What a cutover plan must contain
Every task needs five things: a planned start time relative to go-live (T-0), a duration, a named owner and backup, the evidence that proves it is done, and the rollback step if it fails.
- Timeline from T-72h through hypercare exit
- Named owner and backup for every task
- Verification evidence (record counts, reconciliations, sign-offs)
- Rollback path and point of no return
- Go/No-Go checkpoints with written criteria
The nine workstreams
Group tasks into tracks so each lead owns a clear slice and the cutover manager can read the critical path at a glance.
- Governance & Go/No-Go
- Technical / Basis
- Data migration
- Finance & period-end
- Integrations & interfaces
- Security & access
- Business readiness & communications
- Legacy freeze & decommission
- Hypercare handover
Go/No-Go checkpoints
Hold at least three formal decisions: readiness at T-72h, data loaded at T-24h and business verification at T-2h. Write the criteria before the weekend so the decision is evidence-based, not optimism-based.
Rehearse with mock cutovers
Run two or three full mock cutovers. Record the actual duration of every task, then plan the real cutover on those numbers plus contingency. Close every mock lesson before checkpoint 1.
Track it live, not in a spreadsheet
Spreadsheets break at 2am when twelve leads update the same file. A live cutover board shows variance to plan, blocked tasks and the next checkpoint in real time, and gives executives one reliable view.
Frequently asked questions
How far ahead should a cutover plan start?
Draft it at least three months before go-live, freeze it after the final mock cutover, and execute from about T-72h.
Who owns the cutover plan?
A dedicated cutover manager owns it. Each workstream lead owns their tasks, and the executive sponsor owns the final Go decision.
What is the point of no return?
The moment after which rolling back to the legacy system costs more than fixing forward, usually when the business starts transacting in the new ERP.