Developing a competitive edge in electronic markets via institutional and social based quality signaling mechanisms

Information & Management 51 (2014) 532–540

Contents lists available at ScienceDirect

Information & Management journal homepage: www.elsevier.com/locate/im

Developing a competitive edge in electronic markets via institutional and social based quality signaling mechanisms Carol X.J. Ou a,*, Keith C.C. Chan b a b

Department of Management, Tilburg University, The Netherlands Department of Computing, Hong Kong Polytechnic University, Hong Kong

A R T I C L E I N F O

A B S T R A C T

Article history: Received 24 August 2012 Received in revised form 28 March 2014 Accepted 5 April 2014 Available online 18 April 2014

Much recent effort has been put into developing effective electronic markets. However, the research has mainly focused on institutional trust-building mechanisms. Practically, sellers lack guidelines in shaping competitive edges in electronic markets where institutional mechanisms have been applied to all sellers. In order to fill this gap, we examine the impacts of institutional and social mechanisms on seller differentiation, drawing from quality signaling theories in economics. Hierarchical regression analysis of the objective data crawled from TaoBao.com reveals these two competing categories of quality signal mechanisms result in interesting seller differentiation. Findings, implications and future research are discussed. ß 2014 Elsevier B.V. All rights reserved.

Keywords: Quality signaling theories Electronic markets Institutional mechanisms Social mechanisms Sales volume

1. Introduction The technological advancement in information systems has led to the rapid maturing of electronic markets in the last decade. The electronic market pioneer, eBay, reported approximately 116 million active users globally [7]. TaoBao, the leading electronic market in China, reported more than 370 million registered users as of the end of 2010 and is currently hosting more than 800 million product listings [33]. Among those registered users, there are more than 1.5 million store-owners [5]. To facilitate the development of effective electronic markets, much investigation has been made to help build buyer trust. The huge population of buyers and sellers in the world-wide electronic markets nowadays is a testimony to the effectiveness of the mechanisms developed to ensure the trustworthiness between buyers and sellers. As these markets have gradually matured, the latest challenge for the participants of many electronic markets, especially for the sellers, is to stand out from the crowd. This is considered much more difficult than in traditional offline markets because such information as price and customer feedback is essentially transparent to both online buyers and sellers in electronic markets. Unfortunately, compared to the abundant

* Corresponding author. E-mail addresses: [email protected] (Carol X.J. Ou), [email protected] (Keith C.C. Chan). http://dx.doi.org/10.1016/j.im.2014.04.002 0378-7206/ß 2014 Elsevier B.V. All rights reserved.

literature in building buyers’ general trust in electronic markets (e.g., [13,19,26]), relatively little work has been done to explain how online sellers can stand out in the intensive competition in electronic markets. Investigation into how sellers can differentiate themselves and build competitive edges remains under-explored. In order to fill in this research gap, we investigate the problem of seller differentiation based on quality signaling theories [1,31,16], combined with institutional based [26] and social based [22] mechanisms. Specifically, quality signaling theories [1,31] suggest that sellers’ and products’ quality information can influence buyers’ decisions. Furthermore, Li et al. [16] identified a typology of Internet auction features as quality signals and investigated how these signals help alleviate uncertainty in eBay. Since the emergence of electronic markets, many transaction platforms have devised innovative features in order to enable sellers to provide more quality-related information so as to attract more buyers. Given that the availability of many such quality signaling features are at the discretion of the sellers, their effectiveness in materializing transactions warrants an examination. Following these quality signaling theories related to seller and product characteristics (e.g., [16]), in this study we classify the quality signals into institutional and social based quality signaling mechanisms and examine their effectiveness in differentiating sellers in electronic markets. We argue that the characteristics of sellers and products affect buyers’ purchasing behavior such as where to buy in the electronic markets because seller

C.X.J. Ou, K.C.C. Chan / Information & Management 51 (2014) 532–540

differentiation occurs when sellers and products have better quality signals. Specifically, we take individual sellers’ sales volume of one particular product as the proxy of seller differentiation. We argue that seller differentiation can be measured by both the process (such as product/shop popularity) and the end results (such as sales volume). In this study, we focus on the end results of seller differentiation, i.e., achieving high sales volume. This focus is consistent with quality signaling theories in Internet auctions (e.g., [16]) that emphasize the impacts of quality indicators on consumers’ decisions regarding where to bid (i.e., which sellers to purchase from). Institutional mechanisms, such as reputation system, escrow services, credit card guarantees, and intermediary protection, have been proven to be effective in building consumer trust [26]. These mechanisms are actually applicable to all sellers in an electronic market. While institutional mechanisms function in an electronic market as a whole and are unlikely to be effective in differentiating sellers, it should be noted that some emerging voluntary institutional mechanisms have been well utilized in practice. For example, in an attempt to develop some unique mechanisms, TaoBao permits sellers to voluntarily subscribe to a consumer protection scheme, a 7-day-no-condition product return policy and a 30-day free repair scheme. Considering that sellers’ subscription to these schemes is on a voluntary basis and that it induces more costs for ‘‘bad sellers’’ with low quality products than for ‘‘good’’ sellers with high-quality products, such emerging institutional mechanisms appear to be able to render online sellers a differentiating position in the focal electronic market. Social mechanisms in this study refer to the influence of trends [28] and virtual presence [22] on buyer behavior. Although social mechanisms have received little attention in the literature on electronic markets [3,21], we argue that they are important in persuading buyers to choose one particular seller over others. For example, TaoBao can display the number of people tagging a specific product web page as a favorite item, or the number of people tagging a specific seller as a favorite seller. In addition, sellers can also utilize a platform-embedded instant messaging (IM) tool, namely WangWang, to communicate with buyers. We argue that these factors can influence buying decisions because they require sellers to devote time and efforts to obtain high numbers of product and seller taggings or be online for instant communication with buyers. Therefore, the impact of such social mechanisms on seller differentiation is worthy of investigation from both theoretical and practical aspects. Based on the above observations, we propose to integrate two perspectives, viz., institutional mechanisms and social based mechanisms, to examine their impacts on seller differentiation in electronic markets. Following this introduction, we detail the theoretical development based on the quality signaling theories [1,16,31] and justify the hypotheses in the next section. We then explain how we collect objective data using a web crawler, tailored to TaoBao’s website structure, in this study. In the data analysis section, we present the results of statistical analysis based on 16,994 distinct records of product web pages crawled from TaoBao. We then discuss the key findings and conclude this paper with implications and future research. 2. Theoretical development 2.1. Building a theoretical framework based on the quality signaling theories Electronic markets are characterized by the separation of buyers and sellers and thus high uncertainty. In order to help the buyers identify ‘‘good’’ from ‘‘bad’’ products and sellers, electronic markets have developed quality signaling mechanisms to promote

