Journal of Undergraduate Research
Volume 6, Issue 8 - July/August 2005

Avatars in Global E-Commerce: A Cross-Cultural Analysis of the Effects of Avatars on Online Consumer Behavior in Germany and the U.S.

Shasta Johansen

ABSTRACT

More and more firms are using virtual assistants such as Lori, the virtual skin consultant for L’Oréal (http://lorealparisusa.com/skincare/purezone/home.asp), as communication instruments on their Web sites. The effects of these so-called “avatars” on online shoppers, however, have barely been researched. This paper introduces an empirical study of the communication effects of avatars. Because the Internet is a global medium, a cross-cultural comparison of German and American users was conducted to examine cultural differences in avatar perception. Two online experiments showed that avatars had a positive effect overall on both cultures and revealed few significant differences between German and American consumer perception, which indicates that avatars can be successfully implemented globally.

INTRODUCTION

The Internet provides the opportunity for more effective and efficient target marketing at a global level. Due to the decreasing entry barriers for international markets on the Internet, competition is increasing quickly, along with the increase in the number of firms in the global marketplace (Baack and Singh 2004). Amidst such strong online marketing competition, firms must distinguish themselves from other suppliers by creating extra value through a positive online consumer experience, in addition to catering to normal consumer needs. The use of avatars on commercial Web sites has, therefore, become a popular marketing strategy to answer this growing need for extra value. Avatars, or virtual assistants such as the well-known Clippit the Paperclip from Microsoft Office or Cybelle from http://www.agentland.com, can help users navigate a program or Web site, often providing personalized information to facilitate a more meaningful and satisfying experience. Because the Internet is an international communication medium with no geographic boundaries, it is critical for firms using avatars on their Web sites to be aware of the preferences that exist among their customers. Evidence of such cultural differences between online consumers will prove beneficial to global firms by providing new insight into the use of avatars as a practical communication instrument to build trust and boost user satisfaction.

Currently, over two-thirds of Americans and half of the total German population over the age of 14 use the Internet (Eimeren et al. 2003). Germany has experienced particularly dramatic growth in Internet use in recent years, as shown by the fact that between 2000 and 2003, the number of Internet users doubled in Germany (Eimeren et al. 2003) and the domain name “.de” (for Deutschland) now ranks second in use after “.com” (Berstein 2004). Because of the strong relevance of both countries in the Internet economy, it is important for global firms to understand the characteristics and needs of both the American and German consumer cultures.

Although research about cultural differences has increased over the past years, very little research has been conducted in general about cultural differences in Web site perception, and in particular about avatar perception between culturally different online shoppers. The following research examines the underlying cultural differences in the online consumer behavior of Germans and Americans, and it offers insight into the effectiveness of avatars on commercial Web sites for global firms. Because attractiveness and credibility are the most relevant attributes analyzed in empirical studies of communicators (Ohanian 1991), these two attributes were used to test for avatar effectiveness. For example, are attractive avatars on Web sites more effective than credible avatars? Furthermore, do German and American online shoppers perceive attractive avatars and credible avatars differently? Knowledge of such culture-related perception differences in online consumer behavior can provide global firms with a competitive edge to create avatars most suitable for their target culture.

BACKGROUND

The Wharton Virtual Test Market conducted by the Wharton e-Business Initiative, using participants from the U.S., Europe, and Asia, recently found that the most important predictor of online buying behavior was product information search (Chau et al 2002). It is, therefore, advisable for firms to insure effective Web site navigation to aid their customers in their information search online.

Consequently, in using avatars as virtual customer service representatives to help consumers navigate commercial Web sites and search for product information, firms must take into consideration the cultural differences of the consumers viewing the site. Barber and Barde (1998) claim that, “As different ethnic cultures display different cognitive reactions, they may react differently to different environmental stimuli, such as sound and video clips, intelligent shopping assistants, 3D animation, and so on. What is user friendly in one culture may be different for another group of consumers in another culture.” For example, in countries like the U.S where time is considered precious, heavy graphics with long download times are seen as a nuisance by consumers, whereas many Asian cultures prefer screens heavy with animated graphics, video, and audio.

Chau et al. (2002) claim that in order to succeed, global Web sites must have an appropriate consumer interface, which is defined as “the Web environment in which consumers interact to conduct transactional exchanges” (p. 139). Although a global consumer interface may need a certain degree of internationalization or standardization to maintain a global image, Barber and Barde (1998) argue that a global interface can only be effective if it is localized, or designed to reflect the cultural nuances of the target audience. They also emphasize the importance of using “cultural markers” on Web sites, such as the native language, national symbol, or color. For example, although Americans may perceive white as a sign of purity, in Japan it represents death. Similarly, the Chinese may associate a red background on a Web page with happiness, whereas Americans could perceive the color as danger. A well-designed consumer interface should therefore be localized in order to support the way consumers shop and transact online (Chau et al. 2002).

CULTURAL THEORY AND DERIVATION OF HYPOTHESES

Hofstede (1998) defines culture as “the collective programming of the mind which distinguishes the members of one group of people from another” (p. 8). Culture can be divided into four categories: symbols, heroes, rituals, and values. These categories can be thought of “as the skins of an onion, with symbols representing the outermost, superficial layer and values representing the deepest manifestations of a culture, with heroes and rituals in between” (p. 8). Two large research projects conducted by Hofstede showed that national cultures differ mostly at the level of values, the deepest level. His work has also showed that there are four main variables of culture distinction: power distance, uncertainty avoidance, individualism, and masculinity.

