eBay Online Auctions Research Articles

Incorporating the Time–Order Effect of Feedback in Online Auction Markets through a Bayesian Updating Model

Online auction markets host a large number of transactions every day. The transaction data in auction markets are useful for understanding the buyers and sellers in the market. Previous research has shown that sellers with different levels of reputation, as shown by the ratings and comments left in feedback systems, enjoy different levels of price premiums for their transactions. Feedback scores and feedback texts have been shown to correlate with buyers’ level of trust in a seller and the price premium that buyers are willing to pay (Ba and Pavlou 2002; Pavlou and Dimoka 2006). However, existing models do not consider the time-order effect, which means that feedback posted more recently may be considered more important than feedback posted less recently. This paper addresses this shortcoming by (1) testing the existence of the time-order effect, and (2) proposing a Bayesian updating model to represent buyers’ perceived reputation considering the time-order effect and assessing how well it can explain the variation in buyers’ trust and price premiums. In order to validate the time-order effect and evaluate the proposed model, we conducted a user experiment and collected real-life transaction data from the eBay online auction market. Our results confirm the existence of the time-order effect and the proposed model explains the variation in price premiums better than the benchmark models. The contribution of this research is threefold. First, we verify the time-order effect in the feedback mechanism on price premiums in online markets. Second, we propose a model that provides better explanatory power for price premiums in online auction markets than existing models by incorporating the time-order effect. Third, we provide further evidence for trust building via textual feedback in online auction markets. The study advances the understanding of the feedback mechanism in online auction markets.

Two-sample inference for sparse functional data

In this paper, we develop an asymptotic χ2 test for detecting differences among mean functions when sparse and irregular observations are drawn from the underlying continuous stochastic processes for each subject. We provide theoretical arguments to justify the effectiveness of the proposed test procedure. Numerical experiments, including simulation studies and applications to a CD4 count data set and an eBay online auction data set, are presented to demonstrate the good performance of the developed test.

Social Big Data Analytics of Consumer Choices: A Two Sided Online Platform Perspective

This dissertation examines three distinct big data analytics problems related to the social aspects of consumers’ choices. The main goal of this line of research is to help two sided platform firms to target their marketing policies given the great heterogeneity among their customers. In three essays, I combined structural modeling and machine learning approaches to first understand customers’ responses to intrinsic and extrinsic factors, using unique data sets I scraped from the web, and then explore methods to optimize two sided platforms’ firms’ reactions accordingly. The first essay examines “social learning” in the mobile app store context, controlling for intrinsic value of hedonic and utilitarian mobile apps, price, advertising, and number of options available. The proposed model extracted a social influence proxy measure from a macro diffusion model using an unscented Kalman filter, and it incorporated this social influence measure in a mixed logit choice model with hierarchical Dirichlet Process prior. Results suggest significant effects of social influence, which underscores the importance of choosing different marketing policies for pervasive goods. The comparison of mobile app adoption parameters suggests that among several classical goods mobile app adoption pattern is very similar to that of music CDs. The simulation counterfactual analysis suggests that early targeted viral marketing policy might be an optimal strategy for the app-store platforms. The second essay investigates bidders’ anticipated winner and loser regret in the context of the eBay online auction platform. I developed a structural model that accounts for bidders’ learning and their anticipation of winner and loser regrets in an auction platform. Winner and loser regrets are defined as regretting for paying too much in case of winning an auction and regretting for not bidding high enough in case of losing it, respectively. Using a large data set from eBay and empirical Bayesian estimation method, I quantify the bidders’ anticipation of regret in various product categories, and investigate the role of experience in explaining the bidders’ regret and learning behaviors. The counterfactual analyses showed that shutting down the bidder regret via appropriate notification policies can increase eBay’s revenue by 24%.The third essay investigates the effects of Gamification incentive mechanisms in an online platform for user generated content. I use an ensemble method over LDA, mixed normal and k-mean clustering methods to segment users into competitors, collaborators, achievers, explorers and uninterested users. Then, I develop a state-dependent choice model that accommodates the effect of number of badges, the rank in the leaderboard, reputation points, inertia, and reciprocity, and allow for heterogeneity by Dirichlet Process prior. The results suggested that estimating the model on small samples generate biased estimates. Furthermore, they suggest that the effects of Gamification elements are heterogeneous, significantly positive or negative for different users. I found sensitivity patterns that explain importance of certain Gamification elements for users with certain nationalities. These findings help the Gamification platform to target its users. The simulation counterfactual analysis suggests that a two sided platform can increase the number of user contributions, by making earning badges more difficult.

