Skip to main content
Tenant to tenant, India

The data is the straightforward part. Identity is what makes these projects hard, and the domain can only live in one tenant at a time.

Mail and files move between tenants with mature tooling. What makes a tenant-to-tenant migration different from any other is that user identities cannot come with them: the accounts are new objects, devices joined to the old tenant have to be dealt with, and every application authenticating against the source identity needs re-establishing. On top of that a verified domain can only exist in one tenant at a time, which turns the cutover into a genuinely constrained event rather than a flexible one.

Microsoft
Microsoft
365
Cloud Solution Partner
  • New identitiesAccounts do not carry across
  • One tenantA domain can only be verified once
  • DevicesRe-registration is a user-facing event
  • Deal-drivenThe date usually is not yours
What the work involves

Eight workstreams, and identity determines the shape of all of them.

A tenant migration is not a larger mailbox migration. It is an identity project with a data migration attached, and treating it the other way round is why these run long and why users have a difficult week.

Identity, which is genuinely rebuilt rather than moved

Accounts in the target are new objects with new underlying identifiers. Nothing about the old identity carries across, which affects anything that referenced a user by identifier rather than by address. Where an on-premises directory is involved, the synchronisation arrangement has to be redirected as well, and that sequencing is one of the more delicate parts of the whole exercise.

The domain, and the constraint it creates

A verified domain can exist in only one tenant at a time, which means it must be removed from the source before it can be added to the target. That is the pivot point of the entire migration and it is why cutover is a genuinely constrained event with a real window rather than a gradual transition. Everything in the plan is arranged around this single fact.

Mail, with a delta pass at the end

Mailbox content migrates with mature tooling, and the pattern is a bulk pass well ahead of cutover followed by a delta pass in the window itself. What needs deliberate handling is the period around the domain move, when mail may be arriving at one tenant while users are being moved to another, and where inbound mail routing has to be planned rather than assumed.

OneDrive, SharePoint and Teams content

File content moves reasonably well. Teams is more involved, because a team is a group, a set of channels, a SharePoint site, a chat history and a membership list, and not all of it migrates equally. Private channels, chat history and Planner plans are the recurring gaps, and the honest approach is to establish what will not come across and communicate it rather than discover it.

Devices, which is the visible part for users

Machines joined to the source Entra tenant have to be re-registered or rejoined, and depending on the approach this ranges from a background operation to something requiring each user to sign in again and reconfigure. It is the part users actually experience, so it needs planning and communication far more carefully than the data migration does.

Applications bound to the source identity

Every SaaS product using single sign-on against the source tenant, every enterprise application registration, every service principal and every automation authenticating with a source identity. These have to be re-established in the target, and the inventory is invariably longer than anybody expects because departments have connected things independently over years.

Policy, compliance and the configuration nobody lists

Conditional access policies, device compliance, sensitivity labels, retention policies, DLP rules and legal holds. None of this migrates: it is rebuilt in the target, ideally deliberately rather than by replicating whatever the source had. For a divestiture there is also the question of what compliance data must be extracted before access to the source is lost.

Coexistence, which is more limited than in hybrid Exchange

Two tenants can share some visibility, and cross-tenant synchronisation can give people access across the boundary during a transition. It is genuinely useful and it is not the same as Exchange hybrid, so expectations need setting: free/busy and collaboration across tenants are more constrained, and the aim is to keep the overlap period as short as the business can tolerate.

The hard parts

Eight things that make tenant migrations different, and how we handle each.

If you have run an on-premises Exchange migration and expect this to feel similar, these are the differences that matter. Each has caused a difficult week somewhere.

The domain can only be in one tenant

Removing it from the source and adding it to the target is a hard pivot with a real window. Everything else in the plan arranges itself around this constraint.

  • Plan the window with the business, since it constrains mail flow
  • Have inbound routing decided for the period around the switch
  • Rehearse the sequence, because it is not something to improvise

Identifiers do not carry across

Target accounts are new objects. Anything that referenced a user by identifier rather than by email address will not follow them, and that includes more than people expect.

  • Inventory anything keyed on user identifiers rather than addresses
  • Third-party SaaS user mappings frequently need rebuilding by hand
  • Plan for permissions to be reapplied rather than transferred

Devices have to be re-established

The most visible part of the project for end users, and the one that generates the support volume. The approach chosen determines whether it is quiet or disruptive.

