> It is 2018 and companies still don't know why most customers buy specific stuff. This is not altogether surprising: between perfectly boring options like "it

> It is 2018 and companies still don't know why most customers buy specific stuff. This is not altogether surprising: between perfectly boring options like "it was on sale", "I tried it [somewhere else] and liked it", and "my parents always used to buy it", there's not a lot of room for novel paths of discovery. But if you're a company who makes stuff, the differences may influence spending on one type of advertising vs. coupons or rebates. The process is messy and not rigorous: future quarters' sales will be used to point to some strategy working, despite dozens of variables they can't control. The whole system toes a fine line between trying something, and looking like you're trying something, while building up an elaborate apparatus to cover your ass. Tooling like this generates data, and works to justify the effort being spent. This extrapolates well to all complex bureaucracies. In the face of mounting competition, efforts like this are to gain an edge over competitors who may also do the same.