lexipixel - 2:54 am on Jun 24, 2011 (gmt 0)
1). User opens browser or browser was otherwise dormant before viewing a page on YOUR site.
2). User navigates to a page on YOUR site either by a referring link or directly typing the address in the address bar (which would mean no referrer in second case).
3). User browses YOUR page --- scrolling up, down, all around -- but not clicking anything else on YOUR page.
4). User clicks the [HOME] button on THEIR browser.
... and the total time on YOUR page can only be calculated as "0 seconds" because once the page is loaded, the user has no other interaction with YOUR site -- the content has been loaded into THEIR browser and even if the page contained 1,000,000 words and they took 6 months to read it.... all Google, or any stats / analytics program makes is (1) entry in the log [it provides YOU], so there is no way YOU can calculate how long the user was on YOUR site.
This is not true for Google itself in many cases... Google can track how long the user was browsing YOUR content in many situtations and discover the next track-able action the user takes.
Lets say the user decided to Tweet about it, or went to YouTube to watch a video related to it, or, after clicking [HOME] went to another page that had a DoubleClick ad related to what they were browsing at YOUR site --- well then, Google (since it owns YouTube and DoubleClick -- among many other web properties), and since Google read social signals from Twitter, Google could make a logical conclusion that:
USER "X" read about "blue widgets" and 2 minutes and 30 seconds later, USER "X" watched a video on YouTube about "blue widgets" for 1 minute and 58 seconds, then when the video ends, Google served a Doubleclick driven ad selling "blue widgets", which USER "X" clicked on...
In the above situation, something like [donottrack.us...] is appropriate to discuss next...
Google knows all, sees all -- we are blind.