Migrate G Suite to Office 365

Microsoft documents an IMAP-based path for moving G Suite mailboxes into Exchange Online, and it works well for what it does. What it does is messages. Calendars, contacts, delegation, shared drives, Google Groups and send-as permissions are all outside its scope, and it completes successfully without mentioning them.

That gap is the reason self-run G Suite migrations tend to be judged a failure a fortnight later. The mail is all there; the shared team calendar is not, the reception desk can no longer send as the main address, and nobody can find the client contact list.

Gmail is a label store, not a folder store. A single message can carry five labels; the same message in Exchange lives in exactly one folder. Every migration tool resolves this by picking a winner, and the rule it uses is rarely the one your users would have chosen.

We agree the mapping before the move: which labels become folders, which become categories, and which are dropped because they were an individual's filing habit rather than a shared structure. Users are told what changed before they see it, which converts a support ticket into a note in an email.

Authentication first. SPF is republished to describe Exchange Online, DKIM signing is enabled on the Microsoft side, and DMARC stays at a monitoring policy through the change so nothing is rejected while both platforms are sending. Then a full pre-sync while Google remains authoritative, then the MX change, then two delta passes.

Because Google stays live through the pre-sync, there is no point at which mail is in flight with nowhere to land. Senders retry automatically during the DNS change under standard SMTP behaviour.

Three decisions shape the whole project. What happens to Google Vault archives, which do not transfer to Microsoft Purview automatically and may carry retention obligations. What happens to shared drives, which are a permission graph rather than a folder tree. And whether any account is staying on Google, because a split estate needs split delivery rather than a clean MX change.

None of these are difficult once decided. All of them are expensive to revisit after cutover.