Research Article

Factors affecting price setting in online auctions Received (in revised form): 12th March 2012

Chun Qiua, Peter T.L. Popkowski Leszczycb,c and Yongfu Hed a d

McGill University, QC, Canada; bUniversity of Alberta, AB, Canada; cRenmin University, Beijing, China; and Monash University, Victoria, Canada

Chun Qiu is an Assistant Professor of marketing at the Desautels Faculty of Management, McGill University in Canada. He holds a PhD degree from the University of Alberta, and a Masters’ degree of Economics from Simon Fraser University. Dr Qiu research interests include competitive strategies, pricing and Internet auctions. Peter T.L. Popkowski Leszczyc is a Professor of Marketing at the University of Alberta and Renmin University in China, and director of the Internet auction Web site CampusAuctionMarket.com. His main research interests are related to Auction design, Bidding behavior in auctions, Charity auctions, Charitable giving. His research has appeared in Journals such as, Marketing Science, Management Science, Journal of Marketing, Journal of Retailing, Organizational Behavior and Human Decision Processes, and Product and Operation Management. Yongfu He is an Assistant Professor at Monash University. She holds a PhD degree from the University of Alberta, and a Master’s degree in Economics from the Shanghai University of Finance and Economics. Dr He’s research interests include empirical and theoretical issues related to (Internet) auctions, new product development and behavioural pricing. She has published papers in Journal of Retailing and International Journal of Market Research. Correspondence: Chun Qiu, McGill University, Desautels Faculty of Management, Montreal, QC, Canada E-mail: [email protected]

ABSTRACT This article studies the effects of product class and seller reputation on price-setting in online auctions. Sellers may offer price information to potential bidders through buy-now prices (BNPs) and starting prices (SPs). In two experiments, the authors show that for products with values that are difficult to assess, such price information affects bidders’ perception and willingness to pay. In addition, the authors model sellers’ decision making in setting the SP and BNP as a two-stage process, using data collected from eBay auctions. Results show that reputable sellers are more likely to set a BNP for their high-end products. Journal of Revenue and Pricing Management (2012) 11, 289–302. doi:10.1057/rpm.2012.11 Keywords: online auctions; buy-now price; starting price

Internet auctions have been one of the biggest success stories online. In particular, eBay.com has become a multi-billion dollar marketplace, selling products ranging from brand new standardized products such as books, CDs and electronics to non-standardized products such

as jewelry and artistic products, as well as used items. Altogether, the total value of items sold on eBay was over US$59.4 billion in 2007 (eBay, 2007). The great increase in online auctions has also had an important influence on business, as more and more businesses use

& 2012 Macmillan Publishers Ltd. 1476-6930 Journal of Revenue and Pricing Management www.palgrave-journals.com/rpm/

Vol. 11, 3, 289–302

Qiu et al

Internet auctions to sell a wide spectrum of products. Thousands of buyers and sellers trade at online auction Web sites such as eBay.com, ioffer.com, webidz.com and ebid.net. Over 724 000 American retailers use eBay.com, the biggest Internet auction Web site, as a major channel of distribution, whereas another 1.5 million individuals supplement their income by selling on eBay (eBay, 2005). Therefore, it is important for these sellers to understand what factors influence bidders’ willingness to pay (WTP) in auctions. This article focuses on two important auction design features – starting prices (SPs) (also called starting bids, minimum prices or open reserve prices) and buy-now prices (BNPs) (also called buy prices, or buy-it-now prices). More specifically, we study how sellers set SPs and BNPs, and subsequently how SPs and BNPs can effectively serve as reference prices, providing bidders with information regarding product quality and/or product value. While most previous research on BNPs and SPs has focused solely on the behavior of bidders, in contrast, this article considers the seller’s decision-making process. Both SPs and BNPs are widely used in Internet auctions. An SP is the minimum bid level that is acceptable to a seller – it is the amount at which bidding starts.1 However, bidding above the SP does not guarantee the winning of the auction, as winning depends on other bidders’ bids. Hence, SPs constitute the floor of the final selling price. In contrast, a BNP is a fixed price offer set by the seller, which, when exercised by the bidder, instantly ends an auction and awards the item to the bidder at the fixed price, thus the bidder does not have to wait until the completion of the auction. Hence, BNPs may form the ceiling of the final selling price.2 A considerable amount of research has investigated the impact of SPs and BNPs in auctions, and the effect of SPs on auction selling prices is well established in the literature (Riley and Samuelson, 1981; Engelbrecht-Wiggans, 1987; Levin and Smith, 1994). Previous research has found that, conditional on purchase, an SP has a positive effect on selling price (Engelbrecht-

290

& 2012 Macmillan Publishers Ltd. 1476-6930

Wiggans, 1987; McAfee and McMillan, 1987). However, this research assumes that valuations are exogenous, implying that bidders’ WTP is not affected by the presence or the level of an SP (Engelbrecht-Wiggans, 1987). Recent work suggests that SPs may serve as reserve prices that influence bidders’ valuations (Ariely and Simonson, 2003; Ha¨ubl and Popkowski Leszczyc, 2003; Kamins et al, 2004). The popularity of BNPs has puzzled researchers, as sellers’ reasons for placing a ceiling on the maximum selling price are unclear, and investigators have advanced many different theories to explain the use of BNPs (for an overview, see Haruvy and Popkowski Leszczyc, 2009). One explanation for the use of BNPs is that riskaverse bidders (or sellers) may want to execute a BNP to avoid the risk of losing the auction or paying too much (Hidvegi et al, 2006; Reynolds and Wooders, 2009). Other possible reasons include impatience on the part of bidders (or sellers) who may be willing to pay (or forgo) a premium to end the auction sooner (Mathews, 2004; Zeithammer and Liu, 2006), or that the BNP constitutes a premium paid when bidding is costly (Wang et al, 2008). Finally, a BNP may serve as an external reference price, particularly when bidders are uncertain about the value of the item (Popkowski Leszczyc et al, 2009). In this article, we extend the work of Popkowski Leszczyc et al (2009) by considering the reference price effect of both BNPs and SPs, and we study the moderating effects of product class (low-end versus high-end products) and seller reputation on price-setting. We propose that both SPs and BNPs may serve as reference prices (RPs) that influence bidders’ WTP, particularly for experience goods, where bidders need to try the product before being able to assess it properly. In addition, in many product categories, such as cameras or jewelry, there is a large variation in product attributes, presenting buyers with difficulty in evaluating product quality. In addition, non-standardization presents buyers with obstacles in searching for external information. Sellers, on the other hand, may

