1534 Nibbling, Sniping, and the Role of Uncertainty in Second-Price, Hard-Close Internet Auctions 0.596, the number of bids is 0.660, the proportion of bids is 0.025, and the proportion of bidders is 0.042. The empirical distributions for each of our variables in Table 1b are analogous to those for the coin auctions summarized in Table 1a. Com- parisons of the mean and median values indicate that a number of variables are likely to be non- normally distributed. As a result, it is necessary to employ nonparametric analysis to control for this possibility. Comparing Tables 1a and 1b, we note differ- ences in nibbling activity across the two types of auctions. When evaluated at the sample mean, the automobile auctions have a higher number of bids and bidders, nibbles and nibblers, and a higher proportion of bids that are nibbles and bidders that nibble. With the exception of bids that are nibbles, these differences also hold true at the sample median as well. These statistics lend support to hypothesis H 1 , as the automobile auctions (which exhibit more uncertainty) have both higher numbers and intensities of nibbling. However, it remains to be seen from the subse- quent hypothesis tests whether these differences DUHLQIDFWVWDWLVWLFDOO\VLJQL¿FDQW We can also make inferences about sniping activity across the two auction formats. The au- tomobile auctions exhibit lower mean and median values for both our sniping activity and intensity variables. The latter implies that, when we control for size of the auction, there is less sniping and more nibbling in the more uncertain market. One possible explanation is that these auctions also exhibit more (and more intense) nibbling, which may lead to higher bid prices, and drive some of the potential snipers (who may not have extremely- high reservation values) out of the market before they have a chance to snipe. Our hypothesis tests are contained in Tables 2 through 4. Table 2 contains the Mann-Whitney test used to address hypotheses 1 and 2, while Tables DQGFRQWDLQWKHFRUUHODWLRQFRHI¿FLHQWDQDO\VLV used to address hypotheses 3 and 4. The results in Table 2 provide clear evidence that the graded coin and automobile auction data show different nibbling behavior for all of our measures (p < 0.01), except the proportion of ELGVWKDWDUHQLEEOHV6SHFL¿FDOO\WKHDXWRPR- ELOH DXFWLRQ GDWD FRQWDLQHG VLJQL¿FDQWO\ PRUH nibblers, nibbles, and portion of bidders who nibble. The automobile auctions also exhibited Variable M-W Test Z-Stat Prob. Number of Bids 192.500 -6.372 0.000 Number of Bidders 332.000 -5.189 0.000 Number of Nibblers 344.500 -5.083 0.000 Number of Nibbles 275.000 -5.671 0.000 Number of Bids per Bidder 345.000 -5.085 0.000 Proportion of Bidders who Nibble 578.500 -3.090 0.002 Proportion of Bids that are Nibbles 909.000 -0.265 0.791 Number of Bids in Last Minute 796.000 -1.342 0.180 Number of Bidders in Last Minute 827.500 -1.055 0.292 Proportion of Bids in Last Minute 647.000 -2.704 0.007 Proportion of Bidders in Last Minute 651.000 -2.667 0.008 Table 2. Nonparametric (Mann-Whitney u-test) hypothesis tests 1535 Nibbling, Sniping, and the Role of Uncertainty in Second-Price, Hard-Close Internet Auctions VLJQL¿FDQWO\PRUHSDUWLFLSDWLRQDVHYLGHQFHGE\ WKH VLJQL¿FDQWO\KLJKHUQXPEHURI ELGGHUV DQG bids per auction. $VDFRUROODU\WRWKHVH¿QGLQJV7DEOHDOVR provides evidence about differences in sniping activity across the two product categories. With regard to the total amount of sniping, the samples DUHQRWVWDWLVWLFDOO\VLJQL¿FDQWO\GLIIHUHQWEXWDUH different in the proportion of bids and bidders in the last minute (p < 0.01). This supports our earlier argument that the amount and intensity of sniping differs within the same auction format. The auctions whose product has a more certain YDOXH FHUWL¿HG FRLQV H[KLELWHG PRUH VQLSLQJ behavior. Tables 3 and 4 present our analysis of hypoth- eses 3 and 4. Analysis of the coin auction data in 7DEOHLQGLFDWHWKDWWKHUHLVQRVLJQL¿FDQWFRU- relation (p < 0.05) between the amount of sniping and the amount or intensity