Estimation of poor quality costs in automotive supply industry

The purpose of this study is to investigate and reveal the quantitative factors that affect the invoice amounts that issued by customer companies for inappropriate products on monthly basis, regarding the inappropriate products produced by a company operating in the automotive supply industry, with the regression model and after this, estimating the poor quality costs for the future periods through this model, and if this is not possible, it is the estimation of the poor quality costs for the future periods by trend analysis methods over the poor quality costs of the previous periods. In the research, the data which monthly product quantities produced and delivered at the relevant plants for the projects of each customer, product quantities rejected due to non-conformity, PPM ratios that calculating by rejected product quantities / delivered product quantities x 106, coefficient values of employee experience increasing monthly, product quantities under warranty and invoice amounts that make out by customer companies for inappropriate products, of the company which operates in the automotive supply industry and producing in its factories in Düzce, Manisa, Bulgaria, South Africa, Spain, Mexico and China, is used. These data, which were recorded by the relevant employees of the company, were obtained from the relevant tables in the database through a program prepared in C# programming language, and grouped on a monthly basis based on customer, plant, and project code. Among them, the data of a project with regular data were taken between January 2015 and November 2018. Regression analysis was performed using the invoice amounts as the dependent variable, the number of delivered items, the number of rejected items, PPM rates, employee experience coefficients, and the number of parts under warranty as independent variables and it has been tried to reveal the relationship between the dependent variable and which independent variables. After that, estimation was made with the trend analysis method by using only the data of the invoice amount, which is the dependent variable. Minitab 17 was used for regression model creation and trend analysis, and EViews 11 programs were used for stationarity control of the series.