Build the practice first, then the explanation
Video is the part instructors spend most of their time on and the part that contributes least to whether anyone can do anything afterwards.
Course production budgets — of money and of effort — go overwhelmingly into explanation. Recording, editing, slides, diagrams, re-recording because of a stumble.
The capability the course promises is not produced by any of that. It is produced by the learner doing the thing, repeatedly, with feedback. Explanation is what makes the practice possible; it is not a substitute for it.
The distinction between real engagement and superficial activity also appears in workplace monitoring; this page explains one way software tries to identify simulated input.
Design the exercise before the lesson
Reversing the usual order changes what gets built. Decide what the learner will do, then write only the explanation required to let them do it.
This is a strong filter. Material that does not support an exercise is material that has no immediate purpose, and most of it turns out to be background the instructor found interesting. It also keeps explanations shorter, because you stop when the learner has enough to start.
For broader evidence-based guidance on teaching and course design, the Carnegie Mellon Eberly Center publishes university teaching resources.
What can the learner do after this that they could not do before? If the answer is 'know about', the lesson is a reading and should be labelled as one.
Make exercises produce something
Multiple choice questions verify recall and produce nothing. An exercise that leaves the learner holding an artefact — a working script, a completed analysis, a document, a configured account — does three things at once: it demonstrates the capability, it is inherently memorable, and it is evidence the person can show someone.
Artefacts also solve the motivation problem more reliably than encouragement. A half-built thing pulls people back; a half-watched video does not.
Graduate the difficulty deliberately
Exercises tend to cluster at one level: worked examples that are too easy, or a final project that is too big. The gap between them is where people stall.
A workable progression is three stages. Guided, where the learner follows steps and sees the mechanism. Adapted, where the same task is presented with a change that requires them to think. Independent, where they are given a goal and no steps. Most courses provide the first and the third and omit the middle, which is where the learning actually happens.
Keep explanations short and near the work
Long videos are watched at speed, partially, and once. Explanation broken into short pieces placed immediately before the exercise that needs it is watched properly, because it has an immediate purpose.
Text has an underrated advantage here: it is skimmable, searchable and re-readable in ten seconds. A learner stuck at eleven at night can find one line in a transcript far faster than they can locate a moment in a fifty-minute recording.
Budget the learner's time by exercise
When estimating how long a course takes, count the practice rather than the runtime. A twenty-minute video attached to a ninety-minute exercise is a two-hour lesson, and it should be described that way.
Courses advertised by video length systematically understate the commitment by a factor of several, which sets up the mismatch that produces abandonment in week three.