Architecture of Demand Forecast for Online Retailers in China Based on Big Data

The paper designs a demand forecast system for online retailers based on big data and the theory of consumer behavior on the purpose of improving regional forecasting accuracy as well as shortening the forecasting periods. The necessity and the urgency of strengthening e-commerce enterprises' forecasting abilities are emphasized from the perspective of enterprises and consumers. Analysis and design process make use of the concepts of Engel-Kollat-Blackwell Model to put forward a framework composed of four phased prediction, which are the Initial Demand, the Possible Demand, the Core Demand and the Effective Demand. The further discussion then details how online retailers could implement this big-data-based system to cope with various shopping scenarios, such as the New Arrival periods and the November 11 Singles' Day shopping spree. By restating theoretical innovation in the field of demand forecasting research, the paper concludes with an outlook of potential improvement of the architecture and the direction of related practices.