Simple System Model
In a simple system model with inputs and outputs it is useful to consider it in the context of the human element and the desired deliverable. Ideally the output will be processed so that it provides just relevant and useful information so the person can achieve their deliverable. If there is too much input or not enough processing the person can become flooded with less than optimum or even paralyzed output. In the larger context preplanning with the deliverable in mind to start can help to fine tune the system to deliver what is needed for the intended result. This is why it is useful to start with goals that focus the collection of input. This may seem obvious but it is at the root of the problems of information overload and the knowing-doing gap.
If this overall context is forgotten, it can lead to mistaken metrics. For example, the preoccupation with and striving for an empty inbox (email inbox zero) rather than a full deliverable outbox. What really matters is what you deliver and ship. Not whether you processed all the inputs, which more often than not, are not all needed to get the required results.