Personalized Avatar, A New Way to Improve Communication and E-Service.

Abstract The literature has demonstrated that sorting through the mass of information available on the Internet is time consuming and sometimes imparts misinformation. In this context, the tendency to customize the content of Web pages by interactive tools will soon become a standard in the development of retailer websites. A new form of customization of the information is emerging in the apparel industry: an avatar (i.e. virtual model) that reflects the consumers body. The goal of this research is to determine the influence of a personalized avatar on perceived website quality, attitude and stickiness. Our findings show that it has a significant influence on stickiness and on perceived quality. Moreover, in some cases the degree of this influence depends on the level of assortment proposed.

Keywords: avatar; virtual model; websites; attitude; quality; stickiness; assortment. Track: Advertising, Promotion and Marketing Communication.

Introduction A study revealed that 39% of consumers confirm that the Internet influenced their purchasing decision (CEFRIO, 2006). In addition, US online sales for the year 2007 are expected to reach $131.3 billion, a 20.8% increase compared to the previous year. Sales of apparel, footwear and accessories should reach $22.1 billion in 2007, a 21% increase over 2006 (E-marketer, 2007). Today, US shoppers spend more on online shopping for apparel than for computers, allowing that market to become the largest online retail sales category (excluding travel). Consequently, many retailers try to claim their share of the market by enhancing their retailer website to give their online consumers the best experience possible and to increase the conversion rate (Childers et al., 2001). Thorson and Duffy (2005) state that people have four basic communication needs: information, entertainment, connection with others and buying goods and services. More and more of these needs are being met online. Organizations seek to provide quality and to satisfy customers needs in general and expectations about the quality of the communication in particular. In this context, the tendency to customize the content of Web pages by interactive tools will soon become a standard in the development of retailer websites. 30% of online retailers already offer this type of service to their customers. A new form of customization of the information is emerging in the apparel industry: an avatar that reflects the consumers body. Thus, users can input information regarding measurement, body shape, hair style, etc and create a 2D or 3D virtual model to visualize and try on clothing on a body similar to their own (Calhoun, LymanClarke, & Ashdown, 2007). We refer to this as a personalized avatar. In a broad sense, using this type of avatar is among the virtual tools that allow websites to improve their communication with consumers. However, to the best of our knowledge, no article has focused on the influence of the customization of the information involving the use of a personalized avatar on the perception of the website. The goal of this research is to determine the influence of customization of information by means of a personalized avatar on perceived quality, attitude and stickiness of a website. While the issue of information level and overload is important on the Internet (Spencer, 2001), we will examine our results according to the level of assortment provided on the website. We conducted a 2 x 3 between-subject experiment, by manipulating customization of the information with the use of a personalized avatar and level of assortment provided on the website. This paper also contributes to a better understanding of the impact of customization of the information through the use of a personalized avatar on the perceptions of a website. Literature Review Avatar and e-customization Customization is an advanced form of personalization which is done by the customer (Wind & Ragaswamy, 2001; Cner, 2003). In an interactive environment where the quantity of information is overwhelming, it becomes particularly interesting to customize the contents of Web pages. One of the objectives is to limit exposition to too much information and to offer them

