Analytics: Technical Knowledge versus Business Experience
When I was in elementary school we were doing word problems in math class to learn applications for long division. A memorable example went like this: “If a school has 1406 students going on a field trip and the capacity of a bus is 40, how many buses do they need?” Every single student in class said “35 with a remainder of 6,” proving that they understood long division… and also proving they didn’t understand the question. Unless they were planning to leave six students behind, the correct answer is 36 buses.
I remembered this example as I worked in a client location recently. They had a team of extremely technically proficient professionals that could dazzle with their knowledge of a host of data analytics techniques. The problem was that none of them really understand the business they were in. What if the absence rate at the school was 10%? Would 35 buses be enough? Sure. But what if the absence rate on field trip days was much lower, because parents that had paid for a trip were far less likely to let their kids stay home in borderline health situations? Then maybe 36 are needed. But what if the school board passed a resolution the year before that the school paid for all field trips, so parents didn’t feel the same sense of urgency to have their kids attend?
We could go back and forth all day, but here’s the point: There’s no substitute for understanding your business. Certainly the above problem could be attacked with discrete event simulation and machine learning and artificial intelligence and the like. But someone with an understanding of the business situation and the risk and cost associated with various options could make this decision without a lot of fanfare, and be correct far more often than the programmer trying to ensure all the factors are properly identified and accounted for.
Make sure you take the time to develop your technical people’s understanding of your business- or you may wind up with 6/35ths of a bus.