Multilingual SEO: hreflang, Yandex and Localisation

Multilingual sites are my daily work: I manage page structures serving different countries, in different languages, on different search engines. One clear truth has distilled from that experience: in multilingual SEO, mistakes are silent. The site works, nobody complains — but the wrong page shows in the wrong country and conversion quietly evaporates. This guide exists to prevent those silent failures up front.

1. URL architecture: the first and most permanent decision

There are three common options — and this is the most expensive decision to change later:

Structure	Example	Pro	Con
Subdirectory	site.com/ru/	Single domain authority, easy management	Weaker regional signal
Country TLD (ccTLD)	site.ru	Strongest local signal and user trust	Each domain builds authority from zero; costly
Subdomain	ru.site.com	Easy technical separation	Authority sharing unclear; usually a halfway house

My practical advice: with limited resources, start with subdirectories; when a market becomes strategic, consider a ccTLD for it. If you mix approaches, hreflang discipline becomes even more critical.

2. hreflang: a small tag with big consequences

The rules of hreflang are simple but unforgiving:

• Reciprocity: if the TR page points to the EN version, the EN page must point back. One-way tags are ignored.
• Self-reference: every page must also list its own language version.
• x-default: declare the fallback version for users matching no language.
• Correct codes: language (tr, en, ru) plus optional region (en-GB); invented codes are the classic trap — "uk" means Ukrainian, not the United Kingdom.
• Canonical harmony: hreflang must not contradict canonical — if canonical points elsewhere while hreflang lists the page, the signals strangle each other.

You can implement via HTML <link> tags or the sitemap; for large sites the sitemap is easier to manage. What matters is generating it automatically, from a single source, consistently — hand-maintained hreflang breaks sooner or later.

3. Yandex: the separate reality of Russian-speaking markets

Everyone expanding into Russian-speaking markets learns this: strength on Google does not guarantee visibility on Yandex. The differences I watch:

• Yandex.Webmaster is set up separately: verification, sitemap submission and index monitoring are independent of Search Console.
• The tempo differs: new content enters the index and rankings settle at a different rhythm than Google; it takes patience and regular monitoring.
• Language quality decides: Russian content that smells of machine translation performs notably poorly on Yandex; a native Russian editor makes a felt difference.
• Regionality matters: Yandex is sensitive to region signals (target country/city, local content); make your target market explicit in the tool and in the content.

4. Localisation: begins where translation ends

Content that ranks is written for a market, not translated into it. My checklist:

• Don't translate keywords — re-research them: the same service is searched with different queries in every language; keyword research in the target language is mandatory.
• Local trust marks: currency, contact habits, the certificates and references that market actually recognises.
• Cultural tone: some markets expect formal, titled language; others expect warm directness — one tone does not fit all.
• Local visuals and examples: Russian text with English screenshots erodes trust.

Multilingual SEO in one sentence: technical correctness enters you in the race; localisation quality wins it.

Conclusion

Multilingual SEO is not set up once — it is a system kept alive with discipline: the right URL architecture, automatic and reciprocal hreflang, market-by-market webmaster monitoring (Google + Yandex) and localisation that goes beyond translation. A site with these four legs opens each new language not at increasing cost, but on accumulated authority.