DEMAND FORECASTING AND INVENTORY MODEL ANALYSIS CASE STUDY: PT ABADINUSA USAHASEMESTA

• Main Author: Muhammad Fauzan, Fazli
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/64417
• Institution: Institut Teknologi Bandung

PT Abadinusa is engaged in the healthcare industry, founded in 1980, as a trading company in this sector. PT Abadinusa regularly experiences a bloodline product overstocking up to 88.225 pcs by the end of 2019. This resulted in the the total inventory cost that must be paid by the company at Rp 767.661.570. PT Abadinusa uses only one method in forecasting their demand data, namely the naive approach which uses the demand data of the previous year to decide the amount of procurement in the following year. PT Abadinusa did not have the usual inventory management policies. That is why the company continues to experience overstocked from 2018 until early 2020. Before now the company's warehouses have an overflowing amount of bloodline. Therefore, the company needs to find the most appropriate demand forecasting approach and the most efficient inventory management technique to enforce. The objectives of this study are to determine the most appropriate demand forecasting approach and the most effective inventory management technique to be implemented in Bloodline product-based PT Abadinusa. This research gave the theoretical implications in term of a research framework on the analysis of Forecasting Demand and Inventory Management theory. Also, the framework that used to measure the forecasting error to determine the most effective forecasting method and measure the Total Inventory Cost to determine the most effective Inventory Model. Therefore, this research is expected to be useful for the other researchers that wanted to determine the best forecasting method and inventory model in minimize cost for the company to increase their profitability. Forecasting is an estimation of what will happen next. Simple Moving Average (SMA), single exponential smoothing, Holt’s and Winter's model are forecasting methods used in this research. The findings of these methods can be compared with the 3 formulas for errors including Mean Squared Deviation (MSD), Mean Absolute Deviation (MAD), and Mean Absolute Percentage Error (MAPE). To find out which method has the least forecast error. From the error evaluation, the Single Exponential Smoothing has the smallest error with MAD 4647, MSD 39563283, and MAPE 16 percent by fulfilling the condition of demand characteristic compared to the other methods like the company models, a 6-month moving average, Holt’s Model and Winter’s. This means that Single Exponential Smoothing is the most appropriate demand forecasting tool for the use of PT Abadinusa in Bloodline product. After choosing the best demand forecasting method to apply, the researcher using the data to implement an inventory management strategy using the POQ, Lot-for-Lot and EOQ model. The researcher calculating total inventory cost of each method in a year for 2018 and 2019 period. The data from forecasting results from the best method that already chosen which is Winter’s model will be used to estimate expected EOQ, frequency of order, and TIC for the 2020 period. The Total Inventory Cost based on the Lot-for-Lot, EOQ model, and Company model each year will be compared in order to determine which one is the most efficient approach in minimizing cost in a year and balancing input and output. Based on the EOQ (Economic Order Quantity) calculation on the economical number of the bloodline for each purchase in 2018 totaling 31891 pieces, in 2019 there were 30738 pieces, and 32190 pieces in 2020 (forecasted). Total Inventory Cost according to EOQ calculation in 2018 amounted to Rp 204.502.780, in 2019 amounted to Rp 197.295.026, in 2020 (forecasted) amounted to Rp 206.612.934 compared with the other methods. Thus, from the calculation comparison of the economic number of purchases and Total Inventory Cost between the EOQ model, Lot-for-Lot, POQ and the company model, EOQ method is the most appropriate one.