a glance
| Factor | ccTLD | Subdomain | Subfolder |
|---|---|---|---|
| Geo-targeting clarity | Strongest, automatic | Moderate, configuration-dependent | Weakest on its own, relies on hreflang/settings |
| Authority consolidation | None, each domain separate | Partial, inconsistent across setups | Full, single root domain |
| Crawl budget efficiency | Lowest, new domain per market | Moderate | Highest, shared crawl history |
| Resourcing burden at scale | Highest, N independent authority efforts | Moderate | Lowest, one domain to maintain |
| Time to competitive ranking per market | Slowest for new ccTLDs | Variable | Fastest, inherits existing trust |
A hypothetical illustration
Consider a hypothetical example: a mid-sized SaaS company called Northwind Analytics, currently only serving English-speaking markets from northwindanalytics.com, decides to expand hreflang coverage into German, French, and Japanese markets. Suppose Northwind’s leadership initially considers ccTLDs, northwindanalytics.de, .fr, and a Japan-specific domain, reasoning that local domains would build the most trust with local buyers.
Hypothetically, Northwind’s SEO lead points out that the company has spent several years building backlinks and topical authority under northwindanalytics.com, none of which would transfer to three brand-new ccTLD domains; each would need to build crawl history and link authority essentially from zero, a significant undertaking for a team without dedicated headcount in each region. Given that resourcing constraint, suppose Northwind instead launches northwindanalytics.com/de/, /fr/, and /ja/ as subfolders, paired with correct hreflang return tags across all four language variants and Search Console’s International Targeting settings reinforcing the German and French markets. In this hypothetical, the German subfolder is able to start ranking for regional queries within weeks rather than months, inheriting the root domain’s existing crawl relationship and backlink profile, illustrating the crawl-efficiency and authority-consolidation tradeoff described above.
What this means for implementation at scale
If the organization is expanding into many markets over time and doesn’t have the resourcing to run independent SEO and link-building efforts per domain, subfolders are the pragmatic choice: new markets launch with the benefit of the root domain’s existing authority and crawl relationship, hreflang handles the cross-referencing between language/region variants, and Search Console’s International Targeting report and settings can reinforce country association where the URL structure alone doesn’t make it obvious.
ccTLDs make sense when a market genuinely requires a distinct legal, brand, or hosting entity (common in regulated industries, or where local presence requirements exist), or where user trust signals specifically favor a local domain (this varies by market and audience, and is a business decision more than a pure SEO one). Organizations choosing ccTLDs should budget for the reality that each new ccTLD is close to a fresh SEO project, needing its own link acquisition and content maturity timeline, hreflang alone won’t compensate for that.
Subdomains are the least commonly recommended of the three for pure hreflang-at-scale purposes precisely because they capture neither extreme cleanly: they don’t get the unambiguous geo-signal of a ccTLD, and depending on setup they may not get the full authority consolidation of a subfolder either. They tend to make sense mainly when there’s a technical or organizational constraint (a different CMS, a different hosting environment, or a different team) that makes running true subfolders impractical.
Regardless of structure, the hreflang implementation itself, return tags between every URL in a language/region cluster, self-referencing annotations, and correct language-region codes, is what actually connects the variants for Google. The architecture decision affects how much authority and crawl efficiency each variant starts with; it does not change what hreflang has to do once that architecture is in place.
