Skip to content
TranslationAtlas
All articles

Multilingual SEO Guide: Ranking Translated Pages

Published July 5, 2026 · 8 min read

A translated page and a page that ranks in a new language are two different outcomes. Multilingual SEO is the layer between them: the technical signals and keyword work that tell Google which page belongs to which market, and which terms that market actually searches. Get it wrong and translated content sits invisible; get it right and it competes on equal footing with local-language sites.

Localize keywords, don't translate them

The literal translation of your English head term is often not what people type into Google in the target language. "Cheap flights" translated word-for-word into French doesn't match how French speakers actually search for the same intent, since search volume and phrasing both shift. Before translating a page, run keyword research natively in the target language and market, then write (or translate) the page around what people actually search, not around your source keyword list.

This step alone is often the difference between a translated page that ranks and one that doesn't, regardless of how well the prose reads.

There's a related trap worth naming directly. A translator working purely from the English source, with no keyword brief, will naturally produce grammatically excellent target-language copy built around the wrong terms. This isn't a translation-quality problem; the sentences are correct. It's a briefing problem. Keyword research needs to happen before translation starts, and the resulting term list needs to be handed to the translator alongside the source text, the same way you'd hand over a glossary of brand terms.

Duplicate content and canonical tags across languages

Multilingual sites sometimes trigger duplicate-content flags they didn't earn. Search engines occasionally treat near-identical page structures across languages as more similar than they are, especially early in a new market rollout before enough unique signals accumulate. Canonical tags should point each translated page to itself, never back to the English original; a canonical pointing at the source language tells search engines the translated version isn't the primary page, which defeats the purpose of translating it. This is a separate mechanism from hreflang. hreflang says "here are the alternates"; canonical says "this is the authoritative version of this specific page." Mixing the two up is a common setup mistake.

hreflang: the tag that tells Google which page is which

hreflang annotations tell search engines which language and region a page targets, and which alternate versions exist for other markets. Get it wrong and you risk two translated pages competing against each other instead of each ranking in its own market.

Rules that matter in practice:

  • Every page needs to reference itself in its own hreflang set, along with every alternate.
  • Include x-default for the page shown to users whose language/region doesn't match any variant.
  • hreflang must be reciprocal. If your English page links to the French version, the French version must link back to the English one, or the annotation is ignored.
  • Use consistent locale codes (fr-FR vs fr-CA if regional content actually differs; plain fr if it doesn't).
  • Implement via sitemap or <link> tags, not both inconsistently. Pick one method and audit it after every deploy.

hreflang errors are one of the most common causes of multilingual pages ranking poorly even when the translation itself is solid.

Translate the metadata, not just the body

Title tags, meta descriptions, Open Graph tags, and image alt text all get indexed and shown in search results, so a page with translated body copy but an English <title> looks broken to both users and crawlers. Metadata translation needs the same keyword research as the body: a literal translation of your English meta description won't necessarily include the phrase people search for in that language.

URL structure

Pick one pattern and stay consistent:

  • Subdirectories (/fr/, /de/): the pragmatic default for most sites, keeping domain authority consolidated.
  • Subdomains (fr.example.com): useful if regional teams need separate infrastructure, but splits authority more than subdirectories.
  • ccTLDs (example.fr): the strongest local-market signal, but the most expensive to maintain and the hardest to switch away from later.

Whichever you choose, don't mix patterns across languages on the same site. It complicates hreflang and confuses both crawlers and analytics.

Internal linking between language versions

A language switcher in the header helps users but does little for search engines on its own. Internal links between the language versions of related content, a translated blog post linking to the translated version of a related article rather than the English one, help crawlers discover and associate the alternate pages faster than hreflang alone. This matters most right after launch, when a new-language section of the site has few external backlinks and relies more heavily on internal link equity to get indexed and crawled regularly.

Measuring whether it's working

Track rankings and organic traffic per language segment separately rather than as one aggregate number. A strong English site can mask a struggling French section if you're only looking at total organic sessions. Google Search Console lets you filter by page path or hreflang-tagged URL groups, which is usually enough to spot a market that's underperforming before it's been live for months. Watch specifically for pages indexed in the wrong language (a sign of hreflang or canonical errors), unexpectedly low impressions for a translated page relative to its English equivalent's history, and a low click-through rate despite a decent rank position, which often means the meta description isn't matching what that market searches for.

Content depth still matters, per language

A thin, direct translation of a thin English page will rank like a thin page. Multilingual SEO doesn't lower the bar on content quality; it just adds a translation and technical layer on top of the same fundamentals: depth, relevance, and matching search intent for that specific market.

Where this pays off fastest

If you're prioritizing, start with your highest-traffic pages and the markets with the clearest commercial intent, translate metadata alongside body copy from day one, and verify hreflang before you index. Cleaning it up after a page has already been crawled wrong takes longer to recover from than doing it right the first time.

We translate with native-market keyword awareness rather than literal rendering, and can review your existing hreflang setup alongside the translation work through our localization service. Request a quote for your translated pages and tell us which markets you're targeting first.