Earlier this week Google’s Webmaster Central blog wrote about how to effectively geotarget your website. In that article they promised to discuss multilingual websites as well.
And now they have. In a new Webmaster Central blog post published today they discussed the best practices for multilingual websites. Some key take-aways are:
One page, one language
Don’t put two different languages side by side on a single page. Make sure the content on a page is in only one language.
There’s no special need for language-specific URLs from a purely SEO point of view, as Google apparently detects the language straight from the content itself. Though having language-specific URLs is a good usability practice and helps make management of your site a lot less complicated.
Language indicators in the code
Interestingly enough Google says it ignores all language indicators in the code, including the ‘lang’-attribute and doctype declarations.
When using non-English characters in URLs, make sure your document encoding is set to UTF-8. That way URLs containing special non-English characters (such as á, é, ö, etc) are properly escaped to use the encoding of that character.
Full crawling & indexing
Make sure your properly translated content can be fully crawled and indexed by Google. Cross-link your pages in different languages – so for example on your English language page you should link to the German and French versions.
If you can’t afford proper translations and publish content that’s machine-translated, don’t let Google index it. Apparently Google can see automatically translated content as spam due to its low quality.
There’s a couple of other tips in Google’s blog post, so I recommend you read it in detail: Working with multilingual websites.