The impact of different recommendation algorithms on consumer search behavior and merchants competition

Abstract

Recommendation algorithms on platform markets can be categorized into neutral algorithms and non-neutral algorithms. We explore how these two algorithms affect consumer's search behaviors and merchant's competition behaviors based on a consumer search model. We found that as platform transitions from not providing recommendation algorithms to providing neutral algorithms and then to providing non-neutral algorithms, the price dispersion among merchants gradually increases, while the intensity of price competition decreases. When the difference in transaction utilities among merchants is small, providing neutral algorithms can enhance platform profits, consumer surplus, and social welfare. In the meantime, providing non-neutral algorithms always harms platform profits and social welfare, but still enhances consumer surplus. This study recommends that platforms should maintain a balance between neutral and non-neutral algorithms in the development of recommendation systems, where platforms can then guide merchants to focus their efforts and resources on product development and service improvement, rather than engaging in price wars and paid promotions.