Dan, I also enjoyed this book. I took four things away from it that I think will stick with me:
(1) Universities/colleges are more similar than they are diverse. I'm at a large state school, but the envy students (and faculty) reported when passing the well endowed science labs on their hill at the college mimic the attitudes that many feel toward engineering here.
(2) Innovations that work can be mundane and unpublishable. Although high impact practices work, if you're a dean with the pressing problem that you have a curmudgeonly, misogynist professor nearing retirement, the thing to do is to give him an 8am class to teach and hope no one turns up.
(3) The unit of analysis should be the student, not institutional units. It doesn't matter what proportion of a university's classes are small (well, it does to USN&WR), but it really matters what proportion of classes are small that an average student takes. These numbers can be very different.
(4) MOOCs have been around for centuries. They were called books. A HS student today could buy a used calculus text for a couple of bucks, and learn everything that Newton discovered in a couple of months. But the average student doesn't, and that won't change with online materials. What college provides that self-directed reading/viewing does not, is structure/discipline and curation.