Demand Forecasting in Supply Chains during Promotional Seasons: Differences between Judgmental and Data-Driven Approaches

In recent years, promotional events on e-commerce platforms such as the annual “Double 11” and “Black Friday” have become major shopping festivals and crucial periods for boosting sales. During these promotions, demand surges sharply in a short time, creating challenges for supply chains. Inaccurate forecasting may cause shortages or overstocking, leading to financial pressure and high return rates. Traditionally, managers relied on personal experience to predict sales, but in the era of big data, many firms are turning to statistical models and machine learning for greater accuracy. This study focuses on clothing and 3C (computer, communication, and consumer electronics) products, comparing experience-based forecasts with ARIMA time-series predictions. Results show that experience-based forecasting is flexible but less precise for clothing, while data-driven models perform better for 3C products. The research aims to guide enterprises in choosing forecasting methods suited to their product characteristics.