information that precisely meets their needs (Peppers & Rogers, 1993). In this sense, the use of an avatar can be seen as a form of customization of information. Its use enables retailers to provide information that better fits the consumers preferences, but also to humanize the relationship with the website (Holzwarth, Janiszewski and Neumann, 2006; Trongemann, 2003), facilitate communication with users of the website (Salem & Earle, 2000) and increase user confidence (Qiu & Benbasat, 2005). The type of avatars we focus on is personalized avatars. This technology goes on step further by offering the users the possibility of creating a realistic representation of themselves, and to mix and match products on their own virtual body. Website quality Some consumers have a strong need for social interaction in their shopping experience (Swaminathan et al., 1999). The literature reveals that offering entertainment content to consumers should have a positive impact on the perception of the retailers product and the company (Brown & Stayman, 1992; MacKenzie & Lutz, 1989; MacKenzie, Lutz, & Belch, 1986). Holzwarth, Janiszewski and Neumman (2006) supported that the use of an avatar leads to more satisfaction related to retailers websites and greater entertainment value. Hence, avatars can be used to increase the shopping pleasure and improve the aesthetics of web design. Therefore, we hypothesize the following: H1: Use of a personalized avatar contributes to increase the perceived design quality of a website. In addition, customization makes it possible for organizations to improve the quality of information and hence present products that correspond to the consumers specific needs (Peppers, Rogers, & Dorf, 1999; Robert, 2003). Hart (2002) described this notion of relevance on the Web as the ability of a site to present information that meets the user's needs immediately. Therefore, we posit that the use of a personalized avatar will result in a higher perception of the quality of information, leading to the following hypothesis: H2: Use of a personalized avatar contributes to increase the perceived quality of information of a website. Attitude toward the website Many researchers state that one of the key advantages of customization over the Web is the development of customer loyalty (Peppers et al., 1999; Rowley & Slack, 2001). The literature has suggested that the main value of the customization of a website is the positive impact on website attitude (Stevenson et al., 2000). In consequence, hypothesis H3 can be stated. H3: Use of a personalized avatar contributes to increase consumer attitude toward the website. Stickiness To measure the performance of websites, many specialists evoke the term stickiness. We define this concept as the retention of consumers on the website or the ability to attract and hold a visitors interest (Bhat, Bevans, & Sengupta, 2002). Stickiness can be measured by the average

time a user spends on a site, the number of pages visited and the frequency of his or her visits to the site (Guenther, 2004). The more attractive the site is, the more time people will spend there; and hence the more the opportunity for companies to deliver messages to users (Bhat, Bevans, & Sengupta) and to encourage consumers to buy products. Senecal et al. (2005) have shown that the utilization of an online recommendation agent modifies the purchasing process and increases the consumers time spent and number of visited pages. In the same way, we argue that customizing a website with the help of a personalized avatar will have a positive impact on navigation time and the number of visited pages, leading to the following hypotheses. H4: Use of a personalized avatar contributes to increase the total navigation time. H5: Use of a personalized avatar contributes to increase the total number of visited pages. Assortment Offering a wide variety of products or items is a frequent strategy. The Internet provides an interesting platform for every firm for whom assortments and variety are part of the business. Grewal et al. (1999) recalled that assortments are a strategic positioning tool to attract and retain core customers. Oppewall and Koelemeijer (2005) affirm that larger assortments are rated more positively. It is also important to let consumers choose from a large assortment, rather than restricting them to a smaller one. However, this strategy can certainly have a dark side. Large assortments and complexity are likely to increase consumers dissatisfaction and frustration (Jacoby, Speller, & Berning, 1974). Several researchers (Jacoby, Speller, & Kohn, 1974; Keller & Staelin 1987; Malhotra, 1982) found evidence that increasing the number of alternatives or attributes in a choice set leads to a decline in the quality of consumers choices. The use of a personalized avatar can be seen as a way to help the consumer in a large assortment setting. By trying the products on their own body, consumers could make better choices. We also argue that using a personalized avatar can result in more positive attitude and higher stickiness when the level of assortment increases. In consequence, we can state hypotheses H6, H7 and H8. H6: When the level of assortment offered on a website increases, use of a personalized avatar contributes to increase the total navigation time (interaction effect). H7: When the level of assortment offered on a website increases, use of a personalized avatar contributes to increase the total number of visited pages (interaction effect). H8: When the level of assortment offered on a website increases, use of a personalized avatar contributes to increase consumers attitude toward the website (interaction effect). Methodology Our study used an experimental website constructed in collaboration with Home Shopping Network (HSN.com) and My Virtual Model (MVM.com). HSN is an online retailer, while MVM technology can be used to create 3D personalized avatars by manipulating twelve bits of information like measurements, weight, hair colour (Nantel, 2004). Sampling and Procedure Participants were recruited through a mass e-mail sent to 45,000 individuals. A drawing for a $500 prize served as an incentive. As virtual models primarily target women, we retained only 4

