A bilingual website is standard practice for businesses in Tajikistan: customers search in both Russian and Tajik. But if you simply translate the pages and add a language switcher, search engines often show the wrong version or treat the pages as duplicates altogether. Let’s look at how to build a bilingual site so that both versions can be found on Google and Yandex.
A separate URL for each language version
The golden rule: every version of a page needs its own permanent URL. If the language is switched by a script or through cookies while the address stays the same, the search robot will only see one version — whichever opens by default.
Practical structure options:
- folders: site.tj/ for Russian and site.tj/tg/ for Tajik — the simplest and most common option;
- subdomains: tg.site.tj — these work, but need separate setup in webmaster tools;
- separate domains — only justified for different countries, not for two languages within one country.
hreflang markup: telling search engines which page is for whom
The hreflang attribute tells a search engine that two pages are translations of each other. A user with a Tajik-language interface is then more likely to be shown the Tajik version, and a Russian speaker the Russian one. The code for Tajik is tg, and for Russian it is ru.
What to check:
- hreflang links must be reciprocal: the Russian page points to the Tajik one and vice versa;
- each page also points to itself;
- the full address is used, not a relative one;
- an x-default option is added for users whose language matches neither version.
Multilingual WordPress plugins such as Polylang or WPML output this markup automatically — but once they are set up, it is still worth checking it in the page code.
Translation is not word-for-word substitution
Machine translation without editing reads unnaturally, and both people and search engines notice. There is also a less obvious problem: people phrase their searches differently in different languages. A Russian speaker will type “advertising agency Dushanbe”, while a Tajik speaker may search with entirely different words or mix languages.
That is why the keyword set — the list of queries a page is optimised for — has to be built separately for each language. Look at the suggestions in the search bar and the queries in Google Search Console and Yandex Webmaster, and ask your sales team how customers describe what they need.
Meta tags, URLs and images in both languages too
A common mistake: the page text has been translated, but the title tag, meta description, image captions and alt text are still in Russian. The search engine sees a mix of languages and has a harder time understanding who the page is for.
Check on every language version:
- title and description in the language of the page;
- H1–H3 headings;
- image alt text;
- buttons, forms and error messages;
- the lang attribute in the page code.
Don’t mix languages on one page
If half the blocks on a Tajik page are in Russian, users leave and search engines get a contradictory signal. It is better to hold off publishing a section in the second language than to show it half-translated. The exception is brand names and established terms that are normally written in Latin script anyway.
Track the results of each version separately
Google Search Console and Yandex Webmaster let you view statistics by site section. Compare impressions and clicks for the Russian and Tajik versions: if one of them gets almost no traffic, the problem is usually indexing, the hreflang markup, or a keyword set that was compiled as a formality.
A short checklist
- Each language version has its own URL.
- hreflang markup is reciprocal and includes x-default.
- Keywords are researched separately for each language.
- Meta tags, alt text and the interface are translated.
- One page, one language.
- Results are tracked for each version.
If your site already runs in two languages, start with an audit: it will show which pages search engines can’t see and where the versions conflict with each other.
Leave a comment