Narrowing Initial Bargaining Gap in Traditional Offline Transactions: A Machine Learning Approach

SYNOPSIS We examine how well machine learning algorithms can leverage historical pricing data to reduce the initial bargaining gap between sellers and buyers, aiming to improve personalized price effectiveness for traditional offline firms. In collaboration with a Chinese wholesaler, we construct a pricing model and apply it to real transactions. We show that under the model’s prices, the acceptance rate of initial pricing among buyers increases by 9.91 percent, and the adjustment frequency of initial pricing decreases by 23.52 percent, compared to prices set by traditional salespeople. Furthermore, interviews with practitioners reveal the model’s ability to streamline pricing workflows for salespeople and unlock new revenue potential for the company. Altogether, our findings demonstrate that machine learning can help traditional companies make better-personalized pricing decisions by narrowing the initial bargaining gap and concluding transactions more quickly.