It allows for different levels of precision for main plots and sub-plots

It completely eliminates block effects

It requires fewer experimental units than a randomized block design

It removes the need for randomization within blocks