Crawled often but cache &snippet are from 11 Sept

[Mokita]

Background: I redesigned an old site and the new version was launched mid-October.
Google has crawled the whole site frequently and most pages seem to be indexed and cached with no problems.

Except one page. This page is a first level page, and is a category page which gives entry to most of the second level pages, so it is pretty important. Googlebot has successfully crawled the new version of the "problem" page on these dates: 15 Oct, 19 Oct, 21 Oct, 6 Nov, 10 Nov, & 16 Nov.

An example from the raw logs:

66.249.72.237 - - [15/Oct/2006:13:56:59 +1000] "GET /thispage.htm HTTP/1.1" 200 10789 "-" "Mozilla/5.0 (compatible; Googlebot/2.1; +http://www.google.com/bot.html)"

But it still shows the snippet and cache from 11 September. MSN (Live Search) and Yahoo are showing up-to-date cache, so I doubt there is a problem with the page itself. Why would Google be stuck in a time-warp just with this one page and what can I do about it?

[AjiNIMC]

Was the change a very big change for the index page.?

[googlewebmastercentral.blogspot.com...] , it has the best explanation for it.

Thanks,
AjiNIMC

[Mokita]

Hi AjiNIMC,

Many thanks for the reply.

I may have mixed up my interpretation of the levels. This is not the index (default) page I am talking about - but one linked from it (and every page in the site). So probably it is a second level page that provides the links to the majority of third level pages (which are all correctly indexed and current versions cached by Google).

Was the change a very big change for the index page.?

The change for the index (default) page and the whole site was huge. It has gone from being a FrontPage themed site to XHTML 1.0 Strict/CSS plus a totally different look and feel.

[edited by: Mokita at 6:00 am (utc) on Nov. 18, 2006]

[tedster]

Check your server's http header and see what your server is returning for a Last-Modified date. I like to use the Firefox extension "Live HTTP headers" for this, but any online server header checker should show you wht you need to see.

I don't think Google's new method of dating the cached page plays into this picture. Their new method should provide the date of last access by googlebot, whether the page was modified or not. So, if googlebot was there on 16 Nov, it should show that date.

So find a recent googlebot visit in your logs. Did googlebot get a 200 OK response? A visit that resulted in some kind of error would explain what you are seeing.

Also, there is a chance that this is just a bug on Google's end that is not really affecting your ranking. Can you craft a search query that should return only the new version of your page, and see what that result is? Something like this -- site:example.com "new unique phrase".