The short answer
Keep digital genealogy files in a shallow folder structure, assign each source a stable identifier, and begin every filename with that ID. Preserve the original download or scan unchanged, create separate working copies, store full details in a source index, and maintain verified backups in separate locations.
A digital genealogy archive fails when the filename, folder and tree each hold a different version of the truth. `grandma birth cert final.jpg` may feel descriptive today, but it does not say which grandmother, where the image came from or whether the identification was later corrected.
A durable filing system uses filenames for stable identity and an index for changing conclusions.
Start with an inventory, not mass renaming
Before moving files, list where genealogy material currently lives:
- computer folders and Downloads;
- cloud storage;
- genealogy software media folders;
- phone scans and photographs;
- USB drives, CDs and older disks;
- email attachments;
- files shared by relatives;
- exports from online trees.
Copy important material into a safe staging folder before reorganizing. Do not delete the scattered originals until the new archive is complete, backed up and checked.
If old media may contain malware or unreadable formats, work from a copy and seek technical help rather than experimenting on the only disk.
Use a shallow top-level structure
Start with a small number of stable functions:
```text Family-History/ 00_Admin-and-index/ 01_People-and-families/ 02_Sources/ 03_Research-projects/ 04_Photos-audio-video/ 05_Exports-and-sharing/ 99_Inbox-to-process/ ```
The numbered prefixes keep the structure in a predictable order. The Inbox is temporary; it should be emptied during a regular closing routine.
Avoid nesting ten levels deep. A source may concern a person, spouse, child and household at once. Keep one authoritative file in `Sources` and point to its ID from family and project notes.
Assign a stable source identifier
Use a sequence that does not depend on a conclusion:
`SRC-2026-0042`
Put that ID at the beginning of every related file:
```text SRC-2026-0042_marriage-register_original.jpg SRC-2026-0042_marriage-register_transcript.txt SRC-2026-0042_marriage-register_translation-reviewed.pdf ```
The identifier stays fixed if you later discover that “Ellen Carter” was a different person. Correct the catalogue and conclusion; do not break every link by changing the ID.
Use filenames for retrieval, not the full citation
A practical filename pattern is:
`sourceID_recordType_place_date_role_version.ext`
Examples:
```text SRC-2026-0042_marriage-ohio-1888_original.jpg SRC-2026-0043_census-franklin-ohio-1900_original.jpg SRC-2026-0043_census-franklin-ohio-1900_detail-carter.jpg ```
Use a consistent date order such as `YYYY-MM-DD` when the full date is known. Use only characters that work safely across your devices and backup systems. Keep filenames reasonably short; put long citations, people and interpretation in the source index.
Do not rely on capitalization as the only difference between two names. Avoid `final`, `latest` and `new`. Use a dated or numbered version such as `transcript-draft-02` or `translation-reviewed-2026-08-20`.
Keep originals and derivatives separate
The original is the file as downloaded or captured. A derivative is anything altered for access or analysis:
- cropped detail;
- contrast-adjusted copy;
- annotated image;
- transcript;
- translation;
- combined PDF;
- smaller sharing copy.
Never overwrite the original. Keep complete record pages even if only one entry interests you. Margins, headings, adjacent households and page numbers may become evidence later.
The National Archives recommends retaining physical originals after digitization because digital files have their own preservation risks and can be lost.
Maintain a source index
One row per source ID should record:
- complete source description;
- repository, collection and call number;
- volume, page, image or certificate number;
- date and jurisdiction;
- people and events represented;
- original filename and local path;
- access date and URL;
- provenance or person who supplied it;
- restrictions;
- derivative filenames;
- related research question;
- backup status.
This index is what lets a compact filename remain useful. Save it in a format you can export, open and back up independently of one proprietary application.
Separate people notes from source files
Person or family folders can contain:
- a current summary;
- research questions;
- timeline;
- links or source-ID list;
- correspondence about that family;
- a note explaining unresolved identities.
Do not copy the same census image into every person’s folder. Duplicate files accumulate different edits and make it impossible to know which is authoritative.
If your operating system does not support reliable shortcuts across backup devices, store a short text index of relevant source IDs instead.
Keep project work temporary and traceable
A research-project folder is for one question. It may contain:
- research plan;
- research log;
- working notes;
- comparison table;
- temporary extracts;
- draft conclusion.
When the question closes, move authoritative sources into the source collection, update the tree and person summaries, retain the final reasoning note, and remove unnecessary temporary duplicates.
→ Use a genealogy research log
Handle photographs and interviews as collections
Photographs and oral histories often contain richer metadata than a document filename can hold. Give each item or interview a collection-based ID and keep descriptive metadata beside it.
For example:
- `BOX01-PH-0042_front_original.tif`
- `OH-REED-2026-08-13-01_audio_original.m4a`
Use the source index to connect these items to genealogy claims without renaming them around a guessed identity.
→ Name scanned family photo files
→ Organize oral-history recordings and transcripts
Protect private material before sharing
Keep living-person information, DNA data, correspondence, private interviews and identifying records outside general share folders. Use `05_Exports-and-sharing` for intentional copies, not as another master archive.
Before sending an export:
- inspect it as the recipient;
- remove hidden notes and unrelated media;
- confirm permissions;
- state what the package contains;
- label the export date and version;
- avoid links that expire without warning when the recipient is part of your backup plan.
Back up files in separate places
Keep at least:
- working files on the main computer;
- a local backup on separate media;
- another trusted copy off-site.
Cloud sync is useful but may repeat accidental deletion. Confirm whether the service keeps version history. Encrypt sensitive material where appropriate and make sure a trusted person can understand the recovery instructions.
Twice a year, restore sample files to a temporary folder. Open images, PDFs, audio, the source index and genealogy-software export. Record the date, failures and fixes.
The Library of Congress personal digital archiving guidance emphasizes selecting, organizing, describing and making copies of digital materials rather than assuming storage media will preserve themselves.
A rescue workflow for a messy archive
Work in batches of 25–50 files:
- copy the batch into the Inbox;
- identify exact duplicates without deleting yet;
- separate sources, exports and personal correspondence;
- assign source IDs;
- preserve originals;
- create catalogue rows;
- rename and file working copies;
- verify backup;
- record what was removed and why.
Do not try to identify every unknown person before organizing the files. “Unknown woman; supplied by Ruth Carter; likely 1940s” is acceptable metadata. A confident invented name is not.
Connect the digital system to the complete research workflow
Return to How to organize genealogy research for the whole system. Use Genealogy Starter Tools for a no-cost research log and source sheet, or see the spreadsheet-based Family History Workbook if you need structured digital tracking.