of nibbling in the coin GDWD+RZHYHUWKHUHLVDVLJQL¿FDQWFRUUHODWLRQ between the intensity of sniping behavior and both W KH D PR X Q W D QG L Q W HQ VL W \ RI Q LE E O L Q J 6 S HF L ¿F DO O\ the intensity of sniping behavior (when measured as both the proportion of bids in the last minute of the auction as well as the proportion of bidders Number of Bids in Last Minute Number of Bidders in Last Minute Spearman Spearman Variable Correlation Prob. Correlation Prob. Number of Bids 0.072 0.330 0.085 0.301 Number of Bidders 0.139 0.197 0.168 0.151 Number of Nibblers 0.078 0.317 0.109 0.253 Number of Nibbles 0.101 0.269 0.121 0.229 Number of Bids per Bidder -0.061 0.354 -0.151 0.176 Proportion of Bidders who Nibble -0.197 0.111 -0.180 0.133 Proportion of Bids that are Nibbles 0.149 0.180 0.177 0.138 Proportion of Bids in Last Minute Proportion of Bidders in Last Minute Spearman Spearman Variable Correlation Prob. Correlation Prob. Number of Bids -0.299 0.031 -0.286 0.037 Number of Bidders -0.228 0.079 -0.261 0.052 Number of Nibblers -0.281 0.040 -0.314 0.024 Number of Nibbles -0.271 0.045 -0.284 0.038 Number of Bids per Bidder -0.250 0.060 -0.225 0.082 Proportion of Bidders who Nibble -0.468 0.001 -0.491 0.001 Proportion of Bids that are Nibbles -0.168 0.151 -0.204 0.103 Table 3. Nonparametric (Spearman) correlations for graded coin auctions Note: Probability values apply to a one-tailed test 1536 Nibbling, Sniping, and the Role of Uncertainty in Second-Price, Hard-Close Internet Auctions who snipe) is negatively correlated (p < 0.05) with the number of nibbles, the number of nibblers, and the portion of bids that are nibbles. In direct contrast, the results for the automo- ELOHDXFWLRQVGHPRQVWUDWHDVLJQL¿FDQWSRVLWLYH correlation between the amount of sniping and the amount of nibbling, as well as the intensity RIVQLSLQJDQGWKHLQWHQVLW\RIQLEEOLQJ6SHFL¿- cally, the amount of sniping, as measured by the number of sniping bids and the number of snip- HUVLVVLJQL¿FDQWO\DQGSRVLWLYHO\FRUUHODWHGS < 0.05) with the number of nibbles, the number of nibblers, and the proportion of bids that are nibbles. The intensity of sniping, as measured by the proportion of bids in the last minute and the proportion of bidders who snipe in the last minute, LVVLJQL¿FDQWO\DQGSRVLWLYHO\DVVRFLDWHGZLWKWKH number of nibbles, the number of nibblers, and the proportion of bids that are nibbles. When the results contained in Tables 3 and 4 are taken together, our results extend and clarify Ockenfels and Roth’s (2002) assertion that par- Number of Bids in Last Minute Number of Bidders in Last Minute Spearman Spearman Variable Correla- tion Prob. Correlation Prob. Number of Bids 0.486 0.001 0.481 0.001 Number of Bidders 0.528 0.000 0.536 0.000 Number of Nibblers 0.509 0.000 0.507 0.000 Number of Nibbles 0.568 0.000 0.574 0.000 Number of Bids per Bidder -0.156 0.148 -0.167 0.132 Proportion of Bidders who Nibble 0.218 0.070 0.183 0.109 Proportion of Bids that are Nibbles 0.438 0.001 0.457 0.001 Proportion of Bids in Last Minute Proportion of Bidders in Last Minute Spearman Spearman Variable Correla- tion Prob. Correlation Prob. Number of Bids 0.336 0.011 0.357 0.007 Number of Bidders 0.420 0.002 0.374 0.005 Number of Nibblers 0.410 0.002 0.358 0.007 Number of Nibbles 0.437 0.001 0.442 0.001 Number of Bids per Bidder -0.220 0.069 -0.134 0.184 Proportion of Bidders who Nibble 0.178 0.116 0.159 0.143 Proportion of Bids that are Nibbles 0.407 0.003 0.365 0.006 Table 4. Nonparametric (Spearman) correlations for automobile auction Note: Probability values apply to a one-tail test 1537 Nibbling, Sniping, and the Role of Uncertainty in Second-Price, Hard-Close Internet Auctions ticipants resort to sniping more frequently in circumstances of higher uncertainty in an effort to avoid the bidding wars that may occur when several participants nibble in that same auction. In auctions with less uncertainty regarding the value of the good which is being auctioned, the number of nibblers and the intensity of their nib- bling