A weekly content repurposing workflow

The principle: separate creation from distribution

A shoot produces a master. Distribution needs many smaller assets. Those are two different jobs with different mental modes, and doing them on the same day is why the second one gets skipped. Put them on different days and the derivatives become routine instead of inspiration-dependent.

The workflow below assumes one shoot a week and one person. A team can parallelise the same steps without changing the order.

The folder structure

project/
  00-shoot/          original camera files, untouched
  01-master/         the edit, exported once, clean, no watermark
  02-clips/          cut moments, vertical, captions burned in
  03-variants/       platform exports from the batch
  04-published/      what actually went out, with the date
  rights/            releases, licences, music receipts

The point of 04-published is not tidiness. It is the only way to answer “did we already post this?” three weeks later, and it prevents the accidental duplicate that no tool can fix.

The week

Day	Block	Output
Monday	Select	8–12 candidate moments, ranked, with in/out points
Tuesday	Edit clips	4–6 finished vertical clips with captions
Wednesday	Batch	Platform variants for the week, reviewed on a contact sheet
Thu–Sun	Publish	One post per day, plus one archive clip
Friday	Review	15 minutes: retention, follows, saves; note what to repeat

Two hours on Monday, two on Tuesday, one on Wednesday, and fifteen minutes on Friday. If that is too much, halve the clips rather than skipping the batch — a smaller consistent output beats a monthly sprint.

Four presets, not forty sliders

A preset is a decision you make once. Keep a small set and name them for the job:

• Vertical clean. Reframe, light grain, no colour shift. For talking-head clips.
• Vertical loud. Slightly stronger grain and contrast, for fast cuts and hooks.
• Square feed. Slower pacing, less grain, for carousels and static posts.
• Archive refresh. Wider ranges, trim, metadata rebuild, for older footage.

Export the presets as .rt-preset.json and keep them in the project folder. When a freelancer takes over, they load the files rather than guessing at settings. See presets for the mechanics.

What the batch step actually does

The clips are already edited. The batch step handles the mechanical differences between surfaces and between posts:

1. Reframe and resize for each destination.
2. Apply the preset with a different random draw per output.
3. Normalise audio and rebuild the container per copy.
4. Produce a contact sheet and a distance report.

This is where the video uniquifier fits. It does not choose clips or write captions; it makes the files defensible once those decisions are made.

The numbers worth keeping

A single sheet per week, not per post:

Metric	What it tells you
3-second retention	Whether the hook works. The first thing to fix.
Average watch time / length	Whether the clip earns its duration.
Follows per 1,000 views	Whether the clip builds an audience or just views.
Saves and shares	Whether the content is useful enough to keep.
Source (new vs archive)	Whether old footage still performs.

Review monthly, not daily. Daily numbers are noise; a month of retention data tells you which format to make more of.

The three failure modes

• No clean master. Every derivative inherits the problem: a watermark, a clipped mix, a bad export. Fix the master first.
• Twenty near-identical clips. Volume without distinct ideas reads as spam and bores the audience. Fewer, better clips win.
• Publishing without a record. The same clip goes out twice, or a paid variant contradicts an organic one. The 04-published folder exists for this.

Short answers

---

Next: batch video processing, or the podcast version.