The experimental approach of RCTs and the inferential engine of “big data” and A.I.-based learning algorithms are challenging in the context of strategic actions, where the immediate feedback may not be strongly indicative of longer-run outcomes of interest.
