Technical SEO · Analysis

Local wording for visitors: what Google documents about locale-adaptive pages

Some content genuinely differs by country, such as an emergency number or a currency. I set out what Google documents about its crawler on a page that changes by visitor, and what it recommends instead.

Published
Updated
Last fact checked
Read
5 min
New to this?How to check what Google receives on a page that changes by visitorLearn it step by step: what it is, why it matters and how to do it on your own site. Beginner40 minutesOpen the beginner guide →

Key findings

  • Google says that if a page returns different content by the perceived country or language of the visitor, it might not crawl, index or rank all of that content, because its default crawler IP addresses appear to be in the USA and its requests carry no Accept-Language header.
  • Google recommends separate locale URLs annotated with hreflang, and says not to redirect people automatically by a guessed language, but to let them switch.
  • Google also crawls from IP addresses outside the USA, and says to treat Googlebot as you would any other visitor from the country it appears to come from.

Some details are only right for some readers: an emergency number, a spelling, a currency, a tax term. The tempting fix is one page that quietly changes for each visitor. Google documents what its crawler sees in that case, and the answer shapes how any such page should be built.

This article sets out what Google’s How Google Crawls Locale-Adaptive Pages and Managing Multi-Regional and Multilingual Sites say, and where they stop. Where I give my own reading, I say so. It was prompted by a recent build, ManFlu, a satirical guide whose wording and emergency numbers follow the visitor’s country. If you want the checking steps without the background, my beginner guide to checking what Google receives covers them.

What a locale-adaptive page is

Google defines locale-adaptive pages as pages where the site returns different content based on the perceived country or preferred language of the visitor. Its warning is that Google might not crawl, index or rank all of that content for different locales. It gives two reasons: the default IP addresses of its crawler appear to be based in the USA, and the crawler sends HTTP requests without setting Accept-Language in the request header.

My reading: a site that decides what to show from the IP address or that header will, by default, show Googlebot the version for a visitor in the USA with no language preference.

What Googlebot sees on a page that adapts

Switch between the two states and look at the arrow from Googlebot to the UK wording.

One URL that adapts

  • GooglebotAppears to be in the USADefault versionNo location or language cue

    Reachable Googlebot gets the default version. Google says its default crawler IP addresses appear to be in the USA and that its requests carry no Accept-Language header. A page that decides by those signals gives it the version for that case. How Google Crawls Locale-Adaptive Pages

  • UK visitorInvented exampleUK wordingInvented example

    Reachable The UK visitor gets UK wording. The visitor reaches the version made for their country, either because the page adapts or because they open the UK URL. Managing Multi-Regional and Multilingual Sites

  • GooglebotAppears to be in the USAUK wordingInvented example

    Might be missed Googlebot reaching the UK wording. Google says it might not crawl, index or rank all the content of an adaptive page for different locales. It recommends separate locale URLs annotated with hreflang, so each version has an address of its own. How Google Crawls Locale-Adaptive Pages

Separate URLs per locale

  • GooglebotAppears to be in the USADefault versionNo location or language cue

    Reachable Googlebot gets the default version. Google says its default crawler IP addresses appear to be in the USA and that its requests carry no Accept-Language header. A page that decides by those signals gives it the version for that case. How Google Crawls Locale-Adaptive Pages

  • UK visitorInvented exampleUK wordingInvented example

    Reachable The UK visitor gets UK wording. The visitor reaches the version made for their country, either because the page adapts or because they open the UK URL. Managing Multi-Regional and Multilingual Sites

  • GooglebotAppears to be in the USAUK wordingInvented example

    Reachable Googlebot reaching the UK wording. Google says it might not crawl, index or rank all the content of an adaptive page for different locales. It recommends separate locale URLs annotated with hreflang, so each version has an address of its own. How Google Crawls Locale-Adaptive Pages

The page and its two versions are invented for illustration. Google also crawls from IP addresses outside the USA, so a crawl from the USA is the default case, not the only one.

What Google recommends instead

The page carries a recommendation at the top: use separate locale URL configurations and annotate them with hreflang annotations. The multi-regional guidance says the same for language. It recommends different URLs for each version rather than cookies or browser settings to adjust the language. If you prefer to change content dynamically or reroute visitors by language settings, it says Google might not find and crawl all the variations, because the crawler usually originates from the USA and sends no Accept-Language header.

Two related instructions apply to the visitor’s side. Avoid automatically redirecting users from one language version to another based on what you think their language may be, because that could prevent users and search engines from seeing all versions. And consider links to other versions so people can choose. My article on hreflang return tags covers the annotation itself.

Googlebot from more than one country

The documentation adds that Googlebot also crawls with IP addresses based outside the USA, in addition to the US-based ones. Its rule is that when Googlebot appears to come from a certain country, treat it as you would any other user from that country. The example it gives: if you block USA-based users but allow visitors from Australia, your server should block Googlebot when it appears to come from the USA and allow it when it appears to come from Australia.

Googlebot uses the same user agent string in every configuration, and you can verify these crawls with reverse DNS lookups. My reading: do not build a rule that treats the crawler differently from the visitors it resembles, and do not assume the crawl in your logs always comes from one country.

Keep the rules the same in every locale

If a site uses the robots exclusion protocol, Google says to apply it consistently across locales: robots meta tags and the robots.txt file must specify the same rules in each locale. A variation that blocks one version and not another makes the version you see in a browser a poor guide to what is indexed.

A page whose wording changes: where the documentation stops

Google’s page is about content that differs by locale. It does not separate a whole different page from a few local details inside an otherwise identical page, such as the helpline number in a hypothetical page about looking after someone who is ill. It also does not say that local wording is cloaking, which its Spam Policies for Google Web Search define as presenting different content to users and search engines with the intent to manipulate search rankings and mislead users.

My reading, offered as a way to reason and not as Google’s wording: what the crawler receives is what can be indexed. So the version served with no location or language cue should be the one you want represented in search, and local details are best added on top of a complete default. Where the differences are large enough to change what the page is about, such as different products, laws or prices, separate URLs are the route Google documents.

There is one more warning. Google determines language from the visible content of the page. Translating only the boilerplate while the bulk of the content stays in one language, as often happens on pages with user-generated content, can create a bad experience when the same content appears several times in results with different boilerplate languages.

How to check what Google receives

  • Open the page in a private window with no location or language cues and note what appears for a visitor who tells the site nothing.
  • Use the URL Inspection tool to run a live test and view the tested page’s screenshot and page code.
  • Compare what you saw with what the tool shows. Differences point to where the page depends on visitor information.
  • Check that robots meta tags and robots.txt rules are the same in every version.
  • Check there is no automatic redirect by guessed language, and that visitors can choose.

Mistakes when testing a page that adapts by locale

  • Testing from your own browser only. Your location and language settings decide what you see.
  • Assuming Google will see every variation. The documentation says it might not.
  • Changing only the surrounding text of a page while the main content stays in one language.
  • Applying different robots rules to different versions.

Where this fits at Cultured Digital: international and rendering

My international guide covers sites that serve several countries or languages, and the rendering guide covers what a crawler receives compared with what a visitor sees. See also my article on country domains, subdomains and subfolders.

Further reading on locale-adaptive pages

Content that changes by country and Google may not see it?

Tell Dean which parts of the page change and how the site decides. The first answer is what a crawler would receive.