ERP rollout: all countries at once, or one after another?

For most mid-sized companies, a wave rollout with a pilot country is the right approach: prove a representative country first, then roll out in groups of similar countries. A big bang across all countries at once only pays off when a hard date forces it, such as a carve-out or an end of support. It concentrates the entire risk into a single weekend.

When is the big bang still the right choice?

There are legitimate reasons to start all countries on the same day: a carve-out with a fixed date, an expiring legacy contract, an end of support without extension, or intercompany processes so interwoven that mixed operation is impossible. Then the big bang is not a test of courage but the honest answer to a constraint.

The price is clear: no learning curve, no second chance, peak load for key users in all countries at once. Choosing the big bang means more dress rehearsals, more cutover practice and a solid fallback scenario.

When the big bang is forced on you, the preparation moves forward rather than disappearing. It then takes two complete migration test runs instead of one, a cutover dress rehearsal per country, and a hypercare team that does nothing else for two weeks.

The honest sentence about it is: a big bang is more expensive, not faster. It saves the time of parallel operation and pays for it in preparation and risk.

How do you cut the waves correctly?

Not by size, but by similarity. Countries with the same business model and similar localisation belong in one wave, because there the work repeats instead of multiplying. A proven pattern for the DACH mid-market: wave 1 the pilot country, wave 2 the DACH neighbours, wave 3 Western Europe, then the more complex markets.

• Pilot: representative, but not the most complex market and not the headquarters
• Early waves: high similarity to the pilot, shallow localisation
• Late waves: complex localisation, own plants, difficult legacy data
• Per wave: incorporate findings, then freeze the template

The overall programme length is the counterweight: too many small waves means running two system worlds in parallel for years. Every wave costs its own cutover, its own interim interfaces and its own duplicate master-data maintenance.

What happens in mixed operation between the waves?

The uncomfortable truth of a sequential rollout: between the first and the last wave you live with two system worlds. Intercompany documents run via interfaces or by hand, consolidation needs bridging, and master data has to be maintained twice.

This interim state deserves planning like a sub-project of its own: with clear rules on which system leads for which data, and with an end date. Mixed operation that settles in becomes permanent.

Concretely, three things belong settled before the first wave goes live: which system leads for which data object, how intercompany documents run between the old and the new world, and how the group close is produced during this phase.

The end date is not cosmetic here. Without a date, parallel operation becomes a permanent state, and after two years double maintenance counts as normal. Whoever defends that date saves more than any optimisation inside the transition phase.

Who decides the sequence, and how do you recognise the wrong one?

The sequence is a business decision, not an IT decision: it distributes risk, ties up leadership capacity and determines when each company gets its figures from the new system. It belongs at programme level, with the CFO at the table.

Warning signs of a wrong sequence: the pilot is the most difficult country and never finishes. Or the easy countries are live, but nobody dares approach the complex ones, and the programme falls asleep. In both cases an external assessment helps more than another planning round.

A third warning sign is quieter: if the sequence was never justified within the programme but simply followed from availabilities, it is highly likely to be wrong. Risk distribution does not come out of calendar logic but out of a deliberate decision about which country carries which risk.