Learn › SharePoint & OneDrive › 5 · Compliance & lifecycle

eDiscovery of SharePoint and OneDrive

eDiscovery over SharePoint content turns on one mapping question — which sites and OneDrives hold a custodian's data — and one trap: content the index couldn't read. Get the mapping and the indexing caveat right and the collection is defensible.

The shape of eDiscovery here

Microsoft Purview eDiscovery (Standard, and the richer Premium/eDiscovery graph tier) searches, holds, reviews and exports content across Microsoft 365 — including SharePoint sites and OneDrive accounts. The workflow: define a case, add data locations (or custodians), place a hold, run searches (KQL over content + metadata), review, and export.

Custodian-to-location mapping (the core question)

The mapping that makes or breaks a collection:

Custodian's data eDiscovery location
A person's personal files Their OneDrive (/personal/…), keyed by UPN
Team/departmental documents The SharePoint site(s) those files live in
Files shared into a chat The sender's OneDrive (one-content-system concept)
Files posted in a Teams channel The team's SharePoint site

In Premium, adding someone as a custodian auto-associates their mailbox AND OneDrive; SharePoint sites are added as non-custodial data locations by URL because a site has no single "owner" to key on. Getting this right means knowing where the content actually lives — which is again a site question.

Holds preserve regardless of user action

Placing a hold makes the location immutable for discovery: edits and deletes still appear to work for users, but preserved copies are captured (via the Preservation Hold Library, retention concept) so nothing is lost mid-case. A hold overrides retention deletion and recycle-bin expiry.

The failure modes

  • Partially indexed items. eDiscovery searches the index (search concept), so content the crawler couldn't read — image-only PDFs, encrypted files it lacks rights to, unsupported formats, oversized files — returns as partially indexed / unindexed and can be missed by a keyword search. Premium can include and process these; ignoring them is a defensibility gap.
  • OneDrive keyed by UPN — if a custodian's UPN changed (rename, re-licensing) or their account was deleted, their OneDrive may not resolve without the retained personal site (leaver concept) — collect before the orphan window closes.
  • Guest/external content sits in your sites but the external user isn't a custodian you control — capture via the site, not the person.
  • Encrypted-label content — eDiscovery can decrypt for review with the right configuration; without it, encrypted files export unreadable.
  • Standard vs Premium — advanced processing, custodian management and review-set analytics are Premium capabilities; assuming Standard has them causes scoping mistakes.

What to watch (proofs)

  • Locations in the case: the eDiscovery case's data locations list — every relevant OneDrive (by UPN) and site (by URL) accounted for.
  • Hold report: the case hold's status and statistics — locations on hold and the volume preserved (the preservation proof).
  • Indexing coverage: the search statistics' partially indexed items count — the number that must be reviewed separately for a defensible collection.
  • Leaver readiness: the departed custodian's OneDrive still present (Get-SPOSite -IncludePersonalSite $true) within the retention window before collection.

PowerShell for this concept

  • Get-SPOSite Returns one or more site collections.
  • Get-SPOSite -IncludePersonalSite $true

Discussion

No messages yet — start the thread.

Sign in with your email to join the discussion — we send a one-time link, no password.