For a website to be considered multilingual, its pages must provide content in different languages. Certain HTML tags are used to inform Google that a page has different localised versions. The “HTML Lang” and “hreflang” tags are used to indicate the language of a page and the corresponding URLs of its alternative language versions. In addition, the XML sitemap should also be configured in accordance with the multilingual structure.
How Should Page Translations Be Prepared to Support SEO?
For a multilingual website, simply changing the language of the page content is not enough. To build an SEO-friendly multilingual structure, the way the content is translated is just as important as the technical implementation.
Pages should not be translated into the relevant languages through automatic translation or translation tools. Likewise, artificial intelligence tools such as ChatGPT should not be used for translation, as Googlebot may be able to detect that the translations have not been reviewed or prepared by a human, which may create negative signals across the website.
For an SEO-friendly multilingual structure, the content should be written or reviewed by a native speaker during the translation process. This makes it possible to provide familiar and useful content to users in the targeted language or region while also offering valuable content to search engine bots.
What Is the HTML Lang=region_code Tag and What Does It Do?
The HTML Lang code is used to specify the language or region used by a web page. This allows us to inform search engine bots about the document’s language at the initial stage. In a multilingual structure, when the page content and URL change, this code should also change according to the relevant language or region.
HTML Lang Code:
<html lang="region_code">
Example of Use:
On a website containing both Turkish and English, <html lang="tr"> should be used on Turkish pages, while <html lang="en"> should be used on English pages.
Using the Hreflang Tag
The hreflang tag is placed between the <head></head> tags to annotate the corresponding URLs of a page in different language versions. This helps prevent search engine bots from identifying these pages as duplicate content. Although the language version is different, the page content may remain semantically similar in terms of visuals, headings, and text, so hreflang tags should always be used in a multilingual structure.
The x-default value is a tag that should be included in every hreflang implementation. While hreflang specifies the alternative language versions of a page, x-default is used to identify the page’s default URL. In other words, it defines the default language version that should be displayed for a language or region that does not have a specific corresponding version.
<link rel="alternate" hreflang="language_code" href="alternative_language_page_URL" />
<link rel="alternate" hreflang="x-default" href="default_language_page_URL" />
How Should the X-default Value Be Used for Non-Targeted Regions?
To explain with an example, imagine that a website targets the United Kingdom and France. When a user from Italy visits the website, they may be directed to a language selection landing page where they can choose the language they want to use.
In this scenario, the x-default value can reference the language selection page directly instead of a URL corresponding to a specific language version. This allows users to choose among the languages in which content is available.
What Should Be Considered When Using Hreflang?
- The hreflang annotation should be configured between the
<head></head>tags. - All pages with corresponding language versions should be included.
- Pages with language versions should reference themselves, and the x-default value should also be annotated.
- Supported language and region codes should be specified correctly.
Definitions Used in Hreflang
href: The URL of the corresponding language version of the page should be entered here.
hreflang: The language or language-region code should be entered here.
x-default: This should be used for the default language displayed in regions that are not specifically targeted.
Hreflang Usage Examples
Let us examine how hreflang should be configured for an e-commerce company that sells T-shirts and operates in the United States, the United Kingdom, Germany, and Türkiye. The following example is prepared for a page listing T-shirts.
URL Structure:
For English speakers in the United Kingdom:
https://t-shirt.com/en-gb/t-shirts
For English speakers in the United States:
https://t-shirt.com/en-us/t-shirts
For German speakers in Germany:
https://t-shirt.com/de-de/t-shirts
For English speakers in Germany:
https://t-shirt.com/en-de/t-shirts
For Turkish speakers in Germany:
https://t-shirt.com/tr-de/t-shirts
For Turkish speakers in Türkiye:
https://t-shirt.com/tisort
In this URL structure, Turkish is used as the default language in the root path. Now let us examine how the hreflang structure should be configured.
For the https://ornek.com/tisort page, the hreflang structure should be configured as follows:
<link rel="alternate" hreflang="tr-tr" href="https://ornek.com/tisort" />
<link rel="alternate" hreflang="en-gb" href="https://ornek.com/en-gb/t-shirts" />
<link rel="alternate" hreflang="en-us" href="https://ornek.com/en-us/t-shirts" />
<link rel="alternate" hreflang="de-de" href="https://ornek.com/de-de/t-shirts" />
<link rel="alternate" hreflang="en-de" href="https://ornek.com/en-de/t-shirts" />
<link rel="alternate" hreflang="tr-de" href="https://ornek.com/tr-de/t-shirts" />
<link rel="alternate" hreflang="x-default" href="https://ornek.com/tisort" />
The Most Common Hreflang Mistakes
Not using x-default: If the x-default value is not defined, the default language version that should be displayed for a language or region not specifically targeted may not be identified.
Not including all language versions in hreflang: Failing to include the corresponding versions of each page in the hreflang annotation may result in duplicate content issues.
How Should URLs Be Structured on a Multilingual Website?
As is widely known, URL structure is an area that should be optimised for SEO. In a multilingual structure, it is also highly important for grouping pages and content. There are several factors to consider when creating a URL structure for each language and region target. One of the main considerations is the order of language and country codes. Language codes should be specified in ISO 639-1 format. Another important point is that a country code cannot be used on its own. A language code alone is valid targeting, but country-only targeting is not accepted. Google does not automatically determine language targeting based solely on country codes.