Journal of Revenue and Pricing Management

Vol. 11, 3, 289–302

Factors affecting price setting in online auctions

experience difficulties in effectively communicating unobservable product quality to buyers. Hence, by setting SPs and BNPs, sellers may provide bidders with important information about the value and quality of a product, and subsequently influence their WTP. We also suggest that other factors may moderate the effect of BNPs and SPs on bidders’ WTP, such as sellers’ reputation and the level of product attributes (for example, high-end versus low-end products). In particular, we expect that sellers with a good reputation, who are more trusted, may be able to set higher SPs and offer BNPs to influence bidders’ WTP, whereas sellers with a poor reputation cannot. In addition, we expect that SPs and BNPs will have a greater impact on high-end than on low-end products. The results of two experiments show a significant positive relationship between both SPs and BNPs and bidders’ WTP, but only for products whose values are difficult to assess. In addition, seller reputation and product class moderate the effects of SP and BNP on bidders’ WTP. Finally, data collected from eBay auctions support the behavioral implications of sellers’ price-setting. The article has the following organization. Section ‘Related literature’ offers a discussion of theory. Section ‘Moderators of the effect of BNPs and SPs on bidders’ WTP’ presents the results of two experiments, and section ‘Price setting by eBay auction sellers’ describes an eBay study that empirically examines sellers’ price-setting in auctions. Section ‘General discussion and conclusion’ concludes with a discussion of the study and provides suggestions for future research.

RELATED LITERATURE This article is closely related to the literature on external reference price effects. External RPs are stimuli such as posted regular or special prices that exist in the physical environment (Mayhew and Winer, 1992). External RPs tend to decrease consumer search (Urbany et al, 1988) and provide a price signal or reference to bidders, potentially leading to an increase in

& 2012 Macmillan Publishers Ltd. 1476-6930

sales (for example, providing a regular price as a reference for an item on sale). In an auction, the concept of RPs is more complex because no fixed selling price exists. However, sellers can provide subjective reference prices, such as SPs and BNPs, which may influence bidders’ WTP. This practice is consistent with research showing that seller-supplied external RPs (including price displays and advertised price points) exert an important influence on the formation of bidders’ valuations (Mayhew and Winer, 1992; Briesch et al, 1997; Mazumdar and Papatla, 2000), as well as on purchase decisions (Kalyanaram and Winer, 1995; Grewal et al, 1998; Kopalle and Lindsey-Mullikin, 2003; Mazumdar et al, 2005). External RPs apparently provide reference points, or anchors, that consumers use as a basis for their internal reference price (Chandrashekaran and Grewal, 2003). Previous research has suggested that SPs in Internet auctions may serve as reference prices, affecting bidders’ valuations (Ariely and Simonson, 2003; Ha¨ubl and Popkowski Leszczyc, 2003; Kamins et al, 2004; Suter and Hardesty, 2005). In addition, BNPs can serve as RPs in online auctions, particularly when bidders are uncertain about the value of a product (Popkowski Leszczyc et al, 2009). Therefore, we expect that in online auctions, BNPs and SPs can serve as reference points or prices, influencing bidders’ WTP for auctioned items. While prior researchers have found some support for the reference price effects of BNPs and SPs on bidders’ WTP in auctions, our focus is on the variables that moderate this relationship. There is considerable evidence in the literature suggesting that consumers use external RPs as informative indicators to determine future prices (Compeau and Grewal, 1998). Furthermore, they tend to assimilate information from external RPs with information from previously formed internal RPs (Mayhew and Winer, 1992; Mazumdar and Papatla, 2000; Mazumdar et al, 2005). We, therefore, expect that bidders will assimilate the information cues obtained from BNPs

Journal of Revenue and Pricing Management

Vol. 11, 3, 289–302

291

Qiu et al

and SPs, adjusting their valuation accordingly, because such higher levels of BNPs and SPs are expected to have a positive effect on valuations (as long as they are higher than bidders’ internal reference prices). However, a BNP that is set at a level that is too high may fall outside the latitude of acceptance, and consumers are less likely to update their internal RPs (Lichtenstein and Bearden, 1989). In addition, in an auction setting, Popkowski Leszczyc et al (2009) reported that a BNP that was set too high had a negative effect on bidders’ maximum bids. Although these studies provide different explanations for the use of BNPs, they fail to explain the considerable variation in their usage, by both sellers and bidders, across product categories. This raises questions regarding the various motives for using BNPs, for bidders and retailers, and their impact on selling prices in auctions. In particular, how do BNPs affect bidding outcomes across product categories, and what is the underlying mechanism for this effect? Answering these questions is the objective of this article. This study also contributes to the literature on sellers’ decisions over selling format. For example, Desai and Purohit (2004) study when sellers should offer posted price and when they should bargain. Budish and Takeyama (2001) investigate sellers’ decision on whether or not to set a posted price (that is, BNP) in auctions. They argue that sellers can set a BNP to offer as insurance to risk-averse bidders. Mathews (2004) finds that sellers can offer BNPs when bidders demonstrate heterogeneity in their patience. This article provides guidelines to sellers on when to set BNPs and SPs in online auctions, given product class and sellers’ reputation.

