A generic home page to SEs and fine-tune when the punter (you) arrives?

@dstiles so why in that case was it cached for that specific city - Cabor, AR? If they do serve up a generic page for the SEs then that should be what's cached?

Actually, I guess my specific question is this: are the IP addresses googlebot uses from determined set locations or can it be from anywhere? I.e. if a site is using a visitor's (and spider's) IP to determine their location and then serves up that city page, will Googlebot always see the same city(ies) on the homepage or can it be from any random location?

I know you can sometimes trace it to mountainview, CA but are googlebot's IP's restricted to certain locations?

Sorry, didn't notice the cache bit. Does seem suspicious. Either google has an arrangement with them OR they give out several versions of the page? I'd expect the former as otherwise the company might get hit for spamming? It would make sense for google to chat to (inter-)national "local" directories in connection with their Places, I would have thought.

In a recent report I saw, was it google who were looking into taking over yellow pages? Somebody. Can't recall.

Google is increasingly delivering IP-localized results, but they've been doing this for a while with businesses as well as with directories. Directories also modify their pages for local delivery.

The answer to your question depends on whether its Google or the directory that's doing the localization.

I'm not seeing cached results including localization where there's an IP-based content change served up by the directory. I am seeing different pages served up by Google, though, depending on location, and those of course carry caches that correspond to the localized content.

With directory listings that are localized, you can generally tell if the cache is going to be localized by whether the Google listing itself shows signs of localization. If you can see the city name in the snippet, or a different subdomain or page is offered, then Google is doing the localization by serving different pages.

Eg, from where I am, a search for [yellow pages] returns both a main page whose listing isn't localized by Google, and then additional local listings that Google also serves up. On the main yp.com, there is a YP localized display section (probably an iframe in the middle of the page) that does appear when the page is served. In the cache, this area is blank.

Additionally, there are localized listing pages, chosen by Google, for yp.com after the main listing... and those displayed local pages and those caches are of course localized. Yelp works pretty much the same way, with the main site listed first, and then localized internal pages following.

Some of these results are more granular than others, btw. In large metro areas, some domains show only the core metro area in Google, though I know they go more granular than that on the site. This is true of yelp, eg. But yp.com gets very localized treatment where I am, showing practically a local neighborhood subpage.

Since googlebot can be identified by user agent, Google isn't making its initial cache indexing choices on the basis of where googlebot comes from, but it is serving up the appropriate page on the basis of IP, and the appropriate matching cache goes with that page.

I looked at some pages in google cash for yellowpages.com, using timeline feature, and have found that googlebot had different locations, not only Cabor, AR, but Los Angeles, CA too, for example.
So, googlebot was redirected by ip. This is not good practice from yellowpages’ side, and may bring troubles for search engine. Google can not guarantee that user will get the same version of document as it was presented in SERP.
Now Google goes forward for more and more localization. And, I think, Google wants to decide on HIS side where, and which document, local user should be sended. Not on the side of webmaster. And new instruments, appeared in webmaster arsenal, such as rel=alternative tag, could lead to the fact, that redirect by IP on webmaster’s side will leave in the past.

Thanks for the replies.

@Robert Charlton - not sure I follow, I'm talking about going straight to yp by typing in the domain URL in the address bar. That will serve up a localized page based on IP, and if you check the cached version of the homepage it shows a localized page (when I checked it was Cabot, AR)

Google isn't making its initial cache indexing choices on the basis of where googlebot comes from, but it is serving up the appropriate page on the basis of IP, and the appropriate matching cache goes with that page.

What do you mean by "where googlebot comes from" if you mean something other than IP, do you mean the referrer? I agree that YP is serving up the appropriate page on the basis of IP, so how will googlebot see it? Does googlebot access the site from a limited range of IPs or can it be from anywhere (and how often does it change)? Local.com has their homepage cached as localized to Mountainview, CA which is where I thought Googlebot's IP was always localized to, but clearly that's not the case. What I'm trying to figure out here is if Googlebot will see different (localized) homepages each time it visits, or if it will see a static homepage for any length of time (i.e. if Googlebot uses the same IP for 3 months at a time it should see the homepage as localized for that IP for 3 months)? Why do homepages of different local directories show a cache localized to different cities - is it because Googlebot can access sites with IPs from random places?

Sorry if you answered it in the before post but I wasn't quite following...I'm talking only about the homepage of YP, and how it appears to Googlebot at different times

@cls_wired How are you doing the timeline thing? I'm seeing this for a lot of the big directories (Local.com, kudzu, yellowbook etc), i.e. the homepage will be cached as localized to a specific US city. Does this mean they are all doing it the wrong way? How would you suggest doing it - serving up a national (unlocalized) homepage to bots specifically and a localized page to everyone else?

OK, Stu... I think I see the kind of effect you're talking about, but I'm not seeing it on yp.com.

