Despite much hype about electronic payments systems (EPSs), a 2004 survey establishes that close to 80% of between-business payments are still made using paper-based formats. We present a finite mixture logit model to predict likelihood of EPS adoption in business-to-business (B2B) settings. Our model simultaneously classifies firms into homogeneous segments based on firm-specific characteristics and estimates the model's coefficients relating predictor variables to EPS adoption decisions for each respective segment. While such models are increasingly making their presence felt in the marketing literature, we demonstrate their applicability to traditional information systems (IS) problems such as technology adoption. Using the finite mixture approach, we predict the likelihood of EPS adoption using a unique data set from a Fortune 100 company. We compare the finite mixture model with a variety of traditional approaches. We find that the finite mixture model fits the data better, controlling for the number of parameters estimated; that our explicit model-based segmentation leads to a better delineation of segments; and that it significantly improves the predictive accuracy in holdout samples. Practically, the proposed methodology can help business managers develop actionable segment-specific strategies for increasing EPS adoption by their business partners. We discuss how the methodology is potentially applicable to a wide variety of IS research.
Multisourcing, the practice of stitching together best-of-breed IT services from multiple, geographically dispersed service providers, represents the leading edge of modern organizational forms. While major strides have been achieved in the last decade in the information systems (IS) and strategic management literature in improving our understanding of outsourcing, the focus has been on a dyadic relationship between a client and a vendor. We demonstrate that a straightforward extrapolation of such a dyadic relationship falls short of addressing the nuanced incentive-effort-output linkages that arise when multiple vendors, who are competitors, have to cooperate and coordinate to achieve the client's business objectives. We suggest that when multiple vendors have to work together to deliver end-to-end services to a client, the choice of formal incentives and relational governance mechanisms depends on the degree of interdependence between the various tasks as well as the observability and verifiability of output. With respect to cooperation, we find that a vendor must not only put effort in a "primary" task it is responsible for but also cooperate through "helping" effort in enabling other vendors perform their primary tasks. In the context of coordination, we find that task redesign for modularity, OLAs, and governance structures such as the guardian vendor model represent important avenues for further research. Based on the analysis of actual multisourcing contract details over the last decade, interviews with leading practitioners, and a review of the single-sourcing literature, we lay a foundation for normative theories of multisourcing and present a research agenda in this domain.
Online auctions enable market-level interactions or interdependency of outcomes, which were not observed in physical auctions. One such set of interactions takes place when multiple auctions are conducted to sell identical items by an identical seller in an overlapping manner. This research focuses on overlapping auctions, their interactions, and the related impact on bidder behavior. We introduce the notion of auction "overlap" and examine the impact of market-level factors such as the price information revealed from prior auctions, degree of overlap, the auction format, and the overall market supply on a given auction's price. Despite a competitive setting, we find that, ceteris paribus, English auctions, on average, extract roughly 8.6 percent more revenue per unit than multiunit uniform-price Dutch auctions. We discover that the overlapping auctions attract institutional bidders, who bid in a participatory manner across multiple auctions, and that such bidders exert a downward pressure on auction prices. We find that overlap of an auction with other competing auctions has a significant negative influence on prices, and information about following auctions has a stronger negative influence than information about prior closing auctions. By estimating the expected price difference, we provide practitioners, who have private knowledge of their internal holding costs, a benchmark that can be used in deciding between using overlapping single-unit English auctions and multiunit Dutch auctions.
Despite the growing research interest in Internet auctions, particularly those on eBay, little is known about quantifiable consumer surplus levels in such mechanisms. Using an ongoing novel field experiment that involves real bidders participating in real auctions, and voting with real dollars, we collect and examine a unique data set to estimate consumer surplus in eBay auctions. The estimation procedure relies mainly on knowing the highest bid, which is not disclosed by eBay but is available to us from our experiment. At the outset we assume a private value second-price sealed-bid auction setting, as well as a lack of alternative buying options within or outside eBay. Our analysis, based on a sample of 4,514 eBay auctions, indicates that consumers extract a median surplus of at least $4 per eBay auction. This estimate is unbiased under the above assumptions; otherwise it is a lower bound. The surplus distribution is highly skewed given the diverse nature of the data. We find that eBay's auctions generated at least $7.05 billion in total consumer surplus in 2003 and could generate up to $7.68 billion if the private value sealed-bid assumption does not hold. We check for the validity of our assumptions and the robustness of our estimates using an additional data set from 2005 and a randomly sampled validation data set from eBay.
While traditional information systems research emphasizes understanding of end users from perspectives such as cognitive fit and technology acceptance, it fails to consider the economic dimensions of their interactions with a system. When viewed as economic agents who participate in electronic markets, it is easy to see that users™ preferences, behaviors, personalities, and ultimately their economic welfare are intricately linked to the design of information systems. We use a data-driven, inductive approach to develop a taxonomy of bidding behavior in online auctions. Our analysis indicates significant heterogeneity exists in the user base of these representative electronic markets. Using online auction data from 1999 and 2000, we find a stable taxonomy of bidder behavior containing five types of bidding strategies. Bidders pursue different bidding strategies that, in aggregate, realize different winning likelihoods and consumer surplus. We find that technological evolution has an impact on bidders™ strategies. We demonstrate how the taxonomy of bidder behavior can be used to enhance the design of some types of information systems. These enhancements include developing usercentric bidding agents, inferring bidders™ underlying valuations to facilitate real-time auction calibration, and creating low-risk computational platforms for decision making.
We present a simulation approach that provides a relatively risk-free and cost-effective environment to examine the decision space for both bid takers and bid makers in web-based dynamic price setting processes. The applicability of the simulation platform is demonstrated for Yankee auctions in particular. We focus on the optimization of bid takers' revenue, as well as on examining the welfare implications of a range of consumer-bidding strategies--some observed, some hypothetical. While these progressive open discriminatory multi-unit auctions with discrete bid increments are made feasible by Internet technologies, little is known about their structural characteristics, or their allocative efficiency. The multi-unit and discrete nature of these mechanisms renders the traditional analytic framework of game theory intractable (Nautz and Wolfstetter 1997). The simulation is based on theoretical revenue generating properties of these auctions. We use empirical data from real online auctions to instantiate the simulation's parameters. For example, the bidding strategies of the bidders are specified based on three broad bidding strategies observed in real online auctions. The validity of the simulation model is established and subsequently the simulation model is configured to change the values of key control factors, such as the bid increment. Our analysis indicates that the auctioneers are, most of the time, far away from the optimal choice of bid increment, resulting in substantial losses in a market with already tight margins. The simulation tool provides a test bed for jointly exploring the combinatorial space of design choices made by the auctioneer's and the bidding strategies adopted by the bidders. For instance, a multinomial logit model reveals that endogenous factors, such as the bid increment and the absolute magnitude of the...