533

online transactions, while researchers have contributed to the empirical tests of these mechanisms. Notably, the auction and quality signaling theories in economics [1,31] provided a typology of auction quality and credibility indicators, which has been adapted in the research on Internet auctions (e.g., [16,24]). Specifically, Spence [31], as well as Kirmani and Rao [12], noted that two conditions need to be met before a mechanism can serve as a signaling device. First, the mechanism needs to induce costs for the seller to adopt. Second, the signaling costs must fulfill the single-crossing property that ‘‘such costs are higher for bad sellers than for good sellers so that a separating equilibrium occurs’’ ([16], p. 78). Furthermore, Li et al. [16] classified the Internet auction features into the types of product quality and seller credibility and used a hierarchical Bayesian framework to examine consumers’ bidding behavior in eBay. In parallel with the quality signaling theories found in the economics discipline, IS researchers also suggest the effectiveness of institutional based mechanisms, such as reputation system, escrow services, credit card guarantees and intermediary protection, in building consumer trust (e.g., [26]). We argue that although those institutional mechanisms are effective to signal the quality of seller community as a whole, differentiation from ‘‘good’’ from ‘‘bad’’ sellers demands more than the community-wide institutional mechanisms. On this aspect, we suggest that social based quality signaling mechanisms plays an important role in differentiating sellers in electronic markets. Therefore, this study made use of quality signaling theories to classify the quality signaling mechanisms in electronic markets into those related to seller characteristics and product characteristics. In addition, the perspective of institutional and social based quality signaling mechanisms for trust building provides sellers with different quality indicators to opt in or disclose more quality information. In the combination of quality signaling theories and trust-building mechanisms, we propose a typology of seller credibility and product quality indicators from institutional and social perspectives (as shown in Table 1). In this study, we attempt to examine their effectiveness in influencing the seller differentiation in terms of sales volume. We establish and justify the theoretical hypotheses related to this typology in the following section. 2.2. Institutional mechanisms In order to attract buyers to engage in transactions with ‘‘unknown’’ online vendors in electronic marketplaces, much effort has been put into the study of how effective electronic markets can be designed. In particular, trust and trust building mechanisms have received great attention (e.g., [8,19]). Typically, institutional structures, such as the reputation systems, escrow services, credit card guarantees, and intermediary protections, are considered to be effective trust-building mechanisms (e.g., [25,26]). These institutional mechanisms have been commonly adopted in electronic markets such as eBay and TaoBao.

Table 1 A typology of institutional and social based quality signaling mechanisms. Quality signaling mechanisms Institutional-based

Social-based

Seller characteristics

 Reputation score  Rating scores  Consumer protection scheme

 Virtual presence  Shop tagging

Product characteristics

 Seven-day return policy  Product tagging  Repairing service

534

C.X.J. Ou, K.C.C. Chan / Information & Management 51 (2014) 532–540

Many electronic markets provide a reputation system and a rating system for buyers to evaluate online sellers. In electronic markets such as TaoBao, a buyer can rate each transaction using a scoring system made up of +1 (positive), 0 (natural), and 1 (negative). A reputation score for a seller is then automatically computed by the electronic market by taking the sum of all buyers’ individual ratings for the focal seller. Sellers need to devote time and effort to obtain individual buyer’s positive evaluation in each transaction in order to accumulate a good reputation score in the long term. The signal cost of the reputation score is the future revenue at stake [16]. Furthermore, ‘‘bad’’ sellers find that it is more difficult to obtain a positive reputation than the good sellers in the long run and lower reputation is expected to have negative impacts on sellers’ sales volume and profit [2]. For these reasons, a seller’s reputation score is qualified for the single-crossing property and is hence regarded as a quality signaling mechanism. In addition to the reputation score, a buyer can also rate each transaction from 1 to 5 marks in terms of product descriptions, customer service, and delivery speed. Electronic markets, such as TaoBao, usually display some rating scores on the means of all buyers’ ratings specifically on these three dimensions. These scores can range from 1 to 5 in the case of TaoBao. Such reputation and rating scores reflect how successful a transaction is as evaluated by the buyers based on their transaction experience. Although the reputation scores and the rating scores are both derived from past buyers’ evaluations, they are different. The reputation scores are concerned with the accumulation of individual buyer’s summary evaluation of each transaction (made up of +1, 0, and 1). The rating scores, on the other hand, represent the mean score of buyers’ rating on specific product descriptions, customer service, and delivery speed (ranging from 1 to 5). Similar to the reputation scores, online sellers need to demonstrate their ability to produce credible product descriptions, provide good customer service and deliver products on time, in order to earn buyers’ positive rating scores. This means additional costs for a seller and it also means that seller’s future revenue may be affected if it is not managed well. For ‘‘bad’’ sellers, since obtaining the same positive rating scores for them is more difficult, therefore, the signaling costs are greater. In this regard, the rating score fulfills the requirement of single-crossing property and can also be regarded as a quality signaling device. In a summary, a buyer who has never transacted with a seller may actually trust the focal seller based on the trust ‘‘transferred’’ from other buyers [26]. Hence, buyers will be more likely to transact with those sellers with a higher reputation and rating scores. The institutional mechanisms described above target improvements to the effectiveness of electronic markets as a whole and therefore are imposed on all sellers. For some electronic markets such as TaoBao, additional institutional mechanisms are offered for online sellers to decide whether they would like to adopt. For example, a consumer protection scheme is a contract signed between a seller and TaoBao to protect the buyers’ rights as consumers. According to the contractual requirements, the seller needs to deposit a considerable sum of money with TaoBao so that a buyer can be compensated from the deposit under some circumstances. This option therefore incurs upfront signal costs. In order to protect the buyers’ right under the consumer protection scheme, TaoBao can intervene within 48 hours and proceed in 7 days when receiving a valid complaint. If a complaint is successful lodged, the seller, under consumer protection scheme, has to compensate the buyer and therefore this option can also result in future costs. For ‘‘bad’’ sellers, the probability of them being involved in cases of disputes is higher than that of the ‘‘good’’ sellers and hence, they may be less willing to participate in the consumer protection scheme. From this perspective, the consumer

