The Effect of Probabilistic Selling on the Optimal Product Mix

Determination of the merchandize assortment is an important decision for retailers since the composition and depth of the product mix greatly impact both unit sales and costs. This paper considers how Probabilistic Selling (PS), an emerging marketing strategy, impacts the type and number of products a retailer should carry. We find that adopting PS can alter the optimal number of products (i.e., encourage the retailer to offer more or fewer products), depending on demand- and supply-side factors. Furthermore, introducing probabilistic goods sometimes increases the optimal degree of product differentiation and sometimes reduces it. Specifically, less differentiated products are warranted if there are either few or many consumers with extreme tastes, but more differentiation is needed otherwise. Our analysis reveals that PS can serve either as a substitute to new product introduction (because it enables a retailer to serve a diverse market at a lower cost) or as a complement to new product introduction (since, under PS, a new product enables a retailer to offer additional probabilistic goods that utilize this new product as one of its components). In sum, our results indicate that a retailer must adjust its merchandize assortment appropriately in order to fully benefit from probabilistic selling.