A Latent Class Regression Approach to IT Maintenance Outsourcing Service Management

Abstract

Although information technology (IT) maintenance outsourcing is widely used by today's organizations to improve IT productivity, client satisfaction of IT outsourcing service differs among various clusters of service incidents. Using a dataset collected from a leading Chinese steel group, this paper investigates how client satisfaction in IT maintenance outsourcing varies with service time and user involvement by means of a latent class regression (LCR) approach. Distinguished from traditional clustering and regression methods, the LCR approach can simultaneously classify reported incidents into multiple clusters based on incident attributes and examine the influences of service time and user involvement, together with the latent number of class on client satisfaction in each cluster. Our results identify four incident clusters within which influences of service time and user involvement on client satisfaction differ. Thus, we provide empirical evidence of heterogeneity in information system maintenance outsourcing incidents.