Optimal Incentives for Salespeople with Learning Potential

We study a compensation problem for salespeople with learning potential. In our model, both the firm and sales agent are risk neutral and forward-looking; the agent can privately observe his skill, exert effort, and learn from experience; the firm can learn from the agent's choice and revise sales targets over time. The problem entails a dynamic tradeoff between exploiting learning, screening information, and maximizing efficiency. We find the optimal compensation plan differs substantially from the existing ones: it sets aggressive targets for expediting skill development, and pays the information rent for neutralizing the agent's misbehaving temptation over the entire relationship. We find learning drives the long-run outcomes; ignoring it can mislead compensation design and inflict substantial losses. Our results shed light on when and why firms distort sales, favor incumbents, and prefer long-term plans. By highlighting the critical role of learning in long-run performance, this study advances our understanding of salesforce theory and practice