Companies that expand across borders often invest heavily in translation and local content before deciding how the website should be organized. This sequence creates trouble later, because a site whose structure confuses search engines wastes a large share of the effort that goes into the content itself. Language versions compete with each other, the wrong version appears in the wrong country, and crawlers spend time on duplicate pages.

Japan and Western markets add a further layer of complexity. A Western company entering Japan must decide whether Japanese content lives on its existing global domain, on a Japanese country-code domain, or on a separate site. Japanese companies entering the United States, the United Kingdom, or Europe face the mirror-image question, along with the need to distinguish between several English-speaking audiences and between countries that share a language.

This white paper explains how to choose an international site structure, implement hreflang annotations correctly, and avoid the technical errors that most often undermine cross-border search visibility. Each section includes decision criteria, implementation guidance, and notes on differences between the Japanese and Western contexts. A 90-day rollout plan at the end gives teams a schedule they can adapt to their own resources.

Readers who need a refresher on how search engines discover pages can also consult our explanation of how to register a site with Google.

1. Architecture Decisions Are Hard to Reverse

Site architecture determines how search engines discover, group, and rank the language and regional versions of a website. Once a company publishes hundreds of pages, earns links to specific URLs, and builds reporting around a structure, a change requires redirects, annotation updates, and a period of ranking volatility. Early decisions therefore deserve more analysis than their technical appearance suggests.

Architecture also shapes the business processes around the website. A single-domain structure allows one content management system, one analytics property, and a unified brand presence. Separate country domains allow local teams to operate independently, which supports local legal entities, but they multiply the workload for security, maintenance, reporting, and link building.

Tokyo SEO Maker’s own website illustrates the single-domain approach. The site serves Japanese content at the root and publishes English, French, Spanish, German, Traditional Chinese, Simplified Chinese, and Thai versions in language directories on one domain. Companies that manage many languages from one team often choose this pattern because it concentrates authority and simplifies governance.

2. The Three Core Structures

International websites typically use one of three structures. A country-code top-level domain, or ccTLD, assigns each market its own domain, such as example.jp or example.de. Subdomains place each market on a prefix of the main domain, such as ja.example.com, and subdirectories place each market in a folder, such as example.com/ja/.

A ccTLD sends the clearest geographic message. It can strengthen local trust and allows a local team to manage content with independence. Each new domain starts without the authority of the others, so the company must earn links and visibility again in every market.

Subdomains offer technical separation, which helps when regional sites run on different platforms. Search engines may treat them as distinct sites in some circumstances, so the benefit of the main domain’s authority is less certain. Companies choosing this model should plan separate link-building and monitoring efforts.

Subdirectories keep every market under one domain. They share authority, reduce cost, and simplify analytics and governance. The structure requires disciplined hreflang implementation, since all versions sit side by side and search engines need clear instructions about which one serves each audience.

3. Domain Considerations for Japan and Western Markets

Eligibility rules for Japanese domains deserve early attention. According to the Japan Registry Services registration FAQ, companies and individuals with a permanent postal address in Japan can register a general-use .jp name. The .co.jp domain requires a corporation registered under Japanese law, and foreign companies registered in Japan as a Gaikoku Kaisha may also apply.

A Western company without a Japanese entity therefore cannot register a .co.jp domain. For that company, a Japanese subdirectory on the global domain forms the practical starting point. A local entity changes the picture, because a .co.jp or .jp address becomes available and can reinforce credibility with Japanese buyers.

Japanese users often recognize .co.jp and .jp addresses as signals of a company established in Japan, and many associate them with credibility. This tendency makes a Japanese domain attractive once a company has a local entity and a long-term commitment to the market. Until then, a Japanese-language section with a visible company overview, address, and contact information builds trust on the global domain.

Western markets offer more flexibility. Many companies serve the United States, Canada, and global audiences from a .com domain, while European businesses sometimes pair country-code domains such as .de, .fr, or .co.uk with local entities. Japanese companies entering the West usually start with a .com or an existing global domain and add country-level structures when sales, support, and legal operations justify them.

4. A Decision Framework

The right structure depends on the company’s stage, resources, and market commitment. Common scenarios appear in the table below, matched to a recommended structure with the reasoning.

The chart below compares the relative implementation effort of the three structures. Effort includes hosting, security, analytics, link building, and governance across markets.

Hybrid models deserve special mention. A company may begin with subdirectories, then move its largest market to a country domain once local operations mature. Such a move requires a complete redirect map, updated annotations, and careful monitoring, so the company should plan for it from the start.

