Google's poses a brainteaser for us

Looking for help

[McMohan]

Well, where do I start?
Site www.xyz.co.uk is a well established corporate site with Alexa 5k. This site has a mirror site, www.abc.com which doesn't redirect. Searching for "My Keyword" ranks the mirror site www.abc.com. Searching for this mirror site www.abc.com in Google, shows it is merged with www.xyz.co.uk, as should be the case.

Now how is that possible?

Any info appreciated :)

Thanks
Mc

[McMohan]

Another interesting observation. First let me state the facts -

Updated a popular website's Title attribute. Google has updated the change within a day. Now whether the updated Title is returned to user depends on the search query! If the search query is a remote one, then it returns the updated version, where the website is ranked well. If the query is very competitive (more than 100 M results), then Google returns old version, where the site is still ranked same as before making the change.
Almost forcing me to think, if Google delays changes not just to the SERPs, but also for website change updations, depending on the query.

[Reid]

That's easy
google is sending you to a different database for the higher competitive words than it is for the less competitive words. it's working. It make take some time before all the databases get the update.

[McMohan]

Reid - Checked in all IPs.
What has caught my interest mainly though is, the Title in the SERP is an old one for the competitive phrase, but the cache is latest. That is to say, what Google shows in the SERPs and what it has in the cache are not same. Will sticky, if someone likes to know.

Mc

[ciml]

> different database for the higher competitive words

That correlation would seem more likely to be an artifact of cacheing, rather than by design.

McMohan, it is normal for Google to merge two URLs that return duplicate content. Where possible, they seek to avoid listing the same page multiple times on different URLs.

[Reid]

so the cache has updated but the snippet is still the old one. That must be mid-update?.
Or as CMIL said it is combining the 2.

[g1smd]

I have a site where we removed the email address from the page over a year ago.

If you search for some words on that page, you get a snippet with a Updated 2005 text within.

If you search for the email address, then the snippet returned includes the email address that no longer exists on the real page and the snippet contains the text Updated 2000.

In both cases the cache shown is from 3 days ago. Google has a separate database of titles and snippets and which one is returned depends on your search query.

In some cases a page can appear as a normal result for some queries and as a Supplemental Result for other queries, too.

[McMohan]

ciml - In the first example (msg #1), it has correctly merged the duplicate, less popular site into the main site, which is normal and day-to-day behaviour. What stands out is, in SERPs this less popular site which is merged into the main site is shown, not the main site. That is not normal IMHO.

g1smd - Your observation is similar to what I am seeing (msg #2). Just that, unlike an email which was long gone in your example, the page content in both contexts is same, with only the Title attribute being different.
"Is it a glitch or by design? to stabilize the SERPs for popular searches"

[ciml]

> this less popular site which is merged into the main site is shown, not the main site

This is not easy to check after the merging occurred, but can you be sure that the page on the less popular site couldn't have had more PageRank than the page on the main site?

[McMohan]

can you be sure that the page on the less popular site couldn't have had more PageRank than the page on the main site?

It is the homepage thats being shown. The less popular site has far less number of IBLs (can't know the PR though since it is already merged) and is a newer site, the very facts why this site was merged into the main site. The sole purpose of this less popular site was to preserve the .com identity of their .co.uk corporate site.

[jakegotmail]

is there a posting that has all known ips for the different google datacenters?

[Lovejoy]

I have the same problem, two sites on different servers, but both as far as google is concerned are the same. My original page is an old "Free" page provided by my isp, online since 1997 and #1 for keywords since 2002, but there was not enough capacity to host all of my pages. Being loath to lose a #1 spot I left the old index page up and linked it to my new domain, which now hosts all of my pages, including its own index page which has different content, images and text, but they still resolve to one site on google some reason. The only thing I can think of is that the two have links to one another.

[g1smd]

I already reported a case, last month, where a Google search for:

www.domain.it/keyword/keyword/document.html

returns that page, and then clicking the link marked cache Google then shows the cache, but it says:

>> "This is the cache for: keyword.it/folder/document.html"

which is the alternative URL that the content can be obtained from.

Google has merged the two URLs. In this case they do both represent the same physical server.

[Reid]

jakegotmail

There is quite a nice list here
[webmasterworld.com...]

msg#48