(Un)conditional collection policies on used products with strategic customers

Abstract

For generation products characterized by frequent releases of new versions, when a new version is introduced to the market, the current version usually still has a remaining useful life. This creates a challenge for a firm to manage used product collection and upgraded product introductions at the same time, especially with the presence of a secondary market. This paper develops an analytical model to study the design and evaluation of two widely adopted collection policies for used products. Specifically, depending on whether a monetary reward for returning used products is associated with further purchases, we examine both unconditional (buyback) and conditional (trade-in) collection policies that take place in practice. We find that, in the absence of a secondary market, a conditional collection policy can outperform an unconditional one when the base product is durable and the residual value that the manufacturer can obtain after collection is intermediate. However, when an independent secondary market exists, allowing customers to trade used products with each other, any conditional policy cannot outperform the optimal unconditional policy. In particular, the two policies generate the same profit when the residual value is low; otherwise, the unconditional policy dominates as it effectively mitigates the cannibalization of the upgraded product sales by collecting more used products and reducing the supply to the secondary market. We also discuss the environmental impacts of these two collection policies. Our study helps to understand the impact of strategic customer behavior and the secondary market on the choice of used product collection policies, and provides manufacturers with guidance on the design of the optimal collection policy.

Managing demand slowdown: The interplay between trade‐ins and quality improvement

“…Hence, consumers are homogeneous in the first period but heterogeneous in the second period. This assumption follows Zhang and Zhang (2018) and Dong et al (2023). Additionally, we assume that the firm only knows the distribution of v in both periods.…”