The mention of Groundhog Day conjures up two interrelated references in North America. First is the 100+-year-old tradition of predicting the length of winter based on the early morning experiences of a large rodent. The second is a popular 1993 movie of the same name that tells the story of a weather forecaster played by Bill Murray who the universe forces to live the same day over and over until he achieves a greater understanding of his life and the lives of others. Oddly enough, I see lessons in both of those references for manufacturing and the supply chain community.
Since 1887, a groundhog (Marmota monax) by the name of Punxsutawney Phil and his predecessors have been awakened on February 2 by a mob of Northwest Pennsylvanians. Phil’s slumber is interrupted in a search for insights into how much longer winter might last. The legend is that if Phil sees his shadow on that morning, then there will be six more weeks of winter. If it is cloudy and Phil is unable to see his shadow, spring will come earlier. Since Punxsutawney is the Native American Lenape tribe name for “town of the mosquitoes,” it’s not exactly clear which outcome is the preferred scenario. Is it more desirable to endure extra weeks of harsh, lake-affect winter or hasten the arrival of a swarm of biting insects?
The analogy to supply chain and manufacturing is the legacy bias in the decision-making process. If you have ever witnessed the Groundhog Day event, you would quickly see that the human dignitaries around Phil have pre-decided about whether he does or doesn’t see his shadow. Unsurprisingly, Phil himself, like most small mammals, shows a vapid indifference to weather prognostication. There is no clear correlation to the strength of the shadow-generating sunshine on that respective morning and the prediction. In fact, since 1887 Phil has predicted an extended winter roughly 85% of the time. Something called the “Stormfax Weather Almanac” has been tracking Phil’s predictions since 1887. Phil’s biased prediction has been correct only 39% of the time.
What we have here is a decision-making process that is clearly not data-driven. If Phil flipped a coin in his tiny little rodent paws, he would be more accurate. It seems like over a hundred years of data points might guide a better success rate. Unfortunately, there are still a fair amount of manufacturing forecasts and planning systems that use the same decision-making rules year over year. Rules based on stagnant safety stock levels and expected order volumes that come from a copy-and-paste of last year’s values are likely to result in groundhog-like accuracy. Many manufacturers are most likely grateful that the Stormfax Weather Almanac is not tracking their accuracy.
The movie requires the protagonist weather forecaster to live the same day over and over with experimentation and repeated failure before he can go on with his life. The good news for manufacturers is that the path to a better reality is somewhat more straightforward. Use data and modern analytics to improve your predictions and your overall enterprise effectiveness. The emphasis here is on modern analytics. Using an average based on the last 100 years of data might seem logical but is likely to overlook the more recent trends. Clearly, Phil would want to weigh the more recent data and climate change trends in his weather evaluations. Anyone that thinks that the most recent calendar year is anything but an outlier must have been hanging with Phil in the groundhog’s den and missed the news of 2020.
The good news is that most manufacturers have a ton of recent and older data that may yield insights through more in-depth data analytics including artificial intelligence. Data evaluation can and should be nearly continuous and not based on at least an annual reevaluation. There is virtually no manufacturer that would not embrace better forecast accuracy for the next six weeks and beyond. It is time for all of us to get out from the shadows and make better data-driven planning and manufacturing decisions.
