A company can manufacture, sell and employ people across several regions and still have one operating dependency that determines whether the wider system can respond.

The dependency may sit in a factory, a software stack, a cloud environment, a supplier, a regulatory approval path, a customer-data layer or a decision right. Geographic diversification does not remove that concentration automatically.

That distinction matters because the organisation can look diversified on a footprint slide while the option to move, substitute or localise remains narrow.

Country exposure is easy to see. Operating dependencies are usually harder because they cross functions and legal entities.

A production site may be replaceable while the tooling, engineering release, software integration or homologation path is not. A regional sales organisation may exist while customer identity, payments or data still depend on one central platform.

The resilience question therefore starts with the dependency, not the number of dots on the map.

Building a second version of every capability would destroy much of the scale benefit that made the original operating model attractive.

The decision is selective. Management needs to know which dependency could interrupt a strategically important market or product, how long substitution would take, and what preserving an alternative would cost before the disruption happens.

That can mean a second supplier, a regional interface, an alternative data route, contractual portability, local approval capability or simply an agreed exit path. The right answer depends on the value of the option being protected.

Regionalisation is often described as moving activities closer to markets. That is incomplete if the regional organisation still has no authority over the decisions that matter.

A hub with people but without data access, supplier rights, budget authority, customer visibility or the ability to change a local operating process remains dependent on the original centre.

The operating design therefore has to connect location with decision rights and interfaces. Otherwise regionalisation adds another organisational layer without creating meaningful optionality.

A useful review identifies the small number of dependencies whose failure would constrain a material market, product or cash flow.

For each one, management can then separate four things: the current advantage, the failure mode, the time and cost to restore an option, and who has authority to execute the remedy.

This creates a practical basis for deciding where resilience is worth paying for and where concentration remains economically rational.

Global operating models are built for scale, not for maximum reversibility. That is sensible until a dependency becomes strategically expensive to unwind.

The management discipline is to make that trade-off visible before the option is needed. A multi-region footprint is useful evidence of reach. It is not evidence that the critical operating system can move.