Pasta should be cooked in salty water even if the numbers say otherwise. Olive Garden in order to get lifetime warranties on their pots, eliminated salt in the water and they still don’t salt the water as of today (April 2016). As a decent cook, and others agree, you need salt in the water if you want the pasta to taste good. Only some junior analyst who thinks Olive Garden is the pinnacle of Italian food would think saving about $80 a year, the price of pot or the loss of 3 customer meals, would have no impact. Plus to cover the lack of taste in the pasta, you have to add more salt and sugar, often corn syrup in commercial products, to the sauce.
When we use numbers, we need to make sure the numbers, analytics, or metrics correspond to a meaningful output. I’ve seen consulting companies cut consultants to increase savings, not recognizing that if you have fewer consultants, you’ll have less revenue since consultants are the product. In a classic case, the US government made selling ice cream and visits to beaches illegal in an effort to stop polio. The increased ice cream sales and visits to the beaches was due to the heat. The same heat that enables the polio virus to stay alive long enough to spread.
At SAP TechEd a few years ago, I watched with fascination as Nate Silver kept telling the interviewer that the thing we really needed was to have highly knowledge analyst who understand the complexity of real life topics – actual experts. The interviewer kept wanting him to expound on SAP tools, but Mr. Silver held his ground. It is not the tool, the metrics, or anything that is technical. Some tools are better or easier than others, and SAP has some great tools was all he’d say, but paraphrasing the interview, he said “you have to understand the subject, the relationships, and why a number might go up or go down”. Once you know the basics of the real system, you can use the numbers to look deeper, but not until.
If grammar can save lives (“let’s eat, grandma” vs. “let’s eat grandma”) then analytics can ruin them. Many decisions, such as layoffs or substituting equivalent products, are made on the basis of numeric analysis. As responsible analyst, we must make sure our metrics of success, failure, and our discoveries makes sense. Does the data really explain the results? Is it reasonable? What is scientific, logical reason for the outcome?
As we make more data available online and accessible, we must make sure it is clear what the data represents. Finally, we have a duty to be ardent guardians of proper use of analytics and be aggressive prosecutors when others misuse data and analytics. Numbers, and the conclusion we derive from them, directly impacts people’s quality of life. Number’s can hurt.