Change-Over: Parallel, Phased, Abrupt

Switching from old to new

Think about switching to a new phone. You could carry both for a week until you trust the new one (parallel). You could move your contacts first, then your photos, then your apps (phased). Your partner could try the same model first (pilot). Or you could drop the old phone in a drawer and go (abrupt).

Organisations face exactly this choice when a new system replaces an old one. It’s a trade-off between risk and cost.

Parallel, pilot, phased and abrupt changeover from lowest to highest risk.

The four methods

Method	How it works	Advantages	Disadvantages
Parallel	Old and new run at the same time on the same data; results compared until the new one is proven	Lowest risk; old system is a fallback; results can be checked	Costly; double workload; not always technically possible
Pilot	New system goes live in one location, department or group first	Problems found with limited impact; lessons for roll-out	Takes longer; the pilot site may not be typical
Phased	New system introduced in stages, by module or function	Risk spread over time; people learn gradually	Temporary interfaces between old and new parts; longer overall
Abrupt (direct, big bang)	Old system stops and new one starts at a set moment	Cheapest and quickest; no interfaces between old and new	Highest risk; failures hit everyone at once; hard to go back

Choosing a method

How critical is the system, and how much disruption can the business tolerate?
Can the old and new systems run side by side?
Is there budget and staff for double work?
How many locations and users are there?
How modular is the new system?

Example. A bank replaces its core banking system. Running both in parallel is impractical at its transaction volumes, so it uses a phased approach by product (savings first, then loans), pilots each phase at three branches, and has a tested rollback plan for every cutover weekend.

Always have a way back

Every changeover needs a documented rollback (fallback) plan, with the conditions for using it agreed in advance. It’s critical for abrupt changeover, because once data has moved on, the old system may be hard to restart. Even parallel changeover needs one, for example in case data gets corrupted.

Other go-live controls

a formal go/no-go decision by the business owner and IT
a detailed cutover runbook with checkpoints
converted data reconciled before users start
extra support in the first weeks (hypercare)
clear communication to users

How the exam asks about it

If the question asks…	The answer is usually…
Changeover with the highest risk	Abrupt / direct
Changeover with the lowest risk	Parallel
Main disadvantage of parallel	Cost and double workload
Main risk of phased changeover	Interfaces between old and new components
Implement at one branch first	Pilot

In short

Parallel changeover runs old and new systems together until the new one is proven; lowest risk, highest cost.
Phased changeover introduces the new system in modules or stages; risk spread out, but interfaces between old and new are needed.
Pilot changeover implements in one location or group first, then rolls out.
Abrupt (direct, big bang, cutover) switches completely at once; cheapest and fastest, highest risk.
Whatever the method, a fallback or rollback plan is needed.
The choice depends on criticality, risk tolerance, cost and whether the systems can run in parallel.

Exam traps

Abrupt changeover has the highest risk because there's no fallback once the old system stops.
Parallel is safest but needs extra resources and double data entry.
Phased needs temporary interfaces between old and new parts, which add risk.
A rollback plan is needed even for parallel changeover, for example if data is corrupted.

Slides

Click the slide and use the arrow keys. F toggles full screen.