It’s okay if the same URL appears on multiple sitemap files
It’s fine to have the same URL included in multiple sitemap files. The only caveat is ensuring that there is no conflicting information being provided across the different sitemaps. For example, having a URL in a ‘regular’ sitemap and an hreflang-specific sitemap (for different language versions of your site) is perfectly acceptable, as long as any hreflang annotations given to that page are consistent across both sitemaps.
Mixed-language pages can be confusing to Google
In general, Google tries to use the primary content of a page to determine which language a page is targeting. However, it’s also recommended to make title tags and headings match the page’s primary language. Having various elements on one page in different languages makes it hard for Google to know how the page should appear in the index.
You can ask Google not to translate your pages with a “notranslate” meta tag
Google’s translation feature aims to make your content accessible to a larger group of users, but there may be scenarios where you’d prefer pages not to be translated. It’s possible to prevent titles and snippets from being translated in search engine results pages with the “notranslate” meta tag. This signals to Google that a translation for that page is not required and will also stop users from being shown the ‘Translate this page’ prompt when they open a URL.
Noindexing pages with geo IP redirection is not ideal
One user asked about the use of geo IP redirection in conjunction with noindex tags. The example was having separate pages targeted at users in multiple locations, but using noindex tags to ensure just one is indexed.
John raised the point that Google typically crawls from one location (mostly using a Californian IP address). If the IP address directs Google to one of the URLs you have set to noindex, it might result in those pages not being indexed full stop. This approach, therefore, isn’t recommended. Instead, you should focus on making location-specific content easier to find once the user has landed on the site.
Language is evaluated on a per-page basis for SEO
Does an entire website need to be translated to rank well in an alternate language? John responded to a question about whether it would be ok to only translate some pages in a website rather than the entire site. He answered that language is looked at on a per-page basis, rather than evaluating whole parts of a website, so this approach would be fine. He recommended making sure that internal linking is in place to these translated pages so that they can be found.
It is not possible to specify which countries and regions content should rank in
There is no way to prevent Google indexing content in specific countries and regions, even if it’s not targeted to that audience. The example given was a user who wanted English pages to rank only in the US and the UK. If Google deems the content as relevant to users in other locations, there’s every chance it will be indexed there too (and nothing that webmasters can do to prevent this).
Try to Have the Fewest Internationalised Versions of a Site
Adding many internationalised versions of a website increases the complexity, so you’re better to have fewer versions if possible, especially if you have pages where you can rely on geo-targeting without hreflang. You can use Search Console and web analytics to see if people are landing on the wrong pages, and implement hreflang for those. For different language versions, hreflang can help Google to show the right version, but searches for long tail content are easily recognisable for which language version to show, so hreflang may not be necessary.
Hreflang May Not be Necessary For Translated Versions of Pages
Pages with different language content don’t necessarily need hreflang because the text is different so Google would treat them as unique pages and not as duplicates.
Google Tries to Detect User’s Language Preferences if Using Default Version of Chrome
Google is aware that sometimes users download the default version of Chrome in English rather than their preferred language, and will try to detect this and show the correct language versions of pages to help with localization.