International SEO: Hreflang, URL Structure, Canonicals, and Content Localization for Reaching Overseas Markets

Many businesses decide to translate their website into English or Mandarin and then hope that traffic from abroad will arrive on its own. A few months later the result disappoints: the English pages do not show up in Google for users in Singapore or Australia, the Indonesian pages are shown to English-speaking visitors instead, and search engines treat some pages as duplicates. The problem is rarely translation quality. More often it lies in a technical structure that does not tell search engines that each page has a counterpart in another language and region.

International SEO is the set of practices that ensures the right page appears to the right user, according to their language and, where relevant, their country. It is not just adding a language menu. It involves decisions about URL structure, hreflang annotations, canonicals, sitemaps, and, most often neglected, content that is genuinely adapted to how people in that market search.

This article explains those decisions in language that business owners and marketing teams can use, so you can judge the work of a technical team or vendor and avoid the most expensive mistakes.

Summary

• Decide the target first: language only, or language plus country, because this determines the URL structure and the number of pages you must maintain
• Use reciprocal hreflang on every page, plus x-default, so search engines show the version that suits the user
• Each language version references itself through a canonical tag; do not point the canonicals of all languages to a single page
• Put all versions in the sitemap and link between languages with a crawlable language switcher
• Localizing content matters more than literal translation: keywords, examples, currency, and tone must follow the market

Start from the target market, not the technology

Before touching any code, answer three questions. First, who do you want to reach: people who speak a certain language wherever they are, or people in a specific country? English speakers are spread across many countries, while Mandarin speakers in Indonesia, Singapore, Malaysia, and mainland China have different search habits, including different search engines.

Second, does your offer differ by country? If prices, shipping terms, or regulations differ, you need pages per country. If the offer is the same, pages per language are enough and much easier to maintain. For service and software businesses serving clients across countries from one team, one version per language is usually realistic.

Third, how much capacity do you have to maintain content? Each language version is a long-term commitment. A page translated once and left to go stale does more harm than no page. Decide how many pages and articles you can keep accurate. A broader strategic discussion is in multilingual websites for regional markets.

URL structure: subfolder, subdomain, or separate domain

There are three common patterns for separating language versions, and each has trade-offs.

Subfolders on one domain

For example the main domain with /en and /zh paths. This is the most commonly recommended choice for small and medium businesses, because domain authority is gathered in one place, management is simpler, and one certificate and one analytics setup are enough. The downside is that you are tied to one hosting infrastructure for all languages.

Subdomains

For example en.domain.com. This gives infrastructure flexibility, but search engines may treat a subdomain as more separate from the main domain, making authority harder to build. Choose it when the teams or systems of each region really are different.

Country-code domains

For example domains ending in a country code. This gives the strongest geographic signal and builds local trust, but costs the most: each domain builds authority from zero, and you manage many certificates, hosting setups, and link profiles. It suits companies that truly operate as a local entity in each country.

Whatever the choice, be consistent. Changing the structure later demands tidy redirects so rankings are not lost, as discussed in website redesign and migration without losing SEO rankings. Also avoid choosing language only by URL parameter or cookie, because search engines cannot always crawl such variations and each language should have its own fixed address.

Hreflang: telling search engines about page counterparts

Hreflang is an annotation stating that this page has other versions in other languages or regions. It can be placed in the HTML head, in the HTTP header, or in the sitemap. The rules most often broken are as follows.

• Reciprocity: if page A points to B, then B must point back to A. Annotations without a pair are ignored
• Self-reference: every page includes itself in the hreflang list
• Correct codes: languages use two-letter codes such as id, en, and zh, and a region is added only when needed, for example en-SG. A country code cannot stand alone without a language
• x-default: one fallback entry for users whose language is not in the list, usually a language selector page or the main version
• Absolute, canonical URLs: use full addresses identical to those used in the canonical tag

Another common error is placing hreflang on pages that have no real counterpart. If an article is only available in Indonesian, do not create an English hreflang entry pointing to a page that does not exist or to a substitute that is not equivalent. Hreflang states equivalence of content, not hope.

Canonicals, sitemaps, and language switchers

The canonical tag tells search engines the main address of a page. On a multilingual site, each language version must be canonical to itself. The classic mistake is pointing the canonicals of all versions to the Indonesian page, which in practice asks search engines to ignore your translations. Canonicals and hreflang must agree: the URL listed in hreflang should be exactly its canonical URL.

The sitemap is another very helpful place. Include every URL from all languages, then list the hreflang alternates in each entry. That way, if there is a mismatch between a page and the sitemap, you can audit it in one place. A sitemap also speeds up discovery of new pages, which matters for sites with an active blog such as in the company blog as a lead engine.

