LAUSR

Abstract Demand variability is prevailing in the current rapidly changing business environment, which makes it difficult for a retailer that sells multiple substitutable products to determine the optimal inventory. To combat demand uncertainty, both strategies of inventory substitution and probabilistic selling can be used. Although the two strategies differ in operation, we believe that they share a common feature in combating demand uncertainty by encouraging some customers to give up some specific demand for the product to enable demand substitution. It is interesting to explore which strategy is more advantageous to the retailer. We endogenize the inventory decision and demonstrate the efficiency of probabilistic selling through demand substitution. Then we analyze some special cases without cannibalization, and computationally evaluate the profitability and inventory decisions of the two strategies in a more general case to generate managerial insights. The results show that the retailer should adjust inventory decisions depending on products' substitution possibility. The interesting computational result is that probabilistic selling is more profitable with relatively lower product similarity and higher price-sensitive customers, while inventory substitution outperforms probabilistic selling with higher product similarity. Higher demand uncertainty will increase the profitability advantage of probabilistic selling over inventory substitution.