I think shove as implemented tries to solve two problems: making the vertical axis and bonus actions more relevant.
As far as I know, the bonus action in 5e isn't used by every class, and certainly not at beginner levels. "Unlocking" a character's bonus action is meant to be a satisfying power up moment. I'm not sure why Larian made bonus actions ubiquitous (maybe it's to give players options, maybe it feels weird to have a bonus action and not be able to use it...) but they clearly feel being able to use them every turn is important. This cascades into a myriad of balancing issues. Classes built around the bonus action get comparatively worse and turning some actions into bonus actions opens the door to all kinds of cheese and exploits. I'd like to see the bonus action being more of a bonus than a second action. It's a sliding scale, of course. You could keep potions and offhand attack (without a main attack first) and ditch the rest.
I understand trying to make the vertical axis as relevant as possible. It's a huge part of the engine and Larian want to make the most of it. It's probably why height advantage was so strong at the start of early access. Besides high ground and shove there are only a handful of mechanisms that use height: line of sight, flight, fall damage, ladders (specifically, breaking wooden ones). Meanwhile, some low hanging fruit remains unpicked. I don't think high ground extends weapon range. Cone spells aren't actually conical and won't hit anything on a lower plane. Jumping on enemies works sometimes but the UI needs to be clearer to help avoid situations where I jump next an enemy and trigger combat damaged and prone.
In conclusion, I hate bonus action superman shove as much as the next player. There are other, better ways to accomplish Larian's goals.
