Everyone Quotes "10 to 20% Completion." The Page That Published It Just Admitted It Has No Source.

You have said it. In a webinar, in a pitch deck, in a sales page explaining why your cohort program costs $1,497 instead of $97.

"Self-paced courses only get 10 to 20% completion."

It is the load-bearing statistic of the entire online education industry. And one of the pages that has been publishing it just posted an editor's note saying it cannot find where the number came from.

The note

Ruzuku's widely circulated statistics roundup, a page built from the company's production database of 32,000+ courses plus reports from Kajabi, Thinkific, HolonIQ and Statista, now carries a correction about its own figure. Of the 10-to-20-percent range, the page says: "we have no source for that range."

That is unusually honest. It is also awkward, because the same range is still sitting in the FAQ markup of multiple other pages on the same site, where "self-paced courses average 10-20% completion" is presented as settled fact.

And this is not one publisher's sloppiness. It is the whole genre. Trace the citation chain on almost any course-marketing blog and you land in the same place: a round number, attributed to nobody, repeated until it became true.

Where the real number comes from

The statistic that does survive scrutiny is 12.6%.

It comes from academic work revisiting MOOC completion, built from a dataset assembled using MOOC-aggregating websites such as Class Central and internet searches for enrolment and completion figures, covering a total of 221 MOOCs. Completion, defined as the percentage of enrolled students who finished, varied from 0.7% to 52.1%, with a median of 12.6%.

Read that spread again. Zero point seven to fifty-two point one.

The median is a real number. The range is the actual finding. A 75x gap between the worst course and the best one is not a benchmark, it is a warning that "completion rate" describes almost nothing on its own.

Most of the industry quotes the median and throws away the range, because the median makes self-paced learning look hopeless and the range makes it look like a design problem.

The denominator nobody mentions

Here is the part that should bother anyone who has used this stat to sell something.

Completion rate depends entirely on who you put in the denominator. Researchers comparing methods on the same courses found completion ranging from 1.4% to 50.1% when calculated as the percentage of active students who finished, versus the traditional enrolled-student method. In their own four-course sample, completion based on enrolled participants came to 12.30%, while completion based on active learners came to 23.47%.

Same students. Same courses. Nearly double the number.

Free MOOC enrollment is a click. Buying a $497 course is a decision. Comparing your paid program's finish rate to a free Coursera catalog is not a benchmark, it is a category error that happens to flatter you.

The numbers with receipts

If you want figures that come with a methodology attached, there are three worth knowing.

12.6% is the MOOC median, from 221 courses, with a documented range of 0.7% to 52.1%. Use it to describe free, open, no-commitment enrollment. Nothing else.

53.4% vs 41.9%. Across 32,000+ courses on Ruzuku, scheduled cohort courses achieve a 53.4% completion rate compared to 41.9% for open-access self-paced courses. That is a real gap, and it is far smaller than the 85-vs-15 folklore the cohort-course industry has been selling.

50.9% vs 37.3%. Across 26,632 courses and 1.3 million enrollments, measured per enrollment, courses with lesson-level discussions completed at 50.9% versus 37.3% without. Structure beats format.

The caveat matters, and the publisher states it plainly: all of those statistics describe courses and enrollments on the Ruzuku platform specifically, and are not nationally representative samples or randomized controlled trials. It is a paid-hosting population with instructor relationships attached. Of course it outperforms anonymous free marketplaces.

Why the fake number refuses to die

Because it is useful.

"10 to 20%" brackets 12.6%, so it never looks wrong. It is vague enough to be unfalsifiable and specific enough to sound researched. And it does exactly one job in a sales conversation: it makes self-paced look broken so that whatever you are selling, cohort, community, coaching, accountability, looks like the cure.

You can watch the inflation happen in real time. The same body of work gets cited as MOOC completion "averaging 3-6%" with self-paced marketplace courses at 10-20% on one page and as "12 to 15 percent, per Class Central's tracking" on another. Elsewhere the comparison gets stretched to cohort-based courses reporting 85-96% completion against 10-20% for self-paced, a claim that would represent the largest intervention effect in the history of education research if anyone could produce the study.

What to do before your next webinar

Delete the 10-to-20 slide.

Replace it with your own number. Not an industry benchmark, yours, measured per enrollment, with a stated definition of what "complete" means in your course. You have the data. It is in your platform dashboard, and it took the rest of the industry a decade of recycled statistics to avoid looking at it.

Then compare like with like. Free against free. Paid against paid. Cohort against cohort.

The industry has spent years explaining why students don't finish courses, using a statistic nobody can source, to sell solutions nobody has measured. One page just said the quiet part out loud. The rest of the citation chain is still live.