This paper identifies the drivers of the phenomenal growth in productivity in hydraulically fractured horizontal oil wells producing from the middle member of the Bakken Formation in North Dakota. The data show a strong underlying spatial component and somewhat weaker temporal component. Drivers of the spatial component are favorable reservoir conditions. The temporal component of well productivity growth is driven by increasing the number of fracture treatments and by increasing the volume of proppant and injection fluids used on a per fracture treatment basis. Random Forest, a non-parametric modeling procedure often applied in the context of machine learning, is used to identify the relative importance of geologic and well-completion factors that have driven the growth in Bakken well productivity. The findings of this study suggest that a significant part of the well productivity increases during the period from 2010 to 2015 have been the result of improved well-site selection. For the more recent period, that is from 2015 through 2017, part of the improved well productivity has resulted from substantial increases in the proppant and injection fluids used per stage and per well.