How to set up the language versions of your site: preparation, creation, and experience

With an integrated approach to website development, the question of language versions is considered at the planning stage. You should point out which set of languages ​​on the site best suits your project: its positioning, georeferencing, audience requests.
If you have decided to create a multilingual site and the process is at the very start, great, you have the opportunity to do everything right. This is the best option that allows you to correctly build the site structure, crosslinking, and make technical settings. In addition, by preparing content and meta tags in advance for each array of pages, we avoid the situation when any of the versions contains empty template pages or even worse, simply duplicated content in another language.

Let us consider each aspect that plays a role in the implementation of a multilingual site.

Choice of implementation method: subdomains or directories

In the site structure, arrays of pages in different languages ​​can be arranged in 2 basic ways:

• Each language has a separate subdomain, such as en.site.com, ua.site.com, etc.
• The base language version is located in the main directory of the domain, and each additional language has a separate subdirectory in the root of the site, for example site.com/en/, site.com/pl/, etc.

On the one hand, subdomains provide more marketing advantages for positioning and promotion as they allow you to build a localized site for its target audience and more autonomously promote it through advertising and SMM. This option is suitable if the language versions are intended for different countries and the business is multi-regional. A more radical version of this approach is separate sites (domains) for the different countries/regions, but it is beyond the scope of this analysis.

On the other hand, multilingual subdirectories are technically easier to implement and provide a noticeable advantage for language SEO, as the site’s integrity is complete from the point of view of search engines. Such a site is a larger resource with a branched structure and is filled with content adapted to the needs of specific users. In this scenario, your domain surely wins. This approach also allows you to somehow prioritize the content of the main language by placing it in the base directory of the domain.
We consider the variant of multilanguage through language directories as a basis.

Linking logic between language versions

Obviously, a multilingual site should have a language switcher in the header. A common annoying mistake is when choosing a different language version does not update the current page, but opens the main page of the site in the selected language. Of course, the language version of the page on which the switch takes place should open. The links to other languages ​​in the switcher should lead to the other versions of the same page (most often, this is the only place where links to pages in other languages ​​are located in the site interface).
To summarize, the following is important:

• The language switcher contains links to versions of the current page.
• In the logo in the header, the main page of the current language version is linked, and not the root main page. Of course, the same applies to logos in any through blocks.
• Page menus in the header, links in the footer, links in link blocks (best products, latest news, etc.) lead to pages in the current language version.
• When we link to site pages in news text, product descriptions, and similar content, the links must be in the same language version as the referring material, unless other is a conscious action.

You need to keep an eye on such details because both the clear structure of the site and the user experience are depending on them.

User experience on the site

A good website always focuses on user experience. Of course, the texts must be written or translated and edited by a real person, then the content is appreciated by users and search engines.
In addition, when preparing content, it is important to consider that all alerts, forms, etc. on the pages are given in the language in which the site is displayed to the user.

Temporary close from indexing

So, you have decided on the structure of a multilingual site, prepared the content for future pages, and are ready to give your developers the technical requirements for implementation. If the situation is that a new language version is added to an existing one, it is important to close the subdirectory and nested pages of the new language from indexing until the structure is fully implemented and all the basic content is filled in. There is an important nuance here:
We close the raw version from crawling and indexing only through disallow in robots.txt at the directory level, while DO NOT add noindex tags to the page code. We also do not add this array of pages to the sitemap.
This approach will allow search engines better understand our intentions and quickly index ready-made pages after the further opening. If you close all pages only through noindex in the code, and then remove these tags, then each page will recover for a long time, because robots will likely have time to scan them and store the status. In turn, closing through robots.txt simply reduces the chances of an array of pages getting into the index before it is justified.

Correct markup of page URLs in different languages

It is a well-known fact that duplicate content and repeating the same semantics on pages worsen their positions and do not respond to the needs of each user in the best way. It is important to perform some technical optimization for search engines to understand that language versions contain content in a different language, and not just duplicate the same semantics again:

• The correct markup with alternate hreflang tags must be configured in the code of each page in all languages, which will indicate the language code in the ISO 639-1 standard for the current page (for example, hreflang=”uk”) and in the same way for its variations in other languages. If it is necessary to emphasize that the page is a priority for users from a particular country, then the country code should be used along with the language code (for example, hreflang=”en-US” for English in the USA). For unspecified cases and the default, the page variant is specified with hreflang=”x-default”.
  > Sample code
• The list of URLs in the sitemap.xml file should be organized according to a similar logic for each page in each version, so you need to write the markup of language tags, which will contain a list of alternate hreflang.
  > Sample code

Both of these settings have to be automatically applied to newly created pages in the future.

Opening a language version for indexing

When all the language versions of the site are ready for indexing (content, page meta tags, relinking, technical settings – everything is checked and correct), you need to adjust robots.txt again so that it no longer prohibits scanning arrays of language version pages.
Make these edits, then check that all the desired pages are in the sitemap.xml file and add (update) this file in the Google Search Console panel.
Bravo, your site now has one more language in its set! Crawling and indexing of new pages can take from 2 weeks to 6 months, depending on the size of the site and the complexity of its structure.

Leave the rest to search engines!

Finally, if your marketing strategy relies on at least 20% organic traffic, never redirect users based on their geographic location or system language! The search engine itself will select the optimal page for the user if the technical settings are correct and the content matches the request.
The technology of redirects based on geolocation and language does not work correctly, causing inconveniences and delays in page loading. Moreover, this mechanism is explicitly not approved by Google.

Note: Ukrainian legislation specifies that the Ukrainian version of the site must be the main one for Ukrainian users. Taking into account the technical matters and the letter of the law, it can be considered a sufficient balance that the site content in Ukrainian is not inferior to others in terms of completeness and is open for indexing in order to be a priority in the results for Ukrainian-language search queries. Ideally, it is also recommended to place the Ukrainian version at the root level of the domain, and all others in subdirectories.