Forum Moderators: Robert Charlton & goodroi
The current situation:
The site has multiple language options.
The default home page for the domain root has a javasript in its HEAD that detects locations based on the IP addresses coming in with the requests, and redirects to the appropiate sub-page. There's no other difference in information than the language it's presented in.
I'd like to hear your opinions on this. Especially if you have dealt with anything of similar kind in the past, for I have many concerns about the current method. And while I can pretty much GUESS what would happen if the site went live like this, I'm not sure if saying "told you so" afterwards would made me feel any better.
- How would this method serve the pages to googlebot? Would it default out to English? ( IP from the US )
- With the detection and the redirect in the HEAD of the home page, will it even get indexed?
- I assume sooner or later it'll be as good as non-existent, with only the pages it redirects to being in the index.
- What would be the implications on pagerank if it's not indexed?
- Wait a minute, what would be the implications on indexing, in general?
- Language selection links are on every page, thus all of the sub-pages would be indexed at their proper URLs. However...
- Would the redirect get them indexed through the homepage URL as well? That's duplicate content right there, which would put the site down before it's even indexed for the first time.
- Would this javascript method be considered as something evil?
This entire issue makes me want to say "have it default to english, period." and leave it at that. But that's not quite what "built for users" means I'm afraid.
So, anyone?
Sorry for the blunt all-out request, but this is vital for the site, and I'm not experienced in multiple language sites nor geo-targeting.
- How would this method serve the pages to googlebot? Would it default out to English? ( IP from the US )
I've asked the guys and they said that the english version doesn't redirect, for that is the home page itself. So, if the IP is from the US/UK/Australia/...etc, the root level query loads the default index for the domain, and that's it.
However this would only solve the problem if they knew for sure, that googlebot will always arrive from an IP address in the US. Or UK. And if this is allright with Google in the first place, for technically speaking, this would be cloaking(?)
( Are there any local googlebots operating from non-US IPs at all? )
So, in your opinion is there/is this a way to stay geo-targeted with the home page and not angering Google at the same time? Or there's a reason why I didn't find any relevant examples on the net? The stance is, if we can't get it done properly, we won't do it at all.
Seems like a lot of jumping through hoops when all you really need to do is focus on Content Negotiation. Seems your site is a perfect fit. You can still attempt some form of geo-ip targeting, but you may even decide against it. It all depends on what you want your home page to do, how the site gets indexed and how your end user ends up getting to your site. If they are searching in their own language, they should end up in your language-specific document upon landing. Of course, you would may still want to offer your reader the ability to switch languages.
Yes, content negotiation was one of the options we considered.
The whole idea was introduced pretty late to the development of the web site.
All versions of all pages are on separate URLs exactly because we want them to be indexed properly. Basically it's like six sites in one, with the internal navigation evening out the weight of each version of each page to the same level. People will most likely be arriving from organic search, querying things in their own language, and AdWords ads which can be targeted towards language-specific keywords pretty well. The site has the language selection links pretty high in its hierarchy, I'd assume that they will be the tier 2 pages, direclty below the home page. There's no problem with this part.
The entire deal comes down to the home page only, the default page for the domain. Which we want to default to english, except if let's say, the browsers or the IPs would suggest a user for whom a different language ( one of the tier 2 pages ) would be more fitting.
So we'd need content negotiation/redirects in place only at the domain root index.
The problem is, how would google see this.
Would it see only the english version ( if so, there's no problem at all ) or will it sometimes be served one of the subpages' content? Which could lead to googlebot thinking THAT is the home page content, and index it as so ( after all it requested www.example.com and got the data ).
This would be a problem.
Duplicate content to say the least. ( For the tier 2 subpage for the given language is still available with 100% identical content at its original URL )
Do you happen to know whether googlebot would arrive with a setting for language preference? Or would it come without any idea for what content to negotiate for, and thus be served the default, English version, every time.
Do you happen to know whether googlebot would arrive with a setting for language preference? Or would it come without any idea for what content to negotiate for, and thus be served the default, English version, every time.
If there is no language setting or if the language setting is not one you have setup in your variant files then fallback to the LanguagePriority [httpd.apache.org] (Apache >= 2.0).
If you are worried about it missing the language variations and indexing them then you just set that up in your sitemap.
