Correlating a visitors actions with satisfaction.

[GreenDog18]

If you were Google how would you correlate a visitors actions with their satisfaction level?

Obviously the visitor clicks on a link in SERPS and then returns right back that would be negative...

Different types of visitors actions to analyze...

1. Stays on the website for more than let says 20 seconds with no scroll and returns to SERPS
2. Stays on the website for more than a couple second with scroll and returns to SERPS
3. Stays on the website, no scroll, leaves via a internal/external link
4. Stays on the website, no scroll, leaves via a internal/external link

[Sgt_Kickaxe]

You don't look for signs of satisfaction, you learn to spot dissatisfaction such as is likely with a backpage or new search after the initial clickthrough.

[deadsea]

Google would also be limited by the information that it can actually collect. Last time I checked, the toolbar reported which pages on a site you visited back to Google, but did not report additional activity information including scrolling.

It appears to me that Google is NOT paying attention to things like "number of pages per session" or "time on site" when measuring user satisfaction. I have a site where users can get the info they need on the landing page (1.2 pages per visit average) and usually only spend 50 seconds doing so. Despite these very low engagement metrics, satisfaction is very high. Google seems to have no problem with this and the site continues to enjoy good rankings. Users usually do interact with the page a bit. If Google is measuring anything about the site other than the dis-satisfaction as sgt-kickaxe suggested, it would have to be interaction for this site.