Development of Segmentation Algorithm for Identifying VOCs (Voice of Customers) with Sales Potential and Those with Negative Attitude at Call Centers

VOCs (Voice of Customers) coming through call centers involve a variety of contents such as complaints about products and services, requests for product information and questions about maintenance services, among others. By now, it has been widely recognized that VOCs constitute a valuable source of information for enhancing customer services. Since receiving more than 100,000 VOCs per month is not rare, however, it is virtually impossible to read all of them. Accordingly, it is important to mechanize a procedure for distinguishing VOCs of potential business importance from others, so that valuable information can be extracted from the selected VOCs with speed and cost efficiency. For this purpose, this paper develops a segmentation algorithm for identifying VOCs with Sales Potential and those with Negative Attitude based on a neural network approach combined with text mining.