What is Data Worth to Higher Education?

As I mentioned in my last post, the ET4Online conference just ended and I received some very good feedback from attendees. However, I also received some comments that annoyed me just a little bit. Specifically, regarding some of the sessions I did on data driven decision making I had both face-to-face and email follow-up questions in which I was told something to the effect of the following: "I am from a state school with a tight budget and we simply can't afford to implement large scale data analysis."

You can't? Really?

For the sake of getting to the point I am going to play a little fast and loose with the numbers because I don't want to turn this post into a spreadsheet, so forgive the 40,000 foot analysis. This is not something I usually do. But here we go....

Let's assume you are from a state school where some combination of tuition and state funds bring in $10,000 per student per year. That's actually on the low side, as an average, but for this example lets run with it. Now, lets assume that you build out a data analysis framework, targeted to retention, that costs $100k / year and that that framework will give you numbers that account for 15% of the variance in the tendency of students to disenroll. By the way, from experience I know that that is about what this type of investment will yield. Translated into actionable intelligence this means your institution could utilize the data to target students who are most likely to disenroll (who have attributes that fall within that 15% of variance) and dedicate existing student services to help retain them. Further, lets assume that you are at an institution where 1000 students drop out annually. That means that your actionable intelligence only has to impact 1% of students to break even on your investment. At 2% you have 100% ROI.

Granted 15% of variance accounted for isn't that robust, so lets think about a system that costs $500k / year to implement and you can account for about 35% of the variance in the likelihood that a student will drop out (again a realistic estimate). In this case you would have to use the data to convert 2.5% of the target population to break even and 5% to have a 100% ROI. If our universities can't use hard data to have even this amount of impact then our system is most assuredly broken beyond repair.

As a note - these are just the hard numbers associated with retaining students. Since I am talking to educators is it even necessary to talk about the social costs and loss of intellectual capital associated with students dropping out?