Forecasting Seasonal Products: A Case of Ice-Cream Sales

Abstract

The objective of this research is to develop a model to forecast sales for an ice-cream company. In order to achieve this objective, we evaluate sales data of three ice-cream flavors namely vanilla, chocolate, and Tally Ho (mixture of chocolate and vanilla) from January 2016 to 25 November 2019. To determine which model worked the best, we tested different models such as moving averages, simple exponential smoothing, Holt's method, Winters' method, method modeling seasonality and trend, and an ensemble method. We found Winters' method and modeling seasonality and trend performed well in terms of lowest error rates compared with other methods.

Citation

Trakroo, D. (2021), "Forecasting Seasonal Products: A Case of Ice-Cream Sales", Lawrence, K.D. and Klimberg, R.K. (Ed.) Advances in Business and Management Forecasting (Advances in Business and Management Forecasting, Vol. 14), Emerald Publishing Limited, Leeds, pp. 29-34. https://doi.org/10.1108/S1477-407020210000014003