Comparison of Single Exponential Smoothing and Simple Moving Average Algorithms in Sales Coffee Forecasting

Abstract

Increasingly fierce competition, operational management, especially raw material management, has become a major challenge for many businesses. Therefore, companies or businesses that want to achieve maximum profits need a good and accurate sales prediction strategy for the coming period. Good predictions not only help in anticipating market needs, but also optimize inventory management to minimize the risk of losses due to excess or shortage of stock. The use of good forecasting algorithms is the main key so that companies can analyze historical data in depth to identify relevant patterns and trends so that they can improve operational efficiency. This research aims to compare two forecasting algorithm methods, namely Exponential Smoothing and Moving Average, in determining which method is superior in terms of sales prediction accuracy. The data used in this research comes from historical sales of Kopi Cucu Eyang Coffee Shop. Performance evaluation of the two algorithms was carried out using three main metrics, Mean Absolute Deviation (MAD), Mean Squared Error (MSE), and Mean Absolute Percentage Error (MAPE). The research results show that the Moving Average method is superior in MAPE accuracy with an average of 23%. On the other hand, Single Exponential Smoothing shows superiority in balancing MAD and MSE in certain products. It is hoped that this research can provide useful recommendations to improve inventory management efficiency and support better business decision making.