Businesses aren???t black boxes. Knowing the inputs doesn???t guarantee that you will know the outputs. They are complex and dynamic systems where changing one variable can have unexpected changes on the whole system.
I was remind of this truism while reading some research from a Wharton professor, Katherine Milkman, on how customers of an online grocery change their behaviour based on the availability of delivery slots. The findings are surprising. The timing of delivery slot effects the size and composition of the overall order:
Milkman’s research found that — for the most part — the longer the delay between placing an online grocery order and delivery, the less a given customer spends in general and the greater percentage of that order is allotted for healthier items, like produce, than for junk food. “Spending decreases as we order food further in the future,” Milkman says. “But the more immediate the gratification, the more freely we spend.”
The authors focused on the potential marketing implications of this insight but I think that it has important operational and process management ramifications as well.
Making a decision to extend or compress delivery schedules is an operational decision that might easily be made in isolation.
The impact would appear well-understood: You know the number of trucks and how many deliveries they can make; you know how many orders you get and their value; you can make assumptions for spare capacity and you know the pattern of the orders so you can plan for your peaks and troughs.
But, businesses aren’t black boxes. You can’t predict how customers will change their behaviour to repond to your decision. In this case, accepting longer average delivery waits could provide dramatic operational efficiencies but it would also reduce your revenue. People would buy less.
Sometimes, the best you can do it plan, measure and test and then get ready to do it again.