Facebook recently has been talking about not optimizing for time spent on site but also the quality of this time. They even went and showed in one of the quarterly results that the average time spent on the service went down.
I heard the same argument yesterday in a Q&A with Mark Hull, a product director there.
I was thinking about this and I think there are two sides of this story, one that Facebook tells publicly and one that sounds more like the reality to me.
The reality I expect is that there is no way to quantify the quality of time spent on Facebook. The closest they can get to that is measuring if users will come back after spending less time on the service. Or by doing some controlled experiments where they ask users about their feelings before and after the algorithm change.
Less time on site means less ads to be shown. It also means ads prices will go up as a result of the number of spots going down. Until a certain limit.
Probably Facebook found the margin they can play with the time spent on site without hurting revenue growth, and sold it externally as time well spent.