MODERATORS OF THE EFFECT OF BNPs AND SPs ON BIDDERS’ WTP Ease of value assessment Internet auctions differ from off-line auctions in that bidders cannot actually inspect the product,

292

& 2012 Macmillan Publishers Ltd. 1476-6930

creating value uncertainty. Although value uncertainty plays an important role in auction theory (for example, common- versus privatevalue models), these models generally assume that uncertain bidders obtain information from the bids of other bidders. Other information sources such as SPs and BNPs have not really been considered (an exception is the paper by Popkowski Leszczyc et al, 2009). We propose that bidders will use extrinsic information cues in the form of BNPs or SPs to form their valuations. This process of bidder’s valuations formation is similar to the notion that individuals construct their subjective valuations based on contextual factors, which has been well documented in the literature (for example, Ariely et al, 2003; Bettman et al, 1998). More recently, there has been a growing body of auction literature that suggests that bidders may construct their valuations based on what they observe or experience in the course of the auction (Dholakia et al, 2002; Ariely and Simonson, 2003; Ha¨ubl and Popkowski Leszczyc, 2003), and that their valuations may be influenced by different pieces of information revealed in the auction. Previous research has shown that bidders’ valuations may be influenced by number of bids (Heyman et al, 2004) and the auction format (for example, charity versus non-charity auctions; Popkowski Leszczyc and Rothkopf, 2010), as well as SPs (Ariely and Simonson, 2003; Ha¨ubl and Popkowski Leszczyc, 2003; Kamins et al, 2004) and BNPs (Popkowski Leszczyc et al, 2009). Therefore, we expect that bidders’ valuations will be influenced by different information cues provided in the auction. In this article, we focus on the influence of retailer-supplied BNPs and SPs on bidders’ valuations. To determine the value of a product, bidders must visit the auction’s Web site and read the product description, look at pictures of the product and use other information cues (that is, the BNP and the SP) provided by the seller. We expect these information cues to play a more important role, in particular, when

Journal of Revenue and Pricing Management

Vol. 11, 3, 289–302

Factors affecting price setting in online auctions

product value is more difficult to assess; for example, when supply is scarce, or for non-standardized products such as collectibles, jewelry or paintings. In such instances, strategies aimed at influencing customers’ price perceptions are more effective (Alba et al, 1994). In addition, in an auction setting, higher SPs may have a greater influence on sellers who did not provide a price guide (Brint, 2003). Finally, BNPs can serve as an external reference price, particularly when bidders are uncertain about the value of the item (Popkowski Leszczyc et al, 2009). Therefore, we expect that the impact of the BNP and SP on a bidder’s WTP will be moderated by the ease of value assessment – that is, the reference price will be stronger in cases when product value is more difficult to assess.

In addition, we expect this effect to be moderated by the ease of value assessment. Owing to differences in product characteristics, the assessment of product class differs across product categories (Weathers, 2002), particularly in Internet auctions where physical product examination is not possible before bidding. For certain products, such as the screen size of a monitor or the storage space of a memory card, differences in key product attributes are easily observable and verifiable even without a physical examination. For other products, especially less standardized products such as diamond earrings, this will not be the case, and bidders will be less certain of value. Thus, we hypothesize an interaction effect between the ease of value assessment and product class (high-end products).

Product class

Seller reputation

Product performance and attribute levels are a major source of uncertainty related to assessing product value (Heiman et al, 2002). Differences in product attribute levels affect decision making and create difficulty in constructing consumer valuations (Shugan, 1980; Weathers, 2002). As a result, product class or, more specifically, a set of low- and high-end products that are vertically differentiated in key attributes can be an important moderator of the effectiveness of BNPs. Attributes that differentiate products to a greater extent create difficulty for bidders in assessing the trade-off between increased price level and increased attribute level (Dellaert et al, 1999). Hence, attribute levels play an important role in assigning the value to an item. For high-end products, this trade-off is more consequential because the risk of making a ‘wrong’ purchase decision increases, adding to the perceived uncertainty. In addition, the possible range (lower and upper bounds) of the WTP for high-end products is considerably greater than for low-end products, making high-end products more difficult to assess. For this reason, we hypothesize that the reference price effect of a BNP will be stronger for high-end products than for low-end products.

Because bidders cannot inspect the item before bidding, trust plays a critical role in online auctions. Therefore, feedback ratings are bidders’ main source of information about seller reliability. On most auction Web sites, following a purchase, the winning bidder has an opportunity to rate the experience with the seller. The bidder can leave positive, neutral or negative feedback, which is generally used as a measure of seller reputation (Resnick and Zeckhauser, 2002; Bruce et al, 2004).3 As bidders base their trust in the seller on feedback ratings, we expect that bidders are more likely to ‘accept’ the value signals from more trustworthy sellers. Consequently, we expect that seller reputation will moderate the effect of BNPs and SPs on bidders’ WTP. In addition, more specifically, we expect that the effect of BNPs and SPs will be greater as a result of more positive seller feedback ratings. We next report the results of two experiments that study the relationship between SPs and BNPs and bidders’ WTP, as well as several variables that moderate this relationship. In Study 1, we consider the moderating role of ease of value assessment and product class (low- versus high-end products) on the effect

& 2012 Macmillan Publishers Ltd. 1476-6930

Journal of Revenue and Pricing Management

Vol. 11, 3, 289–302

293

Qiu et al

of BNPs on bidders’ WTP. In Study 2, we investigate the moderating role of seller reputation on the effect of BNPs and SPs on bidders’ WTP. Finally, using data collected from eBay auctions, we study the behavioral implications regarding sellers’ price-setting.

STUDY 1: REFERENCE PRICE EFFECTS OF BNPS AND MODERATING ROLE OF PRODUCT CLASS In Study 1, we investigated the reference price effect of BNPs, focusing on two moderating factors: the ease of value assessment and product class. To obtain a clearer separation based on this dimension, we used the results of a pre-test to select two product categories with different degrees of assessment difficulty: memory cards and diamond earrings. In addition, we manipulated product class within each product category by varying the level of the main attribute. Thus, both memory cards and diamond earrings were manipulated to obtain either a high-end or low-end product. To test our hypotheses, we conducted a two (product category: memory cards versus diamond earrings) by two (product class: high-end versus low-end) by two (BNP: presence versus absence) mixed design with product category as the within-subject factor.

