Maybe not as real as we think. And certainly less real than our everyday conversations generally suggest. Naming something as a system does not make it a system. Nor does it validate the use of system thinking in the situation.
However for various purposes, and under certain conditions, we can proceed 'as if' we are dealing with a system, at least for the time being. The danger lies in assuming that some aspect of a situation can be treated as if it is a system. As Marshall McLuhan said, when we name something we stop thinking about it.
The central idea underpinning systems thinking is that, to a greater or lesser extent, we can rely on the existence of certain patterns of interaction between the elements involved.
In the physical world interactions are shaped by largely universal natural laws, hence the consistency of many interactions is usually very high. Drop a raw egg from a height of two metres onto concrete and the acceleration of the egg and the outcome can be readily predicted - the interactions between the egg and gravity and the subsequent interaction between the egg and concrete are highly consistent for eggs in this situation. Hence, we can address the situation with systems thinking.
Highly consistent patterns can also emerge in human interactions. Not so much because of natural laws but because of shared belief systems, shared goals and agreed practices, social regulations, corrective consequences... Shared beliefs, agreements, regulations, and consequences can act as 'attractors' around which consistent actions and interactions emerge.
Such situations can often be treated 'as if' they are systems, especially if
- the goals/expectations are known and agreed
- inputs are consistent
- processes are clear and practiced
- 'outcomes' are measured and reported against the goals/expectations
Systems thinking can then be applied so that changes and improvements are achieved
- increasing the consistency of the interactions
- changing the interactions to
-> use different (new, less...) inputs
-> reduce variation 'errors'
-> make interactions easier (so that the 'system' is more efficient)
-> improve outcomes (so that the 'system' is more effective)
That is, by continuously monitoring a situation and using informed judgement it may be valid to treat the situation as if it was a system.