Two topics in online auctions

This thesis studies two operations management topics in online auctions, and is divided into two parts. Motivated by the increasing use of ShopBots to scan Internet auctions, the first part of the thesis analytically examines whether or not two competing auctioneers selling the same commodity should share, or pool, some or all of their bidders. Under pooling, the bidding population is represented by three compartments: bidders dedicated to auction 1, bidders dedicated to auction 2, and pooled bidders participating in both auctions simultaneously. Under a bidder strategy shown to induce a Bayesian equilibrium, a closed form expression for the auctioneers' expected revenue under pooling is found, and pooling is recommended where it produces a greater expected revenue than no pooling (i.e., our objective is revenue maximization). Pooling is generally found to be beneficial as long as the two auctions are not too asymmetric and the underlying valuation distribution has certain concavity characteristics. Asymptotic order statistic arguments are used where explicit characterizations are intractable. The second part of the thesis considers a manufacturer who uses a reverse, or procurement, auction to determine which supplier will be awarded a contract. Each bid consists of a price and a set of non-price attributes (e.g., quality, lead time). The manufacturer is assumed to know the suppliers' cost functions (in terms of the non-price attributes). We analyze how the manufacturer chooses a scoring rule (i.e., a function that ranks the bids in terms of the price and non-price attributes) that attempts to maximize his own utility. Under the assumption that suppliers submit their myopic best-response bids (i.e., they choose their minimum-cost bid to achieve any given score), our proposed scoring rule indeed maximizes the manufacturer's utility within the open-ascending format.