A cross-tenant mailbox migration moves an Exchange Online mailbox between Microsoft 365 tenants. The difficult work is usually outside the mailbox itself: matching identities, establishing migration permissions, maintaining address continuity and deciding when users should start working in the destination tenant.
In Part 9 of Microsoft 365 Architecture & Exchange Enterprise Deep Dive, we treat migration as a controlled service transition. The objective is not merely a completed migration batch. It is a functioning target identity, predictable message delivery and evidence that users can resume work. By Muhammad Ahsan, Solution Savy.
1. Architecture: understand what actually moves
Microsoft’s native cross-tenant mailbox migration uses Exchange Online migration infrastructure. The source tenant contains the mailbox being moved. The target tenant contains a prepared mail-enabled user, a migration endpoint and the migration batch. Organisation relationships and an authorised application establish the migration path between the tenants.
This is not a tenant merger. Moving a mailbox does not migrate the user’s Microsoft Entra ID authentication configuration, Conditional Access policies, devices, OneDrive files or SharePoint permissions. Teams dependencies also require separate planning; do not interpret a successful mailbox move as completion of a Microsoft 365 workload migration.
A useful architecture separates four workstreams: identity provisioning, mailbox movement, mail routing and user access. Each has its own acceptance criteria. For example, mailbox synchronisation can succeed while target sign-in remains blocked, or while applications continue sending to an address that has not been established in the destination.
After a successful move, Exchange converts the source mailbox into a mail user with a routing address pointing to the destination. Preserve that source object for the required coexistence period. Deleting it as routine migration clean-up can remove the routing continuity on which remaining source users depend.
Confirm licensing before scheduling: Microsoft documents a Cross Tenant User Data Migration add-on, assignable to the source or target user, alongside the required service licensing. Follow the documented provisioning sequence rather than treating the migration add-on as an Exchange Online service entitlement.
Identify blockers early. Microsoft’s cross-tenant guidance states that mailboxes on any type of hold cannot be migrated through this process. Do not remove a hold merely to enable migration without an authorised compliance decision.
Define scope explicitly. A mailbox migration project needs separate decisions for shared resources, applications, directory objects and other workloads; moving individual user mailboxes does not resolve those dependencies automatically.
2. Identity mapping: prepare a mail user, not a second mailbox
The target recipient must be a correctly prepared MailUser. It represents the destination identity before the source mailbox arrives. Accidentally provisioning an independent target mailbox creates a different state from the one required by the native migration workflow; this is not a mailbox-merging operation.
Create a controlled mapping register with one row per person. Record source and target user principal names, recipient addresses, source ExchangeGUID, archive information, LegacyExchangeDN, relevant X500 addresses and the intended target routing address. Keep immutable identity identifiers alongside display names: names alone are not reliable matching keys.
ExchangeGUID: stamp the target MailUser with the source mailbox’s ExchangeGUID. This is a mailbox identity requirement, not an optional descriptive attribute.
ArchiveGUID: where the source has an archive, prepare the target with the corresponding archive identity as required by Microsoft’s migration instructions.
LegacyExchangeDN and X500 addresses: add the source LegacyExchangeDN as an X500 proxy address on the target and preserve existing X500 proxy addresses. These addresses support resolution of historical Exchange recipients, including replies to older messages.
ExternalEmailAddress: before migration, configure the target MailUser to route to the source mailbox. Use a source address that remains valid during coexistence.
Target addresses: use addresses in domains available in the target tenant. A retained corporate SMTP domain requires a separate domain-transfer plan; it cannot simply be used as though it were already configured in both tenants.
Follow Microsoft’s licensing and attribute-stamping order. In particular, do not enable Exchange Online for an unprepared target identity and assume the resulting mailbox can be overwritten. Where identities are synchronised from an on-premises directory, establish the supported attribute-management path before making changes that synchronisation could later replace.
Use read-only checks to compare both sides. Run each connection in a separate PowerShell session so there is no ambiguity about which tenant you are inspecting:
# Source tenant session
Connect-ExchangeOnline -Organization source.onmicrosoft.com
Get-Mailbox -Identity user@source.onmicrosoft.com |
Format-List UserPrincipalName,PrimarySmtpAddress,ExchangeGuid,
ArchiveGuid,LegacyExchangeDN,EmailAddresses
# Target tenant session
Connect-ExchangeOnline -Organization target.onmicrosoft.com
Get-MailUser -Identity user@target.onmicrosoft.com |
Format-List UserPrincipalName,PrimarySmtpAddress,ExternalEmailAddress,
ExchangeGuid,ArchiveGuid,EmailAddressesReview the actual values, not just whether both commands return an object. A recipient can exist while still having the wrong mailbox GUID, an incorrect external address or missing historical addresses.
3. Migration trust: separate permission from scope
Microsoft’s documented setup uses an application registered in the target tenant with the Exchange Online application permission Mailbox.Migration. The source tenant must consent to the application. Follow the current Microsoft Learn procedure for application configuration, consent and credentials; the application permission name and consent path matter.
The target migration endpoint uses the application credentials to connect to the source environment. Treat its client secret as a privileged migration credential: protect it, track its expiry and exclude it from scripts committed to repositories, screenshots and support evidence.