Experimental procedure We created different Web pages with an auction for each product category and product class. Each auction provided a product description and pictures of the product, and a BNP was either present or not. We manipulated the product class on the most salient attribute, the storage space for the memory cards (4112 megabytes versus 256 megabytes) and the weight for the diamond earrings (0.14 carat versus 2 carats). A total of 87 undergraduate business students from a major university in North America participated in the study in exchange for course

294

& 2012 Macmillan Publishers Ltd. 1476-6930

credit. The experiment was conducted in a laboratory equipped with 15 computers. Upon arrival, participants were randomly assigned a code (for one of the eight conditions) that gave them access to a Web site with an experimental auction. Participants were told that they had to evaluate the product in the auction. After browsing the Web page for approximately 3–5 min, participants were instructed to click on a button and answer several questions related to the auction. We asked participants’ WTP for the item in dollar terms. To measure the ease of value assessment, we asked participants to rate the statement ‘It is difficult to judge the value of the product’ on a 9-point Likert scale anchored by 4 ¼ strongly disagree and 4 ¼ strongly agree. Participants then proceeded to another auction for the other product category. Product class was randomly selected, but the BNP condition remained constant across the two auctions.

Moderating effect of product class We conducted two separate three-way ANOVAs on the two dependent variables, WTP and estimate of retail value (RV). Analyses yielded a significant two-way interaction effect between product class and presence of BNPs for WTP (F1166 ¼ 7.13, Po0.01) and a significant three-way interaction effect for WTP (F1166 ¼ 4.86, Po0.05). The three-way interaction effect implied that participants’ WTP varied according to whether a BNP was present. This effect differed for diamond earrings as compared with memory cards and for high-end products versus low-end products. To further interpret these results, we conducted planned contrasts. The planned contrasts involved product category and product class (mean values of WTP of experiment conditions are summarized in Table 1). We found a significant difference for WTP for high-end diamond earrings (MBNP ¼ 1267.59 versus MNO BNP ¼ 825.59, F1166 ¼ 14.90, Po0.001), whereas BNPs did not influence bidders’ WTP for low-end diamond earrings (F1166 ¼ 1.38, P40.2).

Journal of Revenue and Pricing Management

Vol. 11, 3, 289–302

Factors affecting price setting in online auctions

Table 1: Summary statistics for Study 1 willingness to pay and estimated retail value

Experimental condition 0.14 carat diamond earrings without BNPs 0.14 carat diamond earrings with BNPs 2 carat diamond earrings without BNPs 2 carat diamond earrings with BNPs 256 MB memory card without BNPs 256 MB memory card with BNPs 4 GB memory card without BNPs 4 GB memory card with BNPs a

Willingness to Pay (WTP) US$206.59a (199.13) 52.00 (42.02) 825.59 (761.35) 1267.58 (518.00) 30.59 (17.22) 23.53 (10.86) 98.33 (80.81) 148.50 (93.33)

WTP in dollars and variance in parenthesis.

The planned contrast for memory cards did not reveal any significant effect because of BNPs for either high-end products (F1166 ¼ 0.21, P40.6,) or low-end products (F1166 ¼ 0.003, P40.9). These results indicated several interesting patterns. First, BNPs did not affect auctions of memory cards (whose value is easier to assess). Second, in none of the cases did BNPs have an influence on low-end products. We found a significant effect only for the high-end diamond earrings. Hence, as expected, we found that the effect of BNPs for high-end products is stronger when a person has more difficulty assessing the value of the product. We also expected that a BNP should facilitate the formation of valuations, making assessment of the product’s value easier by reducing people’s uncertainty. Therefore, we next examined the effect of BNPs on the dispersion of WTP and RPs.

Effect on dispersion of estimates of WTP If BNPs serve as effective reference points or prices that influence bidders’ WTP, we should also observe a reduced variance in the estimate of WTP when a BNP is present. For this reason, we conducted Levene’s F-test to assess the equality of variance in different samples (Levene, 1960). Results indicated that variances for memory cards were not statistically significant, whether or not a BNP was

& 2012 Macmillan Publishers Ltd. 1476-6930

present. In comparison, for high-end diamond earrings, the variance for both WTP and RV was significantly smaller when a BNP was present (F ¼ 4.31, Po0.05 for WTP, and F ¼ 9.06, Po0.01 for RP). For low-end diamond earrings, only the variance of WTP was significantly smaller when BNPs were present (F ¼ 13.44, Po0.001).

Discussion of Study 1 The results of Study 1 provide support for the reference price effect for BNPs. In particular, we found support for two moderating variables that influence this relationship – the ease of value assessment and product class (lowversus high-end products). More specifically, we found that BNPs had a positive effect on bidders’ WTP for the high-end diamonds, which were difficult to assess. We also found that the presence of a BNP reduced the dispersion in the WTP, providing additional support for the supposition that people use BNPs in constructing their valuations.

STUDY 2: REFERENCE PRICE EFFECTS OF BNP AND SP AND MODERATING ROLE OF SELLER REPUTATION Study 2 extends Study 1 in two important ways. First, we focused on the effects of BNPs and SPs on bidders’ WTP for high-end products whose values were difficult to assess.

Journal of Revenue and Pricing Management

Vol. 11, 3, 289–302

295

Qiu et al

Second, we investigated the moderating effect of the seller’s reputation on the joint effect of BNPs and SPs. We conducted a two (BNP: presence versus absence) by two (SP: high versus low) by two (seller reputation: good versus poor) between-factorial design. In Study 2, we used only the 2-carat diamond earrings as the stimulus, as Study 1 indicated that this is a product category for which bidders experience difficulty in assessing the value. To measure seller reputation, we used rating, which is equal to the number of positive feedback statements received by the seller. A high rating indicates that the seller received many positive feedbacks, and vice versa. We manipulated the presence of BNP by providing an average online retail price as described in Study 1. We manipulated SP by providing a higher or a lower SP. In the high-SP condition, the SP was set equal to 75 per cent of the average retail price, and in the low-SP condition, the SP was set at $1. Finally, in the high-reputation condition, rating was set to above 15 000. In the low-reputation condition, rating was less than 10. The dependent variable was bidders’ WTP. We adopted the same experimental design and procedure as in Study 1. A total of 209 undergraduate business students from a major university in North America participated in the study in exchange for partial course credit. A three-way ANOVA identified a significant two-way interaction between SP and rating (F1202 ¼ 17.19, Po0.001) and three significant main effects (BNP: F1202 ¼ 7.58, Po0.01, SP: F1202 ¼ 62.51, Po0.001; rating: F1202 ¼ 46.76, Po0.001). The planned comparison revealed a significant moderating effect of seller reputation. When the seller reputation was high, we found a significant effect of BNP and SP on bidders’ WTP. To show that effect, we further conducted a two-way ANOVA on consumer WTP in the high-reputation conditions. We found a significant main effect for BNP (F1100 ¼ 6.85, Po0.01) and for SP (F1100 ¼ 70.80, Po0.001). The planned comparison suggests that the effect

296

& 2012 Macmillan Publishers Ltd. 1476-6930

of offering BNP is significant (MBNP ¼ 1299.28 vs. MNO BNP ¼ 971.98, Po0.01). The effect of setting a high SP is also significant (MHigh SP ¼ 1660.97 vs. MLow SP ¼ 610.99, Po0.001). In comparison, in the conditions where the seller reputation is low, SP does not affect bidders’ WTP when BNP is present (MHigh SP ¼ 731.42 vs. MLow SP ¼ 196.96, P40.48).

