Sequence the migration behind topology, costing and runtime approval, rehearse the cutover against the three risk areas that decide whether balances reconcile while the operating model still fails, and hold the six evidence gates.
The playbook describes the safest pattern in one sentence: a rehearsed, dependency-ordered sequence that freezes source activity, extracts final topology-sensitive and cost-sensitive deltas, validates balances and operational totals, loads target waves, and proves reconciliation before business release.
The load order is determined by what references what, and every step in it has appeared as a dependency somewhere earlier in this path.
The playbook's minimum go/no-go evidence is short enough to quote and specific enough to hold.
A dry run that loads clean data in the right order proves the packages work. That is worth proving, and it is the first rehearsal. It is not what the later rehearsals are for.
The freeze creates a window between the last full extract and the moment business activity stops, and the playbook is specific about what moves in it: topology-sensitive and cost-sensitive deltas.
The cutover posture states that rollback boundaries must be owned by wave, by source contract, and by authoritative reconciliation pack. Three axes, and they answer three different questions.
The target holds a deliberately established opening position. Historical detail goes to an archive, and the executive summary lists deciding archive boundaries among the immediate mobilisation actions — which places the decision at the start of the programme rather than at the end.
The period after go-live is where the deferred consequences surface, and on this programme they surface in a predictable order.
Sequence behind the approvals: topology, then finance foundation, then parties, then products and costing, then manufacturing structures, then open documents, then inventory and subledgers, then opening balances, then delta and the go call. Each step has a gate stated as a pass or fail with a tolerance agreed in advance.