The authors conduct a field experiment in which they vary the sales force compensation scheme at an Asian enterprise that sells consumer durable goods. With variation generated by the experimental treatments, the authors model sales force performance to identify the effectiveness of various forms of conditional and unconditional compensation. They account for salesperson heterogeneity using a hierarchical Bayesian framework to estimate the model. They find conditional compensation in the form of quota bonus incentives to improve performance; however, such compensation may lead to lower future performance. The authors find little difference in effectiveness between a quota bonus plan and punitive bonus plans framed as a penalty for not achieving quota. They find that unconditional compensation, in the form of reciprocity, is effective at improving sales force performance only when it is given as a delayed reward; however, the effectiveness of this plan decreases with repeated exposure. The authors also find heterogeneity in the impact of compensation on performance across salespeople, such that unconditional compensation is more effective for salespeople with high base performance, whereas conditional compensation is equally effective across all types of salespeople.