A thread at SEW today has some back and forth on the issue of Google's analysis of how long users spend on websites and whether this can affect rankings. Danny Sullivan notes that this metric has been affecting AdSense ads for some time - Google simply measures when and if the user returns to the SERPs page. This tactic could be used for organic results as well, although it's unclear as to whether Google's interpretation of the user action would be accurate.

One great comment was posted by Chris Boggs, whose humor I've been enjoying more and more:

surprised that no one has mentioned Urchin in this conversation yet. I get the funny picture of the evil Google scientists taking apart Urchin right now to make it divulge all the secrets of any site running Urchin stats

There are several ways that Google might try to track down the time spent on a website or page and incorporate that into the rankings. Some are mentioned in the recent Google patent, while others are simply intuitive.

1. Measure users that return to the SERPs page and the time they spent at a result
2. Use analytics from tracking software such as Urchin
3. Measure via the Google Toolbar
4. Estimate via visits to Google properties, placement of cookies on popular 3rd party sites, etc.

PhilC notes that this has been tried before - with the DirectHit search engine that used this as part of its primary metric to determine popularity. Many in the thread note the danger of this becoming public knowledge, as proliferation of hitbots to boost rankings would quickly eliminate the value of such a metric.