female subjects for the analysis. Of the 45,000 e-mails that were originally sent, 400 female subjects agreed to participate (84% of them were aged between 19 and 54 years old). The study design involved a 2 (use of a personalized avatar: personalized avatar proposed or not) x 3 (level of assortment: 10, 20 or 30 blouses) between-subject experimental design. Each participant was randomly assigned to one of the six experimental conditions and was exposed to the fictitious website. We asked each of them to buy a blouse, with no time constraint. We offered to 193 out of 400 participants the option of using the virtual model. Of the 193 subjects, 56 did not use it and were eliminated. Finally, 344 participants are included in the analysis (see Appendix). After the task completion, subjects were directed to the questionnaire site. Then, they were debriefed and thanked. Measures Perceived design and information quality were measured on a 7-point scale developed by Bressolles and Nantel (2005). The mean of the three items was used (Cronbachs alpha = 0.92 for layout dimension and 0.90 for information). Attitude toward the website (Awebsite) was measured on a 5-point scale (Chen & Wells, 1999). The mean of the six items was used (Cronbachs alpha = 0.92). Finally, stickiness was measured according to experimental navigation data: task duration and number of visited pages. Results Analysis shows that participants that used the virtual model had a better opinion of the design quality of the site than those that did not use it (F (1, 338) = 3.320, p < .05). H1 was thus supported. Furthermore, participants who used the virtual model concluded that it provides indepth information about the product (F (1, 338) = 3.039, p< .1). H2 was thus marginally supported. Concerning stickiness to the website, our findings revealed that subjects who used a personalized avatar took more time to complete the task (F (1, 338) = 4.375, p < .05) and visited more pages (F (1, 338) = 23.627, p< .000). H4 and H5 were thus supported. Moreover, the level of assortment presented has an impact on both of these dependant variables. Using the virtual model results in a longer task duration only when 30 blouses are offered. It is worth noting that subjects that chose from among 30 blouses used the virtual model 2.71 times compared with 1.58 times for those that had a choice of 20 blouses (F (1, 338) = 7.076, p<.05). This increased use explains the variation in the task duration and showed that the level of assortment acted as a moderating factor. Therefore, the results support H7. In addition, there was a significant interaction between the number of blouses and the use of the virtual model (F (2, 338) = 3.277, p<.05). Consequently, H6 was supported. Use of a personalized avatar had a mixed effect on the attitude toward the website. The results point in two directions. First, use of the virtual model does not influence Awebsite (F (1, 338) = 1.617, p> .05). H3 was not supported. However, an interaction effect existed between the number of blouses and the use of the virtual model (F (2, 338) = 4.550, p<.05), which demonstrated that as the number of alternatives increases, the impact of the use of a personalized avatar on the attitude varies. However, this effect (interaction) seems to obtain for

a certain amount of information only. A t test showed that use of a personalized avatar has a significant influence on attitude toward the website only for 20 blouses (t = -3.106, p<.001). H8 was supported. Discussion, limitations and future researches Our research examined the impact of an innovative way to shape the perceptions of a website: the use of a personalized avatar created by the consumer. Based on an experimental study, we identified the impact of using this kind of avatar on the perceived quality, attitude and stickiness of a website. Although research has been conducted on avatars and customization, we conducted, to the best of our knowledge, the first study concerning the merits of using an avatar related to the consumers self-image to enhance communication and consumers' experience on a website. Our findings suggest that the use of a personalized avatar related to the consumers self-image influences users perception of the design of the website. Norman (2004) asserts that an interface that is more aesthetically pleasing will leave customers with a better impression and can increase consumer retention and loyalty. Consequently, using this kind of personalized avatar could be an asset for a website. In addition, we show that consumers using a personalized avatar have a more positive perception of the information generated by the website. This result is particularly interesting because we know that one of the main reasons for consumers to shop online is to obtain in-depth information on the product (Li, Kuo, & Russell, 1999; Swaminathan, LepkowskaWhite, & Rao, 1999). Other objectives of this study were to explore the impact of customization on website stickiness and attitude. Following the manipulation of the level of assortment, we observed several significant variations in the behavior and attitudes of the respondents. Customizing the information presented on a website by a personalized avatar contributed to increasing total navigation time as well as the number of pages visited. In this study, it is possible that respondents that used a personalized avatar spent more time on the site because of the enjoyment of trying on the clothing offered, whereas those that did not have access to the tool did not put forth the same effort in analyzing all the blouses. Thus, using an interactive system as a personalized avatar could lead to a state of mind where consumers lose track of time (i.e. flow state) (Belleman et al., 2004). However, this influence is significant only for the higher assortment condition (30 blouses). The use of a virtual model is conceivably a relevant service that influences stickiness only when the level of assortment is high. If assortment is lower, consumers may not need such a service. Nevertheless, given that websites tend to increase their level of assortment, personalized avatars could be a very worthwhile tool for retailers. We have also observed that when the amount of information presented on a website increases, customizing this information helps and improves the users attitude. Indeed, we showed that when consumers are exposed to a certain amount of information (i.e. 20 blouses) they develop a form of gratitude toward businesses that try to facilitate information processing through the use of an interactive tool. The scope of this observation is especially important as we assume that a positive attitude to a transactional site can have a definite influence on the desire to return and consequently on purchasing intentions. Nonetheless, these results suggest that up to a certain level of information (30 blouses), the benefits of using a personalized avatar on the attitude toward the website is no longer significant. It is probable that at this stage, other factors contribute to this result.

