Migrating a live charging or billing platform is one of the riskiest things an operator can do. Every minute of failure means lost revenue, wrong bills or customers who cannot top up. “Zero downtime” appears in almost every plan. Far fewer plans show how it will be achieved.
What successful cutovers have in common
- Rehearsals with real data volumes. At least one full dress rehearsal, timed, using production-sized data. The first rehearsal almost always finds something.
- Reconciliation before and after. Balances, rated events and bills compared between old and new platforms, with agreed tolerances.
- Rollback criteria written in advance. Clear, measurable triggers for going back, and a rollback that has itself been tested.
- One go/no-go owner. A named person who makes the call on the night, based on criteria agreed in advance, not on who is most tired or most confident.
- A hypercare period. Extra monitoring and a dedicated team for the first billing cycles after go-live.
The decisions to make early
Most cutover risk is decided weeks before cutover night: how much data is migrated and how, whether old and new platforms run in parallel, which customer segments move first, and who signs off each step. Making these decisions explicit, and writing them down, is most of the work.
On the night itself, a good cutover should feel almost boring. That is the result of the planning, not luck.