Using Google Trends to track retention seems weird to me. My experience is that I google for stuff i don't remember, not the stuff that I do. But beyond that... 
 
You may be right about the value of undergrad non-pre-professional study; I go back and forth in my own mind on the topic. But to the extent that part of the curriculum is valuable (and today I'm a believer), its value outside the academy is not in its immediate applicability to one's everyday life (which you might be able to see in Google Trends). I took a history course in Thomas Jefferson as a soph, for example. I have never since needed to know anything about Jefferson beyond what I knew coming into that class (that he was accomplished and interesting in a wide variety of fields, not merely as a Founding Father). But by following the range of Jefferson's interests, I gained some understanding of concepts new to me at the time ("romantic" as opposed to "classical", for instance), and that understanding lets me observe and participate in everyday life with a perspective I'd lack otherwise. I think that perspective is important and improves both my quality of life and the quality of my participation. That's not something you can measure with Google Trends, but it's not "flusssshhh" either.

I, a Chemistry Ph.D., offer: 
Real asset price histories (USA) are overwhelmingly compelling, see first chart here: 
http://homepage.mac.com/ttsmyf/RD_RJShomes_PSav.html 
And the histories are evidently serial herd behavior: 
http://homepage.mac.com/ttsmyf/3warnsRD.html 
The profession's defensible behavior has these histories overwhelmingly apparent. 
The opposite apparence is the case, by the profession's indefensible behavior.