  • Decide the device approach early, since it drives the communication plan
  • Pilot on a real cross-section before committing to a method
  • Schedule waves so the support desk can absorb the sign-in questions

Every integrated application needs re-establishing

Single sign-on integrations, app registrations, service principals and automations authenticating against the source. The list is always longer than the one you are given.

  • Enumerate from the source tenant rather than by asking departments
  • Some integrations need vendor involvement, which has a lead time
  • A number will turn out to serve a process nobody needs any more

Teams does not move as one thing

A team is a group, channels, a SharePoint site, chat history and membership. Files and structure migrate reasonably; private channels, chat history and Planner are the recurring gaps.

  • Establish and communicate what will not come across, before cutover
  • Chat history is the item users ask about most and expect least
  • Rebuild team structure deliberately rather than replicating clutter

Compliance data is lost when source access ends

Legal holds, audit logs, retention and eDiscovery content live in the source tenant. For a divestiture, access to that tenant may end on a contractual date.

  • Establish retention obligations before the migration, not during
  • Export what must be retained while you still have access
  • Involve legal early, since this is their decision rather than IT

The date is usually set by the deal

Mergers, acquisitions and divestitures come with contractual dates and transitional service agreements. The migration schedule is a constraint rather than something you propose.

  • Work backwards from the contractual date to find what must be dropped
  • Scope reduction is usually the only real lever available
  • Agree explicitly what moves later, rather than pretending it all fits

A divestiture is not a merger in reverse

Separating a business out has its own problems: shared content that must be split, mailboxes with mixed history, and defining what the departing entity is entitled to take.

  • The entitlement question is legal and commercial, not technical
  • Shared document estates rarely split along clean lines
  • Plan for a period where neither side is fully self-sufficient
Why bring us in

We plan around the domain constraint rather than discovering it.

We sequence the whole project around the pivot

The domain can be verified in one tenant at a time, and that single fact determines the shape of the plan. We build the sequence around it, rehearse the switch, and make sure inbound mail routing is decided for the period around it rather than improvised on the night.

We enumerate integrations from the tenant, not from a list

Asking which applications use single sign-on produces an incomplete answer every time, because departments connect things independently. Enumerating app registrations, service principals and consented applications from the source tenant produces the real inventory, which is what stops a post-cutover application outage.

We treat the device experience as the user-facing project

Users judge a tenant migration on whether their laptop still works on the Monday. The device approach determines that entirely, so we decide it early, pilot it on a genuine cross-section rather than on IT machines, and build the communication plan around what people will actually experience.

We are honest about what will not fit the date

When a transaction sets the deadline, the only real lever is scope. We would rather agree explicitly at the start what moves later than deliver a cutover that technically happened on time and left the business struggling for a fortnight.

What drives the project

Why tenant migrations happen.

Acquisition into a parent tenant

The acquired entity moves into the buyer tenant, usually with a contractual date and a transitional services agreement defining how long the old environment remains available.

GCC consolidation into a group tenant

Common in India, where an entity set up independently is brought into a global group tenant. Frequently combined with an on-premises Exchange migration, which makes sequencing the main design problem.

Divestiture and carve-out

A business unit separating out, which is harder than a merger because shared content has to be split and entitlement to data is a legal question rather than a technical one.

Consolidating tenants created by accident

Organisations that ended up with several tenants through acquisitions, regional autonomy or a shadow signup that grew. Usually driven by security and licensing rather than by a transaction.

Regulatory or residency restructuring

Restructuring so that data sits in a tenant with the right residency and governance position, which is increasingly a live question for Indian entities of global groups.

Rebranding and legal entity changes

A change of legal entity or brand requiring a new domain and a clean tenant, where the migration is a consequence of a corporate decision rather than an IT initiative.

Who decides what

The decisions in a tenant migration, and whose they actually are.

These projects stall when a decision sits with the wrong function. Several of the most consequential choices are legal or commercial rather than technical, and identifying that early is what keeps the schedule intact.
DecisionWhose decisionWhen it has to be made
What the departing entity is entitled to takeLegal and commercial, informed by ITBefore any scoping. Everything else in a divestiture depends on this answer and it is the item most often left until it blocks progress.
What must be retained from the source tenantLegal and complianceBefore the source becomes inaccessible. Audit logs, eDiscovery content and legal holds cannot be recovered once access ends.
The cutover date and windowThe transaction, with the businessSet by the agreement. IT works backwards from it, which is why scope rather than date is the negotiable variable.
What moves now against phase twoBusiness sponsor, with IT adviceDuring scoping, in writing. Leaving it implied is how a deferred item becomes a surprise omission.
The device re-registration approachIT, with the business on timingEarly, since it drives the user communication plan and is what people actually experience.
Target conditional access and compliance policyIT and security, not replicated by defaultDuring design. A rebuild is an opportunity to fix what accumulated in the source rather than to reproduce it.
How it differs

Hybrid Exchange migration against a tenant-to-tenant move.

