IMHO Variant Manager was a very unfortunate attempt to attract non-gamedev visualisation folks to the engine by creating way too task-specific tool that doesn’t really fit well into the philosophy of flexibility game engines tend to employ.
Something like variant manager is relatively easily scriptable with Blueprints. So what should have been done is a combination of tutorials explaining new users how they can script such functionality in BP themselves, along with a ready made blueprint in form of a content pack anyone can migrate into their project and start using, if they refuse to learn BP scripting.
This way, there would not be any hardcoded variant manager, and at the same time, users who would not find the BP variant manager sufficient could easily get in and adjust the BP or extend it with their own functionality. An issue such as actor still colliding with player pawn despite being hidden would be trivial to address if it was an editable Blueprint, but it’s almost impossible (for beginners) to address when it’s hardcoded engine feature.