A normal lens kit for dramatic cinema is 18mm, 25mm, 35mm, 50mm, and 85mm or 28-135mm on FF.
I find for music videos (on the 5D) I am likely to go very wide to make a space dramatic and make dolly moves feel "big" so I often go as wide as 14mm or 17mm, but then I usually crop to widescreen. For beauty I like to shoot as long as possible, 135mm or 200mm to flatten faces. So I don't think this is an easy answer. You need it all. For standard coverage of dance I like a UWA of the space, axial punch in that's just tighter than normal, mild WAs for the diagonals, and a portrait lens for the talent, straight on the face.
I also feel ZE lenses are very overrated (they're still good, but terrible values); that said they rent well and clients like them so they are good for pros who want a high end brand. But they're just average performers (with nice coatings, admittedly) and for shooting video at normal stops (f2.8-f8) you'll get the same results with anything. Since getting a 28mm, 35mm, 85mm, and 135mm to complete the kit is prohibitively expensive, I'd just go by feel. Take out a zoom for a while and see what you want. Focal length is all that matters; performance will be fine with anything. I'd go 28mm f2 since I like that focal length, but I also love 35mm on FF a whole lot. But then you don't have enough variation between that and the 50mm (which I also like on FF; 85mm is not as nice, but I use it occasionally).
So buy what you like and rent more when you shoot. If I am shooting a music video I'd like the 16-28mm or a 14mm L for the space, the 24-70mm zooms for coverage, and the 70-200mm zoom for additional coverage and beauty CUs. And three cameras, one with each lens. But I am not a pro so this is just...preference.