You’ve gathered your data and built a case. All of your metrics are in order. The evidence is clear. During the summer sales of ice cream go through the roof. Additionally crime rates go up. You continue gathering information. From these large data sets gathered you begin throwing charts together: Legitimate looking, incredibly spectacular charts and immaculate pivot tables. You now have a presentation along with your recommendations. You have all of the information you need in order to pitch your brilliant findings. You are the data detective, finding the root cause of problems in the multitude of processes around you.
After gathering your data you’ve determined that the root cause of the crime has to be the ice cream. Ice cream causes crime. The data correlates perfectly, it is full proof. The action needs to be swift. You get the buy in of all of the major players around you. Ice cream will be outlawed. You’ve guaranteed the return of civility and order to this mad summer of crime.
Alas, tragedy strikes. The crime rate continues to rise the following summer. Crime rates are worse than before. With a fist shaking to the sky you cry out, “But WHY!?” Defiantly you hit the books again. You pour over your data. You dust off your charts. Surely something must have been missed. Then you overhear a girl in the next cubicle asking her friend if she would like to go out for some “fro-yo”.
“Eureka!” you exclaim, and after dancing around the girl in joy you realize the root of the problem is now, frozen yogurt.
After several cycles of this and eliminating Italian ice, freeze pops, frozen bananas, and lemonade stands, the crime soars higher than ever before. You are relieved of your position the following summer.
Now you are both unemployed and generally irritable as there are no frozen treats available during the hot summer months. You are out of money, downtrodden, and desperate, and decide to go to the tourist area downtown to rob a store, knowing that the registers will be full because many people tend to travel during the summer months…
The moral:
Data analysis, problem solving, and root cause analysis: These are all major tasks, important to most large organizations these days. Although the whimsical ice cream story above is an extreme example, I see these sorts of mistakes being made all the time in analysis. While it is true that many times the solution may be obvious to an observer, it is always important to have the whole picture before jumping to conclusions. I’m not going over anything new, just more of a warning. Jumping to the wrong conclusion can cause devastating effects later on. “Fixing” the wrong part of a process has a tendency to only make things worse and it can be embarrassing for the person who identified the problem in the first place.
So how can we avoid these pitfalls? I think that this is simple. Be a data detective. The truth is in numbers usually, but make sure you have a true understanding of how a process works. Understanding that there are many interdependent parts in any organization or process makes analysis challenging sometimes to be sure. However, as long as all of the parts are accounted for we can truly see all of the variables at play. Once the picture is fully realized, make sure you write out all the variables and non-obvious hypotheses prior to jumping to your most obvious. You may end up running with your initial idea, but certainly you’ve given it the benefit of the doubt.