Google Workspace email archiving

The order that prevents disasters is: document what the organisation is required and choose to keep, per record type, with the legal or regulatory basis; then map those periods onto org units; then apply rules. Applying rules first and reconstructing policy afterwards is how organisations discover their retention was set by whoever happened to open Vault.

Different categories rarely share a period. Employment records, client correspondence, financial documentation and routine internal mail can carry six-year, seven-year and one-year expectations simultaneously, and Vault applies rules to org units rather than to content types, so the org structure has to carry the distinction.

When someone leaves, their mail is often still needed — for a handover, an audit trail or a dispute that has not surfaced yet. Deleting the account destroys it; keeping a full licence is wasteful at scale. An Archived User licence retains the mailbox as searchable Vault data at a materially lower monthly cost with no interactive sign-in.

The decision rule we use is simple: any leaver in a client-facing, financial or regulated role goes to an archived licence for the length of the relevant retention period. Everyone else is transferred and deleted after a fixed window recorded in the offboarding evidence pack.

A legal hold overrides retention and prevents expiry for matching accounts and terms until released. Holds must be placed the moment litigation or investigation is reasonably anticipated, not when it begins, and the placement should be dated and recorded because the date is itself evidence.

Exports produce PST or mbox with a metadata manifest suitable for a solicitor or an assessor. Large exports take hours and count against export quotas, so a disclosure deadline is planned backwards from the export rather than discovered at the end of it.

Retention configuration is not a backup. Vault preserves and searches; it does not restore a deleted shared drive structure or roll a mailbox back to a point in time. Organisations that believe they have backup because they have Vault find out during the incident, which is the worst moment available.

Twice a year on retainer we run a dated restore test: pick a real scenario, attempt recovery with the tooling actually in place, record what came back and what did not, and put the result in writing. That document is what auditors and insurers ask for, and it is the only proof that the arrangement works.