Position Auctions with Endogenous Product Information: Why Live-Streaming Advertising Is Thriving

Live-streaming advertising in e-commerce is soaring. Both Amazon and Alibaba have employed this novel marketing model to engage consumers by sequentially exhibiting different products through live-streaming videos. In this paper, we adopt a mechanism design framework to model live-streaming e-commerce as a position auction with endogenous provision of product information. We prove that finding the mechanism that simultaneously optimizes position allocation and information provision is NP-hard. Thus, we develop several approximation algorithms. Building on the connection to the order selection problem, our analysis establishes that heuristics relying solely on product information provision can achieve up to 66.9% of the optimal revenue in the worst case. In contrast, heuristics exploiting position allocation alone may result in arbitrarily large revenue losses. We attribute the efficacy of product information provision to its enhancement of the value of ad spots in position auctions. Our findings suggest that advertising that focuses on differentiating product information—as in live-streaming e-commerce—is more lucrative than conventional position-based sponsored search. This managerial insight remains valid when accommodating multiproduct purchases with dependency or accounting for random consumer attention spans. This paper was accepted by Omar Besbes, revenue management and market analytics. Funding: This work was supported by the National Natural Science Foundation of China [Grants 72201234, 72192805], the Hong Kong Research Grants Council [Grants 16211619, 16502219, 16212821, 16501722, C6032-21G], and the Guangdong Provincial Key Laboratory of Mathematical Foundations for Artificial Intelligence [Grant 2023B1212010001]. Supplemental Material: The online appendix is available at https://doi.org/10.1287/mnsc.2021.01299 .