Discussion of Study 2 The results of Study 2 support our conjecture of the moderating role of the seller’s reputation. That is, for high-end products, when a seller has a good reputation, either offering BNPs or setting high SPs, this can significantly influence bidders’ WTP. In addition, we found that when seller reputation is poor, offering BNPs does not affect bidders’ WTP. However, setting a high SP can influence bidders’ WTP. These findings offer important implications for sellers in setting prices in auctions.

STUDY 3: PRICE SETTING BY eBAY AUCTION SELLERS In Study 3, we collected data from eBay to examine whether auction sellers set BNPs and SPs in a manner consistent with what we learned in Studies 1 and 2. In particular, we expected that sellers of high-end products be more likely to offer BNPs than sellers of low-end products, as we found that BNPs only affect bidders’ WTP for high-end products. We also expected that sellers with a good reputation be more likely to offer BNPs and set high SPs. This expectation is consistent with results from Study 2, which indicated that BNPs and SPs have no effect on bidders’ WTP when the seller reputation is poor. For Study 3, we again selected diamond earrings, as bidders have difficulty assessing the value for this product category, particularly for high-end items (as indicated in Study 1 and Study 2).

Data collection We wrote a computer program to automatically collect data from eBay.com. In total, we

Journal of Revenue and Pricing Management

Vol. 11, 3, 289–302

Factors affecting price setting in online auctions

collected information from 1410 auctions. To determine the product class of diamonds, we collected information on the product features: the type of the diamond (natural versus artificial), weight, clarity and color, as well as whether an appraisal certificate was provided. We collected information on seller characteristics, which included the seller’s reputation (both the counts and percentage of positive and negative feedback), whether the seller is an eBay power seller or not, whether the seller owns an eBay store and the seller’s years of membership with eBay. Finally, we collected information on SPs and BNPs.4

We now formally specify the model for the stage-two decision, in which sellers determine SPs, as SP ¼ l 0 X þ Z 0 S þ mBNPD þ u;

ð2Þ

where X and S are the vectors, respectively, containing the variables introduced above. l and Z are the corresponding coefficient vectors. The intercept is included in l. m is the coefficient of BNPD (an indicator variable for the decision to set a BNP), and u is the error term. In Table 2, we introduce the variables. In Table 3, we provide summary statistics for these variables.

Sellers’ decisions We modeled sellers’ decision making in setting the SP and BNP as a two-stage process. In stage one, sellers decided whether to set a BNP; in stage two, they decided on the level of the SP. As the BNP determines the ceiling of the possible ending price of an auction and the SP determines its floor, the SP can never exceed the BNP. In setting SPs too high, sellers allow only a small range of bids. Hence, sellers’ decisions in setting BNPs will affect their decisions to set SPs.

Model specification We first modeled the stage one decision, in which sellers decided whether to offer BNPs, as BNPD ¼ b 0 X þ g 0 S þ e;

ð1Þ

where BNPD is the seller’s intention to offer a BNP, which is a latent variable. S is the vector containing the variables of seller characteristics and X is the vector of variables with the features of the auctioned product. b and g are the corresponding coefficient vectors. The intercept is included in b, and e is the error term. When BNPD 40, we observed that sellers offer BNPs. When BNPD p0, sellers do not offer BNPs. Hence, the observed action of setting BNPs, denoted as BNPD, is a dichotomous variable (either 0 or 1), which can be modeled via a binary choice model.

& 2012 Macmillan Publishers Ltd. 1476-6930

Estimation and results We used a standard two-stage estimation procedure, where we first estimated the model for BNPD, and then used the predicted probability from the binary choice model for the SP regression model.5 Overall, the model for the decision to set a BNP fits the data well. The hit rate is 83.2 per cent. That is, out of 1410 auctions, the model correctly predicted the decision to set a BNP in 1174 auctions (with 401 auctions not offering BNPs and 773 auctions offering BNPs). The estimates of the BNP decision model (presented in Table 4) are consistent with our expectation that sellers of high-end products are more likely to offer BNPs. We also find support that more trustworthy sellers are more likely to offer BNPs. The regression model for SPs also fits the data well. The coefficient of determinant R2 is 0.48, which is high for cross-sectional data. Table 5 summarizes the regression outcome.

Discussion of Study 3 In the BNP decision model, we find support for the effect of sellers’ reputation on the decision to set BNPs. Sellers with a superior reputation are more experienced (as they tend to have sold products on eBay for a longer time), and hence the use of BNPs may also be related to sellers’ experience. We found the effect of the duration

Journal of Revenue and Pricing Management

Vol. 11, 3, 289–302

297

Qiu et al

Table 2: Variable names and interpretations for eBay study

Variable name

Interpretation

Dependent variables: BNPD SP Product features (X): Carat Clarity Color Certificate Type Type  Carat Type  Clarity Type  Color Seller characteristics (S): eBay power seller eBay store owner Duration of eBay membership Number of negative feedback statements Percentage of negative feedback statements

Whether the seller sets a buy-now price for the product The starting price for an auctioned product The weight of the diamond in carats. 1 carat=200 mg The clarity of the diamond The color of the diamond Whether a certificate accompanies the product Whether the diamond is natural (Type=1) or lab made (Type=0). The interaction of diamond type and diamond weight The interaction of diamond type and diamond clarity The interaction of diamond type and diamond color Whether the seller is rated by eBay as a power seller Whether the seller owns a store on eBay The number of weeks since the seller became an eBay member The absolute number of negative feedback statements the sellers received The percentage of negative feedback statements the sellers received

