Decision optimization in service supply chain: the impact of demand and supply-driven data value and altruistic behavior

Abstract

The rapid development of information technology has promoted the digital transformation of the service supply chain. Members can collect, store, and transform the value of data to gain profits. Due to the different roles during the service delivery, service integrators (SIs) and service providers (SPs) transform the data value from different sources, which leads to the demand and supply-driven data, respectively. As the leader of service supply chain, the SI may show altruistic behavior and share the data value with SPs. This study constructs a service supply chain consisting of two SPs and one SI and establishes five analytical models. Several important conclusions are obtained. First, the demand-driven data value leads to a decrease in the SI's optimal pricing and the SP's optimal value-added service level, leading to the “paradox of demand-driven data value”. Second, supply-driven data value leads to the increase in SI and SPs’ optimal decisions, and SI can get higher expected utility at no cost, achieving the "free-riding effect". Finally, there is a "transmission effect" among the altruistic behavior, demand-driven and supply-driven data value. When the parameters meet certain condition, customers can obtain an "optimal purchasing area" and obtain higher-level value-added service at a lower price.

Appendices

Appendix 1: Proof of Proposition 1

We first analyze SP₂′s optimal strategy. It can be calculated by the Eq. (3) that $\frac{{\partial^{2} U_{{SP_{2} }} }}{{\partial v_{2}^{2} }} = - 2ck < 0$, SP₂′s utility function is concave in v₂. In order to ensure that the SP’s optimal utility has the maximum value, $p > 2ck\eta v_{2}$ needs to be satisfied, otherwise, when the SI's pricing is too low, the SP will not provide the value-added service. For a given optimal p and v₁, SP₂′s best response can be derived as $v_{2}^{*} (v_{1} ,p) = \frac{p}{2\eta ck}$. Similarly, SP₁′s utility function is concave in v₁ and SP₁′s best response is $v_{1}^{*} (p) = \frac{p}{2k\eta }$ through the first-order-condition. By substituting the response functions $v_{1}^{*}$ and $v_{2}^{*}$ into Eq. (1), we have $\frac{{\partial^{2} U_{SI} }}{{\partial p^{2} }} = \frac{ - (\eta - 1)[4b\eta ck - (1 - \theta )(c + 1)]}{{ck\eta^{2} }} < 0$, the optimal retail price p^* of the SI can be derived through the first order optimal condition $\frac{{\partial U_{SI} }}{\partial p} = 0,p^{NN*} = \frac{2ack\eta }{{4bck\eta - (1 - \theta )(c + 1)}}$, and the optimal value-added service level $v_{SP1}^{NN*} = \frac{ac}{{4bck\eta - (1 - \theta )(c + 1)}},v_{SP2}^{NN*} = \frac{a}{4bck\eta - (1 - \theta )(c + 1)}$ can be obtained. The proof idea for the other four analytical models is similar, and the equilibrium solutions are shown in “Appendix 2”.

Appendix 2

See Table 4.

Table 4 Equilibrium solutions of the DN/SN/DS/DA model