Configure the organisation relationships for the intended direction: outbound mailbox moves from the source and inbound moves into the target. The source relationship associates the authorised application with the move configuration. A mail-enabled security group defines the source mailboxes published for migration through MailboxMovePublishedScopes.
These controls answer different questions. Application consent authorises the application, the organisation relationships establish the move relationship, and the published scope restricts which source mailboxes are available. A valid secret does not compensate for an incorrectly scoped mailbox or a missing relationship.
Have administrators from both tenants review tenant IDs, application ID, relationship direction and migration scope before testing.
Start with an explicitly approved pilot scope, rather than publishing the entire source population.
Use Microsoft’s documented migration endpoint validation before creating production batches. Record the result without recording credential material.
4. Pilot and synchronisation: prove the complete workflow
Choose a small but representative pilot. Include an archive where applicable, a user with delegated access, and a user who relies on desktop Outlook as well as Outlook on the web. The purpose is to expose architectural dependencies before moving a large population, not simply to demonstrate that a small mailbox transfers.
Create the migration batch in the target tenant using the configured source endpoint. Microsoft’s cross-tenant procedure uses the target recipient address in the CSV EmailAddress column. Validate this mapping carefully: source and target addresses can differ, and an incorrect assumption here can obscure an otherwise correctly prepared identity.
Set TargetDeliveryDomain to the intended destination routing domain. Keep this distinct from the later transfer of a shared corporate domain. Where appropriate, synchronise ahead of the change window and control completion separately instead of automatically completing users as soon as they are ready.
Monitor both batch-level and per-user results. A batch summary is useful operationally, but an individual migration report is the better place to investigate a particular user’s failure:
# Run in the target tenant session
Get-MigrationBatch -Identity "XT-Pilot" |
Format-List Identity,Status,TotalCount,SyncedCount,FinalizedCount,FailedCount
Get-MigrationUserStatistics -Identity user@target.onmicrosoft.com -IncludeReport |
Format-List Identity,Status,Error,ReportInvestigate failures by layer: recipient preparation, application consent, endpoint configuration, migration scope and mailbox eligibility. Avoid repeatedly recreating batches before understanding the error. Preserve reports with timestamps so the team can distinguish a persistent configuration fault from a resolved issue.
For permissions and collaboration, validate the actual access scenario after the pilot. Delegate access, shared calendars and resource interactions involve other directory objects; do not assume that moving mailbox content proves those relationships work across the new tenant boundary.
5. Cutover: coordinate mailbox completion, domains and clients
Define cutover as a sequence with named owners and explicit go/no-go criteria. First confirm target sign-in, licensing readiness, mailbox mapping, migration status and helpdesk readiness. Then complete the approved migration wave, verify destination access and test routing before declaring the users operational.
Moving mailboxes and transferring a custom domain are separate operations. During staged migration, use tenant-specific routing addresses to maintain coexistence. The source mail-user conversion supports delivery towards the destination, but the surrounding address design still needs testing in both directions.
If the organisation retains its corporate domain, inventory everything that prevents its removal from the source tenant. Microsoft’s domain-removal guidance covers references such as user sign-in names, email addresses and group addresses. Removing a domain is not simply an MX record change, and mailbox completion does not remove those dependencies automatically.
Plan domain release, verification in the target, recipient address updates and the required DNS changes as their own controlled change. Review MX, SPF, DKIM and DMARC configuration against the destination design. Do not assume the destination inherits the source tenant’s connectors, transport rules or Microsoft Defender for Office 365 policies.
Test internal delivery within the target and cross-tenant delivery in both directions while coexistence remains in place.
Test external inbound and outbound mail, including replies to messages received before migration.
Validate archive access, calendars, delegates and business applications that send to or access the mailbox.
Plan Outlook profile recreation and destination sign-in, following Microsoft’s documented client requirements. Give users precise instructions rather than promising a silent tenant switch.
Define recovery before the final move. Before completion, a failing wave can be held while the source remains the operational mailbox location. After completion, do not describe rollback as a DNS reversal: the mailbox has moved and the source recipient has changed. Any return move requires its own supported preparation and migration plan.
6. Short operational checklist
Eligibility approved: migration licences, mailbox holds and workload boundaries reviewed.
Identity prepared: target MailUser, ExchangeGUID, archive identity, routing and X500 addresses verified.
Trust validated: application consent, credentials, organisation relationships, source scope and endpoint checked.
Pilot accepted: mailbox access, historical replies, archive and delegate scenarios tested.
Cutover authorised: completion schedule, domain dependencies, DNS ownership and client instructions agreed.
Evidence retained: per-user migration results, delivery tests and support issues recorded securely.
Closure controlled: retain necessary coexistence recipients; retire migration-only credentials and configuration after all waves and recovery requirements are closed.
7. Conclusion: migrate the service, not just the mailbox
A reliable cross-tenant migration depends on disciplined identity mapping and deliberate cutover design. Prepare the target recipient correctly, constrain migration permissions, prove the workflow with a representative pilot and separate mailbox movement from domain transfer. Completion is the point at which the user can authenticate, work and exchange messages reliably—not merely the point at which Exchange reports that the move has finished.
Microsoft Learn: New-MigrationBatch
Microsoft Learn: Get-MigrationBatch
Microsoft Learn: Get-MigrationUserStatistics
Microsoft Learn: Remove a domain from Microsoft 365
Originally published on Solution Savy. Need help with this? Book a consultation.