5. How hreflang Works

Hreflang annotations tell search engines which language or regional versions of a page exist, so that each searcher receives the most suitable one. Google’s documentation on localized versions of a page states that localized versions count as duplicates only when the main content remains untranslated. The same documentation describes three ways to supply annotations: link elements in the page head, HTTP headers, and XML sitemaps.

A set of link elements for a page available in Japanese, American English, British English, and German, with a fallback version, appears below. Each line names the language or region and the full URL of the matching version.

<link rel="alternate" hreflang="ja" href="https://www.example.com/ja/page/" />
<link rel="alternate" hreflang="en-us" href="https://www.example.com/en-us/page/" />
<link rel="alternate" hreflang="en-gb" href="https://www.example.com/en-gb/page/" />
<link rel="alternate" hreflang="de" href="https://www.example.com/de/page/" />
<link rel="alternate" hreflang="x-default" href="https://www.example.com/en-us/page/" />

Every version of the page must carry the complete set of annotations, including a reference to itself. Each link must also be reciprocal, which means that if the Japanese page points to the German page, the German page must point back to the Japanese one. Search engines may ignore annotations that fail either test.

The URLs in the annotations should be absolute, indexable, and consistent with each page’s canonical tag. Pages that redirect, carry a noindex directive, or point their canonical tag to a different URL send conflicting signals. Teams that generate annotations through templates or plugins should audit the output regularly, since small template errors can repeat across thousands of pages.

6. Choosing Language and Region Codes

Hreflang values begin with a language code and can add a region code. Language selection follows ISO 639-1, and region selection follows ISO 3166-1 Alpha 2. A region code on its own is invalid, and a country code cannot substitute for a language code, which is why Japanese content uses ja and never jp.

Companies should create separate English versions for different countries only when the content differs in a meaningful way. Differences in price, currency, shipping, legal notices, product range, or terminology justify the extra pages. Near-identical English pages for several countries add maintenance cost and increase the risk of duplicate-content confusion.

7. The Role of x-default and the Language Selector

The Google Search Central blog explains in its post on how x-default can help you that the value can point to a language and country selector, to a page for users without a matching region, or to the version a company regards as its default. Users whose language matches none of the listed annotations reach the x-default page. This makes it a useful safety net for companies that serve only some markets or only some pages in several languages.

A language selector should use ordinary, crawlable links so that search engines and users can reach every version. Selectors that depend entirely on JavaScript or on IP detection may hide alternate versions from crawlers. The selector should also remember the visitor’s choice and apply it on later visits.

Automatic redirects based on location or browser language create problems for users who travel, use shared networks, or prefer a different language. They also prevent crawlers from reaching alternate versions, since crawlers often request pages from a limited set of locations. A dismissible banner that suggests the local version respects the visitor’s choice and keeps every version accessible.

8. Common Implementation Errors

The errors below appear frequently in cross-border site audits. Each one weakens the signals that tell search engines which version belongs to which audience, and each has a straightforward fix.

Technical audits benefit from clear priorities. The chart below suggests how a team might weight its attention across the main areas of an international SEO review.

9. Canonicalization and Near-Duplicate Content

Canonical tags tell search engines which URL represents a piece of content, and hreflang tells them which version serves which audience. The two signals must agree. Each language version should carry a canonical tag that points to itself, and the hreflang set should list those same canonical URLs.

Near-duplicate versions appear most often among English pages for several countries. A page for the United States and a page for Canada may differ only in a price and a phone number, which creates a risk that search engines will cluster them and show one version in both markets. Meaningful differences in content, such as local shipping terms, regulatory statements, testimonials, and examples, help each page justify its own place in search results.

Japanese pages face a different issue. Machine-translated text that follows English sentence structure can read awkwardly to Japanese users, and weak engagement can limit the page’s performance. Native editors who review terminology, tone, and formatting improve both user experience and search visibility.

10. URL Conventions, Slugs, and Character Encoding

Japanese URLs can contain kanji, hiragana, and katakana, which browsers display natively. When the same URL is copied into an email, a spreadsheet, or a report, it often appears in percent-encoded form, which produces long and unreadable strings. Romanized or English slugs avoid this problem, remain easy to share, and simplify analytics reports.

Whatever convention a company selects, it should apply it across all pages and keep it stable over time. Changing slugs after launch requires redirects, and each redirect adds risk. Every page should declare UTF-8 encoding and the correct language attribute, such as lang=”ja”, in the HTML element.

11. Performance, Fonts, and Hosting

Japanese web fonts contain thousands of characters and can reach large file sizes, which slows page rendering. Teams can reduce the load by subsetting fonts, limiting weights, using font-display settings that show text immediately, and defining fallback stacks that include common Japanese system fonts. A content delivery network with locations near the target audience reduces latency for users in Japan and in Western markets.