have a negative impact on sniping. This is consistent with the idea that when the true value of the good has little uncertainty, it is optimal to conceal your reservation price. CONCLUSION We randomly selected items from completed auctions on eBay for two types of goods: one that H[KLELWVDVLJQL¿FDQWGHJUHHRIXQFHUWDLQW \DERXW the product’s value (used cars); and one where WKHUHLVVLJQL¿FDQWO\OHVVXQFHUWDLQW\DERXWWKH SURGXFW¶VYDOXHFHUWL¿HGFRLQV8VLQJQRQSDUD- metric (Mann-Whitney) analysis of variance and Spearman correlation analysis, we test hypotheses based on bidder behavior in the same auction format, but across different product groups. Our results indicate that auctions with a high amount RIXQFHUWDLQYDOXHH[KLELWVLJQL¿FDQWO\PRUH QLEEOLQJ 6SHFL¿FDOO\ PRUH QLEEOLQJ RFFXUUHG both in terms of the number (and proportion) of participants who use a nibbling strategy, as well a s t h e n u m b e r (a n d p r o p o r t i o n) o f i n c r e m e nt a l b i d s that are submitted over the course of the auction. Auction participants were also more prone to en- gage in sniping in auctions with more uncertain value. Moreover, sniping occurred most often in high-risk auctions when several other participants attempted nibbling strategies. This is consistent ZLWKWKH¿QGLQJVRI2FNHQIHOVDQG5RWK which suggest that many bidders use sniping to avoid getting into bidding wars. For lower-risk auc- tions, where the value of the good is known with PRUHFHUWDLQW\ZH¿QGWKDWDXFWLRQSDUWLFLSDQWV are less prone to nibble, and thereby withhold information concerning their reservation price, engaging in sniping behavior. It is also interesting to note that, even though participants did not have reliable information about other participants’ reservation prices, coin auc- tion participants’ behavior deviated only slightly from that predicted in conventional second-bid, hard-close auctions. Our analysis indicates that WKHUHPD\EHOLWWOHEHQH¿WIURPREVHUYLQJRWKHUV¶ bidding behavior in this circumstance. However, in car auctions, where there was a large amount of uncertainty concerning the good being pur- chased, auction participants’ behavior deviated VLJQL¿FDQWO\IURPWKDWSUHGLFWHGIRUWUDGLWLRQDO second-bid, hard-close auctions. Our results are consistent with the idea that bidders engage in nibbling and sniping to overcome a lack of infor- mation concerning the valuation of the product. In this case, the lack of reliable knowledge con- cerning the reservation price of other participants may have induced alternative bidding strategies (nibbling and sniping) in an attempt to overcome LQIRUPDWLRQGH¿FLHQFLHV :KLOH RXU ¿QGLQJV SURYLGH DQ LQWHUHVWLQJ FODUL¿FDWLRQRI,QWHUQHWDXFWLRQWKHRU\RXUVWXG\ is preliminary in nature and could be extended in a number of ways. First, the spectrum of goods included in the study could be expanded in or- der to give a clearer picture of how uncertainty LQÀXHQFHVRQOLQHELGGLQJEHKDYLRU,QDGGLWLRQ LW ZRXOG EH LQWHUHVWLQJ WR VWXG\ WKH UDPL¿FD- tions of other online auction rules, such as those employed by Amazon.com, on the behavior of auction participants. Another potential limitation of our study is the use of Spearman correlations. While correlations are a parsimonious technique, an alternative might be the use of maximum likelihood-based regres- sion techniques. In this case, many other factors would be needed to control for omitted variables. If important control variables are not included, our results would be biased and inconsistent. Most of the control variables needed to perform this 1538 Nibbling, Sniping, and the Role of Uncertainty in Second-Price, Hard-Close Internet Auctions type of analysis are not available through e-Bay or other online auction providers. We leave these topics for future research. REFERENCES Anonymous (2004). At the drop of a hammer. Economist, 371(8375), 10. Ariely, D., & Simonson, I. (2003). Buying, bid- ding, playing, or competing? Value assessment and decision dynamics in online auctions. Journal of Consumer Psychology, 13(1&2), 113-123. Baatz, E. (1999). Online