A Structural Demand Analysis for Wireless Services under Nonlinear Pricing Schemes

Service industries are a rapidly growing sector in the US economy and account for over 60% of its total output. Wireless communications in particular have become ubiquitous and there are an estimated 160 million subscribers in the US alone. In this paper, we focus on the wireless communications industry and model the impact of the nonlinear pricing schemes employed by service providers on customers’ choice of calling plans, their consumption patterns and more broadly their decision to stay with the provider or defect. Within the wireless industry, pricing schemes are typically characterized by a combination of access fees, included free minutes (i.e. a marginal price of zero) and a per-minute marginal price for any consumption above the free minutes. Such pricing schemes are called increasing block pricing schemes as the applicable marginal price increases with increasing consumption. While this paper focuses on the wireless industry, increasing block pricing schemes are widely prevalent in other service industries and the modeling framework in this paper can be readily adapted to these other contexts. Within a wireless context, firms are interested in understanding a) how the different components of the pricing structure can differentially impact customers’ decisions of choice, consumption and defection and b) how these customer responses influence long-term profitability and customer value. Modeling consumer decisions for wireless services presents new challenges. For example, the presence of nonlinear pricing schemes creates an endogeneity between price and consumption. The pricing scheme of calling plans influences the consumption and the level of consumption in turn determines the applicable marginal price. Thus, there is a need to accommodate the entire nonlinear budget set comprised of the fixed access cost and differing marginal prices. In addition, consumers form an expectation of usage that influences their choice of calling plans and the pricing scheme of the chosen plan then affects consumption. Thus, the model also needs to allow for the choice and consumption to be temporally separated and yet be interlinked. In this paper, we develop a structural model for explaining consumer decisions within wireless services and address the above described challenges. We apply the model to longitudinal customer data from a large service provider and hierarchical Bayesian methods are used for estimation. The model gives several substantive results on how past consumption dynamics impact current consumer decisions. For instance, past under-utilization of free minutes either increases current consumption or influences customers to downgrade plans. We also use policy experiments to simulate the effects for price changes on plan switching and customer defection and link these changes to Customer Lifetime Value (CLV). There is evidence that increases and decreases in marginal prices lead to an asymmetry in customer choice responses. In contrast, no such asymmetry is present when access prices are varied. We also find that changes in access price have a greater impact on customer defection and long-term profitability compared to that of changes in marginal prices. Moreover, changes in access prices have greater impact on the customer lifetime value of “light users” while changes in marginal prices have a bigger influence on the long-term value of “heavy users”.