This article possesses several methodological limitations. First, the sample employed for this study consists of consumers that had already used the My Virtual Model website. Secondly, it would have been useful to examine more than three levels of assortment, notably a higher level of assortment, i.e. more than 30 blouses. Finally, this research studied the influence of the use of a personalized avatar on perceptions of the website. A further step will be to take into account several other variables related to the experience on a website like experiential value satisfaction. It would also be interesting to conduct this study with men and adolescents.

References Ansari, A & Mela, C.F. (2003). E-customization. Journal of Marketing Research, 40 (2), 131145. Belleman, S., Johnson E.J., Lohse, G. L., & Mandel, N. (2004). Designing marketplaces of the artificial : Four approaches to understanding consumer behavior in electronic environments. Conference: Marketing and Interne, MIT, June 21st. Bhat, S., Bevans, M., & Sengupta, S. (2002). Measuring usersweb activity to evaluate and enhance advertising effectiveness. Journal of Advertising, 31 (3), 97-106. Bressolles, G., & Nantel. J. (2005). Toward a typology of web sites for consumers. Revue Franaise du Marketing, 213, 41-56. Brown, S., & Stayman, D. (1992). Antecedent and consequences of attitude toward the ad: A Meta analysis. Journal of Consumer Research, 19 (6), 4-51. Calhoun, E.L., Lyman-Clarke, L., & Ashdown, S. (2007). Virtual Fit of Apparel on the Internet: Current Technology and Future Needs. The 2007 World Conference on Mass Customization & Personalization, MIT Cambridge/Boston, October. CEFRIO (2006). NETendances. http://www.infometre.cefrio.qc.ca/ Chen, Q., & Wells W. D. (1999). Attitude toward the site. Journal of Advertising Research, 39 (5), 27-38. Childers, T.L., Carr C.L., Peckc, J, & Carson, S. (2001). Hedonic and utilitarian motivations for online retailing. Journal of Retailing, 77 (4), 511-536. Coner, A. (2003). Personalization and customization in financial portals. Journal of American Academy of Business, 2 (2), 498-504. Grau, J. (2007). Apparel E-Commerce. E-Marketer, August, 1-17. Grewal, D., Levy, M., Mehrotra, A., & Sharma, A. (1999). Planning merchandising decisions to account for regional and product assortment differences. Journal of Retailing, 75(3), 405-424. Guenther, K. (2004). Pull up a chair and stay awhile: Strategies to maximize site stickiness. Online, 28 (6), 55-57. Hart, G. (2002). Nice web site-But is it relevant? Computer World Canada, 18 (1). Hallerman, D. (2006). Search Marketing: Spending and Metrics. E-Marketer, 1-28. 8

