
The plan is not the spreadsheet. The plan is what happens when the third rehearsal fails at two in the morning and somebody has to decide. We rehearse end to end, time the rollback, and publish who makes the call before the weekend starts.
What is in it
Five capabilities, and one named decision-maker per phase.
Three rehearsals
The first finds faults. The third is theater. The second is the one that changes the plan.
The second rehearsal is where authority gets tested
By the second run the sequence survives, so people stop testing the steps and start testing what they are allowed to decide. That is when somebody says out loud that they are not sure whose call it is. They only ever say it once, so if you skip the second rehearsal you never hear it.
The go/no-go criteria
| Criterion | Who owns it |
|---|---|
| Data reconciliation inside tolerance | Your data lead |
| Rollback timed inside the window | Us |
| Third rehearsal's open items closed | Joint |
| Downstream consumers confirmed | Your integration lead |
| Support rota staffed and awake | Us |
| The named decision-maker present | Named before the weekend |
Questions we are always asked
Four of them, answered the way we answer them on the phone.
Who makes the go or no-go call?
One named person per phase, published before the weekend and awake for it. The most common failure on the night is not a technical one — it is that nobody knew who was allowed to decide.
What if a rehearsal fails?
That is what rehearsals are for. The first finds faults, the second tests whether the authority holds when something goes wrong, and the third should be dull. A cutover whose rehearsals were all clean has not been rehearsed hard enough.
Do you rehearse the rollback too?
Executed and timed, not written down. A rollback that exists only as a paragraph in a document is a hope with a heading on it.
How long do you stay afterwards?
Through hypercare and out on a date agreed before the cutover, not a date negotiated after it. Staying indefinitely would be good for us and bad for you.
Case studies
Programs in this area, described by what they were rather than by who paid for us.