The idea that an unlimited number of students would be able to access the very best instructors in the nation and take classes at a fraction of the cost—or even for free—of traditional tuition was enough to spark the imaginations of many people. And, yet, massive open online courses (known as MOOCs) never quite lived up to their hype. Here are four reasons why:

First, the completion rate for MOOCs is usually in the single digits. This is, to put it mildly, disappointing. On the one hand, it might just be a reflection of the fact that the barrier to entry is so very low. In other words, very few people pay thousands of dollars in tuition for a traditional college course if they are not committed to completing it. But, if you can enroll in a MOOC for free and with just a few clicks, students far less committed to completing the course are likely to give it a try. So a lower completion rate is to be expected. At the same time, even students who signal a firm intention to complete a MOOC only finish about half of the time. This is very discouraging.

Second, there are problems of quality control. While planners envisioned having only the top scholars in the field deliver MOOCs, the reality is that edtech tools make it possible for nearly anyone to deliver a MOOC, and it turns out that the top performers in a field are usually not as well-incentivized to create a MOOC as are charlatans who are hoping to improve their own brand. By definition, a student in a field is unlikely to be able to determine who the best practitioners are in that field, leading to an enormous problem in quality control for MOOCs.

Third, despite an abundance of high tech tools, the most frequently used model for MOOCs is a lecture and, if there is any assessment at all, questions that can be machine-graded. And yet education scholars have known for decades that lecturing is one of the least useful instructional methods and that the best assessments are authentic, open-ended investigations of higher-order thinking skills. This mismatch between what is easy to deliver in a MOOC and what is most beneficial to students is a significant problem.

Fourth, there is a known social and emotional aspect to learning that is nearly impossible to replicate in a MOOC. It is the warm relationships between students and instructors that lead to the kind of engagement to prompts excellent learning outcomes. Until and unless MOOCs figure out how to replicate that, they are unlikely to merit a significant place in education.