Organisations that have done an on-premises migration expect this to feel similar. It does not, and the differences all trace back to identity.
Feature
Dimension
On-premises to Exchange Online
Tenant to tenant
User identity
The same identity, synchronisedA new object, with nothing carried across
The domain
Stays with you throughoutMust leave one tenant before joining the other
Coexistence
Full hybrid, mail flow and free/busy sharedMore limited, and best kept short
Devices
UnaffectedRe-registered or rejoined, visible to users
Integrated applications
Largely unaffectedEvery one re-established in the target
Policy and compliance
Migrated or extendedRebuilt in the target from scratch
What sets the date
You do, within reasonThe transaction does
Rollback
Per mailbox, reasonably cleanVery limited once the domain has moved

When the deal sets the date, scope is the only lever you have.

Tenant migrations are almost always driven by a merger, an acquisition or a divestiture, which means the date arrives with the transaction and is not negotiable. Pretending the full scope fits into a fixed window is how these projects produce a bad cutover. The productive conversation is the opposite one: agree explicitly what moves by the date, what moves in a later phase, and what does not move at all. A transitional services arrangement usually gives more room than people assume, and using it deliberately is far better than discovering at the last week that something has to be dropped anyway.

  • Work backwards from the contractual date and find the real constraint early
  • Agree a phase two in writing rather than leaving it implied
  • Historical archives and low-use content are the usual candidates to defer
  • A transitional services agreement is a planning tool, not just a legal formality
The engagement

Five stages, arranged around a single constrained window.

  1. 1

    Assess both tenants and the transaction constraints

    Users, mailboxes, file volume, Teams estate, and crucially every integrated application enumerated from the source rather than from a list. Alongside that, the contractual date, the transitional services arrangement and what the business can tolerate in the cutover window.

  2. 2

    Design the target and agree the scope that fits

    Identity model, device approach, Teams structure, and the conditional access and compliance policies to be built rather than replicated. Then the honest scoping conversation about what moves by the date and what is explicitly phase two.

  3. 3

    Prepare, pre-seed and pilot

    Target accounts provisioned and licensed ahead of time, bulk mail and file passes running well before cutover, integrations re-established and tested, and a pilot group taken through the full experience including the device step, on real machines.

  4. 4

    Execute the window

    The rehearsed sequence: final delta passes, domain removal and addition, mail routing switched, accounts activated, devices handled per the chosen approach. A named decision maker present throughout, because judgement calls will be needed and they cannot wait.

  5. 5

    Stabilise, then close out phase two

    Concentrated support in the days after cutover when volume spikes, integrations verified in production rather than assumed, compliance data confirmed in the target, and the deferred scope delivered rather than quietly forgotten.

Before cutover

Twelve things to have settled before the domain moves.

The domain switch is the point of no easy return. These are the items we confirm before scheduling it, and each corresponds to a problem we have seen when it was assumed.

Identity and access

  • Target accounts created and licensed, ready to activate
    Provisioned ahead, not on the night
  • Every integrated application enumerated from the source tenant
    Not from a list somebody wrote
  • Device approach chosen and piloted on a real cross-section
    This drives the whole user experience
  • Directory synchronisation redirection sequenced
    Where an on-premises directory is involved

Content

  • Bulk mail and file passes complete, delta pass planned
    Only the delta belongs in the window
  • Teams structure agreed, and gaps communicated to users
    Chat history and private channels especially
  • Compliance and legal hold data exported from the source
    While you still have access to it
  • Shared content split agreed, for a divestiture
    A legal question before it is a technical one

The window

  • Domain removal and addition sequence rehearsed
    Not improvised on the night
  • Inbound mail routing decided for the transition period
    Mail arrives during the switch regardless
  • Conditional access and compliance policies built in the target
    Rebuilt deliberately, not replicated
  • A named decision maker available throughout the window
    Because judgement calls will be needed
Questions we get asked

Tenant to tenant, answered plainly.

