Google cache not correct.

[jlander]

I bought some ads on an authority site that went live around the 10th of the month. Yahoo and MSN cache show the new page with the ads, G does not, even though the cache dates are the 13th and 17th.

I'm getting great response from the ads, but what gives? Is Google editing out ads now?

[jlander]

I just checked again today and the cache dates are still after my ad went live, and they still are not showing up.

[g1smd]

Google has an interesting flaw regarding cache dates. I found it like this:

Upload a page "page.html" in May. Google indexes and caches it.

In October rename that page to page.x.html and upload a new page.html. Google indexes and caches the new page content.

In November, delete the newer page.html and then rename the old page.x.html back to be page.html again. Make these actions on the server via FTP. By not uploading the file again, merely renaming it, the filedate is still May!

Google continues to update the cache date every few days, but the cache is of the October file contents, and the page ranks for the October content, even though none of that content exists now.

This happens because when Googlebot requests the file, the filedate is NOT newer than that which it saw at the last visit. If the filedate is unchanged or has gone backwards Google treats the file as being unchanged, and does not respider the actual page content.

I suspect that since your ad is merely inserted into an existing page (the ad being an external file, an include, orsome such scheme), that the filestamp of the page has not altered so Google isn't actually reindexing the on-page content, just telling itself that since the date didn't change then the content could not have changed either.

[AlexK]

The flaw is not with Google. The flaw is with the webmaster due to an incorrect implementation of the response to If-Modified-Since.
.
If the content changes, the file-date should change, else many of your customers will be receiving old content, search-engines or no search-engines.

[g1smd]

The flaw is that there are many Black Hat things that can be done with that... but, if Google also checked, say, the file size or the checksum then there would be much less chance of getting away with the many and various schemes that I can think of.

[jlander]

I see what is happening. Thanks for the explination. I don't believe the file is an include because it has an htm extention.

Will G ever pick up the new page? Does only G do this because both MSN and Yahoo have the correct cache.

[minnapple]

Yep, cached dates have been messed up for awhile now. It is not a webmaster's issue, it is google's.
Perhaps google decided to offer cache data like it offers other toolbar data. Outdated and not very useful.

[jlander]

So then, is G aware of the page changes or not? Could this be used as a way of cloaking? If so, I don't see how it could be against their TOS.

[g1smd]

No, they are blind to the fact that you have changed the file back to be an older version. That version might have only existed for several minutes originally and never even been indexed by them (before being replaced with the new version with a later file date that they did then index).