protection scheme satisfies the single-crossing property and can be regarded as a quality signal device. In addition to a consumer protection scheme, TaoBao also offers a seven-day return policy for the sellers to voluntarily subscribe. When shopping from the sellers with a logo of ‘‘seven-day return’’, buyers have the right to return the products back to the sellers within seven days after the purchase without any other conditions. In addition, TaoBao also provides a repairing service for the sellers to voluntarily subscribe. When the purchase is made in the seller’s website with a logo of ‘‘repairing service’’, buyers can enjoy free repair services provided by the focal seller within thirty days after the purchase. For a seller to participate in these three voluntary institutional mechanisms, the seller needs to be a subscriber to such institutional mechanisms. After the subscription, a certain approved logo is displayed on the seller’s online shop as well as on each product webpage. That means online sellers need to pay for certain fees in order to be approved to display such logos and therefore induce upfront signaling costs. Meanwhile, as the return policy and repairing service incur future costs (as the sellers need to pay the price of handling disputes or for product maintenance). For those sellers with low-quality products with higher return or repair ratings, it will be more expensive for them to adopt these two mechanisms (c.f., [16]). With the institutional mechanisms described above, the electronic market concerned can intervene if a seller fails to comply with the requirements of these institutional mechanisms. The seller can face serious penalties from the electronic market if they behave improperly. Hence, these mechanisms will lead to the requirements for deposits or future costs if a seller is involved in any disputes, whichsatisfies the single-cross property of quality signals according to the quality signal theories (c.f., [12,31]). In general, buyers prefer those sellers who have subscribed to institutional mechanisms [16,27]. This is because of the economic calculative reasoning about the incentives and penalties involved in these institutional mechanisms [29,34]. So in an integrative view, the institutional mechanisms described above can increase buyer confidence and have the potential to increase the transaction volume. We thus propose that: H1. Institutional mechanisms, covering a seller’s reputation score, rating scores, customer protection scheme, seven-day return policy, and repairing service, contribute positively to sales volume in electronic markets. 2.3. Social mechanisms Compared to institutional mechanisms, researchers have not paid much attention to the social mechanisms, i.e., the influence of trends and virtual presence, in electronic markets [3,21]. However, it is noted that some electronic markets have already been incorporating many different social mechanisms to facilitate online transactions. For example, TaoBao provides several such methods, utilizing sellers’ virtual presence via IM, as well as product and seller popularity via buyers’ tagging. We explain how these social mechanisms work and qualify as quality signals in an electronic market below. Virtual presence in this study refers to the extent to which a seller is present in a virtual environment, i.e., an electronic market in this research context. Some electronic markets such as TaoBao incorporate IM technology in their platform design. Buyers can select the seller they wish to talk to via the platform embedded IM– WangWang–and can maintain control over the initiation of conversations. Sellers are not required to be online to communicate with the buyers in electronic markets. As sellers need to put in time and effort or pay customer service personnel to be online to communicate with buyers, this option incurs costs. However, in

