For the old-school vocal performance doubling stuff, it’s a weird combo of knowing just what u wanna do ahead of time, nailing it in the studio, being sure whether u like the take, then learning/internalizing/mimicking what you actually did and repeating it well on the overdub.
Someone like Paul MacCartney seems to have had a very clear idea what he wanted to do, and then did it efficiently and did the above process for his take and overdub immediately - while the inspiration, energy, and studio conditions were unchanged.
with more sophisticated techniques and digital, u have a lot of options - for example, u can take the vocal phrase by phrase, doing the take and double immediately, instead of waiting - but I find that’s harder because u don’t have the time to like/learn/mimic the take for the double, it’s more like archery on skis.