If you were setting up a register bit by bit it makes sense to work and think in binary. At that point you could use the binary value in the code. But there is one subtle reason not to. It is easy to look at a 2 digit hex pattern and reconize it. It is a lot harder to do with eight bits. A seasoned program might look at a register value of 0x5A and know that it is right. The brain would have a more difficult time with 0b01011010.
Also exclude 1, 2, 4 and 8 - now you only have 8 to go. 1110 is kind of a gimme, so is 11, 101 (mentioned) and 1001 - now you have four to go, B, D, C, and 6 - which you memorize by remembering that B and D are those two inverse ones, and the 6 is the one with the two in the middle.
But setting bits in registers using hex is SILLY. It's like dates in roman numerals or something.
Most ASM code and almost all C code I've ever seen use bit shifts never hex declarations.
Only time I see hex declarations are for eeprom/flash tables or headers.