C.X.J. Ou, K.C.C. Chan / Information & Management 51 (2014) 532–540

some cases, TaoBao shop owners employ multiple service personnel in order to provide the customers with around-theclock service, including communicating directly with the shop owner and talking with individual online representatives in handling pre-sales enquiry, checking order status, logistics arrangements, and after sales service. The purpose of providing such a high level of virtual presence is to mimic the face-to-face transactions by bridging a buyer’s perceived psychological distance (social presence) and perceived physical proximity (telepresence) to an online seller [23]. According to social presence theory [30], a medium’s social effects are principally caused by the degree of social presence which it affords to its users. The warmer, the more personal, sociable, and sensitive the social interaction is, the stronger the feeling of social presence perceived by the interlocutors (i.e., the interaction participants). We argue that since an IM tool signifies the communication availability of online sellers and enables two-way synchronous communication, it facilitates the transactions in an electronic market. For sellers with low-quality products, being online may incur more costs because of the higher possibility of requiring to provide justification before successful transactions, or to come up with convincing arguments or be involved in disputes in the future than the sellers with high-quality products. Hence, virtual presence provides another signalling device and an effective means to the end (i.e., achieving high sales volume by providing good online service). The popularity level in this study refers to how popular a specific product web page or a shop is in an electronic market. In order to demonstrate the popularity level of a specific product web page or shop in an electronic market, platforms such as TaoBao offer two social mechanisms, namely shop tagging and product tagging. The former refers to the number of web surfers who are in favor of, and thus tag, the shop in their bookmark collections. The latter refers to the number of web surfers who are in favor of, and thus tag, one specific product webpage in their bookmark collections. We argue that a large number of taggings can induce positive reactions from potential buyers,1 thereby bringing about an effect similar to the extent which a large number of visitors arouse online users’ purchase intention [15]. According to marketing theories, word-ofmouth is a powerful force in the marketplace and it can change a buyer’s behavior [9,14]. In electronic markets, shop taggings and product taggings are both important indicators of the popularity as well as quality level of individual sellers and individual product web pages. A seller must devote time and effort to design their web store and also their products in order to attract buyers’ attentions and taggings. The signaling costs are related to future sales at stake because higher shop and product taggings inform the reputation of the seller and also the focal products. It is more difficult for sellers with low-quality products to attain the same popularity level and therefore their signaling costs are higher. A buyer’s trust in a seller can be built based on trust ‘‘transferred’’ from other buyers [26]. In electronic markets, a large number of taggings suggests that many other buyers are in favor of one specific product webpage or one particular shop. Such a positive sign can have an influence on a potential buyer’s purchase decisions, following the trust transfer theory [26] and the quality signaling theories [1,16]. We thus propose: H2. Social mechanisms, covering virtual presence, shop tagging, and product tagging, contribute positively to sales volume in electronic markets.

1 Considering that some potential buyers, who have tagged a focal shop or product, may not finally become actual buyers, we therefore took shop tagging and product tagging as independent variables, rather than dependent variables, in order to study their impacts on the consumers’ actual buying decision.

535

2.4. The moderating effects of product type Other than the above institutional and social mechanisms, we have also considered product type which refers to different categories of products in terms of prices. Due to the fact that buyers generally perceive higher uncertainty and risk for more expensive products sold online (c.f., [18,26]), we argue that product type can exert a moderating effect in the influencing powers of institutional mechanisms and social mechanisms on sales volume. When a product is more expensive, buyers will put more weight on institutional mechanisms in their decision to buy a more expensive product, considering that those mechanisms can protect the buyers from heavy losses when disputes arise. Likewise, for an expensive product, buyers may be inclined to communicate with a seller so as to better judge the product and the seller via the direct conversations on IM. Similarly, larger numbers of product and shop taggings suggest more web surfers are in favor of one specific product webpage or one specific shop. The collective social influence indicated via a large tagging number appears to be able to dramatically reduce the perception of uncertainty in buying expensive products. In this case, the quality signals can be more heavily weighted for those expensive products because these signals can serve effective mechanisms to reduce uncertainty in the costly payment. With such arguments, we thus propose: H3. The weightings of institutional mechanisms (H3a) and social mechanisms (H3b) are higher for more expensive products when predicting sales volume in electronic markets, suggesting the moderating role of product type. We summarize the above hypotheses in Fig. 1. 3. Methodology 3.1. Measures We rely on the objective data (see Appendix A) available at TaoBao to measure the conceptual constructs. With respect to institutional mechanisms, we first include the individual seller’s reputation scores [i.e., the overall score accumulated based on all past evaluations by buyers who assigns a value of +1 (positive), 0 (neutral) or 1 (negative) after each transaction]. The rating scores cover product descriptions, services, and delivery, ranging from 1 to 5 for each item. The other three institutional mechanisms include consumer protection scheme, the seven-day return policy and repairing service. For these three factors, we coded them as binary variables (either 0 or 1), according to the availability of such mechanisms offered by individual sellers on their websites. Social mechanisms include virtual presence and popularity level. We measured an individual seller’s virtual presence by the number of customer service personnel available online on WangWang. For some shops, there is only one single WangWang (coded as 1). While for some other shops, multiple online sales representatives are used to respond to buyers’ questions in prepurchase enquires, order status checking and product return issues. We thus use the actual number of online sales representatives to measure the virtual presence of the focal seller. Considering that online sales representatives’ online status may be different at different time of a day, we collected sales representatives’ online status in three time periods during a day, viz., the morning (06:00–12:00), afternoon (12:00–18:00), and night (18:00–06:00), for data analysis at a later stage. With respect to popularity level, the number of shop taggings and the number of product taggings indicate how many distinct web surfers have tagged one particular shop and one particular product webpage as their favorite collections, respectively. The moderator,

536

C.X.J. Ou, K.C.C. Chan / Information & Management 51 (2014) 532–540

Fig. 1. The research model.

