I worked on a grocery store inventory system a while back, and it was very hard to get it rolled out. Produce can go bad, and get thrown out, so it is very hard

I worked on a grocery store inventory system a while back, and it was very hard to get it rolled out. Produce can go bad, and get thrown out, so it is very hard to accurately predict how much you will need. Further, your predictions depend upon how quickly it goes bad, which may not be easy to predict. Further, your predictions and orders depend upon people entering accurate data into the system. It is clear the employees are resisting putting accurate data into the system. Further, when you design the system you need to determine a probability of running out. The naive answers seems to be about 0%, but from an analytics standpoint a higher number may seem preferable.