I keep coming back to the same observation: the argument about global standardisation never really goes away. It just changes costume. One year it’s about cybersecurity. The next it’s about cost. Then it’s about data, tooling, reporting, efficiency, speed. Same tension underneath.
In a global group, it’s a permanent line of friction between two legitimate instincts. Group wants coherence because coherence is what makes scale survivable. Regions want autonomy because autonomy is what keeps execution close to reality.
Both instincts are healthy. The problem starts when we treat them as enemies.
From where I sit, full standardisation across everything is not a plan. It’s a fantasy with a budget. Countries operate under different legal frameworks. Clients expect different rhythms and behaviours. Local operating models are not identical, and they shouldn’t be. Forcing uniformity where the business is structurally diverse tends to create workarounds, quiet resistance, and shadow solutions that nobody proudly owns.
At the same time, letting every region solve everything its own way is also not a strategy. It feels fast locally, right up until you try to connect the dots. You end up paying multiple times for the same capability, you multiply integration complexity, you compromise the reliability of group reporting, and you create uneven security posture. Fragmentation is expensive. It just hides its invoice well.
So the question I try to answer is not “standardise or decentralise.” I try to be more concrete. What must be common, and what can remain different without breaking the Group?
That is where my mental operating model has evolved. I use what I call a dandelion pattern: a multi-hub-and-spoke system rather than a single headquarters core with everything radiating out. In practice, that means a central backbone that is clearly steered and protected, and multiple regional hubs that are empowered to adapt, deliver, and run close to their markets.
I like the dandelion image because it forces a useful discipline. There is a stem that must be strong, because it carries the essentials. And there are multiple heads that can thrive in different conditions, because they are allowed to adapt without severing themselves from the stem.
The backbone is not negotiable. Identity and access, cybersecurity baselines, core data definitions, foundational infrastructure patterns, and the few platform choices that enable interoperability across the Group. If those drift by region, the Group becomes a collection of loosely connected companies. That might still work for a while, but it will not scale cleanly, and it will not be resilient under pressure.
Then there are the layers I deliberately leave open to regional hubs. Client-facing operational processes, certain workflow decisions, local execution patterns, and parts of the application layer where market reality differs. This is where autonomy creates speed rather than chaos. It is also where regions can innovate without waiting for a central queue.
The dandelion only works if the roles are explicit. Central teams must be accountable for the backbone and have the legitimacy to define the standards that make it coherent. Regional hubs must be accountable for adoption, local delivery, and the operational outcomes that actually matter in their markets. If that accountability is blurred, standardisation turns into endless negotiation and local autonomy turns into polite fragmentation.
This is also where communication matters more than people admit. I have learned that you cannot publish standards and expect adoption. If a region experiences the central solution as slow, unclear, or detached from real constraints, they will route around it. Not out of rebellion, but out of survival. So I have to explain the “why” behind what we standardise, be honest about trade-offs, and be equally clear about where flexibility is available.
I am not looking for uniformity. I am looking for interoperability. I want regions to keep their strength without turning the Group into a patchwork of incompatible systems and inconsistent risk exposure.
Standardising at global scale is not a one-time transformation project. It is an operating discipline. It is governance, architecture, and behaviour, repeated every month, under budget pressure, under delivery pressure, and under the temptation to take shortcuts.
When it works, it looks boring. That is usually a good sign.
