Paradox is pretty talented at making everything shittier with each iteration. It's like they reinvent the wheel with each game, but somehow end up with worse and worse wheels. It's not just the dogshit UI or the map that seems to get uglier and uglier with each game, but also things like diplomacy – how the fuck do they keep remaking that same fucking system over and over and over again with every game, every time running on the exact same concepts, yet always come out with terrible dogshit that they need to iterate over, just to end with something that feels like a discount version of EUIV's? Is this some sort of an ego issue? "No way I'm copying stuff from past games, I will make everything myself, and better!"
It's because they have separate teams working on each game, rather than the whole studio working on one game at a time. There's probably some amount of crossover, but for the most part it's just one group of people working on HoI, and a different group of people working on CK, etc. This is partly why their engine was so fragmented and non-standardized prior to the Jomini subsystem (and to a degree, still is), and it's also why many of the games have wide variations in performance. This also likely means that most of the game design is separate for each game, and it's probably why every game has one thing it really focuses on and develops that the others all neglect (characters for CK, pops for Vicky, warfare for HoI... not sure what EU's focus is, diplomacy maybe? it feels like the most "generalist" game). Like in theory if you come up with a system for characters and internal sub-national governments in CK2 then you should be able to have that SAME system in Victoria 2 or 3. The reason they don't isn't that it would be too complicated, it's that it's a different team working on the system and they're on different versions of the engine and stuff can't be ported without a lot of effort even if they did want to collaborate to that extent, which they likely don't.
https://www.youtube.com/watch?v=M6rTceqNiNg
This is mostly a talk about how they implemented multi-threading, but there's some bits here and there that reveal the fragmented nature of the studio.