Google and behavioral analysis

I have read this article about google behavioral analysis, but had no time to post my thoughts. To sumarise, results of experiments are discussed with 2 implications:

a) google uses click on SERP’s analysis to change placements for the results ( It is discussed in the google patent as well)

b) Google uses Analytics analysis for modifying SERP’s, more exactly the time users spend on the site or their behavior on the site.

Although the first one could be guessed from the google patent, the second one is much more interesting and a bit concerning. I tend to disagree with Visio on the cause of the results a bit. But lets explain this from the begining.

First, lets assume that the importance of page (or result) is the value it gives to the user. This is mentioned in some posts/discussions at google. The value might lie not in the whole site, but in the single page of results as well (when people look to solve specific problem). An example:

A site that provides some instructions to solve a problem. Lets take http://www.wikihow.com as an example. Maybe all other guides are not important for the users. Thus, average pageview count is small. There should be no bounce rate, though. Another case would be a site with POOR navigation where users stumble to find right things. Also, this would have strong implications to affiliate sites that provide useful content, but tend to send users to their partners by explaining how to solve their problem using someones product. More often then not product manufacturers focus on product, and not presenting all the problems users can solve with it.

So, why the SERP change?

Perhaps thing that matters is the original traffic increase itself, similar to the original experiment. People started using the site more -> Site is better and worth increased results.

I am not saying that navigation does not matters, but I kinda hope that it is not the main factor on itself.