Conceptualizing means-end chains of user goals as networks

Goals are desired states that an individual tries to attain. The process of achieving a goal can be represented as interlinked means-end chains of user goals that have been traditionally visualized as hierarchies. Evidence in recent literature suggests that a network structure would be more appropriate and provide insight into a user's process of seeking a goal. We investigated user goal means-end chains for the eBay online auction system, and produced its structure as a goal network. To analyze this network and assess the importance of various goals, social network analysis measures were used (specifically, degree and flow-betweenness centrality). In addition, goal networks for users with low and high IS value were created and differences in goal importance in the two groups were considered. Results revealed that the most important user goals are closely related to key features of the auction system; users with high IS value want to use eBay to buy, sell, and bid for products, while users with low IS value seem to avoid using eBay because of uncertain price bidding. As such, the results of our study suggest that differences in IS value may be due to differences in IS usage. IS designers, marketers, and providers of online auction system can use our findings to design and promote better systems for their users.

Nonparametric panel estimation of online auction price processes

In this study, we examine the price process of eBay online auctions. We pool sparse and unevenly spaced bidding histories of individual auctions, resulting in an unbalanced panel of bids. Since the price processes of online auctions are monotonically increasing within individual auctions and exhibit a substantial degree of heterogeneity, we propose a monotone series estimator for panel data with auction-specific slopes for a common relative price growth curve. We generalize Ramsay’s (1998, J R Stat Soc B 60(2):365–375) monotone series estimator to fit our panel model of the price process. We apply the proposed model and estimator to eBay auctions of a popular hand-held device (a Palm PDA). The results are shown to capture closely the overall pattern of observed price dynamics. In particular, the early bidding, mid-auction bidding drought, and end-auction bid sniping are well approximated by the estimated price curve.

Distance-based clustering of sparsely observed stochastic processes, with applications to online auctions

We propose a distance between two realizations of a random process where for each realization only sparse and irregularly spaced measurements with additional measurement errors are available. Such data occur commonly in longitudinal studies and online trading data. A distance measure then makes it possible to apply distance-based analysis such as classification, clustering and multidimensional scaling for irregularly sampled longitudinal data. Once a suitable distance measure for sparsely sampled longitudinal trajectories has been found, we apply distance-based clustering methods to eBay online auction data. We identify six distinct clusters of bidding patterns. Each of these bidding patterns is found to be associated with a specific chance to obtain the auctioned item at a reasonable price.

Modelling Price Paths in On-Line Auctions: Smoothing Sparse and Unevenly Sampled Curves by Using Semiparametric Mixed Models

Summary On-line auctions pose many challenges for the empirical researcher, one of which is the effective and reliable modelling of price paths. We propose a novel way of modelling price paths in eBay's on-line auctions by using functional data analysis. One of the practical challenges is that the functional objects are sampled only very sparsely and unevenly. Most approaches rely on smoothing to recover the underlying functional object from the data, which can be difficult if the data are irregularly distributed. We present a new approach that can overcome this challenge. The approach is based on the ideas of mixed models. Specifically, we propose a semiparametric mixed model with boosting to recover the functional object. As well as being able to handle sparse and unevenly distributed data, the model also results in conceptually more meaningful functional objects. In particular, we motivate our method within the framework of eBay's on-line auctions. On-line auctions produce monotonic increasing price curves that are often correlated across auctions. The semiparametric mixed model accounts for this correlation in a parsimonious way. It also manages to capture the underlying monotonic trend in the data without imposing model constraints. Our application shows that the resulting functional objects are conceptually more appealing. Moreover, when used to forecast the outcome of an on-line auction, our approach also results in more accurate price predictions compared with standard approaches. We illustrate our model on a set of 183 closed auctions for Palm M515 personal digital assistants.