product type, is operationalized as an ascending number from lowest-price products to highest-price products with the corresponding coding number starting from 1 for the lowest price product. The dependent variable, sales volume, was measured by the number of pieces of a single focal product sold by an individual seller. Therefore sales volume is product and seller specific. This is a number automatically recorded by the electronic market and displayed in each individual seller’s webpage of the focal product. Appendix A shows an example of a TaoBao seller. The details of the objective data collected in this study are explained in Appendix B. 3.2. Control variables In addition to the variables mentioned above, we also collected data related to seller location (variable called ‘‘Location’’), setup duration (variable called ‘‘Setup_Years’’, meaning how long the shop has been established), and price premium and treated as control variables for data analysis. We argue that some buyers prefer transactions in the same city in order to reduce postage and increase transaction convenience. We coded the most popular city for online transaction in TaoBao as 1, the second most popular one as 2, etc. As a seller with a long history in a marketplace must have the ability to attract new and maintain old customers, we suggest that the setup duration has a positive influence on buyers’ decisions. Meanwhile, empirical research (e.g., [10,27]) has demonstrated the impacts of price on buyers’ decision. In the electronics markets, some buyers perceive a relatively higher price to be an indication of good quality [2]. Premium pricing, meaning setting the price of a product higher than similar products [2,26], has often been used as a pricing strategy [10] where the price is set close to the high end in order to attract quality- or statusconscious buyers. Following [2], we thus control the effects of price premium in our data analysis and measure price premiums based on the difference between the actual sales price and the industrial average sales price which is calculated in terms of percentage. 3.3. Data collection In order to collect the objective data, we have written a web crawler tailored to TaoBao’s website structure in Java. With this web crawler, the information contained on the web pages of TaoBao can be collected and processed automatically. The information collected was then written by the web crawler into XML format. The XML data are then converted into the data format

acceptable by Statistical Package for the Social Sciences (SPSS) for the purpose of data analysis. To test our hypotheses, we chose to obtain sellers’ data on TaoBao related to electronics products which an average industrial price was imposed by TaoBao and indicated in sellers’ webpages of such products. We focus on Nokia, Samsung, and Apple products considering that they are the market leaders in cell phones and tablet PCs. Specifically, we designed our web crawler to cover eight possible different models including Nokia N8, N97 mini, 5800XM, Lumia 920, Samsung SIII, Samung Note 2, iPad 4, and iPhone 5. There are several reasons why we chose to focus on these products:We set the sample frame on electronic products, specifically, cell phones and tablet PCs, because there are many sellers of such products. The number of shops at TaoBao selling these products, as well as the variation of their prices and sales volumes, is sufficient to provide a reasonable sample for meaningful data analysis.We found that TaoBao actually provides a reference price, namely the industrial average price, for these cell phones and tablets PCs. The industrial average price of each cell phone and tablet model is automatically displayed on all sellers’ individual product webpage, which is imposed by TaoBao. This industrial average price allows potential buyers to determine the price difference and also enables us to calculate the price premium for the focal seller.We collected data for three of the most popular brands, viz., Nokia, Apple, and Samsung, in the data analysis. The influences of brand effects can be randomized for our studies to so that any bias can be canceled out with inclusions of all such different brands and products.For comparison, it should be noted that the price of these cell phone and tablet PC models vary from RMB1701.48 to RMB4870.76. They can be regarded as different types of products in terms of pricing as well. Such a classification allows us to empirically examine the moderating effects of product type as proposed in the research model. Given the above, we used our web crawler to search all distinct web pages from TaoBao that sells Nokia, Samsung, and Apple products covering the eight models. The crawler returned a total of 16994 valid records of distinct web pages after we removed all broken links or those leading to the message that ‘‘this seller no long exists at TaoBao’’. We summarized the frequency of the sellers who sell these products in Table 2. 4. Data analysis As explained in the previous section, the sales representatives’ online status was collected according to three different time periods within a day, viz., the morning (06:00–12:00), afternoon (12:00–18:00), and night (18:00–06:00). Specifically, the

C.X.J. Ou, K.C.C. Chan / Information & Management 51 (2014) 532–540 Table 2 The frequency of web pages selling the 8 product types.

Total

Product Type

Industry average price

Nokia 5800XM Nokia N97 Mini Samsung S3 Nokia Lumia 920 Samsung Note2 Nokia N8 Apple iPad4 Apple iPhone5

1701.48 2543.96 2900.16 3444.96 3737.90 3824.46 3835.65 4870.76

8

–

Frequency

Percent (%)

391 435 3790 828 3714 579 3359 3898

2.3 2.6 22.3 4.9 21.9 3.4 19.8 22.9

16994

100.0

percentages of sellers who maintain sales representatives online in WangWang are 65.1%, 66.5%, 47.1% respectively in the morning, afternoon and night sessions. In order to obtain an overview regarding possible effects that the online seller’s virtual presence may bring about, we calculated the sales representatives’ online status in the three time periods using virtual presence as a latent variable, and ran a regression in Partial Least Square (PLS) following the method introduced by [23]. We discovered that for each time slot, the effect of virtual presence was 0.186 (p < 0.01), 0.396 (p < 0.01), and 0.177 (p < 0.01), respectively. In order to check if there is any potential difference between possible effects on weekdays and that on weekends, we also separated our analyses according to these two time slots, but no major differences were found. We suspected the reason why the data were not significantly distinguishable between weekends and weekdays was due to the facts that all the shops and salespersons, regardless of whether or not they are online and offline staff, work in both weekdays and weekends in China. As a result of all these considerations, we use the factor score of the latent variable generated in PLS to represent virtual presence for the later analysis. Before conducting the regression analysis, we first made use of the collinearity-checking function at SPSS to examine the potential