Speed matters for users and for search visibility, and the effect increases on mobile devices. Companies should test pages from locations in each target market, since a site that loads quickly from Tokyo may perform poorly for visitors in Europe or North America. Core Web Vitals reports in Google Search Console help teams identify problem pages by market.

12. Structured Data and Trust Signals

Structured data helps search engines understand the organization behind the site and its relationship to each market. Organization markup should include the legal name, address, contact information, and links to official profiles, and Japanese pages should present the company name in Japanese and Latin characters where both apply. Consistency between the website, business listings, and registration records strengthens the signals that support credibility.

Trust signals carry particular weight for foreign companies. Our explanation of E-E-A-T and specific SEO evaluation criteria describes how experience, expertise, authority, and trustworthiness shape quality assessments. A technically correct international structure supports those signals by ensuring that users and search engines find the right content in the right language.

13. Search Console, Sitemaps, and Monitoring

Monitoring begins with verification. Teams should verify the domain, or each language directory, in Google Search Console so they can review indexation, performance, and technical issues by market. XML sitemaps for each language version help search engines discover pages, and sitemaps can also carry hreflang annotations for large sites.

Reports should separate performance by country and by query language. A Japanese page that receives impressions from users in Japan and clicks from Japanese queries shows the structure working as intended. Unexpected traffic from the wrong country or language points to annotation errors, incorrect redirects, or content that fails to match the target audience.

Japan also includes search behavior outside Google, with Yahoo! Japan holding a recognized place among users. Companies should review visibility on each relevant engine and follow the guidance in our overview of SEO and the Google algorithm in Japan.

14. A 90-Day Rollout Plan

The plan below suits a company introducing a new international structure or restructuring an existing one. Smaller sites can compress the schedule, and large commerce platforms may need additional time for template work and quality assurance.

15. Governance Mistakes to Avoid

Translating content before deciding the architecture creates expensive rework. Teams that publish hundreds of pages under a temporary structure must later redirect and re-annotate them. A short architecture workshop before the first translation project prevents most of this waste.

Leaving ownership undefined causes drift over time. Marketing teams add pages, developers change templates, and local partners publish content, yet nobody maintains the annotation set. Assigning a single owner for international SEO, with a monthly audit schedule, keeps the structure consistent as the site grows.

Ignoring local teams and partners limits the quality of the result. Japanese staff can review terminology, tone, and layout, and Western regional teams can confirm pricing, legal notices, and examples. Their input should enter the process during the build phase and before launch.

Summary

International site architecture gives every later SEO effort a foundation. The choice among ccTLDs, subdomains, and subdirectories shapes authority, cost, local trust, and operating complexity, and the right answer depends on the company’s stage and market commitment.

Western companies entering Japan usually begin with a Japanese subdirectory on a global domain and consider a Japanese domain once a local entity exists. Japanese companies entering Western markets usually begin with an English subdirectory or a .com and add country-level structures when operations justify them. In both directions, hreflang annotations, canonical tags, and a crawlable language selector turn the structure into clear signals for search engines.

Consistent governance sustains the result over time. A named owner, a regular audit, and a clear rollout plan allow the site to grow across languages and markets without losing technical coherence.

Frequently Asked Questions

Should a Western company use a subdirectory or a .jp domain for its Japanese site?

A subdirectory on the global domain suits companies that lack a Japanese entity and want to begin quickly, since eligibility rules restrict who can register .jp and .co.jp names. Japanese domains become attractive once the company has a local entity, local operations, and a long-term commitment to the market.

Does hreflang improve rankings?

Hreflang helps search engines serve the correct language or regional version to each user, which improves relevance and reduces the chance of the wrong page appearing. The annotations do not raise a page’s authority by themselves, so content quality, links, and technical health remain essential.

Can one English page serve the United States, the United Kingdom, and Australia?

A single English version works well when the content, pricing, and legal information apply to all three markets. Separate versions make sense when currency, shipping, regulation, spelling, or product availability differ in ways that matter to customers.

Is it acceptable to use machine translation for the Japanese site?

Machine translation can support a first draft, but native editors should review every customer-facing page. Awkward phrasing, incorrect terminology, and unnatural formatting reduce credibility with Japanese users and can limit search performance.

How long does it take to see results after restructuring an international site?

Search engines typically need several weeks to process redirects, new annotations, and updated sitemaps, and larger sites may take longer. Teams should monitor indexation and traffic by market weekly during the first three months and expect some fluctuation during the transition.

What is the most common hreflang mistake?

Missing reciprocal links and missing self-references appear in many audits. Both errors weaken the annotation set, and a template-based generation process with regular crawls catches them early.