Learning When to Quit in Sales Conversations
A workshop with Eva Ascarza, William Joyce Professor of Marketing from New York University Stern School of Business.
Learning When to Quit in Sales Conversations
Emaad Manzoor, Eva Ascarza, and Oded Netzer (2025)
Salespeople frequently face the dynamic qualification decision of whether to persist in a sales effort or abandon it to pursue the next lead. Yet, little is known about whether these decisions are efficient or how firms can improve them. We study high-volume outbound sales, where leads are ample, time is scarce, and failure is common. We formalize dynamic qualification as an optimal stopping problem and develop a large language model (LLM)-based sequential decision agent --- a stopping agent --- that learns when to quit from historical conversations via offline reinforcement learning. When applied to calls from a large European telecommunications firm, our stopping agent reduces total call time by 67% while retaining 83% of observed sales. If the time saved is reallocated to new calls drawn from the same distribution, expected sales increase by up to 47%. Upon examining the linguistic cues associated with salespeople's quitting decisions, we find that their decisions are disproportionately associated with a few salient expressions of prospect disinterest and appear to overlook early indicators of eventual failure. These patterns are consistent with cognitive bounds on salespeople's ability to integrate and act on conversational information in real time. Our findings illustrate how offline reinforcement learning and agentic AI tools can be deployed to harness the natural language understanding capabilities of LLMs to improve the allocation of scarce salesperson time.
The European Quant Marketing Workshop
The European Quant Marketing Workshop i supported by EMAC's Quantitative Marketing SIG and Center for Retailing at Stockholm School of Economics.
This workshop series aims to help Ph.D. students, junior researchers, and everyone else who is interested in doing and publishing quantitative marketing works. We cover a diverse range of topics with the intent to spark your inspiration and motivation through learning about quantitative models and methods from established scholars in the field.