
Adjusting division and stop volumes is better done in Organ settings | Rank voicing, rather than doing special things to your audio system that may only make sense for a small subset of your organs in the long run. You'll want your audio setup to be generic and robust enough to handle a wide variety of samplesets without a lot of fiddling.
Stereo samples generally sound better than mono, giving a better spacial impression of the organ. Some people do mono, but usually they use a lot more than 4 speakers.
Are your 4 speakers the same model? If so, 2 stereo pairs in one group can make good use of their capacity, letting Hauptwerk allocate pipes to stereo pairs.
Another common approach with 4 speakers is a stereo pair in front and a stereo pair in back, as separate groups for surround samplesets, with another group of both pairs as above for non-surround samplesets like Salisbury.
Do you intend to add a subwoofer and/or expand to more speakers? Then you could look at a larger audio interface, but for now I'd test and learn with what you have.
I do send each division to separate audio groups on many organs, but I have 3 - 4 pairs of speakers in each group. I wouldn't chop up a small number of speakers by division, personally. But you can certainly try SW CH on one pair and GT PD SO on the other (or any split you like) to see what it does for the sound.
Edit: In Danny's terms below, I use divisional routing with a Hauptwerk array for each division. But only 4 groups (arrays of speakers), so I combine divisions where applicable as GT + BD, SW + SO, CH + EC. PD separate. But for some smaller organs I use pitch-based layers with manual 16 - 8' to one group, 4'+ to another, pedal to another. In those cases, I group speakers differently. You can have overlapping groups, so I have more than one group scheme in the same preset.