Abnormal Event Analysis and Forecasting of Product Sales: The Case of Semiconductor Distributor in Taiwan

• Main Authors: Hsuan-Ping Peng, 彭煊萍
• Other Authors: Hsiu-Wen LIu
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/04636244826970656987

碩士 === 東吳大學 === 企業管理學系 === 100 === Usually the major abnormal events will cause the miserable impact to the global corporations and markets. For example: the US subprime mortgage crisis became the global financial crisis in the end. This study is using a domestic semiconductor distributor as the case study and through the outliers model in the time series analysis to evaluate the effects on the company revenues when it faced the abnormal events. By using the theories of Products Life Cycle Theory in Marketing Management and the outliers model in the Time Series Analysis from the statistics point of view to review its sales data which began from October 2005 to June 2011, total sixty nine months and found this study case did have abnormal events during this period and this is also cause the effect to its sales volume. In this study, comparison before and after the abnormal event, through the Time Series Outlier Analysis to detect outliers and then deleted, to identify and establish to the ARIMA (1,1,1) model. ARIMA (1,1,1) model for the assessment of the sales impact of abnormal events, calculate the cases of semiconductor distributor in three of its abnormal event, a total loss of USD 1,522,541 and its impact sales volume fell 25 percent down. Its three abnormal events were：First, the global financial crisis in September 2008, which caused the sales volume of semiconductor distributor has dropped a lot. Second, the white-box handset has been banned the importation by local governments in December 2009, which caused the sales volume of semiconductor distributor has been raised up in the short term. The third, due to unclear strategy of its major customer in February 2011, which caused the sales volume of semiconductor distributor has been declining .The empirical results show that: First, considering the mode of outliers is better than not considering the mode of outliers. Second, the model of outliers could assist the decision maker to capture the effects of abnormal events more accurately, and improve sales forecasting and inventory management capabilities. Expecting this study can be a reference for decision maker to manage the abnormal events by the experiences accumulated and to have more capabilities to predict and react to the abnormal events.