Who you are or whom you know? Consumption interdependences in social networks

Abstract

This paper presents an analysis of whether a consumer's decision to switch from one mobile phone provider to another is driven by individual consumer characteristics or by actions of other consumers in her social network. Such consumption interdependences are estimated using a unique dataset, which contains transaction data based on anonymized call records from a large European mobile phone carrier to approximate a consumer's social network. Results show that network effects have an important impact on consumers’ switching decisions: switching decisions are interdependent between consumers who interact with each other and this interdependence increases in the closeness between two consumers as measured by the calling data. In other words, if a subscriber switches carriers, she is also affecting the switching probabilities of other individuals in her social circle. The paper argues that such an approach is of high relevance to both switching of providers and to the adoption of new products.

Keywords:

• social networks
• diffusion
• network effects
• survival analysis
• mobile telecommunications

Acknowledgements

The author gratefully acknowledges financial support from the University of Nottingham and the ESRC. He also thanks Peter Marsden, David Paton, Paul Stoneman, Peter Swann, Christophe Van den Bulte and participants at various conferences and seminars for helpful comments. Finally, he is grateful to two anonymous referees and the journal editor for their valuable comments. All remaining errors are solely mine.

Notes

Involuntary churn is, for example, caused if a subscriber does not pay his bills and his mobile phone is disconnected by the carrier. Although this type of churn is also correlated in a social network, it is not the focus of this paper, as coordinated decision-making is more likely if the consumer switches a carrier.

This paper focuses on direct network effects, as consumers are unlikely to widely spread information on their churning decision and there is no obvious way of determining that somebody has churned. This is different from, for example, from Granovetter (1983) or Gulati and Gargiulo (1999). Gulati and Gargiulo (1999), for example, study organizational networks where companies are likely to actively search for partners.