Forty lectures, eight weeks, one term start.

The term starts in eight weeks, there are forty lectures, and they are all recordings of a room. Somebody has asked whether they can be made presentable. The honest answer is that all of them can be made better and none of them can be made perfect, so the only useful question is what order to work in.

This is the order. It is drawn from what the evidence says changes behaviour, not from what looks most impressive in a review meeting.

First: the opening ninety seconds of every video

More than a third of all abandonment happens in the first 3% of a video's runtime. On a fifty-minute lecture that is the first ninety seconds — typically a room settling, a projector warming up and somebody asking whether the microphone is on.

Cutting the dead opening and replacing it with a title card naming the course, module and topic takes minutes per video and addresses the single largest loss point in the whole archive. Forty videos, one afternoon. Nothing else on this list has that ratio.

Second: audio, on all forty

Room echo and hum are the cheapest thing to fix and the most expensive thing to leave. In a controlled study where only the audio quality of a recorded talk was degraded, the same talk was rated substantially worse and the same speaker judged less intelligent. Your faculty are being assessed on a room acoustic they could not hear because they were standing in it.

It is also a compounding fix: clean audio produces better captions, which is the next item.

Third: captions and transcripts, on all forty

Not optional. Annexure VII requires a transcript alongside every e-Tutorial video, and Annexure VI Part III requires accessible content. If you skip this you have not produced a rough course, you have produced a non-compliant one.

It is also the item most likely to be separately budgeted already, so it may not compete for the same money as the rest of this list.

Fourth: segment the longest ten

Here the order becomes selective, because segmentation is the item that costs real editorial judgement — somebody has to decide where an idea ends.

Sort the forty by length and take the longest ten. Problem-attempt rates fall from 56% for videos under three minutes to 31% for videos over twelve, so the long ones are where the loss is concentrated. Cutting a fifty-minute lecture into five topic-length pieces does more than cutting a twelve-minute one into two.

Fifth: branding, if there is time

Logo, consistent lower-thirds, opening and closing bumpers. This is what makes a course feel like one course rather than forty files, and it matters — but it is fifth because a student who abandoned at ninety seconds never saw the logo either.

What to leave alone

• Animated b-roll. It is the most compute-intensive effect, it needs subject-expert review on every cutaway, and review capacity is the thing you have least of. Save it for the term when you are not triaging.
• Re-recording anything. Eight weeks does not survive a faculty scheduling round. Work with what exists.
• The flagship course. Counter-intuitive, but it is usually the one already in reasonable shape. The gain is in the neglected thirty.
• Perfection on any single video. Forty videos at seventy percent beats eight at ninety-five and thirty-two untouched.

The schedule

That last step is the one that ends the triage cycle. Every term somebody does this from scratch because nobody measured what happened last time.

Sources