Also a lay person. Maybe it's because of the time complexity. Let's say you have data of size 10 (any unit) which becomes size 8 after compression.

If you encrypt first: Encrypt size 10 of data -> Conpress size 10 of data

If you compress first: Compress size 10 of data -> Encrypt size 8 of data

So second way is faster as the compression still takes the same size of input but the encryption takes a smaller input.