Most people who scroll past your short have never seen your channel. They haven’t heard the forty minutes of context before the moment you clipped, and they don’t know your guest. They have about two seconds and one question: is this worth my time?
So the job isn’t to find your favourite part of the episode. It’s to find the parts that still work for a stranger. These are the rules I rely on, learned from our early pilot edits.
Look for complete ideas, not great quotes
A great line with nothing around it usually falls flat. A complete idea has three parts:
- A setup: the question, the claim or the situation.
- A turn: the surprise, the disagreement, the story beat.
- A payoff: the punchline, the answer, the line people will repeat.
When you scan a transcript, don’t just mark punchy sentences. Mark whole exchanges where all three parts happen within roughly half a minute to a minute. That’s your shortlist.
A good test: read the clip’s transcript on its own, without the episode. If you need to explain what’s going on first, it isn’t a complete moment yet. Either widen the cut to include the setup, or move on.
Open on the strongest line
The best clips open in the moment, not before it. If the payoff line also works as a hook, a “cold open” helps: start with that line, then play the exchange that leads to it.
What to cut from the start:
- greetings, “so, um, anyway” and throat-clearing
- the host re-reading the question the audience just heard
- the bit where someone asks for a sip of water
What not to cut: the one sentence of setup that makes the payoff land. It’s tempting to go shorter and shorter, and it’s the most common way to ruin a good clip. In one of my early pilots I cut clips to fifteen or twenty seconds to make them punchy, and the creator told me straight away that the setups were gone. They were right. Going back to 30–45 second clips, with the context left in, fixed it.
End on the point, not after it
Clips die in their last three seconds as often as their first. End right on the payoff, the laugh or the final word of the answer. Don’t end:
- on dead air or a reaction that trails off
- on a transition (“alright, let’s move on”, a stinger, a segment bumper)
- halfway through the next thought
If the payoff is a laugh, leave a beat for it. If it’s a statement, cut close.
Let the content set the length
There’s no single correct length. A comedic exchange might work at twelve seconds. A story or an argument might need fifty. Different creators have different preferences too: one might want every clip in the 30–45 second range, while another is happy with anything from 8 to 45 seconds as long as it stands on its own.
Pick a range that suits the creator and the format, then let each moment find its natural length inside it. Stretching every clip to the same duration is how you end up with padding.
Skip the parts that only work in context
Some segments are great in the episode and useless as clips:
- Explanations of rules or formats (“so for this round we each pick…”).
- Stat readouts and lists with no reaction attached.
- Inside jokes that depend on an earlier bit.
- Moments you’ve already clipped, or that appear in the episode’s own cold open. Check what’s already been posted.
Keep a couple of spares
When you choose clips for someone else, plan for a veto or two. They might not want a certain take out there on its own, or a guest might not like a moment. Picking one or two backups at the start is much cheaper than hunting again later.
A quick checklist
Before a moment makes the cut, it should pass all of these:
- Makes sense to someone who’s never seen the channel
- Has a setup, a turn and a payoff
- Opens on something worth staying for
- Ends cleanly on the point
- Doesn’t depend on an inside joke or an earlier segment
- Hasn’t already been posted as a short
Finding the moment is only half the job. Framing it, captioning it and titling it the way the creator would is the other half. That’s a topic for another post.