Other possibilities include profiting from in-course mentoring services, career counseling — and charging universities for licensing. In October 2012, for example, edX licensed a circuit-theory MOOC designed by Agarwal to San Jose State University in California, where it was used as the online component of a flipped classroom experience. In return for the licensing fee, “the professors can offer the course on campus, tweak the course however they please, get access to students' grades and online activity, and all the analytics a teacher would want to see”, says Agarwal. In this particular experiment, he adds, the San Jose course's usual 40% failure rate fell to 9%.
Analytics are another example of the Silicon Valley style, potentially allowing the MOOC companies to do for education what Internet giants such as Google or Amazon have done for marketing. In Coursera's case, says Koller, the platform monitors the students' every mouse click — “quiz submissions, forum posts, when and where a student pauses a lecture video, or rewinds, or moves to 1.5 speed”.
The company is constantly using these data as feedback, says Koller, both for refining the platform's user interface and for improving the course content. If 90% of the students start stumbling over the review exercises for a certain lecture, for example, then maybe it is time to revise that lecture.
“But anything we do is just the tip of the iceberg,” says Koller. When data from individual students are multiplied by tens or hundreds of thousands of students per course, they reach a scale big enough to launch a whole new field of learning informatics — “big-data science for education”, Pea calls it.
Learning informatics could provide an unprecedented level of feedback for colleges and universities, says Stevens: “We haven't measured learning in higher education very often, very consistently or very well — ever.” Academics have endlessly studied factors that are associated with university enrolment and success, such as race, parental income and school achievement. They have also studied what happens after graduation: the higher earnings and other benefits that college confers, on average, over a lifetime.
“What we don't know is how college performs this magic,” says Stevens. “We certainly don't know the extent to which digitally mediated college experiences will deliver the same returns as a four-year residential experience.” Now, however, he and his colleagues can begin to see what education science will look like as it merges with data analytics. Instead of looking at aggregate data about students on average, for example, researchers can finally — with appropriate permissions and privacy safeguards — follow individual students throughout their university careers, measuring exactly how specific experiences and interactions affect their learning. “It's thrilling,” he says, “a huge intellectual frontier.”
What remains to be seen is how higher education will change in response to the new technology. Maybe not much, says Dede. Yes, the major universities will extend their courses beyond their own campuses; the MOOCs have already shown them that they can do so with relatively little effort and potentially large profits. But the MOOC founders' other goal — fundamental reform in on-campus teaching — is a much tougher proposition.
“Universities think of themselves as being in the university business, not the learning business,” explains Dede. That is, they mostly take their existing structures and practices as given, and look to MOOCs and other online technologies as a way to do things more cheaply. But experience with earlier innovations such as personal computing shows the limits of that approach, he says: real gains in the productivity and effectiveness of learning will not come until universities radically reshape those structures and practices to take full advantage of the technology.