Leadanta launched in 9 languages — English plus Spanish, Arabic (right-to-left), French, German, Brazilian Portuguese, Hindi, Kannada, Tamil, and Telugu. In the 10 days since the language pages went live, our English search impressions kept growing — Sevantay (one of our portfolio sites) is now at 71.1k Google impressions, up 40% in two weeks.
That outcome is rare. Most multilingual rollouts crater either the new languages, the original one, or both. Here's why — without giving away the specific playbook we charge for.
Multilingual is harder than it looks
Looks like 'translate the homepage'. Is actually 6+ hard decisions made in week 1 that you'll be living with for 2+ years. Get any one of them wrong and the cleanup is expensive.
- URL structure. Sub-directories, sub-domains, or separate domains? Each option distributes SEO authority differently. The wrong choice fragments rankings.
- Search-engine canonicalisation. Google has to be told which language version is which. Get the signal wrong and Google picks one and ignores the others.
- Content quality. Machine-translated pages without human review are flagged as low-quality and excluded from results. Pure auto-translation is worse than no translation.
- Right-to-left handling. Arabic and Hebrew aren't just text changes — they need full layout flips, mirrored navigation, and careful handling of mixed-direction content (English brand names inside RTL paragraphs).
- What to translate and what to defer. You usually can't afford to translate everything at launch. Choosing the right pages — and being honest about the rest — keeps trust intact.
- Local intent signals. Hreflang annotations, regional hosting, locale-aware OG metadata — small mechanical decisions, all of them invisible to the casual eye, all of them visible to Google.
How most agencies get this wrong
We see the same three failure modes when auditing competitor sites:
- Translation-plugin sites. Drop in a plugin, magic-translate everything, ship. Google reads the same machine-translated text every plugin produces and de-ranks the lot.
- Separate-domain sprawl. Buy ten country domains, install ten sites, watch every domain take 2 years to start ranking from scratch.
- Missing language signals. Languages exist but Google never figures out which version belongs to whom. The wrong one shows up in search results.
Our approach (without giving you the playbook)
We've made the URL-structure choice, the language-signal choice, the translation-quality-review process, and the RTL handling decisions across multiple sites. Each piece is small; the combination is what protects English rankings while opening real organic growth in other languages.
That combination is what you pay us for. It's not posted online — and there's a reason. The decisions interact: making one without the others is worse than not multilingual at all.
What the result looks like
On leadanta.com today: 9 languages live, Arabic rendering correctly in RTL, English search impressions on a +40% trajectory at Sevantay (our reference portfolio site), and a foundation that scales to more languages without re-architecting.
All on the same single domain. All with one canonical English version. All deliberate.