Saw the Petri piece on key decisions to avoid data loss in M365 tenant-to-tenant work....

Saw the Petri piece on key decisions to avoid data loss in M365 tenant-to-tenant work. It's a fair list, but the part that gets underweighted in most planning docs is testing. Not unit tests on the migration tool. Actual end-to-end rehearsals against a representative slice of mailboxes, OneDrive content, Teams chat, and shared channels, with the destination tenant configured the way it will be on cutover day. Most of the data loss conversations I've been in started long before cutover. They started when somebody approved a wave list without checking license assignment in the target, or assumed a mail-enabled security group would carry permissions cleanly, or didn't validate retention policy alignment between source and destination. By the time you're staring at a delta sync, your options are narrower than they should be. The other quiet one is chat. Teams private chat history is still the thing that makes runbooks longer than they used to be, and it's the thing end users notice first if you get it wrong. I'd rather spend an extra week in pilot than explain to a VP why their thread with general counsel is gone. Decisions matter, but the decisions are only as good as the rehearsal that validates them. Pre-cutover testing is where data loss gets prevented, not on the bridge call.