Building a Data Management Workflow for a Small Crew: Who Backs Up What and When

The Short Film That Lost a Day Because Nobody Was Assigned to the Drives

A 5-person crew shoots a 3-day short film. There is no DIT. The DP shoots on a BMPCC 6K. At the end of day one, everyone assumes someone else has offloaded the cards. The director assumes the DP did it. The DP assumes the producer did it. The producer left early for a location permit. At 11pm, the 1st AC discovers the cards are still full. She connects a laptop, starts the copy, and goes to bed. In the morning, the copy completed but no checksum was run. No one verified the backup. On day two, a CFexpress card read error corrupts one clip. The unverified backup turns out to be the only copy. The corrupt file is unrecoverable. The production reshoots 45 minutes of coverage on day three instead of shooting the planned content.

The data management failure was not a technology problem. It was an assignment problem: no one was specifically responsible for the backup window, and no protocol existed for what "backed up" meant.

This post gives a complete data management framework for 2-5 person crews without a dedicated DIT -- covering who does what, when the backup window happens, what tools are required, and how to structure the process so the protocol survives the chaos of a real shoot day.

The Three Non-Negotiable Requirements

Before the workflow, three rules that apply to every production regardless of size:

Rule 1: Two copies before a card is formatted. A card is not safe to format until there are two verified copies of every file on it, on two physically separate drives. "Verified" means checksummed -- a bitwise confirmation that the destination file is identical to the source. A file copy that completes without error checking is not a verified copy.

Rule 2: The backup is one person's job. On a small crew, assign this responsibility explicitly before day one. It does not matter who -- the producer, the 1st AC, or the director -- as long as one named person owns the task and is not also expected to be doing something else during the backup window.

Rule 3: The backup window is scheduled, not improvised. The backup happens at the same time every day. Typically: during lunch (partial offload) and at wrap (full offload). If the wrap offload will take longer than the crew's availability, address this with the Codec Storage Calculator during prep -- know the offload time before the first shoot day.

The Minimum Hardware Setup

For productions using the Backup Strategy Calculator, enter the daily acquisition volume and the offload time window to confirm that the hardware setup can complete the backup within the available time. A BMPCC 6K shooting BRAW 5:1 for 6 hours generates approximately 475 GB per day. A USB 3.2 drive at 600 MB/s throughput copies 475 GB in approximately 13 minutes. A USB 3.0 drive at 100 MB/s copies the same data in approximately 80 minutes. Drive speed matters.

The Daily Workflow, Step by Step

At Lunch (Partial Offload)

Responsible: Designated data manager (producer or 1st AC on a small crew).

1. Receive all filled cards from the camera department. Note the number of cards and their labels in the camera report.
2. Copy each card to Primary Drive using checksumming software. The software generates an MD5 or SHA-256 checksum for each file during the source read and verifies the destination matches.
3. Once all cards are verified on the Primary Drive, copy the verified files from Primary Drive to Redundant Drive. Use the same checksumming tool to verify the second copy.
4. When the Redundant Drive copy is verified, the cards are safe to return to the camera department for formatting and reuse.
5. Log the card labels, file counts, and total data size in the production's daily camera report.

Do not return a card to the camera department at step 4 unless you have personally verified both copies. This is the single most important procedural step.

At Wrap (Full Offload)

Same process as lunch for all cards shot in the afternoon session. Additionally:

1. Confirm the total day's data against the expected volume from the daily shooting report.
2. Label both drives with the shoot day number and date. Keep Primary Drive and Redundant Drive in separate physical locations overnight: one with the producer, one with the DIT or camera department contact. Drives stored in the same bag that gets left in the same vehicle are not meaningfully redundant.
3. Generate a proxy transcode from the verified Primary Drive files if the editorial workflow requires offline proxies. This can run overnight. The film production storage guide covers the complete proxy workflow from DCI files to editorial-ready H.264.

Crew Assignment Map for a 3-Person Crew

For 2-person crews, the producer/director doubles as data manager. The protocol is identical but the lunch offload may need to happen during camera setup time rather than an actual lunch break.

