There are several pieces to this change: new features to write and deploy, bits of infrastructure to add or modify, tests to perform. We should try to coordinate it all, to ensure a reliable new system is in place within a reasonable timeframe.
For this task, we might list all the changes and tests that need to occur, roughly scope each out as best as possible, and plan when they could ideally happen. We could also try to determine whether coordination is needed with live FR campaigns or tests for any of the steps.