Podcast File Organization That Survives a Full Season
A producer wrote to me in September asking for help finding ninety seconds of audio. A guest from episode 14 had since become mildly famous, and a news programme wanted the clip. The show was 70 episodes old.

A producer wrote to me in September asking for help finding ninety seconds of audio. A guest from episode 14 had since become mildly famous, and a news programme wanted the clip. The show was 70 episodes old. Every file from all 70 sat in one cloud folder called Podcast, and the names ranged from audio.wav to zoom_0.m4a to Final FINAL v3 USE THIS.wav.
We found it. It took two hours and forty minutes, most of which was spent opening files to hear who was talking. That is the real cost of disorganised audio, and it never shows up as a line item: not lost files, just hours spent re-listening to work you already did. I came to podcasting from documentary editing, where a messy project gets expensive by the second week, so the habits below will look fussy for a show with nine episodes. Run them anyway. The payoff arrives at episode forty.
Why audio projects rot faster than video ones
Video editors inherit structure whether they want it or not. Premiere and Resolve both insist on a project file in a location, both generate cache and proxy folders, and both break loudly when something moves. The nagging teaches you discipline.
Podcast tools do the opposite. Riverside, Descript and Zencastr all hand you a download with a name they chose, usually built from a session ID and a track number, and the file lands wherever your browser puts things. Nothing breaks when you leave it there. The show still publishes. So the structure never gets built, and the only index of what happened in which episode lives in the producer's memory, which is reliable for about eight months.
The second reason is that audio files look interchangeable. A folder of 34 WAV files gives you no thumbnails and no visual cue about which is the host and which is the guest. With video you can at least glance. With audio you have to press play, and pressing play is what ate those two hours and forty minutes.
One folder per episode, and what belongs in it
The unit of organisation is the episode, not the month and not the season. Everything that touches an episode lives in one folder, including the things that feel like they belong elsewhere, such as the guest's headshot and the signed release. Here is the template I set up for shows I consult on.
| Subfolder | What goes in | Keep for |
|---|---|---|
| 01_raw | Untouched recorder and platform downloads, one file per microphone | 90 days, then thin out |
| 02_session | The editor project file and its sidecars | Life of the show |
| 03_assets | Music, stings, guest headshot, sponsor read script | Life of the show |
| 04_exports | The published master, plus the clip exports you cut for social | Forever |
| 05_text | Transcript, show notes draft, chapter list, guest release | Forever |
Numbered prefixes exist so the folders sort in working order rather than alphabetically. It looks like a small thing. It means anyone covering for you opens the right folder first instead of landing in assets and wondering where the audio went.
Keep the transcript in the episode folder even if your host also stores it. Transcripts are the only searchable index of what was actually said, and a plain text file sitting next to the audio turns a two hour hunt into a search across folders that returns a hit in four seconds.
The naming rule, which matters more than the folders
Three parts, always in the same order: date, episode number, role. A file called 2026-04-18_ep014_guest-mcallister.wav tells you everything from the file list, with no player involved.
Date first, in year month day order, because that is the only format that sorts correctly in every operating system and every cloud client. Episode number zero padded to three digits, because ep9 sorts after ep80 and you will not notice until the list is long enough to be annoying. Role last, as a word rather than a number, because track 2 means nothing in six months while guest-mcallister means something forever.
Rename on import, not later. The thirty seconds it takes while the download is fresh is the cheapest thirty seconds in the whole workflow, and the knowledge of which track is whose has a half life of about a day. If you record locally on both ends, which is the approach I argue for in the guide to recording a remote podcast guest, you will have two files arriving at different times from different machines, and that is exactly the moment confusion sets in.
One more rule, unglamorous and worth it: never put a space, a slash or a hash in an audio filename. Spaces break command line tools and batch scripts, slashes break on upload, and a hash can truncate a URL if the file ever gets served. Underscores and hyphens cost you nothing.
Why keeping every raw file forever is the wrong default
The advice you hear is that storage is cheap, so archive everything and never think about it. The first half is true. The conclusion does not follow, because the scarce resource is not disk space, it is the ability to find one file among thousands under time pressure.
Run the arithmetic on a normal show. A 90 minute two microphone session at 48 kHz and 24 bit comes to roughly 1.5 GB of raw audio, and that is before the safety recordings, the video track nobody asked for and the three aborted takes at the top. Seventy episodes puts you near 110 GB, most of it duplicate material that was never used. A 4 TB drive holds that easily. A human scanning for the right file does not.
What I do instead is thin out at 90 days. Once an episode has published and the correction window has closed, the raw multitrack has done its job. I keep the edited session, the stems if the show uses music, the published master as a 48 kHz WAV, and the transcript. The untouched platform downloads go. Over two seasons that cut a client's archive from 180 GB to 31 GB, and the useful part of it now fits on a drive they can take home.
There are three exceptions I make without argument: anything with a legal question attached, anything where a guest asked for a section to be removed, and any interview with someone who has since died. Those raws stay. Everything else is a convenience copy of work you already finished.
The ten minute close out
Rename and file the raws. Straight after recording, before the adrenaline fades. Three files, three names, one folder.
Export the master twice. One compressed file for the host and one 48 kHz WAV for the archive. The second one is what you will need in two years when a radio producer asks, and a 64 kbps mono MP3 from your publishing pipeline will not survive being cut up and remastered. The level to hit on both is covered in the piece on podcast loudness targets before upload.
Write one line in a plain index file. Episode number, date, guest name, two or three topics. A text file at the top level of the show, appended to every week. This is twenty seconds of typing that replaces most of what people pay a database to do.
Check the second copy exists. Cloud sync is not a backup, because a deletion syncs too. One local drive plus one offsite copy, and once a quarter open a file from the offsite copy to prove it opens. I have seen two backup sets that had been silently writing zero byte files for months.
Set the folder template up once as an empty skeleton and duplicate it for each new episode. The whole routine takes about ten minutes an episode, and the first time somebody asks for a clip from two years ago you will get it back in one afternoon. Other workflow pieces in this vein sit in the podcasting section.
FAQ
What is the best folder structure for a podcast?
One folder per episode with numbered subfolders for raw audio, the editing session, assets, exports and text. Episode as the unit beats organising by month or by file type, because everything you need during a reissue sits in one place.
How should I name podcast audio files?
Date in year month day order, then the episode number padded to three digits, then the role of the track as a word. Rename at import rather than later, and avoid spaces and slashes entirely.
Should I delete raw podcast recordings?
Most of them, about three months after publication, once the edited session and an archive quality master are safe. Keep raws tied to legal questions, guest removal requests or interviews you cannot repeat.
Is cloud storage enough of a backup for a podcast archive?
No, because an accidental deletion or a corrupted file syncs along with everything else. Keep a local copy and an offsite copy, and open a file from the offsite set every few months to confirm it is readable.
More in Podcasting
Podcasting
Podcast Music Ducking That Keeps Speech Clear and Even
The most common note I give on a new show is not about the microphone. It is about the music under the host's first sentence.
Podcasting
Podcast Cover Art Size and Design That Reads in Apps
Your show sounds good, but in the app it shows up as a blurry square with text nobody can read.
Podcasting
Podcast Mic Bleed and How to Stop It in a Shared Room
Two people at one table, two microphones, and every word from your co-host shows up a fraction of a second late in your own...