It seems several people here use the term gain staging as a continuous situation.
My opinion, well not just just mine, let's say the opinion or methodology I believe in is as follows.
You "stage" your tracks... then you mix and maintain those levels.
If you have ever seen a drag race they are constantly making little adjustments to the car, then they bring it to the line, burn in the tires, they are "staging" the car for maximum performance. Once the green light come on the staging is over, no going back.
Same with tracks. You prep your tracks and stage them for mixing. Some people do this as they are recording their tracks. They just know how to record at say -10 peak and -20 RMS. Others are not quite so good and maybe get some -3 peaks and -35 RMS. So the first example is basically staged. It's ready to race. The 2nd needs tweaking into that -10 / -20 zone so it too can race.
All or most tracks get that staging treatment so you can push up the faders and it sounds like song and the meters look good. Now maybe the 2 Mix is -5peak / -12RMS. No problem.
GREEN LIGHT --- Staging is over and the MIX is on.
Now you EQ, compress, distort, pan, saturate, reverberate, modulate, chorus, delay, ... all those things are processes on your totally staged track/s. These processes can and will change your levels going to your 2 Mix. So what can you do?
What you do is use your ears and meters to ---- maintain what you staged ----. You correct levels in all your processes so as to maintain you staging.
If you take a track and boost EQ for an Octave, well now you are above your "staged" level, so you need to correct the process level by lowering it back to staged level. An example of this is some EQ emulations that when you turn up say a high frequency shelf for "air" it will actually give a several dB boost all the way back down the frequency line. It becomes a "louder is better". IF you know that and want that, that's fine, but it's still above your "staged" level. Same thing with saturation, and distortions.
So process levels are not "gain staging". They are maintaining the staging. IF you could push all your faders up into a balanced mix in gain staging, you should be able to do the same after all processing and basically hit your 2 Mix meters at pretty much the same level.
If you play guitar or maybe keyboards and use FX. Every guitar player plugs in an FX and if it's role is simply to add an effect but no more gain, you just use your ears click it on and off while adjusting the output level. So now you have colored your sound but not increased what it's sending to your amp.
The difference between a Drag race and a DAW. After the green light you can go back in the DAW and try to re-gain stage varied tracks which is going to be a 'tail wagging the dog' situation.
This is not to say that if gain IN is 5 then gain OUT must be 5 also. But that's a good rule and you are comparing apples to apples and not being fooled by "louder is better". As you get closer to the 2 Mix maybe you boots some things. Also some tracks you might go in at 5 and want to come out at 4 for some reason.
So anyway, you want to stage your tracks so they can then be processed and leveled to make it to the 2 Mix finish line without blowing up. You can't add tons of processing and not re-adjust your levels back to nominal staged values and expect reasonable levels on your 2 Mix.
None of this is carved in stone, it's just a method to get from nicely staged tracks to processed tracks that don;t blow up your 2 Mix buss. This allows you to focus on the 'sound' of the processes you apply and not the mistakes and illusions you might be creating road blocks with.