Every guide, in one list

Courses built from what the instructor knows are comprehensive. Courses built from an outcome are finishable.

Length is the single biggest predictor of completion, and it is the variable instructors are least willing to cut.

For a wider operational comparison beyond a single course, this resource looks at coordinating distributed teams and their workload.

Most stuck learners are not stuck on the current lesson. They are stuck on something assumed three lessons ago.

Video is the part instructors spend most of their time on and the part that contributes least to whether anyone can do anything afterwards.

For open web standards and technical reference material, the World Wide Web Consortium maintains public documentation used across the web.

Most course quizzes measure whether someone was recently exposed to a sentence. That is not the same as measuring capability.

Everything in a digital course dates, but not at the same rate. Structuring for that at the start is what makes updates affordable.

Live sessions create accountability and exclude people. Asynchronous includes everyone and abandons them. Most courses need a specific mixture.

A cohort is not a list of people who bought the same thing at the same time. Making it a group is deliberate work.

Learners form a judgement about whether a course is for them within the first hour, and it is very difficult to revise afterwards.

Silence in a course community is usually a design outcome, not a personality trait of the people in it.

Individual feedback is the highest-value thing a course offers and the first thing that breaks as numbers grow.

Attrition is not evenly spread. It concentrates in a predictable window, which means it can be planned for.

The usual explanation is motivation. The usual cause is that the hours were never going to be there.

Advertised duration describes the content. Actual duration describes the learner. The two differ by a factor most course pages would rather not print.

The plan that fails is the one that assumes a good week. Most weeks are not good weeks.

Almost everyone overestimates how much they studied. A fortnight of honest records usually explains the whole problem.

The right length depends on what kind of work it is, and most advice ignores that entirely.

Reviewing at intervals is one of the few study techniques with strong evidence behind it, and one of the easiest to over-engineer.

The hardest stretch is not the middle. It is the point where the course is nearly done and no longer urgent.

Price determines how much attention each learner can have. Deciding them separately produces a course that cannot deliver what it sold.

Most training operations run on people who teach alongside another job. The schedule has to be built around that rather than in spite of it.

Support load is not proportional to cohort size. It is proportional to how unclear the course is, which means most of it is preventable.

Completion rate is the number most often quoted and most easily arranged. The denominator is where the arranging happens.

A generous guarantee reduces purchase hesitation and removes the commitment that makes people finish.

Internal training has a captive audience, which removes the marketing problem and replaces it with a scheduling one that is harder.

A folder, a payment link and an email list run a course perfectly well up to a point. Knowing that point saves a subscription and a migration.

Production quality has a floor that matters and a ceiling that does not. Most first-time instructors spend heavily above the ceiling.

The tool cannot decide what is worth assessing. It can decide whether marking fifty submissions is a day or an afternoon.

Course economics are decided by hours that nobody counts: preparation, marking, support and the endless small administration.

Course tooling is evaluated by the instructor and used by learners. Those are different tests and only one of them usually gets run.