How Long Should an Online Course Be? Finish-First Guide
The honest answer to “how long should an online course be?” is that length is the wrong number to optimize. A 90-minute course people finish beats a 12-hour course they abandon in week one. The metric that predicts whether your course sells, gets refunded, or earns a word-of-mouth recommendation is not its runtime — it’s its completion rate. So the real question is narrower and far more useful: how long can your course run before people stop finishing it? That depends on three things — the promise you made, who your students are, and how you cut each lesson — and the published research gives us clear guardrails for all three.
Length is a proxy; completion is the target
Creators fixate on total hours because marketplaces and buyers treat hours as a shorthand for value. It is a poor shorthand. Padding a course to “look substantial” is the fastest way to lower your finish rate, and a low finish rate quietly taxes everything downstream: fewer testimonials, more “this wasn’t what I expected” refunds, and no second purchase from the same student.
Flip the metric. Decide the outcome a student should reach, build the shortest honest path to it, then track how far people actually get. If you measure only one thing after launch, measure completion — it is the single clearest signal that your length fits your audience. It is also why we steer creators toward a finishable structure and a visible course completion milestone rather than a longer syllabus. A student who crosses the finish line is a student who writes the review and buys the next thing.
What the data says about attention and length
Two bodies of evidence are worth knowing before you put a stopwatch to your lessons, and they point in slightly different directions.
First, general video behavior. Wistia’s State of Video analysis, drawn from more than 13 million videos, found that clips under a minute average roughly a 52% engagement rate, that engagement declines noticeably after the five-minute mark, and that crossing the 30-minute line brings a further measurable drop in how much of a video people watch. The blunt takeaway: the longer a single video runs, the smaller the share of it anyone sees.
Second, and more relevant to teaching, education-specific research. A large study of MOOC videos covering 6.9 million views found that most students stopped watching after about six minutes — regardless of how long the video actually was. That is the origin of the popular “keep videos under six minutes” rule. But those were free, low-commitment learners. A separate study of enrolled university students found roughly 90% completed their videos, with median viewing sessions of about 12 to 13 minutes. The gap between those two numbers is the most useful fact in this entire debate.
| Audience studied | Scale | What viewers did | Practical lesson ceiling |
|---|---|---|---|
| Free MOOC learners (edX study, Guo et al.) | 6.9M video views | Most stopped around the 6-minute mark | Keep lessons ≤6 min |
| Enrolled university students (Stanford, Lagerström) | College cohort | ~90% finished; median session ~12–13 min | Lessons up to ~12–13 min hold |
| General web video (Wistia State of Video) | 13M+ videos | <1 min ≈ 52% engagement; decline after 5 min; drop past 30 min | Avoid single 30+ min videos |
Sources: UC San Diego’s summary of the Guo and Lagerström instructional-video studies and Wistia’s State of Video analysis of optimal video length (both accessed October 2026).
Read together, the research does not say “shorter is always better.” It says attention tracks commitment. That hands you the two levers you actually control.
Lever one: match total length to the promise
Set total course length by the size of the transformation you promised, not by a runtime you are trying to hit. Three rough brackets cover almost every course:
| Course type | Total length | Lesson length | Best for |
|---|---|---|---|
| Quick-win / mini-course | 30–90 min | 3–8 min | One specific result; lead magnet or tripwire |
| Skill course | 2–5 hours | 6–12 min | A concrete capability (e.g., “run Facebook ads”) |
| Transformation / flagship | 6–15+ hours | 8–15 min, chunked into modules | A role or identity shift at a premium price |
Notice the pattern: as the promise grows, total length grows but per-lesson length barely moves. You earn the right to a longer course by breaking it into more short pieces, not by making each piece longer. A flagship program is not one long course; it is many finishable lessons stacked toward a bigger outcome. This is where deliberate curriculum design pays off — the modules carry the length while individual lessons stay watchable.
Lever two: set lesson length by who is watching
Here is the practical translation of that MOOC-versus-enrolled gap. If your audience is cold — free webinar attendees, a first mini-course, people who have not paid — design as if you have six minutes of attention per lesson. Front-load the payoff, cut the throat-clearing, and end each lesson on a small, visible win so the next click feels earned.
If your audience is warm — people who paid real money and enrolled with intent — you have more room. Ten-to-thirteen-minute lessons are fine, because buyers behave like the enrolled students in the research, not like free browsers. A paying student will sit through a longer walkthrough when it is clearly moving them toward the result they already bought.
The expensive mistake runs in both directions: building a paid flagship with lessons so short they feel thin and incomplete, or padding a free mini-course with lessons so long that nobody finishes and nobody upgrades. Price and intent, not the genre of the topic, decide the ceiling.
How to right-size a course before you record
You do not have to guess your way to a number. Four steps get you to a length you can defend:
1. Start from the outcome and a tight outline. List the minimum steps a student must complete to reach the promised result. Anything that does not move them toward it is a candidate for cutting. A disciplined course outline is the cheapest place to find the fat before you have wasted hours recording it.
2. Time each outcome, not each topic. Estimate how many minutes of genuine instruction each step needs. Add them up — that sum is your honest total length, before any padding creeps in.
3. Cut to the attention budget. Split anything over your per-lesson ceiling (roughly six minutes cold, up to about twelve warm) into separate lessons with their own titles. A 20-minute lesson is usually three lessons wearing a trench coat.
4. Ship, then let completion edit the course. After launch, watch where people drop off. A lesson with a cliff in its retention curve is too long, badly placed, or out of order — fix that one lesson instead of rewriting the whole course. Length is not a decision you make once; it is a dial you tune against real behavior.
The length your students actually finish
There is no universal correct runtime, and anyone who hands you one without asking about your promise and your audience is selling a template, not advice. Set total length by the transformation, set lesson length by commitment — roughly six minutes for cold audiences, up to twelve or thirteen for paying ones — and then let completion data trim whatever is left. Build the shortest course that still delivers the result you promised, and you will beat every bloated “complete masterclass” on the one metric that compounds: students who finish, and come back for more.
