本論文為銷售分析與預測之實證研究以紙箱產業為研究對象,使用預測方法為紙箱產業建構市場銷售預測模式,並探討X公司四種產品類別未來市場動向。以預測結果對紙箱產業預估成長銷售之建議。在預測方面,本研究採取主要五種預測模型,分別是簡單移動平均模型(Simple moving average model)、加權移動平均模型(Weighted moving average model)、指數平滑模型(Exponential smoothing model)、最小平方法(Least squares model)以及季節調整趨勢法(Trend adjusted for seasonal effects model)。 本研究對象X公司以四種不同產品的類別,分別蒐集三年的銷售數據(2010年~2012年)進行實證分析紙箱類別銷售對於企業利損的情況以及產業需求的趨勢,本研究發現:(1)紙箱產品的銷售呈現四種不同的特性,分別是週期性、趨勢性、季節性和隨機性;(2)紙箱產品從2010年到2012年都呈現季節的趨勢,其中燈具箱及電線箱為最主力的產品,結果可以看出X公司四種產品中,銷售量最高前兩名分別是燈具箱(約佔整體銷售比將近一半)與電線箱(約佔整體銷售的三分之一),為X公司之核心產品與主要的獲利來源,其他產品之銷售量相對較低(佔整體銷售量五分之一左右),但也都穩定成長;(3)以MAPE值來衡量五種預測方法,其中燈具箱、電線箱、鐵管傢俱箱、羽毛球線等體育用品箱等適用季節調整趨勢法,同時電線箱亦適用最小平方法;(4)選擇適用的預測方法可以使管理者節省時間、成本降低、增加銷售及擴展利潤。
Sales forecasting is an integral part of business management. It refers to the process of evaluating what the firm’s sales will be in the future. A good prediction of sales will help managers make better business decisions on inventory management, cash flow, resource allocations, and firm’s growth. Carton industry is an interesting industry in Taiwan. The industry’s sales are affected by a number of factors, such as seasonality, irregularity, trend, and cyclicality. Thus sales forecasting is especially important for manager’s decision making in this industry. Using sales data on four product categories of company X in the period 2010-2012, the present study empirically analyzes and conducts different sales forecasting models: simple moving average model, weighted moving average model, exponential smoothing model, least squares model, and trend adjusted for seasonal effects model. The results of this study show that carton sales are affected by seasonal effect and present a stable growth trend. The five sales forecasting models show a very similar forecasting result, but trend adjusted for seasonal effects method is the best model that can generate the most precise forecasting results for carton sales. The results of this study provide importance contributions to both academic research and business managers in making better sales forecasting decisions.