Every ERP program eventually reaches the same fork in the road: switch everything on at once, or roll it out in waves. I've watched both approaches succeed and both fail. The deciding factor is rarely the technology. It's how well the choice fits your organization's appetite for risk and change.
Let me be clear from the start: there is no universally right answer. Anyone who tells you big-bang is reckless, or that phased is always safer, is selling you their comfort zone, not your outcome.
What each approach actually means
Big-bang means you cut over to the new system across all sites, functions, and processes on a single date. The old system goes dark. Everyone starts Monday morning on the new world.
Phased means you sequence the rollout — by geography, by business unit, by module, or by a mix. You run old and new in parallel for a period, then retire the legacy system piece by piece.
Simple on paper. The consequences are anything but.
The real trade-offs
Big-bang concentrates risk into one moment. When it works, it's clean: one dataset, one process model, no messy interfaces bridging old and new. You avoid the tax of running two systems side by side, and you get to the benefits faster.
But the downside is brutal if things go wrong. A failed big-bang can stop your ability to ship products, invoice customers, or pay suppliers — company-wide, on day one. There is no fallback that isn't painful.
Phased spreads risk over time. You learn from the first wave and fix problems before they hit the next site. A stumble in one region doesn't paralyze the whole company.
The price you pay is complexity and duration. You maintain temporary integrations between old and new. You carry the cost and cognitive load of two operating models for months, sometimes years. And momentum is hard to sustain — I've seen programs lose energy and sponsorship in wave four that were flawless in wave one.
When each one wins
In my experience, big-bang tends to win when:
- Your processes are highly interdependent and hard to split cleanly.
- The organization is relatively small or centralized, so coordinating one cutover is realistic.
- Running parallel systems would be more expensive or risky than a single sharp transition.
- Leadership can credibly commit to an intense, all-hands stabilization period.
Phased tends to win when:
- You operate across many sites or countries with different local requirements.
- The scope is large enough that a single cutover would be impossible to test properly.
- You can define clean boundaries — a business unit or region that can run semi-independently.
- The organization has limited tolerance for company-wide disruption.
The part everyone underestimates: change management
Here's what I've learned the hard way. The technical cutover is the visible half. The human cutover decides whether it holds.
Big-bang is a change-management sprint. Everyone crosses the same line at the same time, which is actually an advantage — there's no ambiguity about "the new way," no one clinging to the old system because it still exists somewhere. But you get one shot at readiness. Training, super-users, hypercare, and communication all have to peak on the same date. If people aren't ready, the whole company feels it at once.
Phased is a change-management marathon. The upside: you build a group of experienced champions from wave one who carry credibility into later waves. "It worked in Switzerland, and here's what we fixed" is far more persuasive than any slide.
The trap is fatigue and inconsistency. Different waves drift toward different local habits. The narrative gets stale. Sponsors who were front-and-center at the first go-live quietly disappear by the third. You have to actively refresh energy and guard the process standard, or phased quietly becomes many small custom systems wearing the same logo.
The approach you choose is a bet on where your organization is strongest — its ability to absorb one big shock, or its stamina to sustain change over time.
How I actually decide
I don't start with the method. I start with three questions.
First: How interdependent are the processes? The harder they are to separate, the more big-bang makes sense — because a phased split would create fragile, temporary bridges you'll regret.
Second: What's the cost of a company-wide failure versus a prolonged transition? If a bad day one would threaten the business, phased buys insurance. If dragging two systems for a year would drain more value than a sharp cutover, big-bang earns its risk.
Third: Can leadership sustain attention? Big-bang needs one intense, unified push. Phased needs durable sponsorship across many months. Be honest about which one your leadership can genuinely deliver — not the one that sounds better in a steering committee.
Answer those honestly and the method usually chooses itself.
The real failure isn't picking big-bang or phased. It's picking one to avoid a hard conversation — and then under-resourcing the change effort that any serious ERP program demands. The system doesn't transform your company. Your people do. Choose the path that gives them the best chance to succeed, and commit to it fully.
Cédric Bignet is an AI & ERP Change Management expert at Novartis and founder of AInspire. He writes about change management, AI adoption and enterprise transformation.