537

multicollinearity problem, all the VIF values of independent variables (IVs) were found to be smaller than 2, except three rating scores in description, service, and delivery. From a conceptual perspective, these three rating scores appear to be naturally grouped into the same set of factors because of the halo effect (c.f. [32]), meaning that a consumer’s individual evaluation rating on a seller is influenced by his/her overall impression of the seller in the context of this study. In order to further investigate the multicollinearity problem, we conducted factor analysis across three rating scores in description, service, and delivery. This showed that these three rating scores were loaded into one single factor. Following classical approaches to handle multicollinearity problem [17], we obtained the corresponding factor score, labeled as F_RatingScore, associated with the three rating scores as an IV for later regression analysis. Following all these steps, we conducted a 4-step hierarchical regression against sales volume to examine the theoretical hypotheses, starting from the control variables (i.e., ‘‘Location’’, ‘‘Setup_Years’’, and Price Premium calculated in terms of percentage). Such hierarchical regressions allow us to determine the unique variance measured as the increment in R square change and the F value, showing how each set of IVs contributed to the dependent variable [11,6], viz., sales volume, in our study. Table 3 shows that the R square changes of those 4 steps of regressions are all significant at the level of p < 0.01, including the regressions of the control variables (model 1), the institutional based mechanisms (model 2), the social based mechanisms (model 3), and the moderating effects of product type against sales volume (model 4). Although each specific institutional and social mechanism demonstrates does not exert the same influence in seller differentiation, the overall significance of R square changes for models 2, 3, and 4 verify H1, H2, and H3 in principle. The results suggest positive increasing impacts of these two categories of factors–institutional mechanisms (F change = 106.456, p = 0.0000) and social mechanisms (F change = 2599.888, p = 0.000)–on sales volume. The R square change for model 4 is significant

Table 3 Results of the Hierarchical regression (dependent variable: sales volume; n = 16994). Category

Predictors

Standardized coefficients *0.01 < p < 0.05, **p < 0.01) Model 1

Control variables

Location Setup_Years PricePreium

Institutional based quality signaling mechanisms

ReputationScore F-RatingScore ConsumerProctectionScheme SevenDayReturn RepairingService

Social based quality signaling mechanisms

VirtualPresence ShopTagging ProductTagging

Moderator

ProductType ProductType  ReputationScore ProductType  F_RatingScore ProductType  ConsumerProctectionScheme ProductType  SevenDayReturn ProductType  RepairingService ProductType  VirtualPresence ProductType  ShopTagging ProductType  ProductTagging R square Adjusted R square

Model summary

R square change F change Sig. F change

0.019* 0.054** 0.004

Model 2

Model 3

Model 4

0.007 0.036** 0.004

0.006 0.018** 0.003

0.006 0.023** 0.002

0.100** 0.015 0.111** 0.061** 0.039**

0.036** 0.012 0.074** 0.034** 0.026**

0.142** 0.033 0.107** 0.052** 0.044**

0.060** 0.032** 0.557**

0.071** 0.099** 0.256** 0.025 0.234** 0.035 0.042 0.021 0.019 0.014 0.071** 0.364** 0.360 0.359

0.003 0.003

0.034 0.033

0.338 0.338

0.003 18.008 0.000

0.030 106.456 0.000

0.305 2599.888 0.000

0.022 63.914 0.000

538

C.X.J. Ou, K.C.C. Chan / Information & Management 51 (2014) 532–540

(F change = 63.914, p = 0.000), validate H3 where product type was only regarded as a moderator in the whole research model, but not an independent variable. These two sets of quality signaling mechanisms, together with the product type as the moderator, result in 35.9% variance explained to the dependent variable, sales volume, suggesting an adequate goodness-of-fit for the overall research model [4]. We discuss the findings related to specific significant factors found in the hierarchical regression (as shown in Table 3) in the following key finding section. 5. Key findings, implications and future research 5.1. Key findings Although not every factor is as significant as expected, the results of the 4-step hierarchy regression indicate interesting findings. Specially, our data suggest that buyers are not sensitive to sellers’ pricing strategy, but they are sensitive to the seller history and location. Buyers’ purchase decisions are significantly influenced by the institutional based quality signaling mechanisms. These effective institutional based quality signaling mechanisms include the reputation system (b = 0.100, p < 0.01), joining customer protection scheme (b = 0.111, p < 0.01), enforcing seven-day return policy (b = 0.061, p < 0.01), and providing thirty-day repairing service (0.039, p < 0.01). Our data suggest they can serve as an effective starting point for the seller to attract potential customers. However, more importantly, our regression results reveal that online sellers with a better utilization of social mechanisms, specifically virtual presence (b = 0.060, p < 0.01 in model 3), shop tagging (b = 0.032, p < 0.01 in model 3), and product tagging (b = 0.557, p < 0.01 in model 3), can win more customers. As also evidenced by the hierarchical model 4, the influence of reputation score (bProductTypeReputation = 0.400, p < 0.01) and popularity tagging (bProductTypeShopTagging = 0.071, p < 0.01; bProductTypeProductTagging = 0.364, p < 0.01) on sales volume are more substantial for the expensive product type. We discuss the implications of these key findings below. 5.2. Implications This study is among one of the first research on seller differentiation in electronic markets. Unlike the other studies that focus on one particular domain, such as institutional mechanisms, this study emphasizes the competing powers of the two blocks of variables, viz., institution mechanisms and social based quality singling mechanisms, in determining a seller’s position in the electronic market. By extending quality signaling theories [1,31,16], this study established a typology of quality signals related to institutional and social mechanisms in electronic markets with both theoretical and practical insights, as explained below. From the theoretical perspective, this study contributes to the current research on online marketplace by proposing an integrative lens to examine the factors related to quality signals on differentiating sellers. In the discipline of economics, research (e.g., [16]) has provided a basic typology for classifying Internet Auction features as quality signals into product and seller related characteristics. In the IS discipline, previous work (e.g., [2,20,25]) has largely focused on whether the seller’s reputation is a significant factor in determining price premiums that buyers are willing to pay. Instead of merely demonstrating that reputation is an important factor in online transactions, this study shows that social mechanisms, is the most effective quality signals and the most robust predictors of sales volume by contributing substantial variance explained to seller differentiation. By including the product type, under the categories of different prices, as the moderator in our model, we demonstrate that social mechanisms,

