automata – Different PDA design processes — both valid?

This video shows how to design PDA from a CFG:

Basically, we always have 4 basic states, and one of them is a “hub” for loops that implement separate rules like so (“q loop” state in the pic):

enter image description here

The author of the vid claims that any CFG can be converted into PDA using this basic structure.

Yet, I find PDAs that look very different.
E.g., this one:

enter image description here

My question: Is the single-hub method not always applicable? What are the general guidelines for when to use it and when not to use it?

If you could also explain how the PDA from the second image works step-by-step verbally, I would appreciate it.