Depending on your DAW, I have been splitting the vocal track and raising or lowering the section to get my levels right, then using the fades for final adjustment/level of the track. I think it helps by using less compression and not using automation. I don't know why, but it seems automation is such a PITA while a work is in progress.
I wouldn't say my approach is the best, but it seems from a workflow it helps me at least. Makes final mixing less complicated.