At PT DBAS, the main problem faced is the problem of sales and demand for BABY-GRADE-A cotton buds. If a production shortage occurs, the company loses sales leading to overwork and costs. Likewise, if there is overproduction, even though the product is durable with a longer expiration date, warehousing problems may arise, such as limited warehouses, higher inventory costs, damaged products, and environmental problems like damage to the production floor. Thus, this study was conducted to discover a suitable forecasting method for BABY-GRADE-A to forecast its future demand, error numbers for each method, and sales figures for BABY- GRADE-A s in 2021 using the proposed forecasting method.
This research uses a quantitative descriptive by comparing the measurement of sales data errors in the Naïve, Moving-Weighted Moving Average, Single Exponential Smoothing, Holt's and Holt's Winter Exponential Smoothing, Trend Projections, and Polynomial Regression. To test the accuracy of the forecasting method, three measurements were used, namely MAD, MSE, and MAPE. Then, MRC and Paired Sample t-Test are to verify and validate the method.
Based on calculations and analysis, the most suitable forecasting method for BABY-GRADE-A products is Polynomial Regression. MSE and MAPE resulting from the method are 6,103.18 and 9.07%, respectively. Then, the forecast demand in 2021 is predicted to be 11,426 cartons. To implement, there are several aspects that should be considered by the company and future researchers such as market change, marketing strategy, inventory management and other operations activities.
Keywords: Forecasting, Forecasting Error, Moving Range Chart, Paired Sample t-Test, Polynomial Regression