Innovation isn't the most time consuming part of a PhD. If you are doing benchwork there is an amount of time that is simply required to get experiments functioning.
I suppose I am thinking of it from math/coding perspective where the right idea saves a ton of time.