Because identity is rebuilt rather than moved. In a hybrid Exchange migration the user keeps their identity, their device stays joined, and integrated applications carry on working; only the mailbox location changes. In a tenant-to-tenant move the target account is a new object, devices joined to the source tenant have to be re-registered or rejoined, every application authenticating against the source identity has to be re-established, and policy and compliance configuration is rebuilt from scratch. On top of that, a verified domain can only exist in one tenant at a time, which makes cutover a genuinely constrained event rather than a gradual transition.

Assessment and design typically take a few weeks, and the overall duration is usually set by the transaction rather than by the work. Where the date is genuinely fixed, we work backwards to find what fits and agree explicitly what becomes phase two. Where there is flexibility, a mid-sized migration is commonly a couple of months, most of which is preparation: enumerating and re-establishing integrations, pre-seeding content, and piloting the device approach. The cutover window itself is short and intense, and everything else exists to make that window uneventful.

Files and team structure migrate reasonably well. Chat history, private channels and Planner plans are the recurring gaps, and the most useful thing you can do is establish exactly what will not come across during assessment and tell people before cutover rather than after. Users ask about chat history more than almost anything else and expect it least. Where any of that history carries a genuine retention or legal obligation, it needs exporting from the source tenant while you still have access, which for a divestiture may end on a contractual date.

Almost certainly something, and how much depends on the device approach you choose. Machines joined to the source Entra tenant have to be re-registered or rejoined for the target, and the available methods range from largely background operations to something requiring each user to sign in again and reconfigure. This is the part users actually experience and therefore the part they judge the migration on, so we decide the approach early, pilot it on a real cross-section rather than on IT machines, and build the communication around what people will genuinely encounter on the Monday.

Partly, and less completely than hybrid Exchange allows, which is an expectation worth setting early. Cross-tenant synchronisation can give people access across the boundary, and some sharing and collaboration works across tenants. What you do not get is the seamless shared mail flow and free/busy that Exchange hybrid provides. The practical consequence is that the overlap period should be as short as the business can tolerate. Long coexistence between tenants is workable and it is friction that people feel daily, so it is worth compressing rather than extending.

They are rebuilt in the target rather than migrated, and that is an opportunity rather than only a cost. Conditional access, device compliance, sensitivity labels, retention, DLP and legal holds all have to be recreated, which means you can build what you actually want rather than replicating whatever accumulated in the source over the years. The one genuinely urgent item is compliance data: audit logs, eDiscovery content and anything under legal hold live in the source tenant, and if access to that tenant ends on a contractual date then extraction has to happen beforehand. Involve legal early, because what must be retained is their decision.

No, and divestitures are generally harder. A merger moves everything in one direction; a carve-out requires deciding what the departing entity is entitled to take, which is a legal and commercial question before it is a technical one. Shared document estates rarely split along clean lines, mailboxes contain mixed history, and shared systems have to be either duplicated or separated. There is also usually a period where neither side is fully self-sufficient, and that period needs planning rather than enduring. We start these engagements with the entitlement question rather than with the tooling, because everything else depends on the answer.

Integrated applications, consistently. Every SaaS product using single sign-on against the source tenant, every app registration, every service principal and every automation authenticating with a source identity stops working, and the inventory is invariably longer than the one you are handed, because departments connect things independently over the years. We enumerate them from the source tenant rather than by asking, because asking produces an incomplete list every time. Some need vendor involvement with a real lead time, which is precisely why this work belongs in preparation rather than in the cutover window.

Only in a very limited sense, and this is a genuine difference from other migrations. Once the domain has been removed from the source and verified in the target, reversing that is disruptive rather than routine, and it is not something to plan on. What you can do is reduce the need for rollback: pre-seed content so the window contains only a delta pass, rehearse the domain sequence, pilot the device approach on real users, and verify integrations in a test capacity before the night. The defence here is preparation rather than reversibility, which is why we spend proportionally more time before the window than in it.

Yes for the great majority of it. Assessment, design, content pre-seeding, integration work and the cutover window itself are all delivered remotely, since both tenants are in the cloud. The part that sometimes benefits from presence is the device step, particularly where the chosen approach requires user interaction and the organisation would rather have somebody available on the floor for the first morning. We are based in Gachibowli, Hyderabad and work with organisations across India, and we will travel for a cutover where it genuinely helps. Initial enquiries get a reply within 4 business hours.
References

Official documentation

Next step

Start with the integration inventory and the date.

The two things that determine whether a tenant migration goes well are the real list of applications bound to the source identity, and an honest conversation about what fits the contractual window. Both take days to establish. Remote-first from Hyderabad, serving all of India.