Social-Network Influence on Telecommunication Customer Attrition

We investigate the network effects of churn in the telecommunication industry. Under calling party pays regime differentiation between on-net and off-net prices implies that customer's calling cost depends on operators chosen by the clients he calls. We assume that clients minimize their expenses. Therefore, after a single person churn we observe churn induced in his social neighborhood. Our aim is to verify, which measures of individual position in a social network are important predictors of induced churn. We control the results for changes in market prices structure, social network structure and number of operators on the market. Using multiagent simulation we show that (a) network structure and number of operators significantly influence induced churn level and (b) weighted prestige is its important predictor.