Double Moving Average and Double Exponential Smoothing Method in Sales Forecasting

Abstract

At present, predictions concerning sales are predominantly based on historical sales data, a practice that frequently yields inaccurate results. Such inaccuracies can lead to substantial financial losses, compelling organizations to lower the capital expenditures associated with certain products to offset these losses. This issue primarily arises from the failure to employ an appropriate forecasting methodology, resulting in estimations that lack reliable analytical foundations. This research aims to evaluate the efficacy of the Double Moving Average method compared to the Double Exponential Smoothing technique for sales forecasting, specifically through a case study involving herbal products. This study also seeks to analyze and compute the Mean Absolute Percentage Error (MAPE) for each forecasting method based on prior observations and research. The analysis draws on a dataset comprising 200 sales transactions from the five top-selling products collected between April 2022 and April 2024. The outcomes of this investigation provide MAPE values derived from the sales data, followed by a comprehensive summation of the calculated MAPE for each method. For June 2024, the results recorded are slightly higher in some cases. For example, the DES result for HNI Eucalyptus Oil was 42.65 with a MAPE of 0.46, while the DMA was 44.44 with a MAPE of 0.44.