Binary Logistic Regression as a Method of Predicting Customer Dissatisfaction in Resolving Complaints: The Case of Bosnia and Herzegovina, Serbia and Former Yugoslav Republic of Macedonia

The paper presents the results of customer satisfaction survey in the car spare parts industry through the presented analyses conducted in period from 2015 to 2018 on a sample of 3269 professional (target) customers of a company that operates in region of Balkans: Bosnia and Herzegovina (BIH), Serbia (SRB) and Former Yugoslav Republic of Macedonia (FYROM). The Binary Logistic Regression was used as a method of predicting customer dissatisfaction in resolving complaints by combining variables from the survey-questionnaire regarding the relations of employees and customers, customer satisfaction with the assortment, their satisfaction with the expertise of the staff, and the ratio of the value of the product in relation to its price. The study identifies that the ratio of employees, satisfaction with the assortment of goods as well as the quality ratio for the price, are statistically significant in the prediction model whether the customers will be dissatisfied with the resolution of complaints. For customers who estimate the ratio of employees to be competent, the probability is 3,367 times higher that they will be satisfied with the resolution of complaints, compared to those who considered the attitude of the staff as unsatisfactory. For customers who are satisfied with the assortment of goods, the likelihood is 4,435 times more likely to be satisfied with the resolution of complaints, compared to those who rated the goods assortment as unsatisfactory (not fully). For customers who consider the price and quality ratio to be satisfactory, the probability is 2,564 times higher that they will be satisfied with the resolution of the complaints, compared to those who consider this ratio to be unsatisfactory.