Comparison of Single Exponential Smoothing And Holt-Winter Exponential Smoothing Methods in Sales Commercial Business

Sales is an important aspect of running a commercial business. To increase sales, it must be accompanied by adequate availability of goods so that consumer needs can be met. However, the challenge in availability is that it is difficult to predict future consumer needs, so overstock and shortages often occur. On the other hand, the organization must be able to provide goods on time. This can be detrimental to the organization. For this reason, many solutions have been developed to overcome this problem, namely by predicting the availability of goods in the future using the single exponential smoothing and Holt-winter exponential smoothing methods. The single exponential smoothing method was used by previous researchers to provide good predictions on annual historical data. However, by using the same data converted into a monthly format, trend and seasonal factors were discovered. Therefore, we perform ex-inventory predictions using a single exponential smoothing method and then compare their performance. Trend and seasonal factors can be handled with the Holt-Winter exponential smoothing method. To determine the best prediction results from both methods for monthly data, we used the MAPE (Mean Absolute Percentage Error) method. The research results show that the Holt-winter exponential smoothing method has better performance with a smaller error value compared to single exponential smoothing, namely 1.828%.