Economics Faculty Database
Arts & Sciences
Duke University

 HOME > Arts & Sciences > Economics > Faculty    Search Help Login pdf version printable version 

Publications [#346772] of Daniel Yi Xu

Journal Articles

  1. Hu, MM; Yang, S; Xu, DY, Understanding the social learning effect in contagious switching behavior, Management Science, vol. 65 no. 10 (October, 2019), pp. 4771-4794, Institute for Operations Research and the Management Sciences (INFORMS) [doi]
    (last updated on 2023/03/29)

    We study the contagious switching behavior related to a consumer’s choice of wireless carriers, that is, that a consumer is more likely to switch wireless carriers if more of their contacts from the same carrier have switched. Contagious switching (or a positive network effect) can be driven by information-based social learning, as well as other mechanisms related to network size. Although previous marketing literature has documented the social-learning effect, most of the applications studied involve products in which consumers usually do not enjoy any direct benefits from a large network other than from information-based social learning. We explore the importance of the social-learning effect relative to other mechanisms that may also lead to the network effect. We propose a dynamic structural model with interpersonal interactions. To model the social-learning effect, a consumer uses feedback from his or her contacts who have switched from a focal carrier to update his or her quality expectations of alternative carriers. Our model further accounts for two unique aspects of consumer strategic learning: (i) the individual’s perception on the signal of alternative carriers from contacts who switch is systematically different according to whether the signal comes from a loyal contact; and (ii) that the perceived noisiness of the signal on alternative carriers from a contact who has switched depends on the strength of the relationship between the individual and the contact. The remaining network effect not captured through social learning is modeled as a function of the size of the network. We solve the model with a two-step dynamic programming algorithm, with the assumption that a consumer is forward-looking and decides whether to stay with the same service carrier in each period by maximizing the total utility received from that day onward. We apply the proposed model to the data set of a mobile network operator in a European country. We find that churning/switching behavior is contagious in the network context and that one-third of general network effects can be attributed to social learning. We also detect strategic learning by consumers from their contacts in two ways: the experience signal on alternative carriers from a more loyal contact who has switched from the focal carrier is perceived to be more positive than that from a less loyal contact; and the social-learning effect is stronger from an individual’s closest contacts. The simulation analysis demonstrates the value of our model in helping a company prioritize its customer relationship management effort.

Duke University * Arts & Sciences * Economics * Faculty * Research * Staff * Master's * Ph.D. * Reload * Login