Gilbert N. Lewis first proposed a conceptual framework for covalent bonding in 1915 and suggested a system of dots to represent in a simple manner how these bonds might be treated.
In the examples we have seen so far we have used a pictorial process of "fitting" in atoms to make octets. This works well for very simple molecules but it is sometimes difficult with more complex structures. For that reason we can introduce a series of steps which will generally result in a "correct" Lewis structure for most molecules.
Of course, there are always the exceptions..... Sometimes there is no reason to prefer one possible Lewis structure over another equivalent but different one.
And there are some molecules that just don't obey the simple rules.