I may get disagreement here, but this is my view on soundstage. Side to side soundstage is enhanced by fewer drivers or more of a point source (i.e. old Tannoy speakers.) They can still be excellent provided that the drivers are in a single vertical line for each channel, although I think you get a bit of smear from the room because it will reflect back from more than a point source. We are basically looking at recreating the point source of microphones, correct?
Front to back soundstage is enhanced by dynamics, partly by how much louder the close material is compared to the rear instruments. You tend to hear less detail in the further instruments. But the main character of depth is the result of transients that recreate a 'space'. Close instruments are heard more directly with reflections taking a back seat, whereas farther instruments have reflected sound competing more with the direct signal recorded.
I think any part of the system that can reproduce these transients more clearly, will help resolve the front to back soundstage. The room can, of course, interfere. In fact, I often find that havings the speakers a few feet from the back wall creates more front to back.
Perhaps everyone here will have a different perspective as to which part of the reproduction chain will make the most difference. I say whatever the weak link in the system, that's what will provide the most improvement.