how does google know if a page is from (enter nation here)?

[NewKidOnTheBlock]

I was just trying to explain to somebody who was wondering whether a certain phrase was said like A or like B in French.

So I did what I always do when wondering something like this: I went to google.fr, chose pages from France and saw that B got 1,000s of results whereas there were only 3 results for A. Thus B is the more popular way to say it.

The reason why I like to search for pages from France as opposed to pages in French is that if a page is from France it's more likely to be created by a French-speaker (a French person) than a page in French is (not that much of a difference, but teachers or professors or simply people interested in the language might have pages in French, too)....thus it'd be a better indicator.

BUUUUUUUT,

how does the big G decide whether a page is from France? if it is hosted in France? Or do they have a better way to identify whether a page is from a certain country (when showing them as search results from a certain country)?

(If one did the same thing in order to find out how to say something in "American English" (as opposed to British English for example), I guess if they use hosting, it might not be such a good way with the amount of people hosting in the US)

[jimbeetle]

Well, usual suppositions are ccTLD (.fr as opposed to .co.uk), server (IP) location, lang attriibute (rarely used, but say lang=en-us as opposed to en-gb), and why not throw in registration info (maybe a London-registered address more likely to be GB-oriented than a New York City address).

Again, all supposition, but ccTLD and server location appear to be the main factors.

[NewKidOnTheBlock]

thanks, I guess google is probably smart enough to have an algorithm for that that does something like computing the chances of a site being from a certain country (by using indicators like that) instead of a single factor (which I guess I kind of thought when asking this question!lol)

[TryUsOut]

That was a nice explaination jim. I was also curious at what other factors Google looks at. You just educated me a bit more on that subject.

[NewKidOnTheBlock]

@tryusout: some (many) people believe what sites the links pointing to you are from play a role, too (so basically the factors above would be considered for all those sites to find out their possible location and then g would consider from how many pages that are likely in a certain country the links pointing to your site are from)

[jimbeetle]

Good catch NewKid, that factor slipped my mind the other day.

[beu]

Yeah, I agree and after working with lots of Canadian sites in French and English I can say the GEO of the host is a big factor.

Also try Google's new whois feature, click on "Whois record for..." and then scroll down until you see "IP Location" under "Server Data".

IE
http://www.google.com/search?q=whois+google.com

You might try setting a regional GEO preference in Google Webmaster Tools when applicable and where available.

[MattUK]

I have also noticed that the geographical location of the domain registrar can have an impact. The order of importnace seems to be (in my experience)

1. Local TLD
2. Location of host
3. Location of domain regsitrar
4. Location of incoming links
5. langauge, addesses etc etc

[beu]

There was a good post a few days ago that in some ways also relates to this issue:
http://googlewebmastercentral.blogspot.com/2008/06/how-google-defines-ip-delivery.html

;)

[MattUK]

Al so I forget to mention that in Webmaster Tools you can also now set a location for a site. I'm not sure if this overrides any of the other factors though.