The short answer
Give every interview one stable identifier and use it at the beginning of every related filename. Keep the unchanged original recording, an access copy, transcript, metadata log, permission note and related images in one interview folder. Record who, when, where, what and sharing restrictions, then keep verified copies in at least two separate storage locations.
An oral history can become unfindable even when the audio still exists. A filename such as `Voice Memo 47.m4a` does not tell a future relative who is speaking, when it was recorded, what permission was given or whether a transcript belongs to it.
The solution is not a complicated archive system. It is a stable identifier, a predictable folder, a small metadata record and backups you have actually tested.
1. Preserve the original recording first
Copy the file from the phone, recorder or video camera as soon as practical. Do not trim silence, normalize volume, rename speakers inside the file or overwrite the only copy.
Keep the original bytes as the preservation master. Make a separate copy for listening, editing or sharing. If you later improve the audio, label the result as a derivative rather than a replacement.
Record these technical facts before apps or devices change them:
- original filename;
- file format;
- file size;
- duration;
- recording device or app, if known;
- date copied;
- person who copied it.
The Oral History Association describes this as technical metadata: information about the digital file and how it was created.
2. Assign one interview identifier
An identifier should remain stable even if the title or your understanding changes. Use a short collection code, surname or narrator code, date and sequence:
`OH-REED-2026-08-13-01`
Here `OH` means oral history, `REED` identifies the collection or narrator, the date follows year-month-day order, and `01` distinguishes multiple sessions.
Do not put a sensitive topic in the identifier. Do not rely only on a person’s first name. Do not number files according to their current folder position; folders may move.
Record the identifier in the interview log and at the start of every filename connected to the session.
3. Use one folder per interview session
A practical folder structure is:
```text OH-REED-2026-08-13-01/ 01_original/ 02_access/ 03_transcript/ 04_metadata_permission/ 05_related_material/ ```
This separates preservation files from items you can safely edit or share. If your archive is small, the numbered subfolders are optional; the common interview identifier is not.
For several sessions with the same person, make each session a separate folder. Keep a collection-level index above them.
4. Name files by identifier, role and version
Use lowercase role words and ISO-style dates. Avoid spaces, punctuation that behaves differently across systems, and labels such as `final_final2`.
Examples:
```text OH-REED-2026-08-13-01_audio_original.m4a OH-REED-2026-08-13-01_audio_access.mp3 OH-REED-2026-08-13-01_transcript_draft.docx OH-REED-2026-08-13-01_transcript_reviewed.pdf OH-REED-2026-08-13-01_metadata.txt OH-REED-2026-08-13-01_permission.pdf OH-REED-2026-08-13-01_photo-01_front.tif ```
Use an explicit version or status for documents that change:
- `draft-01`
- `reviewed-2026-08-20`
- `share-family`
- `restricted`
The word `master` can be ambiguous when several people make edits. `original` is clearer for the untouched recording; `access` is clearer for the everyday listening copy.
5. Create a one-page metadata record
Descriptive metadata explains what the interview is. Record:
- interview identifier;
- narrator’s full name and preferred display name;
- interviewer’s full name;
- recording date and place;
- session number and duration;
- languages used;
- main people, places and subjects discussed;
- short contents summary;
- related photograph or document identifiers;
- whether facts are firsthand, reported or uncertain;
- permission and access restrictions;
- location of the preservation and backup copies;
- date the record was last checked.
The Oral History Association notes that descriptive metadata allows future users to know who was interviewed, when and where, and what the recording contains. Put the metadata beside the interview, not only inside one person’s private app.
6. Keep permission with the content
A recording without its permission context is incomplete. Store a clear note stating:
- the purpose agreed before recording;
- who may access the file;
- whether quotations or excerpts may be shared;
- whether living people’s names require redaction;
- whether the narrator reviewed a transcript;
- any later restriction or withdrawal;
- who recorded the permission and when.
Do not infer public permission from a private family recording. If a copy is prepared for a website, archive or publication, label that share version separately and retain the restriction note with every transfer.
→ Plan consent and privacy before recording
7. Link transcripts to time in the recording
A transcript becomes much more useful when a reader can return to the voice. Add timestamps at regular intervals and at major topic changes:
```text [00:00:00] Opening identification and permission [00:03:42] Lake Street house and neighbors [00:18:15] First job at the bakery [00:31:08] Photograph OH-REED-PH-0042 ```
Use consistent speaker names. Mark an unclear word as `[inaudible 00:14:22]` or `[uncertain: Marston?]` instead of silently inventing it.
Keep the transcript status visible: automatic draft, corrected transcript, narrator-reviewed transcript or edited reading copy. These are not interchangeable.
→ Transcribe an interview without losing the voice
8. Connect photographs without copying context into filenames
If a photograph appears in the conversation, assign it its own stable identifier. Put that identifier in the transcript and metadata:
“At 00:31:08, narrator discusses photograph `BOX01-PH-0042`.”
Do not rename the photograph `Dad-at-bakery-definitely-1954.jpg` if the identity or date remains uncertain. Record claims, sources and confidence in the catalogue.
→ Name scanned family photo files consistently
9. Separate preservation, access and public copies
One file should not serve every purpose.
Preservation copy: unchanged original recording, highest available quality, restricted from routine editing.
Access copy: convenient format for family listening and transcription.
Public or shared excerpt: only material covered by the agreed permission, with necessary redactions and a label explaining that it is an excerpt.
The transcript should also retain its level of editing. A readable family edition may remove false starts, but it must not be confused with a verbatim research transcript.
10. Back up the package and verify it
Keep at least two copies in different places; three copies using different storage types is safer for valuable family archives. A possible arrangement is:
- working copy on the computer;
- local backup on an external drive stored separately;
- encrypted cloud copy or another trusted off-site location.
A sync service is not automatically a backup: deletion or corruption may synchronize too. Check whether previous versions can be restored.
Twice a year, open a sample of recordings, transcripts and permission files from each backup. Confirm that filenames, duration and folder counts match the index. For a larger archive, record checksums so an unexpected file change can be detected.
11. Maintain a collection-level index
For more than a few interviews, create a simple spreadsheet with one row per session:
- interview identifier;
- narrator;
- date;
- main topics;
- duration;
- transcript status;
- permission level;
- folder location;
- backup verified date;
- follow-up status.
This is an index, not a substitute for the full metadata file. Its purpose is to help a relative discover what exists without opening every folder.
12. Package a copy another person can understand
Before sending an interview to a relative, open the package on another device. Include:
- a clearly named listening file;
- a readable transcript or contents summary;
- the narrator, interviewer and date;
- a note about permitted sharing;
- explanations for any related photo identifiers;
- your contact information for corrections.
Do not send only a cloud link that expires without warning. If the recipient is part of the preservation plan, confirm they downloaded and opened the files.
A 20-minute rescue for disorganized interviews
If you already have many scattered recordings:
- collect copies without deleting originals;
- make an inventory using the current filenames;
- listen to the opening and ending of each file;
- assign a stable identifier;
- record known people, date and restrictions;
- rename a working copy, not the only original;
- group related transcript and photographs;
- make a verified backup.
Unknown details can stay unknown. A file labelled “narrator probably Ruth; recorded before 2018; source: John’s old phone” is more trustworthy than a confident guess.
Build the archive from the moment you record
Use the family history interview checklist before your next conversation, or start with the free Family Story Starter Kit. For the whole workflow, read the family history interview guide.