Table 3: Summary statistics for the variables in eBay study

Variable BNPD* SP Carat Clarity** Color ** Certificate* Type* eBay power seller* eBay store owner* Total number of feedback scores Number of negative feedback statements

Mean

Standard deviation

Minimum

Maximum

0.61 1016.21 0.97 8.65a 5.69b 0.29 0.89 0.61 0.80 11 036.70 199.22

0.49 2587.38 0.86 2.44 1.91 0.45 0.31 0.49 0.40 18 564.20 557.81

0 0.01 0.1 3 1 0 0 0 0 0 0

1 40 000 10 12 12 1 1 1 1 54 410 2711

Notes: *dummy variables; **categorical variables. aaverage grade for clarity is SI2–I1; baverage grade for color is H-I.

of eBay membership to be significantly positive (b ¼ 0.007, t ¼ 3.07, Po0.01). The estimates of the regression analysis partly support our expectation that sellers with good reputations are more likely to set a high SP. We find that power sellers or eBay

298

& 2012 Macmillan Publishers Ltd. 1476-6930

store owners are more likely to set high SPs (b ¼ 411.42, t ¼ 2.87, Po0.05; b ¼ 565.37, t ¼ 3.39, Po0.01). Nevertheless, sellers with low feedback scores are also more likely to set high SPs (b ¼ 99.92, t ¼ 2.99, Po0.05). One explanation for this result is that sellers

Journal of Revenue and Pricing Management

Vol. 11, 3, 289–302

Factors affecting price setting in online auctions

Table 4: Results and interpretations of the logistic regression on BNPD

Variable

Coefficient

t-statistics

Interpretations A product made of large diamonds is more likely to be auctioned with a BNP A product with better clarity grade is more likely to be auctioned with a BNP A product with better color grade is more likely to be auctioned with a BNP A product with a certificate is less likely be auctioned with a BNP A product made of natural diamonds is more likely to be auctioned with a BNP A product made of large natural diamonds is more likely to be auctioned with a BNP A product made of natural diamonds with better clarity grade is more likely to be auctioned with a BNP A product made of natural diamonds with better color grade is more likely to be auctioned with a BNP A power seller is more likely to offer BNPs An eBay store owner is more likely to offer BNPs A long-time eBay member is more likely to offer BNPs Sellers with more negative feedback are less likely to offer BNPs Sellers with a higher percentage of negative feedback are less likely to offer BNPs

Carat

0.495

3.231***

Clarity

0.108

4.109***

Color

0.234

7.291***

0.973

8.477***

Type

1.623

4.735***

Type  Carat

0.520

4.019***

Type  Clarity

0.020

6.660***

Type  Color

0.004

1.509

eBay power seller eBay store owner

0.203 0.592

1.919** 5.453***

Duration of eBay membership

0.007

3.069***

0.002

8.515***

Certificate

Number of negative feedback statements Percentage of negative feedback statements

0.015

2.906**

Notes: *Significance at 10 per cent; **significance at 5 per cent; ***significance at 1 per cent.

Table 5: Results and interpretations of the linear regression on SP

Variable

Coefficient

t-statistics

Carat Clarity

127.804 256.267

0.785 8.350***

162.471

4.015***

Certificate Type

180.968 2329.850

1.248 6.063***

Type  Carat

2074.400

11.725***

Type  Clarity

32.727

8.501***

Type  Color

0.137

2.719**

Color

eBay power seller eBay store owner Feedback scores BNP decisions

411.423 565.373 99.916 825.66

2.871** 3.390*** 2.99** 2.819**

Interpretations A product with better clarity grade is auctioned with a higher SP A product with better color grade is auctioned with a higher SP — A product made of natural diamonds is auctioned with a higher SP A product made of large natural diamonds is auctioned with a higher SP A product made of natural diamonds with better clarity grade is auctioned with a higher SP A product made of natural diamonds with better color grade is auctioned with a higher SP An eBay power seller sets a higher SP An eBay store owner sets a higher SP A low-feedback seller sets a higher SP A seller not offering the BNP sets a higher SP

Notes: *Significance at 10 per cent; **significance at 5 per cent; ***significance at 1 per cent.

& 2012 Macmillan Publishers Ltd. 1476-6930

Journal of Revenue and Pricing Management

Vol. 11, 3, 289–302

299

Qiu et al

with low feedback scores do not set BNPs (from the BNP choice model). Thus, to influence bidders’ WTP, those sellers tend to set high SPs. This result is consistent with the behavioral finding in Study 2 that, when the seller’s reputation is poor, bidders’ WTP is not influenced by BNPs but rather by high SPs. The regression results support this alternative explanation, as we also detected the negative effect of offering BNPs on SPs (b ¼ 825.66, t ¼ 2.82, Po0.05). This finding suggests that a seller’s failure to offer a BNP will significantly increase the SP.

GENERAL DISCUSSION AND CONCLUSION This investigation studied the impact of sellerspecified RPs on bidders’ WTP in auctions, focusing on two different types of reference prices: the SP and the BNP. The SP is a price floor, which is an indicator of the lowest price a seller is willing to accept for an item, whereas a BNP forms a ceiling for the selling price and is the amount a seller is willing to accept to forgo the auction. Both SP and BNP may provide important information to bidders regarding how much a seller values the product for sale. The results of two experiments provided substantial support for our hypothesis that retailers can use BNPs and SPs as reference prices, thereby increasing bidders’ WTP. Moreover, the results showed that several variables moderate the relationship between the magnitude of the BNP and/or SP and bidders’ WTP. Study 1 showed the existence of a reference price effect of BNP, but only for high-end products and when product value is difficult to assess. Results of Study 2 provided further support for the reference price effect of BNP and SP. Specifically, we found that seller reputation positively moderates the reference price effect. Finally, Study 3 showed that price setting by eBay sellers was consistent with the results in Studies 1 and 2 – that is, sellers are more likely to set BNP when they are more reputable and when the product is a high-end one.

