Reconciling internal and external satisfaction through probabilistic graphical models: an empirical study

Purpose: Nowadays, service organizations have become increasingly aware of the fact that satisfied employees can positively affect customer satisfaction. This paper is a first stage of a wider research aiming at holistically reconciling internal customer (employees) and external customers satisfaction by using a statistical tool for multivariate data analysis i.e. Bayesian networks and their Object-Oriented version. Methodology: This study is based on survey data collected in an Italian hospital. For each ward a model has been estimated to evaluate the satisfaction drivers by category and some scenarios for the improvement of the overall variables are developed. A global model based Object-Oriented network is modularly built in order to provide aholistic view of internal and external satisfaction. The linkage has been reached by building a global index of internal and external satisfaction based on a linear combination. Findings: First, results achieved with Bayesian networks are consistent with the results of previous research obtained by using PLS-SEM tool. Moreover, probabilistic models allow to evaluate the impact of some scenarios for pursuing continuous improvement focusing on some variables of interest. Then, with the proposed global model, we aim at implementing some simulations in order to identify the strategic factors for achieving standards of global satisfaction. This is facilitated with an index useful for jointly managing internal and external satisfaction. Originality/value of paper: The novelty of the paper lies in the efforts to link internal and external satisfaction based on a probabilistic expert system which is able to generate scenarios of improvement. From an academic viewpoint in the service operations, the study is pioneer since it proposes a holistic approach to jointly manage factors affecting internal and external satisfaction