Preservation Letters That Actually Hold

Most litigation hold notices are written to be sent, not to be followed. They name a matter, invoke an obligation, and land in inboxes where recipients have no practical idea which of their systems the notice covers.

Name systems, not concepts

"All relevant communications" is not an instruction. A custodian can act on "your Outlook mailbox, the #project-atlas Slack channel, and the shared drive folder under Engineering/Atlas" — and, critically, can tell you when one of those doesn't exist or was already purged.

• Enumerate the repositories by their real names, as the custodian sees them.
• State what to stop doing — auto-delete rules, device trade-ins, mailbox cleanup.
• Give a named human to reply to, and require an acknowledgement.
• Re-issue on a schedule; a hold sent once is a hold that decays.

Acknowledgement is the evidence

When preservation is challenged, the question is rarely whether a notice was drafted. It is whether the organization can show who received it, when, what they were told to preserve, and what they did. That record has to be a byproduct of the process, not something reconstructed later.

Defensibility is a paper trail you generated at the time, not an argument you assemble afterwards.

Scope drifts — plan for it

Custodian lists change as facts develop. New systems surface in the first round of interviews. Departures create orphaned mailboxes. A preservation program needs an owner whose job includes revisiting scope, because the initial list is a hypothesis about where the evidence lives, and it is usually incomplete.

The practical test is simple: if opposing counsel asked today what you preserved and when, could you answer from records rather than memory? If the answer requires a search through email threads, the program is paper.