Feedback when there are more learners than hours
Individual feedback is the highest-value thing a course offers and the first thing that breaks as numbers grow.
Ask people what made a course worth the money and the answer is usually feedback on their own work. Ask an instructor what consumes their time and the answer is the same thing.
The tension is structural: the most valuable element of a course is the one that does not scale, and every growth decision runs into it.
For a workplace example of turning activity records into trend data rather than a single headline metric, this guide covers analytics at team level.
Spend the scarce resource where it is scarce
Not all feedback needs a human. Anything that can be checked mechanically — does it run, does it produce the right output, does it meet the stated constraints — should be, leaving human attention for judgement: is this a good approach, what would a professional do differently, what habit is forming here.
Separating these two is the single biggest available saving, and most courses do not do it, so instructors spend their limited hours pointing out that something does not run.
For wider context on technology and learning, UNESCO publishes public resources on digital education.
A learner can find out whether their answer works without you. They cannot find out whether it is a good answer.
Feedback on a sample, transparently
Reviewing every submission from every learner is often impossible past a certain size. Reviewing a defined sample — one submission per learner per module, or the module where feedback matters most — is honest and workable.
What matters is that it is stated in advance. Learners who were told they get detailed feedback on two of six exercises are satisfied when they receive it; learners who expected six and got two feel short-changed by the same delivery.
Common-error feedback covers most of the ground
Across a cohort, submissions cluster: the same handful of mistakes account for the large majority. Writing those up once, well, and publishing them to everyone serves more learners than individual notes on the same points.
The write-up improves with each cohort and eventually becomes part of the course material, which is the point at which the feedback burden actually falls rather than merely being redistributed.
Peer review works if it is structured
Peer feedback has a poor reputation earned by unstructured versions of it, where learners are asked to comment on each other's work and produce encouragement.
With a rubric, specific questions and a required format, it works well, and it has a second benefit that is often larger than the first: evaluating someone else's work against criteria is itself a strong learning exercise, frequently more instructive than receiving comments.
Be specific and be brief
Long feedback is not better feedback. Three concrete points a learner can act on before the next exercise beat a page of observations that arrives after the moment has passed.
Timing dominates depth. Feedback delivered while the learner still remembers what they were thinking is acted on; the same feedback a fortnight later is filed.
Track how long it actually takes
Feedback time is the main variable cost of running a course and it is almost never measured. Instructors estimate it low, then find the week consumed.
Timing a few review sessions gives a real per-submission figure, which is what makes cohort size, pricing and staffing decisions possible. Without it, those decisions are made on an assumption that is usually wrong in the expensive direction.