One of the challenges we face when thinking about systems is believing they’re real. In an important sense, they’re not; systems are models, abstractions that describe causal relationships between things in the world and predict the outcomes of the interactions between them.
There are many ways to model a complex set of interrelated elements. For example, if you want to understand how a car works, you constrain the model’s boundaries to the car’s constituent parts. If, on the other hand, you care about how people use cars to go from one place to another, your model must also include the network of gas stations and refineries, the road network, and possibly the financial systems that make it possible for people to use private cars. Where you place the system’s boundaries will depend on what you’re interested to know out about the situation.
A good model allows us to intervene competently. On the other hand, an incomplete or poorly defined model can lead to ineffectiveness (at best) or disaster (at worst). Thus the importance of understanding systems: he or she who can produce the best model (the one that most accurately describes the system’s elements, relationships, purposes, dynamics, and so on) can act more skillfully than someone who can’t.