multilingual domains both crawled and come up to the wrong audience

Question

I have a domain with multilingual support through subdomains. Each language version is identical besides the titles which are translated. e.g.:

domain.com
ja.domain.com
es.domain.com

I'm redirecting users to the right subdomain according to a IP country translation table. So, for example if you're accessing the site from Japan, I'm displaying the ja.domain.com version.

The problem is that web crawlers have crawled both domains, and sometimes show the wrong version in search results. For example if I'm from New York I might be displayed the ja.domain.com version instead of the domain.com version.

How can I avoid this?

Answer 1

There are a few problems here:

1. As w3d notes, you're effectively creating a duplicate content problem. That can be perfectly legitimate, but needs to be managed to avoid the obvious problems it introduces.

   However, I'd certainly say that if you think the titles need to be translated, then really all of the content does. Translating only the titles will confuse users and search engines (imagine clicking on what looks like an English result in Google only to land on a page written in Chinese. You'd probably bounce right back again, unless you read Chinese}.

2. Don't redirect users based on perceived language. See this previous question/answer here. Remember that Googlebot, Bingbot, etc. crawl from the US, so if you're serving content based on IP, they won't see all of your language versions.