A Survey-Based Procedure for Measuring Uncertainty or Heterogeneous Preferences in Markets, with Application to a New Test for Common Values

The effects of heterogeneous preferences or uncertainty about item values on the variance of dependent variables (e.g., auction prices; retail price dispersion; or investment choices in stocks, R&D, or education) are usually relegated to the error term, which a) confounds these effects with other drivers of the error term and b) could lead to heteroskedasticity at best or omitted variable bias at worst. This paper shows how surveys can generate a measure of the amount of information or heterogeneity of preferences in a market by measuring the distribution of private information signals in eBay online auctions for computers. Regressions utlizing this measure produce significantly different results than those that rely on hedonic measures to control for heterogeneity alone. Furthermore, this measure provides a new way to test for the presence of a common values component in auctions.

The roles of players and reputation: Evidence from eBay online auctions

The uncertainty about product quality and the traders' credibility may impede the liquidity and efficiency of online auctions. In contrast to the literature on online feedback mechanisms that evaluate sellers' reputation without distinguishing their roles in the transactions, we claim that “not all reputation is created equal” — bidders in online auctions assess sellers' reputation based on their feedback history gained from sales solely. This paper explores the impact of the consumers' above behavior on the reputation system with an analytical model and empirical evidence from a popular online auction website eBay.

The Timing of Bid Placement and Extent of Multiple Bidding: An Empirical Investigation Using eBay Online Auctions

Online auctions are fast gaining popularity in today’s electronic commerce. Relative to offline auctions, there is a greater degree of multiple bidding and late bidding in online auctions, an empirical finding by some recent research. These two behaviors (multiple bidding and late bidding) are of “strategic” importance to online auctions and hence important to investigate. In this article we empirically measure the distribution of bid timings and the extent of multiple bidding in a large set of online auctions, using bidder experience as a mediating variable. We use data from the popular auction site www.eBay.com to investigate more than 10,000 auctions from 15 consumer product categories. We estimate the distribution of late bidding and multiple bidding, which allows us to place these product categories along a continuum of these metrics (the extent of late bidding and the extent of multiple bidding). Interestingly, the results of the analysis distinguish most of the product categories from one another with respect to these metrics, implying that product categories, after controlling for bidder experience, differ in the extent of multiple bidding and late bidding observed in them. We also find a nonmonotonic impact of bidder experience on the timing of bid placements. Experienced bidders are “more” active either toward the close of auction or toward the start of auction. The impact of experience on the extent of multiple bidding, though, is monotonic across the auction interval; more experienced bidders tend to indulge “less” in multiple bidding.

Modeling Concurrency of Events in Online Auctions via Spatio-Temporal Semiparametric Models

We introduce a semiparametric approach for modeling the effect of concurrent events on an outcome of interest. Concurrency manifests itself as temporal and spatial dependencies. By temporal dependency we mean the effect of an event in the past. Modeling this effect is challenging since events arrive at irregularly spaced time intervals and thus the standard definition of time-lags does not apply. Our concurrency model also takes spatial effects into account. We interpret the meaning of in a slightly non-traditional way in order to conceptualize the more abstract notion of space among a set of item-features. We motivate our model in the context of eBay's online auctions. In particular, we model the effect of concurrent auctions on an auction's price. Our concurrency model consists of three components: a transaction-related component that accounts for auction-design and bidding-competition, a spatial component that takes into account similarity among item-features, and a temporal component that accounts for events in the past. To construct each of these model components, we borrow ideas from spatial modeling and from the mixed model methodology. We also develop a new time-lag metric to handle unevenly-spaced time series by interpreting a time-lag as the information distributed over a certain time-window in the past and by incorporating this information via inclusion of appropriate summary measures. We illustrate the power of this model by applying it to a large and diverse set of laptop auctions crawled off 1 eBay.com. We show that our model results in superior predictive performance compared to a set of competitor models. Our model also allows for new insight into the factors that drive price in eBay's online auctions and their relationship to bidding-competition, auction-design, product-variety and temporal learning effects.