Podcast highlight reels: how I cut the best 6 minutes out of a 90 minute episode

A podcast highlight reel is a short edit, usually 3 to 8 minutes, that strings together the strongest moments of one episode or one season into something a new listener can watch cold. It is not a trailer and it is not a clip. A trailer teases and withholds. A clip is one idea, 30 to 90 seconds, built for a feed. A highlight reel delivers actual value on its own: three or four complete thoughts, in an order that makes sense, with enough context that nobody feels dropped into the middle of an argument.

The distinction matters because it changes the edit. A trailer can cheat with music and fast cuts. A highlight reel cannot, because people are going to sit and watch it. If a segment does not earn its 90 seconds, it comes out.

There are two flavours I get asked for. Episode reels condense one conversation. Season or channel reels pull from ten or twenty episodes to make a single "this is the show" piece for a sponsor deck, a website header, or a pinned video. Season reels take longer because someone has to watch everything first, and that is billed as multiple episodes rather than one.

I look for moments where the guest stops performing. That is the whole method, and everything else is bookkeeping.

In practice my editor marks four things on the first pass through the episode. One, the answer that surprised the host, which you can usually see on the reaction camera before you hear it. Two, the specific number or name, because concrete beats abstract every time. Three, the story with a beginning and an end that survives being lifted out. Four, the line where the guest contradicts something obvious, which is the closest thing a podcast has to a hook.

What I ignore: applause, laughter that is not attached to a joke the viewer heard, and the segment the host is proudest of. Hosts tell us their favourite bit is minute 40, and about half the time it plays flat out of context.

Nothing here is auto-cut by silence detection. Silence detection finds pauses. It does not find the moment a guest hesitates for two seconds before admitting something, which is frequently the best two seconds in the episode. That is why a human watches the whole thing, and it is the same reason our single-camera shows get the same treatment as the multi-cam ones.

Six minutes is my default, and I will argue for it. Under three minutes you are making a trailer and you should call it one. Over ten and you have made a shorter episode, which nobody asked for.

Six minutes fits roughly four segments at 60 to 80 seconds each, plus a cold open and a short outro card. That structure is boring on paper and it works. Cold open with the single best 12 seconds, no logo, no music swell. Then the segments, hardest hitting first, because retention on a reel falls off in the first 30 seconds and never recovers if you open soft.

For sponsor decks and website headers I go shorter, 90 seconds to 2 minutes, because the viewer is skimming and has somewhere else to be. For YouTube as a standalone upload I go longer, 8 to 10 minutes, because watch time is the currency and a reel that holds for 8 minutes tells the algorithm more than one that ends at 3.

If you are unsure, pick six and cut down. It is much easier to remove a segment than to find one you already threw away.

A reel built alongside the full episode edit is the cheapest way to get one, because the editor has already watched the footage. Starter is $199 per episode for single camera, 1080p, one revision round. Studio is $299 per episode for 2 to 4 cameras, 4K, and unlimited revisions. That is per episode up to 60 minutes as the base, less for episodes up to 30 minutes and more for episodes up to 90.

A reel is a separate deliverable on top of that. So is packaging: vertical clips are $25 each, show notes and chapters are $35, and rush 24 hour delivery is $75 if you need it before a launch.

If you are running weekly, the packs matter. A 4-episode pack saves 15% and monthly saves 25%, which is what most shows end up on once they stop treating each episode as a separate decision. For the full breakdown of tiers and what sits in each one, see pricing.

Compared to what agencies quote, this sits low. The usual range I hear from hosts shopping around is $300 to $400 per episode plus something like $400 a month for clips. I go into that maths properly on the video podcast editing agency page, including where an agency genuinely beats me.

Make it yourself if you have four spare hours per episode and you enjoy the work. That is the honest answer, and some hosts do enjoy it.

The part people underestimate is not the cutting. It is the watching. To pick four great moments from 90 minutes you have to sit through 90 minutes with your attention on, usually twice. Then the assembly is maybe 90 minutes, colour and audio another hour, captions and titles another. Call it four to five hours for a good six minute reel, and the first three you do will not be good.

Hiring is not automatically right either. If your show is fortnightly and you are still finding the format, doing it yourself teaches you what your audience responds to faster than any brief I could write. The switch usually happens when the show goes weekly and the reel becomes the thing that gets skipped.

I wrote up the full trade, including salary maths and the handover problem, on podcast editing service vs in-house editor. The short version: it is a scheduling decision more than a money one.

Transitions are where most self-cut reels fall apart. Hard cutting between two segments recorded 40 minutes apart is jarring because the room tone changes, the guest has moved, and the energy resets. I bridge with a half second of B roll, a title card naming the topic, or a music duck. Cheap, invisible, and it fixes the whole feel.

Audio needs one pass across the finished reel, not per segment. If you level each clip separately you get four slightly different sounding rooms. Normalise the assembled timeline instead.

Colour matters more on a reel than on a full episode because you are jumping between moments and any drift is visible in seconds. Consistent skin tones across an hour of footage is a real job, and I cover how I approach it in podcast color grading.

Captions on by default. Most reels get watched muted first, on a phone, in a queue somewhere. Burned in captions, not the platform auto ones, because auto captions get names and jargon wrong and your guest will notice.

And whatever you build, the thumbnail decides whether it gets opened. My sister studio ClickTheory does 8 thumbnails for $250, and there is a longer walkthrough of what works on the podcast thumbnail guide.

The reel should be one output of a single edit pass, not a separate project. When my editor is already in the timeline with the full episode assembled, marking reel moments costs minutes. Starting from scratch two weeks later costs hours.

So the order I run is: full episode edit first, then the reel from the same marked timeline, then clips from the reel segments that overperformed. Each stage inherits the colour, audio, and captions from the one before it, which is why the whole set looks like it came from the same show. Because it did.

First cut of everything comes back in 48 hours. Revisions land the next working day. You send notes with timecodes, I make the changes, nobody schedules a call about it.

One thing that is not negotiable on my side: you own the files. Footage, project files and exports are yours, and I keep them alongside you rather than behind a login. If you leave, you leave with everything, including the reel timeline so your next editor can rebuild from it. I have been on the other side of that and it is a miserable way to lose a year of work.