Holzwarth, M., Janiszewski, C., & Neumann, M.M (2006). The influence of avatars on online consumer shopping behavior. Journal of Marketing, 70 (4), 19-36. Jacoby, J., Speller, D., & Berning, C. (1974). Brand choice behavior as a function of information load: Replication and extension. Journal of Consumer Research, 1 (1), 33-42. Jacoby, J., Speller, D., & Berning, C. (1974). Brand choice behavior as a function of information load: Replication and extension. Journal of Consumer Research, 1 (1), 33-42. Keller, K.L., & Staelin, R. (1987). Effects of quality and quantity of information on decision effectiveness. Journal of Consumer Research, 14 (2), 200-213. Li, H., Cheng, K., & Russel, M. G (1999). The impact of perceived channel utilities, shopping orientations and demographics on the consumers online buying behavior. Journal of Computer Mediated Communication. Retrieved from www.ascusc.or.cmc/voI5/issue2/hairon.html Mackenzie, S.B., & Lutz, R. (1989). An empirical examination of the structural antecedent of attitude toward the ad in advertising pretesting context. Journal of Marketing Research, 53 (11), 48-65. Mackenzie, S.B., Lutz, R., & Belch, G.E. (1989). The role of attitude toward the ad as a mediator of advertising effectiveness: A test of competing explanations. Journal of Marketing Research, 23 (05), 130-143. Malhotra, N. (1982). Information load and consumer decision making. Journal of Consumer Research, 8 (4), 419-430. Nantel, J. (2004). My virtual model: Virtual reality comes into fashion. Journal of Interactive Marketing, 18 (3), 73-86. Norman, D. (2004). Emotional design: why we love (or hate) everyday things, Basic Books: New York. Oppewal, H. & Koelemeijer, K. (2005). More choice is better: Effects of assortment size and composition on assortment evaluation. International Journal of Research in Marketing, 22, 4560. Peppers, D., Rogers, M., & Dorf, B. (1999). Is your company ready for one-to-one marketing?, Harvard Business Review, 77 (1), 151-160. Qiu, L. & Benbasat, I. (2005). Online Consumer Trust and Live Help Interfaces: The Effects of Text-to-Speech Voice and Three-Dimensional Avatars. International Journal of Human Computer Interaction, 19(1), 75-94. Salem, B., & Earle, N. (2000). Designing a nonverbal language for expressive avatars. In Proceedings of CVE2000. ACM Press, New York, 93101. 9

Senecal, S., Kalczynski, P.J & Nantel, J. (2005). Consumersdecision-making process and their online shopping behavior: A clickstream analysis. Journal of Business Research, 58 (11), 15991608. Spencer, R. (2001). Use web services wisely to avoid information overload. Employee Benefit Plan Review, 55 (11), 6-8. Rowley, J., & Slack, F. (2001). Leveraging Customer knowledge Profiling and Personalisation in E-Business. International Journal of Retail & Distribution Management, 29 (9), 401-415. Stevenson, J. S., Bruner II, G. C. & Kumar, A. (2000). Web page background and viewer attitudes. Journal of Advertising Research, 20 (1), 29-34. Swaminathan, V., Lepkowska-White, E., & Rao, P.Rao (1999). Browser or buyers in cyberspace ? An investigation of factors influencing electronic exchange. Journal of Computer-Mediated Communication, 5 (2). Available http://jcmc.indiana.edu/vol5/issue2/swaminathan.htm Thorson, E. & Duffy, M. (2005). Newspapers in the new media environment. Paper presented at the NAA Marketing Conference, Orlando. Trogemann, G. (2003). Mit Hand und Fu: Die Bedeutung der nonverbalen Kommunikation fr die Emotionalisierung von Dialogfhrungssystemen. Avatare-Digitale Sprecher fr Business und Marketing, 267290. Wind, J., & Rangaswamy, A. (2001). Customerization: the next revolution in mass customization. Journal of Interactive Marketing, 15 (1), 13-45.

10

Appendix

Cell Composition Yes 43 53 41 137 Virtual Model Use No 74 70 63 207 Total 117 123 104 344

10 blouses 20 blouses 30 blouses Total

11