Voluntary turnover among trained salespeople that have ongoing customer relationships is of concern to managers, due to the high costs involved in recruiting and training replacements and re-establishing relationships with customers. This research examines a new approach to assessing the impact of personal sales expectations on job satisfaction and tendency to leave or remain with a company. The modeling approach is tested in the case of an industrial commodity manufacturer wishing to improve retention of a large business-to-business sales force. Job satisfaction is related to the gap between actual and expected performance, so a model is developed for individual sales expectations and updated monthly using readily available data. Estimation of a hazard model for turnover confirms that salespeople whose performance exhibits larger negative gaps are more likely to leave, whereas those exhibiting larger positive gaps are more likely to remain with the company. This model permits more accurate and more frequent predictions of turnover than is possible using attitude surveys, as well as predictions for employees that do not respond to surveys.