Most people who sign up for an online course never finish it, and many never start. A multi-year study of 5.63 million learners, published in Science in 2019 by Justin Reich and José Ruipérez-Valiente, found that 52% of people who registered for Harvard and MIT open online courses never entered the course, and only 3.13% of those who did start completed it in 2017-18. Low completion is rarely about lazy students: long videos, unclear structure, no feedback and no clear finish line all push people out. Research points to fixes any academy can apply: lessons of about six minutes, short quizzes between lessons, progress that is saved, a certificate at the end and a habit of checking where students stop. Below we explain each fix with the numbers behind it and show how HBS LMS supports them.
How many students actually finish online courses?
Far fewer than course sellers like to admit. Reich and Ruipérez-Valiente analysed 565 course runs with 12.67 million registrations from 5.63 million learners at Harvard and MIT, from 2012 to 2018. Their peer-reviewed paper, "The MOOC pivot", reports that:
- 52% of those who registered for a course never entered the course material.
- Among participants who registered and viewed the course, 3.13% completed it with a passing grade in 2017-18.
- Among learners who paid for a verified certificate, completion in the same year was 46.02%.
The authors write that low completion "has not budged in six years" despite heavy investment in course design. These were free, open courses, and learners who pay tend to do better, as the 46.02% figure shows. But the shape of the problem is the same everywhere: a large share of students leave before they get the result they came for. A companion study by the same researchers found lower completion among learners from Arab countries than among learners from other countries at the same level of human development, so academies in Egypt and the Gulf have no reason to assume they are immune.
Completion matters to the business, not only to teaching. In the academies we spoke to, the students who finished were the ones who left reviews, came back for the next course and recommended the academy to friends. A student who stops at lesson three does none of that.
Why do students drop out of online courses?
Usually for ordinary, fixable reasons. The same Science paper notes that attrition "typically remains high in the first two weeks of a course", which tells you where to look first. The causes we kept seeing in academies were:
- They never start. Buying takes a minute; starting needs time, a quiet moment and a clear first step. If access arrives hours after payment, momentum is gone. We explain how to fix that in our guide to collecting course payments with automatic enrolment.
- Lessons are too long. A two-hour recorded session uploaded as one video asks for attention most people do not have on a phone after work.
- There is no clear path. A folder of forty videos with no order or stated outcome leaves students guessing what to watch next.
- They lose their place. Interrupted on a phone, they come back days later and cannot find where they stopped, especially when lessons live in a long chat, a problem we cover in the hidden costs of running an academy on WhatsApp and Google Drive.
- They get no feedback. Watching without checking understanding feels like progress until the first real task.
- There is no finish line. Nothing marks the end, so there is no reason to push through the last modules.
- Nobody sees them leave. Without data, the academy notices a dropout only when a refund request or a complaint arrives.
Do shorter lessons really keep students watching?
Yes, and this is one of the best-measured findings in online learning. Philip Guo, Juho Kim and Rob Rubin studied 6.9 million video watching sessions in open online courses for a peer-reviewed 2014 paper. Video length was, in their words, "by far the most significant indicator of engagement": the median engagement time was "at most 6 minutes, regardless of total video length", and students often made it "less than halfway through videos longer than 9 minutes." Their recommendation was to "segment videos into short chunks, ideally less than 6 minutes."
In practice, that means you do not need to re-record a whole course. Cut each long session into lessons of one idea each, give each lesson a title that names the skill, and remove the warm-up and the off-topic minutes. A course of thirty short lessons feels lighter than five long ones, even when the total time is the same.
Why do quizzes help students learn and finish?
Because recalling information strengthens memory more than reading or watching it again. In a well-known study, Roediger and Karpicke (2006) found that students who were tested repeatedly on a passage recalled 61% of it a week later, compared with 40% for students who only restudied it, even though the tested group read the passage 3.4 times and the restudy group 14.2 times. The students who only reread also felt more confident they would remember, which is exactly why rewatching feels productive when it is not.