According to Google’s guidelines, language and country codes should be used in the following order:
{language_code}-{country_code}
Examples of URL structures that support language targeting:
/en-us/ for English speakers in the United States
/en-gb/ for English speakers in the United Kingdom
/en-tr/ for English speakers in Türkiye
/de-au/ for German speakers in Austria
/de-be/ for German speakers in Belgium
/nl-be/ for Dutch speakers in Belgium
/fr-be/ for French speakers in Belgium
Examples of URL structures that do not correctly support language targeting:
To target French speakers in Belgium, /be/ should not be used because “be” used alone represents Belarusian according to ISO 639-1. In this example, the targeted language and region would be incorrect.
Methods for Creating URLs in a Multilingual Structure
Subdomain
A language structure can be implemented through subdomains in a multilingual website. For example, the relevant language may be configured on a subdomain such as:
en.ornek.com
Subfolder
A URL path can be created for each language. For example, the relevant language may be configured in a subdirectory such as:
ornek.com/en/
Country-Code Domain Extension
Using different domain extensions is another possible method for creating a multilingual structure. For example, the relevant language may be configured through domain extensions such as:
ornek1.en
ornek2.de
Parameter
Language options can also be grouped using parameters. For example, different languages may be configured with a structure such as:
ornek.com/?dil=tr
These parameters should not be blocked through the robots.txt file.
Among these four options, the most commonly preferred structures are subfolders and subdomains. My recommendation is to use a subfolder structure, as this may support domain authority through an increased number of ranking keywords and allow search engine bots to crawl language-version URLs more easily.
Let us examine an example for an e-commerce company that sells T-shirts and operates in the United States, the United Kingdom, Germany, and Türkiye.
The following example shows a URL structure based on subfolders.
URL Structure for Language and Region Targeting
| URL | Description |
|---|---|
| https://example.com/en/ | A URL indicating only that the language is English |
| https://example.com/en-us/ | A URL targeting the United States in English and displaying prices in US dollars |
| https://example.com/en-gb/ | A URL targeting the United Kingdom in English and displaying prices in British pounds |
| https://example.com/de-de/ | A URL targeting Germany in German and displaying prices in euros |
| https://example.com/ | The root path may target the default language or contain a landing page for language selection. In this example, Turkish is targeted in the root path. |
How Should an XML Sitemap Be Created for a Multilingual Structure?
Using a sitemap helps search engines crawl your website more efficiently. Through the sitemap, we inform Google about the pages we consider important on our website. With the increase in international websites, Google announced sitemap annotations supporting multilingual structures in December 2011.
To optimise XML sitemap files in UTF-8 format for a multilingual structure, a logic similar to hreflang implementation should be followed. The corresponding language versions of each page should be specified using <xhtml:link rel="alternate">. Let us examine a multilingual XML sitemap for the T-shirt sales example above:
<?xml version="1.0" encoding="UTF-8"?>
<urlset xmlns="http://www.sitemaps.org/schemas/sitemap/0.9">
<url>
<loc>https://t-shirt.com/tisort</loc>
<lastmod>2021-06-25</lastmod>
<xhtml:link rel="alternate" hreflang="tr-tr" href="https://t-shirt.com/tisort"></xhtml:link>
<xhtml:link rel="alternate" hreflang="en-gb" href="https://t-shirt.com/en-gb/t-shirts"></xhtml:link>
<xhtml:link rel="alternate" hreflang="en-us" href="https://t-shirt.com/en-us/t-shirts"></xhtml:link>
<xhtml:link rel="alternate" hreflang="de-de" href="https://t-shirt.com/de-de/t-shirts"></xhtml:link>
<xhtml:link rel="alternate" hreflang="en-de" href="https://t-shirt.com/en-de/t-shirts"></xhtml:link>
<xhtml:link rel="alternate" hreflang="tr-de" href="https://t-shirt.com/tr-de/t-shirts"></xhtml:link>
</url>
</urlset>
What Should Be Considered When Preparing an XML Sitemap for a Multilingual Structure?
When configuring a sitemap for a multilingual structure, unlike a standard sitemap, all alternative language versions of the corresponding URL inside each <url></url> element should be specified using <xhtml:link rel="alternate">. Apart from this requirement, the following points should also be considered when preparing the XML sitemap:
- Google expects the standard sitemap protocol to be used in all formats, and each sitemap should contain a maximum of 50,000 URLs.
- The sitemap should be created in XML format, such as
/sitemap.xml. If it is planned to be separated according to data categories, multiple sitemap files may be created. - The sitemap should not contain URLs returning 3xx, 4xx, or 5xx HTTP status codes. Only URLs returning a 200 Success status should be included.
- Pages containing a noindex meta tag should not be included in the sitemap.
- The XML sitemap file should be generated dynamically every day. When pages are deleted or removed from the website, they should also be removed from the sitemap.
- The file should be created in UTF-8 format.
We have examined in detail how websites with multilingual structures should be optimised and how SEO improvements can be implemented in accordance with Google’s guidelines. Websites that market content, products, or services globally can use these methods to target the correct languages and regions.
