The problem and the target hardware will dictate if wasted space need be considered. Here are a couple reasons that would drive us to analyze wasted space:
If memory is a limiting factor (e.g. target device has a relatively small amount of memory)
If the data (element) type is of a relatively small size (see below)
It is common that data size is much larger than "wasted space".
If our structure contains "customers", and for example purposes, we estimate "Customer" objects average 2500 bytes each, then the wasted space (24/2500) is less than 1% of the data memory use, and thus becomes a minor factor.
In this case, we would generally not perform a "wasted space" analysis. (naturally there are always exceptions if later in the project we are desperate to find ways to save even small amounts of memory usage -- then we may undertake the analysis)
However, if we have a specific case where each data elements are relatively small, then may want to consider wasted space.
If our structure contains "lines", and for example purposes, we estimate "Line" objects average 48 bytes each, then the wasted space (24/48) is almost 50% of the data memory use.
In this case, we would want to do a memory space analysis.
Data Structures And Algorithms (DSA)
(Chapter 304 - Dynamic Array)