The barista: A model for bid arrivals in online auctions (original) (raw)

Modeling bid arrivals in online auctions

under review, 2004

We introduce a new family of non-homogeneous Poisson processes (NHPP) that are useful for modeling pure and contaminated self-similar processes which describe arrivals within a finite time period. Our motivation comes from the bid arrival process in online auctions. Modeling bid arrivals in online auctions is challenging since bidding dynamics change over the course of the auction. While the start of the auction typically sees an unusual amount of early bidding which is followed by a period of little activity, the auction end typically experiences an enormous amount of last minute bidding, also known as "sniping." This observed heterogeneity in bidding dynamics commands a very flexible class of models. We address these modeling challenges by proposing a class of 3-stage non-homogenous Poisson processes. We investigate the probabilistic and statistical properties of these models and illustrate their usefulness for fitting and interpreting real data from eBay.com.

23 Models for Bid Arrivals and

The arrival process of bidders and bids in online auctions is important for studying and modeling supply and demand in the online marketplace. Whereas bid arrivals are observable in online auction data, bidder behavior is typically not. A popular assumption in the online auction literature is that a homogeneous Poisson bidder arrival process is a reasonable approximation. This approximation underlies statistical models and simulations used in field studies.

Modeling the dynamics of online auctions: A modern statistical approach

Economics, Information Systems and E-Commerce …, 2006

In this work we propose a modern statistical approach to the analysis and modeling of dynamics in online auctions. Online auction data usually arrive in the form of a set of bids recorded over the duration of an auction. We propose the use of a modern statistical approach called functional data analysis that preserves the entire temporal dimension in place of currently used methods that aggregate over time thereby losing important information. This enables us to investigate not only the price evolution that takes place during an auction, but also the dynamics of the price evolution. We show how functional data analysis can be combined with cluster analysis and regression-type models for data exploration and summarization, and for testing hypotheses about relationships between price formation and other relevant factors.

Modeling first bid in retail secondary market online auctions: A Bayesian approach

Applied Stochastic Models in Business and Industry, 2019

We propose a Bayesian framework to model bid placement time in retail secondary market online business‐to‐business auctions. In doing so, we propose a Bayesian beta regression model to predict the first bidder and time to first bid, and a dynamic probit model to analyze participation. In our development, we consider both auction‐specific and bidder‐specific explanatory variables. While we primarily focus on the predictive performance of the models, we also discuss how auction features and bidders' heterogeneity could affect the bid timings, as well as auction participation. We illustrate the implementation of our models by applying to actual auction data and discuss additional insights provided by the Bayesian approach, which can benefit auctioneers.

Bidding behavior in dynamic auction settings: An empirical analysis of eBay

Electronic Commerce Research and Applications, 2010

We study the impact of dynamic features of eBay auctions on bidding behavior. Due to highspeed internet and practically costless search possibilities, bidding behavior is no longer a function of characteristics of a single auction but depends on auctions running simultaneously, completed auctions, available Buy-It-Now prices as well as various outside options. We study how this dynamic market affects a bidder's choice of participating in an auction or leaving eBay for an outside alternative. We analyze Texas Instruments (TI-83) Graphing Calculator auctions featured on eBay. We estimate a random-effects probit model to study bidders' probability of staying in eBay, while controlling for unobservable individual-specific heterogeneity. Our main result shows that market tightness -the ratio of bidders to sellershas a negative and significant effect on bidders' decision to remain in eBay. Moreover, variables containing information from other eBay auctions significantly affect bidders' participation decision, thus emphasizing the importance of the dynamic, multi-auction environment in eBay marketplace for potential buyers.

A simulation-based model for final price prediction in online auctions

Journal of Economics …, 2007

Online auctions, a profitable, exciting, and dynamic part of e-commerce, have enjoyed increasing public interest. However, there is still a paucity of literature on final price prediction for online auctions. Although Markov process models provide a mathematical ...

The Impact of Online Auction Duration

Decision Analysis, 2010

O ne view regarding auction duration suggests that longer auctions would result in more bidders and more bids, which in turn would result in higher prices. An opposing view is that shorter auctions might appeal to impatient bidders, or alternatively, that shorter duration might lead to more competitive dynamics. To examine these competing notions, we conduct pairwise comparisons of simultaneous auctions identical in all but duration. The auctions are conducted on two different platforms-eBay and a local auction site. We find that in eBay auctions, longer duration increases the number of bidders and bids, and consequently increases final prices by about 11%. On the local auction website, with far fewer auctions and a more steady set of participants, the effect is reversed, and shorter auctions generate higher prices by about 20%. Both sets of effects are robust and significant. We look at bidding activity on both sites to try to get at the root of that reversal. We find that in eBay auctions, the higher price in the longer-duration auction is accompanied by a higher number of participating bidders and a higher number of bids placed in the auction. In the local site, we find that the auction duration does not significantly affect the number of participating bidders or the number of bids placed in an auction. However, the magnitude of jump bids is negatively and significantly correlated with duration. These jump bids are in turn shown to impact final prices.

Modeling price dynamics in eBay auctions using differential equations

Journal of the American Statistical …, 2008

Empirical research of online auctions has dramatically grown in recent years. Studies using publicly available bid data from websites such as eBay.com have found many divergences of bidding behavior and auction outcomes compared to ordinary offline auctions and auction theory. Among the main differences between online and offline auctions is their longer duration (typically a few days). Along with the anonymity of bidders and sellers and the low barriers of entry, the longer online auctions tend to exhibit variable dynamics both in the bid arrivals and in the price process. In this paper we propose a family of differential equations models that captures the dynamics in online auctions. We show that a second-order differential equation well-approximates the three-phase dynamics that take place during an eBay auction. We then propose a novel multiple-comparisons test to compare dynamic models of auction subpopulations, where the population grouping is based on characteristics of the auction, the item, the seller, and the bidders. We accomplish the modeling task within the framework of principal differential analysis and functional data models.

Capturing the structure of Internet auctions: the ratio of winning bids to the total number of bids

Internet auctions have become popular and have a significant role in the further development of electronic commerce. Capturing the structure of Internet auctions, however, has not been sufficiently attained. An important feature in Internet auction is the existence of micro-macro link. As a first step to analyze micro-macro link we investigate the process from the macro-level phenomena to the micro-level behaviors. We focus on bidders' learning ability of the quoting prices and examine the real auction data in a certain period for the identical goods. To enable the analysis, we introduce the ratio of win- ning bids to the total number of bids (RWT). An interesting feature of RWT is that it pays attention to losing bids as well as winning bids. The results showed that the winners are superior in learning the quoting prices compared to the losers.

A Family of Growth Models for Representing the Price Evolution in Online Auctions

Several recent papers in the e-commerce literature have emphasized the importance of the price formation process and dynamics in online auctions . For example, the price velocity indicates how quickly the price is changing at every point during the auction. This knowledge can be used to build powerful dynamic forecasting models for price . It can also be used to visually "mine" a database of auctions for the same or similar products . Despite the usefulness of the price process and its dynamics in practical applications, methodological difficulties remain. Some of these difficulties arise from the lack of a single class of models that adequately captures all the different growth processes that occur during online auctions. Moreover, existing models may not be restrictive enough about the monotone behavior of ascending auctions or are computationally demanding. In this work we propose a new class of growth models for representing the price process in online auctions that is computationally efficient and preserves the monotonicity of the auction.