Some systems are best left alone. For example, a rainforest can function perfectly well without human intervention. That’s a natural system that evolved into its current configuration over a long time, and it’s likely to continue adapting to changing conditions. (Barring some major environmental disruption.)
Most human-made systems haven’t had as much time to adapt; they’re aggregates of design decisions that may or may not effectively serve their intended purposes. Some of these interventions may truly be in service to the systems’ goals, but others may be driven by political motivations. (That’s one reason why you should think small when designing a system from scratch.)
As with the rainforest, conditions around the man-made system will change over time. How will the system address these changes? Designing the system itself is not enough; the design team must also design the system that continues the ongoing design of the system. We call this governance. Governance, government, governing; they all have to do with ongoing interventions aimed at keeping systems functioning as intended. These terms are all derivates from the Greek word kubernan (“to steer”), which is also the root for the word cybernetics. Governing is a quintessential systemic activity.
When do you intervene? How do you intervene? With how much force? How frequently? Who intervenes? If the intent is to keep systems functioning for a long time, these questions are essential. They also imply a corollary: you must know what you’re governing towards. What’s the purpose of the system? What are its intended outcomes? You can’t steer effectively if you’re unclear on the destination.