manifested as a seller’s reputation score and his/her popularity tagging, are the most prevalent factors in pushing up sales volume for expensive products. The extant literature has provided a starting point for us to investigate into the effectiveness institutional mechanisms in building buyers’ trust and consequently the price premium. This study suggests a new integrated theoretical lens with the quality signals to investigate sellers’ differentiation from institutional and social mechanisms, thus providing researchers a springboard to further examine the buying and selling issues related to electronic markets from the perspective of quality signals. From the practical perspective, this study provides both electronic markets and online sellers with compelling design guidelines. The current marketplaces have heavily incorporated the institutional mechanisms such as the reputation system, as well as various guarantees and warranty. As also informed by the data, the seller’s reputation score is still the most effective among all institutional mechanisms. However, more importantly, we suggest that the current design of the marketplaces and individual sellers’ design can evolve from focusing on institutional mechanisms to the emphasis of social mechanisms. With technology advancement and the possibility of virtual presence of the seller, the interaction between sellers and buyers through computer-mediated interaction (via the IM tool, namely WangWang, in the current research context) can in fact serve as a good channel to convince potential buyers to make up their purchase decisions. A better design of individual product’s webpage, such as including detailed product descriptions and more pictures, can attract more buyers to tag the product webpage or the shop, as their favorite collections in bookmarks. As evidenced by this study, it is very likely to transfer potential buyers into actual buyers via an effective website designed with these mechanisms. This is the case especially for expensive products. Therefore, better utilizing these social mechanisms can render online sellers a better position in the electronic market. From the methodological perspective, we offered an innovative way to collect objective data. With the use of web crawler and the analysis of a large dataset of objective data, we can address the potential bias issues resulted from subjective data. We believe the crawling method used in this study can provide both researchers and practitioners with a good alternative in data collection. 5.3. Future research This study opens up several opportunities for future research. First, the study demonstrates the significance of social influence factors in electronic markets. However, the data and such a platform design are only available in China, typically TaoBao. The design factors proposed in this study can be compared, in terms of differences and effectiveness, with those in other electronic markets such as eBay. Second, considering that average industrial price is shown in all individual sellers’ product webpage, foisted by the TaoBao platform, electronic products is the product chosen in our current research. Our current approach to analyze webcrawled data can be used to examine different products. Third, we only include objective data in this study. However, we believe that the incorporation of other subjective data such as trust can further strengthen the current study. Last but not the least, it will be also very interesting for the time effects of all the proposed factors on stimulating sales volume to be studied via longitudinal studies. This can enhance the empirical findings from the current crosssectional study. 6. Conclusion It is difficult to stand out from the crowd of sellers in online marketplace. However, this study demonstrates three categories

C.X.J. Ou, K.C.C. Chan / Information & Management 51 (2014) 532–540

of mechanisms for the online sellers to utilize. Our data indicate their significant explanation powers on sales volume. Although the reputation in online marketplace may act as a barrier of entry for new sellers, sellers can better make use of the social based quality singling mechanisms, typically virtual presence, product and shop tagging, to enter the market as new comers or further

539

strengthen the market leadership as the core players in the competitive electronic markets. With the technological advancement, we believe electronic markets will provide the sellers more effective mechanisms to facilitate the online transactions and thus a long-term sustainability of electronic markets can be maintained.

