They “sound design” the sounds with a lot of comp… ( for drums you can radicaly change the sound with a comp… )and when the sound is designed… they maximize it or detach it with another comp… and another comp at the end of the chain just before the output for the Brickwall effect… There is no rules with effects ^^
You also need to understand the way the order of a chain of effects works.
If you put an effect in the middle of a chain, it creates a (potentially) completely different effect than if you put it at the end, or the front.
After you understand that, you realize that as the people were tracking, and adding effects, these things can just add up. Once you have a long chain of effects, maybe you want to compress that; ad infinitum.
Mostly what you are seeing is the effect of effect experimentation over time, not a precise “point” or “reason”. I’ve personally rarely had an entire effect chain in mind when I started. I just keep adding, tweaking, listening, reiterating.
My personal opinion is there is a lot of over-compression, or savage use of compression, that leads to difficult-to-listen-to music. Too often though I see compression over-use where deft EQ and Volume use would have done fine. Just my opinion.
However, there are some well used neat tricks with using two-stage compression, and it’s something I use often on vocals and drums. Used in a subtle way, I’ll use one comp to fatten the tails of all the sounds, and the other to restrict the transient attacks. I achieve this through different attack settings. Open attack (say 10-50ms), tight release, deep threshold and serious make-up to get the fat sound. Short attack (1ms or less), tight or slack release depending on the sound, and a shallow threshold to tuck-in stray transients fairly transparently. The first of the two methods really benefits in filtered compression mode where the low freqs are ignored. Try it out on a snare for example.