 We're in a really early stage of learning analytics. The numbers have actually gone down from 2012, so fewer places are doing learning analytics than said that they were doing it, and it's a lower level of priority. And there are a couple of reasons for that, I think. One is it's difficult, I mean it's technically difficult, it's a very new market, the solutions really aren't there, and it also is a big issue, it's got all these different moving parts and there's not really a solution, a technical solution, though sometimes people look for one. It's also politically really charged, you know, if you think about it. So you can do institutional analytics because it doesn't stick on the faculty. Once you start talking about learning and gathering learning data, especially at the course level, you run into faculty autonomy issues and you've got to be very delicate. And I think a lot of people are not doing learning analytics right now because they're just staying away from that. I think of myself as a data person, I'm a social scientist, I love data, but I came across this article, it was actually a history article, by some historians who were perplexed by the statistics about surviving a sinking ship, okay, because what do you think of when you think of a sinking ship? It's women and children first, okay, and the crew goes down with a ship. Well, turns out that all comes from one ship, that comes from the Titanic. And so they looked at a larger data set. They looked at all the shipwreck statistics that they could find, so a really large data set. And turns out that men out-survive women by a substantial percentage and the crew out-survives the passengers by a substantial percentage. It's a little depressing from a practical point of view, you know, but it just shows the value of sort of asking really complex questions, you know, asking questions about things that you assume to be the case and then also getting a begin-up data set. So I think, you know, we need to sort of probe those things. You know, I know one of the things I'm frustrated about with predictive analytics is, like, you go there and you talk to people and they say, okay, our number one predictor of student success is GPA. It's like, no, that's an indicator of student success, you know, like, it's like the successful students are successful. We actually need to get sort of at deeper levels and ask tougher questions, I think. We'll see it much more and we'll see it sort of driven by a larger approach to digital strategy at the level of the institution. But in terms of the actual technologies, I think we'll see it get much more into the course level kinds of things because right now it really is that sort of big macro kinds of things. I'm sort of excited by some of the companies, for some reason they're coming out of Scandinavia. So one is a company called Konexus, an Norwegian company that works in K-12, but they've got these great sort of dashboards for teachers where they can, you know, use multiple sources of data to actually see how students are progressing and do some sort of micro interventions at that sort of level. And I think we'll see more of that. I think we'll also see it become much more automated in a sense that, you know, if I'm not doing something, my phone will start to buzz and nudge me to do it.