Statistical Modeling of Win Odds for Sales Opportunities

• Activity Number: 312
• Type: Contributed
• Date/Time: Tuesday, August 6, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Marketing
• Author(s): Ta-Hsin Li*+
• Companies: IBM
• Keywords: Statistical Modeling ; Sales Pipeline ; Markov chain ; Logistic regression

Abstract

Large enterprises use sophisticated information technology to track the progress of sales opportunities as their status evolve over time. A sales opportunity is entered into the database once it has been identified. Between the time when the opportunity is identified and the time when the opportunity is won or lost, the engagement typically undergoes a multistep sales cycle. The status of the opportunity in this sales process is regularly updated in the opportunity database as the engagement progresses. While the main purpose of such a database is for the management of sales activities, there is a growing interest recently in the possibility of using the data to shed some light on what to expect from the sales pipeline in terms of revenue, cost, and demand on resources (human or material). A key question is whether or not the data can be utilized to predict the win odds for each opportunity. In this paper, we propose a statistical framework for modeling and estimation of the win odds as functions of future time. We discuss several possible models with different complexity and sophistication. We compare the performance of these models using real world data.