auctions start to pick up steam. Purchasing, 127(6), 46-56. Bajari, P., & Hortacsu, A. (2003). The winner’s curse, reserve prices, and endogenous entry: Empirical insights from eBay auctions. Rand Journal of Economics, 34(2), 329-355. Bauer, R. (1960). Consumer behavior and risk taking. In R. Hancock (Ed.), Proceedings of the American Marketing Association (pp. 389-398). Brint, A. T. (2003). Investigating buyer and seller strategies in online auctions. Journal of the Op- erational Research Society, 54(11), 1177-1197. Celsi, R., Rose, R., & Leigh, T. (1993). An explo- ration of high-risk leisure consumption through skydiving. Journal of Consumer Research, 20(1), 1-23. Dholakia, U. (2005). Concept discovery, process explanation, and theory deepening in e-marketing research: The case of online auctions. Marketing Theory, 5(1), 117-124. Dholakia, U. (2005). The usefulness of bidders’ reputation ratings to sellers in online auctions. Journal of Interactive Marketing, 19(1), 31-40. Dholakia, U., Basuroy, S., & 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. Dholakia, U., & Simonson, I. (2005). The effect of explicit reference points on consumer choice and online bidding behavior. Marketing Science, 24(2), 206-217. Dholakia, U., & Soltysinski, K. (2001). Coveted or overlooked? The psychology of bidding for comparable listings in digital auctions. Marketing Letters, 12(3), 225-237. Gardner, R. (2003). Games for business and economics, 2 nd ed. Hoboken, NJ: John Wiley and Sons Publishers. Gilkeson, J., & Reynolds, K. (2003). Determinants of Internet auction success and closing price: An exploration study. Psychology and Marketing, 20(6), 537-566. Heyman, J. E., Orhun, Y., & Ariely, D. (2004). Auction fever: The effect of opponents and quasi- endowment on product evaluations. Journal of Interactive Marketing, 18(4), 7-21. Kamins, M., Dreze, X., & Folkes, V. S. (2004). Effects of seller-supplied prices on buyers’ prod- uct evaluations: Reference prices in an Internet auction context. Journal of Consumer Research, 30(4), 622-628. Kauffman, R. J., & Walden, E. A. (2001). Eco- nomics and electronic commerce: A survey and directions for research. International Journal of Electronic Commerce, 5(4), 5. Marcoux, A. (2003). Snipers, stalkers, and nib- blers: Online auction business ethics. Journal of Business Ethics, 46(2), 163-173. McDonald, C., & Slawson, V. (2002). Reputation in an Internet auction market. Economic Inquiry, 40(4), 633-650. Ockenfels, A., & Roth, A. (2002). The timing of bids in Internet auctions: Market design, bid- 1539 Nibbling, Sniping, and the Role of Uncertainty in Second-Price, Hard-Close Internet Auctions GHUEHKDYLRUDQGDUWL¿FLDODJHQWVAI Magazine, (Fall), 79-88. Oh, W. (2002). C2C versus B2C: A comparison of the winner’s curse in two types of electronic auctions. International Journal of Electronic Commerce, 6(4), 115-138. Puto, C., Patton, W., & King, R. (1985). Risk handling strategies in industrial vendor selection decisions. Journal of Marketing, 49(1), 89-98. Roth, A., & Ockenfels, A. (2002). Last-minute bidding and the rules for ending second-price auctions: Evidence from eBay and Amazon auc- tions on the Internet. American Economic Review, 92(4), 1093-1103. Sandholm, T. (2000). Issues in computational vickrey auctions. International Journal of Elec- tronic Commerce, 4(3), 107. Stafford, M. R., & Stern, B. (2002). Consumer bidding behavior on Internet auction sites. In- ternational Journal of Electronic Commerce, 7(1), 135-150. 6WDQGL¿UG665RHORIV05'XUKDP< (2004). The impact of eBay’s buy-it-now func- tion on bidder behavior. International Journal of Electronic Commerce, 9(2), 167-176. Ward, S. G., & Clark. J. M. (2002). Bidding be- havior in on-line auctions: An examination of the eBay Pokemon card market. International Journal of Electronic Commerce, 6(4), 139-155. Wilcox, R. (2000). Experts and amateurs: The role of experience in Internet auctions. Marketing Letters, 11(4), 363-374. This work was previously published in the International Journal of E-Business Research, edited by I. Lee, Volume 4, Issue 1, pp. 69-81, copyright 2008 by IGI Publishing (an imprint of IGI Global). 