The localized content I see on the home page of yp.com is in an iframe, and the Google cache on my system, anyway, does show a blank area where that content is. I need to reset that manually to get my location. That said, they do have me at least in California, so there may be something else going on.

This differs from another iframed localized display I have handy. A client who serves up localized weather on some of his pages also uses an iframe for that display... and while the Google cache appears to show me my local weather, as if Google had cached it that way, it is in fact showing a live data feed from the weather info supplier, which is detecting my location via IP, and displaying it in the iframe and within the cache.

The localized content on Local.com a different setup yet. It is not in an iframe... it's on the page, rewritten by javascript... and I assume served to Googlebot by default as Mountainview. For whatever reason, I'm assuming that Local.com just decided to do it that way, but that's just an assumption.

Within the cache, the page continues to display as Mountainview, but if I click on the cache link to the actual page or on the serps link to the page, the live page localizes instantly to where I am.

I can also further reset the location by clicking "change location" again and entering a new location, different from what's IP-detected. If I do that, the new location is remembered by a cookie, and it apparently overrides the IP detection. I can open and close the page, and whatever location I've set manually is remembered and returned.

However, if I bring up the Google cache again, with my cookies unflushed, as before, the cached page continues to show Mountainview in the cache. But, if I call up the live page, either from the cache or from the serps, the location displayed is no longer my actual location... it's the preferred location I've set and is remembered by my cookies.

So, there's a complex set of individual behaviors, depending on how a particular page is set up and when and how the localization displays are triggered.

Anyone else is welcome to jump in... please. ;)

Ok thanks RC I think I'm following what you're saying. Just to refocus the discussion I'm going to repeat my main question:

Are the IP addresses googlebot uses from determined set locations or can it be from anywhere? I.e. if a site is using a visitor's (and spider's) IP to determine their location and then serves up that city page, will Googlebot always see the same city(ies) on the homepage or can it be from any random location?

Does anyone know this?

IP-based delivery of localized pages with Google's permission is a possibility.

I'm not seeing it on any of the directory pages you mentioned, but it is something that's OK with Google, as long as that's the same thing as the customer in a geo-area will see.

Generally, though, the cache shows some kind of a neutral page. What the user sees would depend on location.

@DiscoStu
>timeline
You can define time period in special form in left options panel. Difined time period google keeps in cgi parameter &tbs.

>Does this mean they are all doing it the wrong way?

We can observe the situation from three sides- webmaster's, Google's
and user's. If document with universal, not local content, has a regional block, and Google index it, the part of this regional block, for instance name of the location, may be used by SE in snippets. And relevant document will be represented in Google's SERP by irrelevant snippet. In that case Google will loose in search quality, user will skip relevant info, and webmaster will suffer from traffic loose. And yes, I think that's wrong.

>How would you suggest doing it - serving up a national (unlocalized) homepage to bots specifically and a localized page to everyone else?

I see several ways how Google may avoid this unpleasant situation. A kind of new cannonical tag, or, maybe, <nosnippet> tag. I don't know what it will be, but, in my opinion, google may easily solve the problem, if it will become critical. And with more and more locality in serps, the problem may become critical soon.


Of course it's beside the point of cloaking, when URL
could show different local content ( for example, list of services or
goods) according to IP.

OK I'm an idiot, I don't know why I didn't think of checking the server logs guess I haven't done it in a while :P Anyway, Google is accessing with IPs from all over the place in a short time span, so I'd feel very comfortable serving up localized content to all users including spiders. Google should be smart enough to figure out that the website is showing internal links and content relevant to whatever IP location they are accessing from.

so I'd feel very comfortable serving up localized content to all users including spiders. Google should be smart enough to figure out that the website is showing internal links and content relevant to whatever IP location they are accessing from.

Let's look at [local events] query. I found local.com on the 2-nd page in google.com whith snippet starts from "Mountain View", and didn't find it in google.co.uk in top50
They loose traffic, dont' them? And Google loose in search quality, and it should be smart from his side to give some instrument to webmaster.

Google should be smart enough to figure out that the website is showing internal links and content relevant to whatever IP location they are accessing from.

The issue I'm having in this is that trying to get different page content to be indexed or rank for multiple locations on one url doesn't make sense. It can't be done.

That's why you use different different localized pages on their own urls... and why for something like a home page, yes, it is a bad practice for Local.com to have the location reflected in its page title and cache at all. But that's local.com's setup, not Google's.

You can deliver localized pages on different urls via IP delivery, and as long as Googlebot and the user see the same content, Google has no problems with that... in fact, Google is apparently helping with that IP localization. Ie, you search for [yellow pages] from mudville, and Google shows both a location-neutral YP.com home page and also a www.yellowpages.com/mudville local page.

But, again, these localized pages are each under a different url... not one url under which you see multiple pages of localized content.

And again, a reasonable amount of localized content in iframes that aren't indexed is OK. But the iframe area will be empty in the cache, and you don't get any SEO benefit from that content.