It seems that my original post has disappeared with the launch of the much more awesome community site, so if it is somewhere already my apologies for the double post.
One of the things my team at Rowan University is getting into is Analytics, not necessarily to solve problems but to identify patterns to generate questions about improving course design. Devlin Daley once said when talking about Analytics "Good Analytics make good questions, not good answers," and that has always stuck with me. By looking at course analytics you can come up with all kinds of questions and possibilities to improve the design of the course.
At Rowan Online (I promise this isn't a plug), we're using both Canvas and Kaltura (media platform) to deliver content to students. Both provide Analytics and they're both proving to be very helpful when discussing course design to faculty. Often as a designer I find I have to sort of pitch my case to an instructor who has been doing things in a face to face environment for a long time. They don't necessarily know the online space is a different environment. By showing them some analytical patterns you can really easily help them with come evidence based course design.
Here is a short talk I did with my colleague at one of our open houses. This was a dry run for a conference presentation at Drexel University, so my apologies if we spoke too fast on some of the info points.