Referral, Learning and Inventory Decisions

Abstract

With the proliferation of digital social networks, businesses increasingly use referral programs to increase market exposure and sales. When customers refer a product to others they naturally disclose their purchase decisions. Thus the referral process introduces a social learning effect. We study the interaction between social learning and referral program structure and examine their impact on a firm's inventory decisions. We find that the presence of customers who lack knowledge of their own preferences introduces demand bias but social learning reduces this bias at the expense of increased demand variance. We characterize the optimal inventory levels for different numbers of referrals allowed by the firm and find that it is governed by the combination of \emph{market exposure effect} and \emph{demand substitution effect}. In a single referral program, the stock-out of one product can diminish the demand of the other product. In contrast, a multiple referral program allows a firm to achieve full market exposure but meanwhile increases the demand variance. Hence, the optimal referral program has to balance the trade-off between market exposure and demand variance, and thus allows either one or two referrals per customer.

Keywords: referral program, social learning, inventory management