Appendix A. An example of a TaoBao seller’s shop interfaces

Appendix B. The objective data collected in this study Number in Appendix A

Labels in Appendix A

Explanations

1

ReputationScore

2

RatingScore-Description

3

RatingScore-Service

4

RatingScore-Delivery

5

Consumer ProctectionScheme

6

SevenDayReturnPolicy

7

ReparingService

8

VituralPresence

9

ProductTagging

10

ShopTagging

11

IndustryAveragePrice

12

ProductPrice

540

C.X.J. Ou, K.C.C. Chan / Information & Management 51 (2014) 532–540

References [1] G.A. Akerlof, The market for ‘Lemons’: quality uncertainty and the market mechanism, Quarterly Journal of Economics 84 (3), 1970, pp. 488–500. [2] S. Ba, P.A. Pavlou, Evidence of the effect of trust building technology in electronic markets: price premiums and buyer behavior, MIS Quarterly 26 (3), 2002, pp. 243–268. [3] J. Chen, C. Zhang, Y.F. Yuan, L.H. Huang, Understanding the emerging C2C electronic market in China: an experience-seeking social marketplace, Electronic Markets 17 (2), 2007, pp. 86–100. [4] W.W. Chin, Issues and opinion on structural equation modeling, MIS Quarterly 22 (1), 1998, pp. vii–xvi. [5] Cisco, Taobao.com Chooses Cisco to Support Growing Database of Users, 2009 available at http://www.cisco.com/en/US/solutions/collateral/ns340/ns517/ ns224/case_study_alibaba.pdf (accessed on 20.07.13).. [6] F.D. Davis, R.P. Bagozzi, P.R. Warshaw, User acceptance of computer technology: a comparison of two theoretical models, Management Science 35 (8), 1989, pp. 982–1003. [7] eBay, eBay Online Media Centre, 2013 available at http://pages.ebay.in/community/aboutebay/news/infastfacts.html (accessed on 20.07.13). [8] D. Gefen, E. Karahanna, D.W. Straub, Trust and tam in online shopping: an integrated model, MIS Quarterly 27 (1), 2003, pp. 321–339. [9] J. Graham, W. Havlena, Finding the ‘‘Missing Link’’: advertising’s impact on word of mouth, web searches, and site visits, Journal of Advertising Research 47 (4), 2007, pp. 427–435. [10] V.O. Kayhan, J.A. McCart, A. Bhattacherjee, Cross-bidding in simultaneous online auctions: antecedents and consequences, Information & Management 47 (7–8), 2010, pp. 325–332. [11] R.T. Keller, Technology-information processing fit and the performance of R&D project groups: a test of contingency theory, The Academy of Management Journal 37 (1), 1994, pp. 167–179. [12] A. Kirmani, A.R. Rao, No pain, no gain: a critical review of the literature on signaling unobservable product quality, Journal of Marketing 64 (April), 2000, pp. 66–79. [13] M. Koufaris, W. Hampton-Sosa, The development of initial trust in an online company by new customers, Information & Management 41 (3), 2004, pp. 377– 397. [14] H.H. Kuan, G.W. Bock, Trust transference in brick and click retailers: an investigation of the before-online-visit phase, Information & Management 44 (2), 2007, pp. 175–187. [15] M.B. Lee, K.S. Suh, J. Whang, The impact of situation awareness information on consumer attitudes in the Internet shopping mall, Electronic Commerce Research and Applications 2 (3), 2003, pp. 254–265. [16] S. Li, K. Srinivasan, B. Sun, Internet auction features as quality signals, Journal of Marketing 73 (1), 2009, pp. 75–92. [17] Loh Lawrence, N. Venkatraman, Determinants of information technology outsourcing: a cross-sectional analysis, Journal of Management Information Systems 9 (1), 1992, pp. 7–24. [18] V.F. Magirou, Stockpiling under price uncertainty and storage capacity constraints, European Journal of Operations Research 11 (3), 1982, pp. 233– 246. [19] D.H. McKnight, V. Choudhury, C. Kacmar, Developing and validating trust measures for e-commerce: an integrative typology, Information Systems Research 13 (3), 2002, pp. 334–359. [20] M.I. Melnik, J. Alm, Does a seller’s ecommerce reputation matter? Evidence from eBay auctions The Journal of Industrial Economics 50 (3), 2002, pp. 337– 349. [21] C.X.J. Ou, R.M. Davison, Why eBay lost to TaoBao in China: the glocal advantage, Communications of the ACM 52 (1), 2009, pp. 145–148. [22] C.X.J. Ou, R.M. Davison, P.A. Pavlou, Y. Li, Leveraging rich communication tools: evidence from online trust and Guanxi in China, Proceeding of the 29th International Conference on Information Systems, Paris, December 14–17, 2008.

[23] C.X.J. Ou, P.A. Pavlou, R.M. Davison, Swift Guanxi in online marketplaces: the role of computer-mediated-communication technologies, MIS Quarterly 38 (1), 2014, pp. 209–230. [24] Y.H. Park, E. Bradlow, An integrated model for whether, who, when, and how much in Internet auctions, Journal of Marketing Research 42 (2), 2005, pp. 470–482. [25] P.A. Pavlou, A. Dimoka, The nature and role of feedback text comments in online marketplaces: implications for trust building, price premiums, and seller differentiation, Information Systems Research 17 (4), 2006, pp. 391–412. [26] P.A. Pavlou, D. Gefen, Building effective online marketplaces with institutionbased trust, Information Systems Research 15 (1), 2004, pp. 37–60. [27] J.S. Raju, The effect of price promotions on variability in product category sales, Marketing Science 11 (3), 2002, pp. 207–220. [28] R.A. Shang, Y.C. Chen, L. Shen, Extrinsic versus intrinsic motivations for consumers to shop on-line, Information and Management 42 (3), 2005, pp. 401–413. [29] C. Shapiro, Premiums for high quality as returns to reputation, Quarterly Journal of Economics 97, 1983, pp. 659–686. [30] J.A. Short, E. Williams, B. Christie, The Social Psychology of Telecommunications, John Wiley & Sons, New York, 1976. [31] M.A. Spence, Job market signaling, Quarterly Journal of Economics 87 (3), 1973, pp. 355–374. [32] E.L. Thorndike, A constant error in psychological ratings, Journal of Applied Psychology 4 (1), 1920, pp. 25–29. [33] Wikipedia, TaoBao, 2013, available at http://en.wikipedia.org/wiki/Taobao (accessed on 20.07.13). [34] O.E. Williamson, The Economic Institutions of Capitalism, Free Press, New York, 1985. Carol X.J. Ou is an Associate Professor of Information Management at Tilburg University. Her research interests include electronic commerce, social networks, computer-mediated communication, website design, knowledge management, and IT adoption in organizations. Her publications have appeared in such journals as Communications of the ACM, Decision Support Systems, Information Technology & People, Journal of AIS, Journal of the American Society for Information Science and Technology, International Journal of Human-Computer Studies, Information Systems Journal, and MIS Quarterly, among others. Carol is also serving as a Senior Editor for Information Systems Journal, Information Technology & People, and Electronic Journal of Information Systems in Developing Country. She is a Certified IS Auditor and an Academic Advocate of IS Audit and Control Association. Keith C.C. Chan obtained his B.Math. (Hons.) degree in Computer Science and Statistics and M.A.Sc. and Ph.D. degrees in Systems Design Engineering from the University of Waterloo, Waterloo, Ontario, Canada in 1984, 1985, and 1989 respectively. He joined the IBM Canada Laboratory as a software analyst soon after graduation and was involved in the design and development of software engineering tools and image and multimedia software products. In 1994, he joined The Hong Kong Polytechnic University where he is currently a Professor in the Department of Computing. He was Head of this Department from 2002 to 2008. Prof. Chan’s research interests are in Big Data Analytics, Cloud Computing and Software Engineering. He is the author and co-author of three books and close to 200 publications in refereed journals and conference proceedings. He is on the editorial board of four journals and is on the program committee of numerous conferences. He is currently also a Guest Professor of the Graduate University of the Chinese Academy of Sciences, Beijing, China.