1540 Copyright © 2009, IGI Global, distributing in print or electronic forms without written permission of IGI Global is prohibited. Chapter 5.10 The Driving Forces of Customer Loyalty: A Study of Internet Service Providers in Hong Kong T. C. E. Cheng The Hong Kong Polytechnic University, Hong Kong L. C. F. Lai The Hong Kong Polytechnic University, Hong Kong A. C. L. Yeung The Hong Kong Polytechnic University, Hong Kong ABSTRACT In this study we examine the driving forces of customer loyalty in the broadband market in Hong Kong. We developed and empirically tested a model to examine the antecedents of customer loyalty towards Internet service providers (ISPs) in Hong Kong. Structural equation modeling (SEM) was used to evaluate the proposed model. A total of 737 valid returns were obtained through a questionnaire survey. The results show that customer satisfaction, switching cost, and price perception are antecedents that lead directly to customer loyalty, with customer satisfaction exert- LQJWKHJUHDWHVWLQÀXHQFH$OWKRXJKZHIRXQGWKDW VHUYLFHTXDOLW\VLJQL¿FDQWO\LQÀXHQFHVFXVWRPHU satisfaction, which in turn leads to customer loy- DOW \ZHGLGQRW¿QGDGLUHFWUHODWLRQVKLSEHWZHHQ service quality and customer loyalty. Our results also reveal that corporate image is not related to customer loyalty. Our empirical investigation suggests that investing huge resources in building corporate image can indeed be a risky strategy for ISPs. INTRODUCTION 'XHWRDUHFHQWVLJQL¿FDQWVXUJHLQWKHQXPEHU of ISPs, the broadband market in Hong Kong has 1541 The Driving Forces of Customer Loyalty EHFRPH YHU\ FURZGHG OHDGLQJ WR ¿HUFH SULFH competition, which has eventually resulted in the elimination of many ISPs from the market. From 2001 to 2006, the number of ISPs in Hong Kong dropped from 258 to 181. As the broadband market matures, the focus of ISPs has shifted from customer acquisition to customer retention. In March 2006, there were around 2.6 million Internet users, including both broadband and narrowband users, representing a 39% penetra- tion rate in Hong Kong. About 64% of these users DFFHVVWKURXJKWKHEURDGEDQG,QWHUQHW2I¿FHRI the Telecommunications Authority, 2006). These ¿JXUHVHVWDEOLVK+RQJ.RQJDVRQHRIWKHPRVW ,QWHUQHWFRQQHFWHGFLWLHVLQWKH$VLDQ3DFL¿F region. 7KHVLJQL¿FDQFHRIFXVWRPHUOR\DOW\FDQQRW be overemphasized because it relates closely to the continued survival, as well as the future growth, of companies. For a company to maintain DVWDEOHSUR¿WOHYHOZKHQWKHPDUNHWUHDFKHVWKH saturation point, a defensive strategy aiming at retaining existing customers is more important than an offensive one, which targets at expanding the size of the overall market by inducing potential customers to subscribe to its services (Ahmad & Buttle, 2002; Fornell, 1992). Previous studies on customer loyalty focused on customer satisfaction and switching barriers (Dick & Basu, 1994; Gerpott, Rams, & Schindler, 2001; Lee & Cunningham, 2001). These studies have found that customers experiencing a high level of satisfaction are likely to remain with their existing service providers and maintain their service subscriptions. Switching barriers, on the other hand, play a moderating role in the relationship between customer satisfaction and customer loyalty (Colgate & Lang, 2001; Lee & Cunningham, 2001). Researchers in this area have further elaborated on the linkages between price factors and perceived value (e.g., Grewal, Monroe, & Krishnan, 1998), as well as between price and customer loyalty (e.g., Voss, Parasura- man, & Grewal, 1998). In addition, the marketing literature supports the general notion that pricing factors affect the price perceptions of custom- ers, which in turn contribute to customer loyalty (Reichheld, 1996). By using SEM, this study empirically ana- lyzes whether customer satisfaction, switching cost, price perception, and corporate image are antecedents of customer