Okay, maybe I've spent too much time surfing Netflix and Amazon, but I've been wondering whether that technology could be used in our advising context as well. Let's say that we included a student's top 5 strengths, then how they rated certain professors on IDEA. You could then compare that information with other students who shared similar traits and course ratings to come up with a "if you loved Professor X for Western Civ, you're likely to be thrilled with Professor Y for Masterpieces of Lit."
This is obviously similar to what Netflix, Amazon, and other websites do. If you have an on-line evaluation system, we'd have the teacher evaluation data, and if we used a personality profile such as Strenthsfinder or MBTI for all of your students, you'd have the requisite student information. Plug both sets of data into an automated registration database, which many institutions already have, and you could have some computer program do the number crunching. You'd then have to do some validity testing to see whether or not this process is giving meaningful results.
Like I said, it's a little wacky, and it would probably make a lot more sense for a much bigger organization, but maybe it's an idea worth exploring?