Adsense bot spidering my local pc?!?

serving contextual ads before page is published

[Adam_C]

Can anyone explain to me how when I'm working on my sites html files on my PC's hard drive with the Adsense code in place, that Google know's which ads to serve?

Surely, they must be spidering the page to get this information.

But should this really be happening?

I first thought that it was due to the GG Toolbar (with advance features on), but I turned them off a few weeks ago.

[Clark]

Is your hard drive accessible on the net?

Google will look at the referrer and send a bot to that address.

[Clark]

Took another look, it may be using keywords in the url.

[Fairla]

What's interesting is that in many cases I've noticed (when viewing my webpages on my hard drive using Homesite's browse mode) that the ads I see are different and often much better suited to the content on the page than the ads Google actually serves after I upload the page.

[Imaster]

Google sees the file name and the directory name (where the file is residing) when it is on your local pc. Just pick up the html file and experiment by keeping it in different directories, and see how the ads change.

For e.g, if your file is in
C:\design\file.html
you will see ads related to the keyword "design"

Try changing the folder name and see how the ads get changed. ;)

[Adam_C]

thats a very interesting observation. I'll check it out.

Still kind of begs the question, how do they get the directory name?

[Clark]

The javascript on the page requests the code to display on your page and sends a referer page to them which probably includes your entire filename structure. Look at your logs for similar examples. Also, with Javascript in your page it can detect a bunch of stuff.

[Imaster]

Looks like one of the factors in the Google AdSense targeting algorithm is the analysis of the URL or path of the page.

It might be possible through javascript to analyze the complete URL or path of a file without even crawling the page. If that is true, it is not necessary for the page to be crawled by the MediaPartners bot to check the URL/Path.