loyalty in the context of the ISP market in Hong Kong. We also seek to identify elements of service quality as anteced- ents of satisfaction, and their levels of impact on satisfaction, and to ascertain whether service quality is a direct antecedent of customer loyalty. We examine the degree to which switching cost and price perception account for the variations in the strength of consumer loyalty to ISPs. Finally, we test if corporate image has any impact on customers’ loyalty to their present ISPs. THEORETICAL BACKGROUND AND HYPOTHESIS DEVELOPMENT Customer loyalty is a purchase behavior, which, unlike customer satisfaction, is an attitude (Grif- ¿Q&XVWRPHUOR\DOW\LVFRQFHUQHGZLWKWKH likelihood of a customer returning, making busi- ness referrals, providing strong word of mouth, as well as offering references and publicity (Bowen & Shoemaker, 1998). Loyal customers are less likely to switch to competitors in view of a given price inducement, and they make more purchases compared to less loyal customers (Baldinger & Rubinson, 1996). Although most research on loy- alty has focused on frequently purchased package goods (i.e., brand loyalty), the loyalty concept is also important for industrial goods (i.e., vendor loyalty), services (i.e., service loyalty), and retail establishments (i.e., store loyalty) (Dick & Basu, 1994). As evidenced in the previous discussions, customer loyalty has been generally described as occurring when customers repeatedly purchase goods or services over time, have word of mouth, and make referrals to other customers. 1542 The Driving Forces of Customer Loyalty Antecedents of Customer Loyalty One of the major factors found to affect customer loyalty is customer satisfaction. Halstead, Hart- man, and Schmidt (1994) considered customer satisfaction as an affective response that focuses on product performance against some prepurchase standard during or after consumption. Mano and O l i ve r (19 93) r ef e r r e d t o s a t i s f a c t i o n a s a n a t t it u d e or evaluative judgment varying along the hedonic continuum focusing on the product, which is evalu- DWHGDIWHUFRQVXPSWLRQ)RUQHOOLGHQWL¿HG satisfaction as an overall evaluation based on the total purchase and consumption experience of the target product, or service performance compared with prepurchase expectations over time. Oliver (1997, 1999) regarded satisfaction as DIXO¿OOPHQWUHVSRQVHRUMXGJPHQWRQDSURGXFW or service, which is evaluated for one-time or ongoing consumption. 6HUYLFHTXDOLW\FDQEHGH¿QHGDVWKHUHVXOW of the comparison between a customer’s expec- tations on a service and their perception of the way the service has been delivered (Gronroos, 1984; Lehtinen & Lehtinen, 1982; Lewis & Booms, 1983; Parasuraman, Zeithaml, & Berry, 1985, 1988, 1994). Perceived service quality is usually measured by two dimensions, namely process quality and output quality. Parasuraman et al. (1985, 1988, 1994) developed the 22-item SERVQUAL instrument, which has been widely used to measure service quality in many indus- tries, such as banking (Mukherjee & Nath, 2005), health care (Choi, Lee, Kim, & Lee, 2005), and airport service (Fodness & Murray, 2007). The SERVQUAL instrument assesses the overall service quality by comparing service expectation DQGDFWXDOSHUIRUPDQFHLQWHUPVRI¿YHJHQHULF dimensions, namely, tangibles, reliability, respon- siveness, assurance, and empathy. When consumers switch service providers, they will incur various costs ranging from the time spent in gathering information about po- WHQWLDODOWHUQDWLYHVWRWKHEHQH¿WVIRUIHLWHGGXH to termination of the existing service. Patterson DQG6PLWKGH¿QHGVZLWFKLQJFRVWDVWKH perception of the magnitude of the additional cost incurred to terminate a relationship and to secure DQDOWHUQDWLYHRQH6HOQHVGH¿QHGVZLWFKLQJ FRVWDVWKHWHFK Q LFDO¿QDQFLDODQGSV\FKRORJLFDO IDFWRUVWKDWPDNHLWGLI¿FXOWRUH[SHQVLYHIRUD customer to change brands. &RUSRUDWHLPDJHLVGH¿QHGDVWKHRYHUDOOLP- pression about a company formed on the minds of the public (Barich & Kotler, 1991; Dichter, 