300

& 2012 Macmillan Publishers Ltd. 1476-6930

Most previous research on BNPs has focused solely on the behavior of bidders. In contrast, this article considers the seller’s decisionmaking process. We propose that BNPs function as external reference prices, and that retailers may benefit from setting BNPs (as well as SPs) in auctions, particularly when bidders are uncertain about the value of a product. These bidders may assimilate the BNP and/or SP levels as useful information in forming their valuations. Our findings are consistent with a growing body of literature that suggests that people participating in Internet auctions construct the valuation of an auctioned product during the bidding process (Dholakia et al, 2002; Ariely and Simonson, 2003; Ha¨ubl and Popkowski Leszczyc, 2003), and that their valuations may be influenced by different pieces of information revealed in the auction. This information may include the BNP, as demonstrated in this article, as well as bids submitted by other bidders (Milgrom and Weber, 1982), the number of bids (Dholakia et al, 2002) and starting bids (Ha¨ubl and Popkowski Leszczyc, 2003; Kamins et al, 2004; Suter and Hardesty, 2005). More research is needed to study the optimum level of a BNP and to what extent a BNP that is too high may have a negative impact. A negative impact may also be related to bidders’ perceptions of price fairness, as setting a high reference point in the form of a starting bid may have a negative effect on long-term profits (Suter and Hardesty, 2005). Some research findings indicate that even exaggerated RPs may have a positive influence on bidders’ perception of value (Urbany et al, 1988), whereas others suggest that RPs at moderate levels have a stronger impact (Kopalle and Lindsey-Mullikin, 2003). However, most research results indicate that when RPs are too high (that is, the difference with a bidder’s internal reference price is large), bidders are more likely to discredit the reference price information (Gupta and Cooper, 1992; Bobinski et al, 1996). This finding suggests that the relationship between BNPs and/or SPs and auction outcomes may be non-linear.

Journal of Revenue and Pricing Management

Vol. 11, 3, 289–302

Factors affecting price setting in online auctions

Future research should also study the impact of different price signals, such as manager suggested retail prices. Another interesting question is: What is the moderating effect of seller reputation on BNPs and SPs? Further research is also needed in studying the level of SPs (for example, the trade-off between a high versus a low SP). Although a high SP may have a direct positive effect on auction outcome, it also has a negative indirect effect through the number of bidders. Future research could also consider secret versus open reserve prices. In the case of secret reserve prices, bidders only observe whether or not a reserve is met. Finally, future research is needed to study the impact of competing auctions (Haruvy and Popkowski Leszczyc, 2010). What is the influence of competing auctions with different BNP and SP strategies? Additional research is also required to investigate the psychological process that underlies the effect of BNPs and SPs on auction outcomes. This process may be a relatively low level, and possibly a non-conscious, mental process consistent with either priming or anchoring and adjustment, or it may be a deliberate inference of the value of the auctioned product based on the retailer-specified BNP. The results of our research suggest that the process is more likely deliberate, as we have not observed reference price effects for low-end products. In contrast to normative economics theory, we find that BNPs and SPs can have a direct impact on bidders’ WTP. These findings have important implications for seller strategy in online auctions and for consumer welfare.

ACKNOWLEDGEMENT This research was supported by the grants from the Social Sciences and Humanities Research Council of Canada 410-2010-1124 and 410-2011-0058, and the University of Alberta GRA Rice Faculty Fellowship. The first author wishes to thank Lv Xianhua for support.

NOTES 1 An auction can also have a secret reserve unknown to bidders until a bid that exceeds

& 2012 Macmillan Publishers Ltd. 1476-6930

2

3

4

5

it has been submitted. However, as secret reserves are unobservable, they are difficult to use as value indicators, and hence are not considered in this study. It is possible for a bidder to bid above the BNP, particularly in eBay auctions where BNPs are temporary and the BNP disappears once a bid (above the secret reserve) has been made. A considerable number of investigations have studied the effectiveness of feedback mechanisms. For an overview, see Haruvy and Popkowski Leszczyc (2009). For an auction, BNP is either present or not, whereas SP is always present: ranging from $0.01 and up. This procedure will account for endogeneity bias, as BNDP is an endogeneous variable included in the SP equation.

REFERENCES Alba, J.W., Broniarczyk, S.M., Shimp, T.A. and Urbany, J.E. (1994) The influence of prior beliefs, frequency cues, and magnitude cues on consumers’ perceptions of comparative price data. The Journal of Consumer Research 21(2): 219–235. Ariely, D., Loewenstein, G. and Prelec, D. (2003) Coherent arbitrariness: Stable demand curves without stable preferences. Quarterly Journal of Economics 118(1): 73–105. Ariely, D. and Simonson, I. (2003) Buying, bidding, playing or competing? Value assessment and decision dynamics in internet auctions. Journal of Consumer Psychology 13(2): 113–123. Bettman, J.R., Luce, M.F. and Payne, J.W. (1998) Constructive consumer choice processes. Journal of Consumer Research 25(3): 187–217. Bobinski, G.S., Cox, D. and Cox, A.D. (1996) Retail ‘Sale’ advertising, store credibility and price rationale. Journal of Retailing 72(3): 291–306. Briesch, R.A., Krishnamurthi, L., Mazumdar, T. and Raj, S.P. (1997) A comparative analysis of reference price models. Journal of Consumer Research 24(2): 202–214. Brint, A.T. (2003) Investigating buyer and seller strategies in internet auctions. Journal of the Operational Research Society 54(11): 1177–1197. Bruce, N., Haruvy, E. and Rao, R. (2004) Seller rating, price, and default in online auctions. Journal of Interactive Marketing 18(4): 37–50. Budish, E. and Takeyama, L. (2001) Buy Prices in Internet auctions: Irrationality on the Internet? Economics Letters 72(3): 325–333. Chandrashekaran, R. and Grewal, D. (2003) Assimilation of advertised reference prices: The moderating role of involvement. Journal of Retailing 79(1): 53–62.