The language switcher should be a normal crawlable link, not a button that works only through script without changing the address. Avoid automatic redirects based on location or browser language that cannot be undone. Forced redirects confuse traveling users, and a crawler coming from one country will see only one version. It is better to show a gentle suggestion, such as a small banner offering another language, and leave all versions accessible.

Localizing content: more than translation

A page translated word for word rarely competes with a page written for local readers. Several things need adapting.

• Keywords: how people in one language search for a topic is not always the literal equivalent of the keyword in another language. Do keyword research per language rather than translating a keyword list
• Examples and context: currency, units, regulations, payment platforms, and company examples should be relevant to readers in that market
• Tone and format: the level of formality, date format, and number format differ between cultures
• Titles and meta descriptions: rewrite them so they fit and appeal in the target language. Mandarin characters and Latin letters have different display limits
• Images: text inside images and alternative text need to be translated too

If you use machine translation or AI assistance to speed up, set a review process by native speakers. Content that sounds stiff or uses the wrong terms damages prospects' trust and can be judged low quality. A discussion of how AI-based search assesses content quality is in AI-based search and the future of SEO for business websites.

Schema, speed, and other technical signals

Structured data should be written per language: name, description, and the inLanguage property follow the page concerned, while organization information stays consistent. Detailed guidance is in schema markup and structured data for business website SEO.

Speed plays a large role in international markets because visitors access from afar. Use a content delivery network and image optimization so pages load quickly anywhere. Metrics such as LCP, INP, and CLS apply in every country; details are in Core Web Vitals for business websites and website speed and its impact. For sites whose content is managed separately from presentation, headless website architecture makes it easier to serve many languages from one content source.

Also make sure the language attribute on the html tag matches the page content. It helps screen readers, browser translators, and search engines, and is part of website accessibility for business.

Measuring international SEO results

Without measurement you do not know whether the effort works. Some realistic ways to monitor:

• Track impressions, clicks, and position per country and language in the search engine's webmaster tool, then compare the pages you target with those that actually appear
• Check the indexing report to find pages excluded as duplicates or whose canonical was chosen by the search engine differently from what you set
• Track conversions per language version in analytics: form submissions, WhatsApp clicks, or consultation requests. Traffic without conversion is not the goal; see [website and mobile app analytics for decisions](/blog/analitik-website-dan-aplikasi-mobile-untuk-keputusan)
• Audit hreflang and canonicals periodically whenever a new article or page is added, because small errors spread fast when a template is wrong

Give it time. New pages in a new language usually take weeks to months to be indexed and begin ranking, and the outcome depends on competition in that market. Do not claim results before the data supports them.

Common mistakes

• Translating the entire site at once without the capacity to maintain it, so non-Indonesian versions fall behind and go stale
• One-way hreflang, or hreflang pointing to a redirected URL or one with a different canonical
• Canonicals of all languages pointing to a single page
• Automatic location-based redirects that block crawlers and users
• Leaving some elements in the original language, such as menus, forms, error messages, and image text
• A sitemap containing only one language
• Internal links from English pages that keep pointing to Indonesian pages, so users and link authority switch language unnoticed

Getting started

Start with a simple audit: list the pages most important to international prospects, for example service pages and the five to ten best articles, then check whether each has a correct counterpart, canonical, and hreflang. Fix templates first because one fix applies to all pages. After that, do keyword research per language for priority pages, rewrite titles and meta descriptions, and schedule a native-speaker review. Finally, set up monthly monitoring. The foundation supporting all this is in basic SEO for company websites, and if you also sell products across borders, continue to selling abroad with a cross-border online store.

Frequently asked questions

Will the same translation be treated as duplicate content?

Different language versions of the same content are not duplicates to search engines as long as the structure is correct: each version has its own URL, is canonical to itself, and has reciprocal hreflang. Duplicate problems arise when several URLs show the same language with slight variations, for example different country pages with identical content.

Do I need hreflang if I only have two languages?

Yes. Hreflang helps search engines choose the suitable version for each user, including with two languages. Without it, the wrong version may show up or the versions may compete with each other.

What about Mandarin: simplified or traditional?

Choose based on the market. Simplified Chinese is used in mainland China, Singapore, and Malaysia, while traditional is used in Taiwan and Hong Kong. If you target Southeast Asia and mainland China, simplified characters are usually more appropriate. Use the zh tag or a specific variant as needed.

Do I need search engines other than Google?

It depends on the market. In mainland China, users commonly use local search engines with their own rules and hosting requirements. If that is a real target, plan separately. For other markets, focusing on Google usually covers most of the traffic.

Conclusion

International SEO is a combination of structural decisions and content commitment. Choose a URL structure you can maintain, set up reciprocal hreflang with consistent canonicals, put all versions in the sitemap, and localize content based on keyword research per language. Then measure results per country and language and audit periodically. Businesses that are disciplined about these basics are far more likely to appear in front of international prospects than those that merely add a language menu.