A conversation with my son Matt confirmed my suspicion. The Google AdSense ads I recently installed on this website are actually giving me an insight into what the Google search engine spider cherry-picks from of my web page content.

It's not hard to imagine: AdSense ads are context sensitive. They exist as scripts on the web page. In order to be context sensitive, the script must initiate an indexing when the page is opened and refreshed.

Is there any reason to think that the indexing process performed by Google AdSense would be different from the process used by Google the search engine? None I can think of. Both indexing processes need to do the same job: extract core meaning from a page and compare it to a database.

In AdSense, the database contains paid ads waiting for a relevancy match. In search, the database holds keywords. But the meaning extracted from the web page could easily be identical.

Therefore, one might get a peek into the Google indexing algorithm by reviewing a series of web pages which display AdSense ads, and studying the ad content.

Studying AdSense Relevancy on Poingo.com

I studied the 30 or so pages on this site and checked the AdSense ads on each page for relevancy to the page content. Results were quite interesting.

The site contains a number of pages which present the features of various software or service offerings. Verbiage on these pages tends to be sparse and oriented toward key concepts.

On these product presentation pages, AdSense did a great job of extracting meaning.

For example, the page offering Poingo Email Printer, software which creates PDFs, was accompanied by AdSense ads which all pertained to PDF conversion. Text on the page was minimal, but the page title contained "create PDF", there were 3 keywords metatags containing "PDF", and the first paragraph contained "convert PDF" in bold.

From an indexing standpoint the page spoon-fed meaning to Google, and obviously there was a wellspring of PDF software advertisers for Google to find in its database. A match (or five matches to be exact) made in heaven!

Similarly, pages offering FTP software and an Outlook add-in received highly relevant companion ads. Again, words on the page were sparse, but page title and paragraph text contained the obvious words FTP and Outlook respectively, and Google AdSense took the bait like a trout succumbing to Robert Redford.

The three pages mentioned above offered essentially single concept offerings. PDF. FTP. Outlook. No confusing multiple choices.

When analyzing the page which offers Lightning Navigator, hotkey shortcut software with multiple features, AdSense picked one feature, screen capture, to orient 3 of the 5 the companion ads.

Interestingly, screen capture is listed seventh on the list of product features. It follows six other features which were all keyword-optimized but ignored by AdSense.

From previous research, I recall that keywords pertaining to screen capture such as "print screen", "screen shot", and "screen grab" receive many more clicks per day than other features such as "automatically create email" and "internet shortcut".

Apparently in this example, AdSense was quickly able to select the key concept for which it had the most ads to apply, and then threw most of its ad eggs in this basket.

Interaction between page and AdSense now becomes more interesting. Inventory of relevant advertisers becomes a factor in selecting key concept. That makes sense. You can't post an ad if it's not in the queue.

The non-screen capture ads on the Lightning Navigator page are as follows:     1 for shortcuts (highly relevant)
    1 for surveillance equipment (huh??)

I have no doubt that there is a reason the surveillance equipment ad appeared, but it was not visible to me in the text of my page, the ad itself, or the page to which the ad linked. Mysteries abound in the "second-guess-the-Google-algo" world.

If your eyes are not bleary yet, stick around. There is more to tell.

Continue to Part 2