Common Small-Crew Data Failures and How to Prevent Them

Failure: Cards formatted before backup verified. Prevention: The camera department never formats a card without a written sign-off from the data manager. A simple text message confirmation ("Day 2 cards verified, safe to format") is sufficient -- but it must happen before the camera is loaded for the next setup.

Failure: Both copies on the same drive or in the same bag. Prevention: Two drives, two locations, every night. This is the only step that protects against physical loss (theft, flood, fire). Two copies on a laptop and a drive in the same backpack is one theft away from being a single copy.

Failure: Proxy creation before backup is verified. Prevention: Never run proxy transcode from a card directly. Copy to Primary Drive first, verify, then run proxy transcode from the verified copy. Running a transcode from a card simultaneously with a copy is one of the fastest ways to introduce read errors.

Pro Tips

Tip 1: Name drives consistently before day one. A naming convention like [PRODUCTION]_A_D01 (project name, drive letter, shoot day) makes it impossible to accidentally overwrite the wrong drive or lose track of which drive holds which day's material. Create the naming convention in pre-production and label physical drives with tape before they go on set.

Tip 2: If the only available checksumming tool is a terminal command (md5sum on Linux/Mac, certutil on Windows), it works. Run md5sum filename on both the source and destination files and compare the output manually. It's slower than dedicated software but provides the same verification. For productions that don't have budget for SilverStack, this is the minimum viable verification method.

Tip 3: Keep the camera report's data log as a separate column in the existing call sheet or in a dedicated Google Sheet that the producer updates from their phone after each offload. The log should record: card label, file count, data size, Primary Drive verified (yes/no), Redundant Drive verified (yes/no), card returned for format (yes/no). Six columns, updated twice a day, prevents every data-related argument at the end of the production.

Frequently Asked Questions

Do I need checksumming software or can I just use a copy tool?

Standard copy tools (Mac Finder copy, Windows drag-and-drop, rsync without verification flags) do not confirm that the destination file is bitwise identical to the source. They confirm that the copy completed without an OS-level error, which is not the same thing. A storage drive can silently corrupt data during a write and report the copy as successful. Checksumming software reads both source and destination, generates a hash of each, and confirms they match. For production footage that cannot be reshot, checksumming is mandatory, not optional.

What checksumming software do professional DITs use?

YoYotta ($995) and SilverStack ($1,200) are the industry standard tools. Both produce verified copies, generate checksum reports, and create standard-format camera reports automatically. For productions that cannot justify that cost, Hedge for Mac ($50/year) provides verified copying and basic checksum reporting. The free command-line tools (md5sum, sha256sum) are technically sufficient but require manual record-keeping.

What happens to the drives after the production ends?

The Primary Drive becomes the editorial drive -- the working source that the editor ingests from. The Redundant Drive is stored offline as the production's master backup. Neither drive should be reformatted until the final deliverables are archived in their own long-term storage (ideally a third copy, offsite or on LTO tape). Drives fail. The production that formats its Redundant Drive after delivery and then discovers a problem with the deliverables has no fallback.

Should proxies be generated on set or in the edit suite?

On a small crew, on-set proxy generation introduces complexity and can slow the backup window. The practical approach: generate proxies once from the verified Primary Drive, either overnight or before the editor begins work. This keeps the on-set workflow focused on the critical task (verified backup) and moves the proxy step to a later time when a computer can run the transcode unattended. For reference, the documentary data storage breakdown covers the full proxy overhead in the total storage calculation.

Related Tools

The Backup Strategy Calculator models the hardware and drive count needed to maintain the 3-2-1 backup standard for any acquisition volume. The Codec Storage Calculator calculates daily acquisition size for your specific camera and codec, which is the input needed to confirm your drives and offload window are adequate before the shoot starts. The Cloud Storage Cost Calculator models the cost of adding a cloud backup layer to the on-set drive workflow.

Conclusion

Data management on a small crew fails when it is treated as something that happens naturally rather than something that is assigned and scheduled. One named person, one defined window, two verified copies before the card is formatted: that is the complete framework. Everything else in this post is the implementation detail around those three rules. The production that sets up this structure in pre-production and repeats it identically on every shoot day will not lose footage. What data management failure have you seen or experienced that could have been prevented with a clearer assignment of responsibility?