Business Forecasting

Foreword

The forecasting spring began with the M4 Competition, where a complex hybrid approach combining statistical and ML elements came first, providing a 9.4% improvement in its sMAPE relative to that of the Comb benchmark.

Makridakis and Petropoulos (2020)

On March 12, 2020, U.S. toilet paper sales ballooned 734% compared with the same day the previous year, becoming the top-selling product at grocery stores by dollars spent. Clearly, forecasting did not predict the huge surge in demand that created panic-buying – demand that became exaggerated once photos of empty store shelves began circulating on social and mass media. Worse, the scarcity lasted for several months even as manufacturers rushed to produce and ship more paper.

This story reveals two important facts about forecasting. The first relates to its phenomenal failure during the pandemic. The second, even more important one, confirms its astounding success for all other periods when there has always been toilet paper available to buy.

Moreover, it is not just toilet paper. Walmart sells more than 50 million items in its 12,000 stores and online, and it is rare that customers could not find an item they want to buy. Its sophisticated forecasting system accurately predicts the demand for all its products, avoiding stockouts in the great majority of cases. The same is true with essentially most retailers that manage their supply chain in such a way that their customers could buy whatever they want at practically ...