A company translates 40 well-ranking English pages into French. The translation is accurate, the tone is appropriate, and the client signs off on it. Six months later, the French pages are attracting almost no organic traffic, because the company did not treat the job as SEO translation.
The explanation of why this happens is usually the same: A search engine result page is a market, and different markets do not always share vocabulary. The French version of your best page can be linguistically flawless and still use phrases nobody searches for.
This article covers where faithful translation and search visibility diverge, which page elements should be rewritten rather than translated, and the technical aspects that can quietly cancel out good linguistic work.
How SEO Translation Differs from Regular Translation
A translator’s mandate is to maintain fidelity to a source text. An SEO writer’s mandate is to preserve alignment with demand. Those two goals coincide often enough that any incompatibility stays invisible until traffic data arrives.
Consider a page ranking for “moving company.” The faithful French rendering is entreprise de déménagement; the phrase Quebeckers actually search up, however, is most frequently déménageur. Both are correct in French, but only one matches the query.
If this happens with every heading, every internal link anchor and every metadata field over 40 pages, you will have a site that reads well but competes for nothing.
The reverse is also possible: A team that chases hits online without a linguist ends up with headings stuffed with terms nobody would actually use, which damages credibility in a market where language is of great importance. The goal is to create a page that satisfies queries and still reads like it was written by a person.
Keyword Research Does Not Survive Translation
This is the root cause behind most SEO translation projects that result in disappointment, and it has two distinct components.
Search Volume Does Not Transfer Between Languages
It’s not guaranteed that an English keyword that drives your traffic has a French counterpart. Sometimes the French equivalent produces a fraction of the volume, sometimes it produces more. Other times still, a concept fragments into three separate queries that each deserve their own page, or two English pages collapse into one French query and start competing with each other.
None of that can be deduced from the source text; it only becomes obvious when someone performs keyword research in the target language, in the target market, before translation begins.
Search Intent Shifts Between Markets
A given concept can carry different intents across languages. A search term that returns comparison articles in English, for example, may return service provider information in French, or vice versa, which changes what the page needs to contain.
A translated informational article targeting a query that now returns transactional pages will underperform regardless of quality, because it does not answer the questions asked by searchers.
What Needs Rewriting Instead of Translating
Faithful translation is suitable for some elements, but others need to be rebuilt around target-language demand, and this is where SEO translation comes in handy.
| Element | Literal translation | Localization |
|---|---|---|
| Meta title | Correct wording, wrong keyword, often over the character limit | Target keyword front-loaded, length verified after translation |
| Meta description | Accurate summary that reads flat | Action verb, target phrasing, click-oriented |
| H1 and H2 | Mirrored English structure | Headings built on real target-language queries |
| URL slug | Translated word for word, or left in English | Short, keyword-bearing, accent-free and hyphenated |
| Internal link anchors | Literal anchor text | Anchors matching target-language search terms |
| Image ALT text | Skipped entirely | Written in the language of the page it appears on |
| FAQ questions | Translated questions | Questions phrased as people search them locally |
Two mechanical points deserve attention. Firstly, French almost always takes up more space than English, often by a fifth or more, so a meta title translated at the same word count typically exceeds the display limit and needs to be recomposed rather than trimmed. Furthermore, accented characters in URL slugs create encoding problems, so slugs are written unaccented even when the language requires accents.
Planning a multilingual content launch? Request a scoped estimate before the translation stage begins.
Technical Aspects That Can Undo Good Linguistic Work
Perfect copy on a badly configured page will still fail to be seen. Google’s own documentation is explicit about the setup requirements, which are less flexible than most teams might assume.
Google’s guidance on multiregional and multilingual sites recommends different URLs for each language version rather than adjusting content through cookies or browser settings. The companion guide on localized versions explains how to establish the relationship between those URLs with “hreflang” annotations, which search engines use to identify a document’s language and geographic data, and confirms that localized pages count as duplicates only when the main content stays untranslated.
These implementation failures occur consistently across projects:
- Missing self-referencing hreflang tags, since every variant must list all variants including itself
- Hreflang pointing to URLs that redirect, return errors or are blocked from indexing
- Language codes that are wrong or invented, such as using a country code where a language code should be used instead
- Canonical tags, which identify a page’s main version, on translated pages pointing back to the source-language version, which suppresses the translation
- Automatic IP-based redirection that prevents the crawler from reaching a language version at all
- No x-default declared for users whose language the site does not support
Any one of these situations can cause a full translation budget to go to waste, and none of them will be caught during a linguistic review of the text.
It’s worth noting for Canadian sites specifically that a French Canadian page and a French European page are separate hreflang targets, and declaring only a generic French variant leaves the search engine to guess which audience a page should serve. Organizations selling into both markets need to specify the regional codes, not just the language code.
Building a Workflow That Protects Rankings
In SEO translation, the sequence of tasks matters about as much as the tools.
Map Keywords Before Translation Begins
Run target-language keyword research first and produce a mapping document containing the source URL, target URL, primary keyword, secondary keywords and target intent. The translator then writes with a defined target in mind instead of mirroring the source, and metadata gets composed rather than converted.
This approach is closer to adaptation than to translation; for further reading, the distinction between faithful rendering and rebuilding for effect is set out in this comparison of transcreation and translation. Web-specific vocabulary carries its own difficulties, which are examined in this look at translating web expressions without losing their impact.
Check the Rendered Page Rather Than the File
Metadata can get truncated in the content management system, or CMS, causing buttons to overflow at tablet width, hreflang tags to fail to render, and ALT text to stay in the source language. None of this data is visible in the delivered document, which is why a final pass in the publishing environment is a necessary part of the work rather than an optional extra.
That verification step is the purpose of a web quality assurance review, and it goes hand in hand with web and digital translation built for content that lives in a CMS rather than a document.
Turn SEO Translation Into a Ranking Asset
A translated site that nobody can find is a cost with no return. Fixing this issue does not mean better translators or better SEO consultants working separately; rather, it involves sequencing the translation and SEO work so that keyword research precedes writing and that technical verification follows publication.
Do the research first, brief the linguists with real target-language data, rebuild metadata rather than converting it, and verify the live page. Sequenced that way, SEO translation stops being a cost sink and starts to generate a return on your investment. The pages you publish should be findable in the language they were written for.
To discuss your multilingual pages and get an assessment of what should be adapted rather than translated, start the conversation here.
FAQ
What is the difference between SEO translation and regular translation?
Regular translation prioritizes fidelity to the source text, whereas SEO translation prioritizes alignment with search demand in the target market, which means keyword research happens before writing, and page elements like titles, headings and metadata are composed rather than converted. The result reads naturally in the target language while using phrases people actually search for, not the literal target-language equivalent of the source-language keyword.
Does translating my website automatically improve my international rankings?
No. Translation alone rarely produces rankings on its own. Pages also need target-language keyword research, rebuilt metadata, and correct technical configurations including separate URLs for each language and valid hreflang annotations. Google treats localized pages as duplicates only when the main content stays untranslated, but a flawlessly translated page targeting the wrong query still competes for search traffic that does not exist.
Should URL slugs be translated for multilingual SEO?
Generally, yes, because a slug carrying the target-language keyword reinforces relevance and reads correctly to users who see it. Write slugs without accents or special characters to avoid encoding problems, keep them short and hyphenated, and settle on them before publication. Changing a slug afterwards requires a redirect, so choosing your slug strategy during keyword mapping saves avoidable technical cleanup later on.
