Online course completion rate: how to raise it
How many students actually finish an online course, what the Harvard and MIT MOOC data shows, and 7 practical ways to get more people to the final lesson.
Selling a course is half the job. The other half is getting buyers to finish it. A student who stops at lesson three doesn't get the result, doesn't leave a testimonial and rarely buys the next course.
This guide reports the most solid data we found on online course completion and offers seven practical levers to raise it.
How many students finish an online course?
It depends a lot on the type of course and student. The most-cited study is by Justin Reich and José A. Ruipérez-Valiente, published in Science in 2019, which analysed all MIT and Harvard courses on edX from 2012 to 2018: 5.63 million learners and 12.67 million course registrations. According to Inside Higher Ed's summary, 3.13% of all participants completed their courses in 2017-18, down from about 4% in the two previous years and nearly 6% in 2014-15.
Among those who paid for the verified track the picture is different: 46% completed in 2017-18, 56% in 2016-17 and about 50% in the two years before. The authors note that low completion rates, the bane of MOOCs, have not improved over six years.
See the data as a table
| Group | Completion |
|---|---|
| All participants | 3.13% |
| Paid verified-track learners | 46% |
Beware the comparison: MOOCs are free courses with one-click enrolment, very different from a creator's paid video course. The useful takeaway is that commitment matters. People who pay and have a goal finish far more often.
Why does completion matter to a creator?
Because it bridges the sale and everything after it. People who complete the course get the result, and people who get the result share their experience, recommend the course and buy the next step. Almost all the testimonials you use in a launch, as we explain in course testimonials and reviews, come from finishers.
How do you measure it?
Pick one definition and keep it fixed: for example, the share of students who complete the last lesson within 90 days of purchase. Then look at the lesson-by-lesson chart: where the curve drops sharply, there's an obstacle.
Which levers raise completion?
- A quick first win. Within the first hour, the student should get something concrete. That's the job of onboarding.
- Lessons with an action. Every lesson ends with something to do, not just watch. See online course structure.
- A suggested pace. "Two lessons a week" helps more than "whenever you like".
- Gentle reminders. An email to anyone who hasn't logged in for ten days, with a link to their next lesson.
- A community. Seeing others progress motivates; see community for your online course.
- Milestones. A small recognition halfway, or a certificate at the end.
- Personal contact. A direct message to someone who has stalled beats five automated emails.
An example
A 24-lesson course has 200 students. The creator looks at the lesson-by-lesson chart and finds that 80% reach lesson 5 but only 45% reach lesson 7: almost half are lost between lessons 5 and 6. Lesson 6 is 38 minutes long, all theory, no exercises. He splits it into three 10-12 minute lessons, each with a small task, and adds an automated email to anyone who doesn't open the next lesson within seven days. In the next cohort, 62% of students reach lesson 7. It isn't scientific proof, because two things changed at once and the students are different; however, it shows where to act. In practice, lesson-level data turns a vague problem ("they don't finish") into a precise, fixable one. The numbers are our example.
Which mistakes should you avoid?
- Only looking at the final figure. Where people stop tells you more.
- Adding content to "add value". A longer course gets finished less.
- Lessons with no action. They're watched and forgotten.
- No contact after purchase. Silence is the fastest way to lose a student.
- Comparing yourself with MOOCs. They're a different product: compare with yourself over time.
Key terms. Completion rate is the share of students who finish within a fixed period. A MOOC is an open, usually free online course. The drop-off curve shows how many students remain lesson by lesson.
How we checked. We read the study summary on MIT's Teaching Systems Lab page and the Inside Higher Ed article reporting its figures. We didn't read the full Science paper. The example is ours. Contact: hi@heytea.app. Editorial note: we check every figure against the sources listed below and correct mistakes openly.
Frequently asked questions
What is a good completion rate for an online course?
There's no single figure that fits everyone. The most-cited data is on free Harvard and MIT MOOCs on edX: in 2017-18, 3.13% of all participants completed, but 46% of those who paid for the verified track did. People who pay and have a goal finish far more often: measure your own rate and compare it over time.
Why does completion rate matter?
Because people who finish get the promised result, and people who get the result leave testimonials, recommend the course and buy the next product. A course where few reach the end generates less word of mouth and more refunds.
How do you measure completion?
The simplest measure is the share of students who complete the last lesson or exercise within a fixed period, say 90 days from purchase. Also look at where they stop: the lesson after which many quit is the first one to revisit.
Do short lessons help students finish?
Often, because each lesson fits into a break and gives a small sense of progress. But it matters more that each lesson has a clear purpose and an action to take. A short lesson without an exercise is forgotten as easily as a long one.
Do deadlines and groups raise completion?
Formats with dates and other people, like cohort courses, create a rhythm and commitment self-paced courses lack. In the Harvard and MIT data, learners who had paid and were therefore more committed completed far more often: commitment and rhythm matter.
Sources
- Justin Reich and José A. Ruipérez-Valiente, "The MOOC pivot", Science 363 (6423), 11 January 2019, summary by MIT Teaching Systems Lab. Accessed 19 October 2026. tsl.mit.edu
- Doug Lederman, "Why MOOCs Didn't Work, in 3 Data Points", Inside Higher Ed, 15 January 2019. Accessed 19 October 2026. insidehighered.com