Journal of Revenue and Pricing Management

Vol. 11, 3, 289–302

301

Qiu et al

Compeau, L.D. and Grewal, D. (1998) Comparative advertising: An integrative review. Journal of Public Policy and Marketing 17(2): 257–273. Dellaert, B., Brazell, J. and Louviere, J. (1999) The effect of attribute variation on consumer choice consistency. Marketing Letters 10(2): 139–147. Desai, P. and Purohit, D. (2004) ‘Let me talk to my Manager’: Haggling in a competitive environment. Marketing Science 23(2): 219–233. Dholakia, U.M., Basuroy, S. and Soltysinski, K. (2002) Auction or agent (or both)? A study of moderators of the herding bias in digital auctions. International Journal of Research in Marketing 19(2): 115–130. eBay. (2005) New study reveals 724 000 Americans rely on eBay sales for income. The eBay Press Release, 21 July, http:// investor.ebay.com/releasedetail.cfm?releaseid=170073, accessed 1 August 2008. eBay. (2007) 2007 Annual Report, http://investor.ebay.com, accessed 1 August 2008. Engelbrecht-Wiggans, R. (1987) Optimal reservation prices in auctions. Management Science 33(6): 763–770. Grewal, D., Monroe, K.B. and Krishnan, R. (1998) The effects of price-comparison advertising on buyers’ perceptions of acquisition value, transaction value, and behavioral intentions. Journal of Marketing 62(2): 46–59. Gupta, S. and Cooper, L.G. (1992) The discounting of discounts and promotions thresholds. Journal of Consumer Research 19(3): 401–411. Haruvy, E. and Popkowski Leszczyc, P.T.L. (2010) Search and choice in online consumer auctions. Marketing Science 29(6): 1152–1164. Haruvy, E.E. and Popkowski Leszczyc, P.T.L. (2009) Internet auctions. Foundations and Trendss in Marketing 4(1): 1–75. Ha¨ubl, G. and Popkowski Leszczyc, P.T.L. (2003) Minimum prices on product valuations in auctions. Marketing Science Institute Reports 3(03–117): 115–141. Heiman, A., McWilliams, B., Zhao, J. and Zilberman, D. (2002) Valuation and management of money-back guarantee options. Journal of Retailing 78(3): 193–205. Heyman, J., Mellers, B., Tishcenko, S. and Schwartz, A. (2004) I was pleased a moment ago: How pleasure varies with background and foreground reference points. Motivation and Emotion 28(1): 65–83. Hidvegi, Z., Wang, W. and Whinston, A.B. (2006) Buy-price English auction. Journal of Economic Theory 129(1): 31–56. Kalyanaram, G. and Winer, R.S. (1995) Empirical generalizations from reference price research. Marketing Science 14(3): 161–169. Kamins, M.A., Dre`ze, X. and Folkes, V.S. (2004) A field study of the effects of minimum and reserve prices on internet auction. Journal of Consumer Research 30(4): 622–628. Kopalle, P. and Lindsey-Mullikin, J. (2003) The impact of external reference price on consumer price expectations. Journal of Retailing 79(4): 225–236. Levene, H. (1960) Robust tests for equality of variances. In: Ingram Olkin, S.G. Ghurye, W. Hoeffding, W.G. Madow, and H.B. Mann (eds.) Contributions to Probability and Statistics:

302

& 2012 Macmillan Publishers Ltd. 1476-6930

Essays in Honor of Harold Hotelling. California, MA: Stanford University Press, pp. 278–292. Levin, D. and Smith, J.L. (1994) Equilibrium in auctions with entry. American Economic Review 84(3): 585–599. Lichtenstein, D.R. and Bearden, W.O. (1989) Contextual influences on perceptions of merchant-supplied reference prices. Journal of Consumer Research 16(1): 55–66. Mathews, T. (2004) The impact of discounting on an auction with a buyout option: A theoretical analysis motivated by eBay’s buy-it-now feature. Journal of Economics. 81(1): 25–52. Mayhew, G.E. and Winer, R.S. (1992) An empirical analysis of internal and external reference prices using scanner data. Journal of Marketing Research 19(1): 62–70. Mazumdar, T. and Papatla, P. (2000) An investigation of reference price segments. Journal of Marketing Research 37(2): 246–258. Mazumdar, T., Raj, S.P. and Sinha, I. (2005) Reference price research: Review and propositions. Journal of Marketing 69(4): 84–102. McAfee, R. and McMillan, J. (1987) Auctions and bidding. Journal of Economic Literature 25(2): 699–738. Milgrom, P. and Weber, R. (1982) A theory of auctions and competitive bidding. Econometrica 50(6): 1089–1122. Popkowski Leszczyc, P.T.L., Qiu, C. and He, Y. (2009) Empirical testing of the reference-price effect of buy-now prices in internet auctions. Journal of Retailing 85(2): 211–221. Popkowski Leszczyc, P.T.L. and Rothkopf, M. (2010) Charitable motives and bidding in charity auctions. Management Science 56(3): 399–413. Resnick, P. and Zeckhauser, R. (2002) Trust among strangers in internet transactions: Empirical analysis of eBay’s reputation system. Advances in Applied Microeconomics 11: 127–157. Reynolds, S. and Wooders, J. (2009) Auctions with a buy price. Economic Theory 38(1): 9–39. Riley, J.G. and Samuelson, W.F. (1981) Optimal auctions. The American Economic Review 71(3): 381–392. Shugan, S.M. (1980) The cost of thinking. Journal of Consumer Research 7(2): 99–111. Suter, T.A. and Hardesty, D.M. (2005) Maximizing earnings and price fairness perceptions in online consumer-to-consumer auctions. Journal of Retailing 81(4): 307–317. Urbany, J.E., Bearden, W.O. and Weilbaker, D.C. (1988) The effect of plausible and exaggerated reference prices on consumer perceptions and price search. Journal of Consumer Research 15(2): 95–110. Wang, X., Montgomery, A.L. and Srinivasan, K. (2008) When auction meets fixed price: A theoretical and empirical examination of buy-it-now auctions. Quantitative Marketing and Economics 6(4): 339–370. Weathers III, P.D. (2002) Purchase channel and product characteristic effects on consumer risk perceptions. PhD dissertation, University of South Carolina, AAT 3059481. Zeithammer, R. and Liu, P. (2006) When is Auctioning Preferred to Posting a Fixed Selling Price? University of Chicago. Working paper.

Journal of Revenue and Pricing Management

Vol. 11, 3, 289–302