A Study of Models for Forecasting E-Commerce Sales During a Price War in the Medical Product Industry

Abstract

When faced with a price war, the accuracy of forecasting sales in e-commerce greatly influences an enterprise’s or a retailer’s merchandise inventory strategies. When faced with a price war, an enterprise might obtain certain consumption patterns by analyzing previous sales data. This case study research was conducted in collaboration with a medical product company to explore which of the various forecasting models can better inform a company’s inventory plan. The study used the company’s data from Amazon.com regarding sales volume, number of views, company ranking, etc. between February 7 2016 and March 28 of 2018. Three potential methods of data mining were selected from the literature: the exponential smoothing method, the linear trend method, and the seasonal variation method. Of these, the most suitable was identified for price war situations to forecast the sales volume for April 2018 and to provide concrete information for the company’s inventory plan. The results showed that the seasonal variation method is more suitable than the other two sales forecasting methods. To obtain a more accurate sales forecast during a price war, the seasonal variation method is recommended to be used in the following approaches: Adjust the seasonal index by using a simple moving average. Remove the seasonal index from the sales volume, and conduct a regression analysis using the data within the last month. The resulting predicted value (with the seasonal index removed) should be multiplied by each period’s corresponding weighted moving average to obtain a more accurate sales forecast during a price war.