The effect also holds for online video. In a study published in PNAS in 2013, Szpunar, Khan and Schacter split a 21-minute online lecture into four segments. Students who took a short test after each segment reported mind wandering on 19% of attention checks, compared with 41% for students who were not tested, and they scored 90% on the final test against 68% for the untested group.
Good course quizzes are short and frequent: three to five questions after each lesson, asking students to recall or apply an idea rather than recognise a phrase, with the right answer explained afterwards. Mix in a question from an earlier lesson now and then, so older material stays fresh.
Do certificates make students finish?
A certificate alone will not save a weak course, but a clear finish line and a real commitment help. In the Reich and Ruipérez-Valiente data, learners on the paid verified certificate track completed at 46.02%, against 3.13% for all participants. Be careful with that gap: those learners chose to pay, so they were more committed from the start, and the certificate is only part of the story. What academies can take from it is that a goal students care about, and have invested in, changes behaviour.
To make a certificate work as a goal, issue it only when every lesson is completed and the final quiz is passed, put the learner's name and the course on it, and make it verifiable so employers and clients can check it is genuine. Then tell students about it on day one, not on the last day.
How do you find the lessons where students drop off?
You measure it, lesson by lesson. A drop-off report shows how many students reached each lesson. A gentle decline is normal; a sharp fall after one specific lesson tells you that lesson is the problem. Typical fixes are to split it, add a worked example, move a hard exercise later, re-record a confusing part or cut material that does not serve the outcome.
Track a few numbers every week:
- Completion rate for each course and each cohort.
- The share of new students who open the first lesson within their first week.
- Drop-off by lesson, with special attention to the first two weeks.
- Quiz scores by question, to find concepts that most students miss.
Change one thing at a time, then compare the next cohort with the last. Without this data, every course change is a guess.
What is a practical plan to raise course completion?
Match each cause to a sign you can see and a fix you can make. None of these fixes needs a specific platform:
| Why students stop | What you see | What fixes it |
|---|---|---|
| They never start | Paid students who never open the first lesson | Instant access after payment, a short first lesson with a quick win and a personal welcome from your team |
| Lessons are too long | Students stop partway through videos | Lessons of about 6 minutes, one idea each |
| No clear path | "What should I watch next?" questions | Modules in a fixed order with a stated outcome for each |
| They lose their place | Students restart or skip lessons | Saved progress, so they resume where they stopped |
| No feedback | Low scores on the final exam | Short quizzes after each lesson with explained answers |
| No finish line | Students stall before the last module | A verifiable certificate issued only on completion |
| Nobody sees them leave | Refund requests come as a surprise | A weekly completion and drop-off report, and a call or message from your team to students who stall |
How does HBS LMS help students finish?
HBS LMS is a course platform that HBS Tech builds under your academy's brand and hands over to you. It supports each fix above:
- Instant start. Students are enrolled automatically when payment completes, and can sign in with Google on a platform designed to work first on phones.
- Clear structure. Courses are organised into lessons, and files and attachments sit beside each lesson instead of in a separate folder.
- Saved progress. Progress is saved, so students resume where they stopped instead of searching for their place.
- Quizzes and exams. Add quizzes after lessons and a final exam to check understanding.
- Certificates on completion. The completion certificate becomes available only when the course is finished, and each certificate carries a verification link.
- Completion and drop-off reports. See the completion rate for each course and where students drop off, so you know which lessons to fix.
Following up with students who stall is a job for your team, and the drop-off report tells them who to call and which lesson to fix. Your completion data stays yours too, as we explain in our article on owning your brand and student data.
This article is part of our HBS LMS series on the real problems of running an online academy. Start with the opening article, why we built HBS LMS and the problems we kept seeing.
Want to see where your students stop and help more of them reach the end? Tell us about your courses and we will show you how it works. Explore HBS LMS.
