In the 1990s, Procter & Gamble’s Product Supply Organization kicked off a major Reliability Engineering program, much like the efficiency initiatives of companies such as Toyota. They institutionalized the use of data collection systems in their manufacturing facilities to understand how products and machines would “behave” and could be optimized. By collecting machine failure data via manual sources as well as PLCs (programmable logic controllers), they were able to plot statistical distribution curves that predicted the failure rates of machines, along with the specific causes. More impressively, by linking all the machines together, they were able to predict — and subsequently improve — overall process reliability and product quality. This was all done through physical, smart, connected devices and sensors, which monitored, controlled, and optimized the units with increasing autonomy via continuous learning.