1985; Kotler, 1982). It relates to the different physical and behavioral attributes of a company, such as business name, logo, corporate values, tradition, ideology, and the impression of quality communi- cated by a customer to a potential customer (i.e., word of mouth). The building of corporate image is a lengthy process. The sensory process starts with ideas, feelings, and previous experience with a company that are retrieved from memory and transformed into a mental image (Yuille & Catch- pole, 1977). Past studies have suggested that a host of factors, including advertising, public relations, physical image, word of mouth, and customer’s actual experience with the goods and services, LQÀXHQFHWKHFRUSRUDWHLPDJHRIDFRPSDQ\LQ the mind of a customer (Normann, 1991). Researchers (e.g., Slater, 1997) and consul- tants (e.g., Gale, 1994) have recommended that companies should adjust their strategies to retain customers in order to achieve superior customer value delivery as customer value incorporates both WKHFRVWVDQGEHQH¿WVRIVWD\LQJZLWKDFRPSDQ\ As such, customers’ perceived value is considered as a strong driver of customer retention. Neverthe- less, some important questions about the role of price in services have remained unanswered. One is whether price perception has a direct effect on overall customer loyalty. If so, it is essential for companies to actively manage their customers’ price perceptions because of their impact on value perceptions. Another question is about the forma- 1543 The Driving Forces of Customer Loyalty tion of price perception in services. Answers to these questions can help clarify the measurement and management of price perception. Conceptual Model and Hypotheses We propose a conceptual model that theorizes the relationships among consumer loyalty, service quality, customer satisfaction, switching cost, and corporate image as shown in Figure 1. In what follows, we justify the postulated relationships in the model and formulate several hypotheses to test the model. Service Quality and Customer Satisfaction Service quality researchers refer to satisfaction DVDWUDQVDFWLRQVSHFL¿FHYDOXDWLRQDQGWRTXDO- ity as an overall evaluation based on a whole set of cumulative evaluations. Parasuraman et al. (1994) recommended examining service quality and satisfaction, and their causal link, from both WUDQVDFWLRQVSHFL¿FDQGJOREDOSHUVSHFWLYHV,QWKH context of the ISP business, which mainly hinges on the ongoing relationship between a customer DQGWKHLUVHUYLFHSURYLGHUWKHFXPXODWLYHVSHFL¿F perspective is more suitable to view this ongo- ing relationship. Moreover, service quality is usually considered as an antecedent of customer satisfaction in the ISP business. Therefore, we hypothesize that H1: Perceived service quality is positively related to customer satisfaction. Customer Satisfaction and Customer Loyalty The marketing literature suggests that customer OR\DOW\ FDQ EH GH¿QHG LQ WZR GLVWLQFW ZD\V QDPHO\WKH³EHKDYLRUDODSSURDFK´DQGWKH³DW- titude approach” (Jacoby & Kyner, 1973). From the behavioral perspective, customer loyalty is LGHQWL¿HGDVWKHDFWXDOUHSXUFKDVHEHKDYLRURID customer (Cunningham, 1961). In contrast, the Corporate Image Service Quality Switching Cost Customer Loyalty Customer Satisfaction H1 H4 H2 H3 H5 H6 Price Perception H7 Corporate Image Service Quality Switching Cost Customer Loyalty Customer Satisfaction H1 H4 H2 H3 H5 H6 Price Perception H7 Figure 1. Theoretical framework . (2001). Eco- nomics and electronic commerce: A survey and directions for research. International Journal of Electronic Commerce, 5(4), 5. Marcoux, A. (2003). Snipers, stalkers, and nib- blers:. number of nibblers, and the proportion of bids that are nibbles. When the results contained in Tables 3 and 4 are taken together, our results extend and clarify Ockenfels and Roth’s (2002). hypotheses 1 and 2, while Tables DQGFRQWDLQWKHFRUUHODWLRQFRHI¿FLHQWDQDOVLV used to address hypotheses 3 and 4. The results in Table 2 provide clear evidence that the graded coin and automobile