In Hofstede’s findings, differences between Germans and Americans were most pronounced in uncertainty avoidance, for which out of 100 possible points, Germany scored 65 and the U.S. scored 46 (Hofstede 1998). Uncertainty avoidance deals with “tolerance for uncertainty and ambiguity. It indicates to what extent a culture programs its members to feel either uncomfortable or comfortable in unstructured situations” (Hofstede 1998). Germany’s particularly high ranking in uncertainty avoidance indicates that Germans like to feel secure and may not be very comfortable in unstructured online situations. In the context of the present research study, this suggests that using credible avatars as navigation tools for Germans on Web sites could be more effective than using attractive avatars, because Germans might associate credibility of the virtual assistant with reassurance of the credibility of the company, Web site, and product being offered. Furthermore, cultural anthropologist Hall contends that the decision-making process for Germans is much more involved and deliberate than for Americans, requiring extensive background research (Hall 1990). Germans might, therefore, feel more secure and at ease if they have a credible, experienced avatar to assist them online. Because the findings of both Hofstede and Hall indicate that using credible avatars could create a particularly meaningful experience for German consumers, the following hypothesis was derived:

H1: The use of a credible avatar has a significantly more positive effect on Germans’ Web site perceptions than on Americans’ Web site perceptions.

Americans, however, are stereotyped in cultural research as being very image conscious. Because the average American is exposed to at least three thousand advertisements every day (Kilbourne 1999), this focus on image is especially evident in American advertising and marketing. Hall’s work emphasizes that Americans like idealized images in advertisements. Women are usually displayed as young, healthy, and beautiful, whereas men are shown to be young, strong, and handsome (Hall 1990). Because this information points to the possibility that attractive avatars could be particularly effective as virtual assistants for Americans, the following hypothesis was derived:

H2: The use of an attractive avatar has a significantly more positive effect on Americans’ Web site perceptions than on Germans’ Web site perceptions.

METHODS

The experiment presented here was a replication of a German experiment that had already been conducted with German participants. This pre-existing German experimental Web site and questionnaire was translated from German into English and then run at the University of Florida with MAR 3023 students. In the American study (N=109), 51% of the participants were female and 49% were male. Participant ages ranged from 18-50 years, and the average age was 25 years. In the German study (N=264), 47% of the participants were female and 53% were male. Participant ages ranged from 18-40 years, and the average age was 23 years.

The hypotheses in the current study were tested by comparing the American data to the existing German data in order to look for cross-cultural differences in the effects of avatars on German and American perceptions of the following eight dependent variables: Web site entertainment value, Web site information value, Web site credibility, Web site graphic design, Web site conceptual design, user satisfaction, attitude towards the product, and buying intention. This study therefore sought not only to examine the effect of avatars on the users’ perceptions but also to explore the differences observed specifically between Germans and Americans exposed to attractive and credible avatars.

Overview of Experiment

In each experiment, participants completed three tasks. First, participants answered a short survey about their general use of the Internet. In the second step, participants viewed the Web site of a fictional company offering online customizable shoes and were guided by an unanimated avatar, which acted as a virtual salesperson and helped them design a pair of shoes. Each participant was exposed to either an attractive avatar, a credible avatar, or no avatar. The avatar that a participant received briefly introduced itself with a text dialogue and then led the participant through the shoe configuration process, providing product information and customization tips at each step. Participants in the control group simply received the same text dialogue information but without a visible avatar. After designing their own pair of shoes, for the third and final step participants completed a follow-up questionnaire to capture their reactions to the Web site and avatar guide. Because participants were unaware of the underlying avatar treatment testing, the responses from the credible group could be compared to those from the attractive group, as well as the control group, to test for differences in avatar effectiveness.

Avatar Design

As mentioned above, each participant received either an attractive avatar, a credible avatar, or no avatar. To avoid gender bias, male and female versions of the attractive and credible avatars were assigned. Participants were therefore randomly exposed to one of the following five possible avatar treatments: an attractive female avatar, an attractive male avatar, a credible female avatar, a credible male avatar, or no avatar. The features of the avatars were manipulated to make the attractive avatar group (Kim and Tom) look younger and more physically attractive, whereas the credible avatar group (Dr. Anne Schneider and Dr. Norbert Oswald) was designed to look older and more credible. Figures 1-3 show the possible avatar treatments.

Figure 1. Attractive avatars (Kim and Tom)

Figure 2. Credible avatars (Dr. Anne Schneider and Dr. Norbert Oswald)

Figure 3. No avatar (control group)

Questionnaire Design

The follow-up survey questions addressed various aspects of the participant’s attitude toward the dependent variables tested. All of the questions were rated on a 7-point Likert scale, with 1 meaning “completely disagree” and 7 meaning “completely agree.” To demonstrate this design, Figure 4 shows questions related to user satisfaction with the company.

Figure 4. Example of questionnaire design

The survey questions addressed various aspects of the participant’s attitude toward the dependent variables tested. The eight dependent variables used for the analysis were Web site entertainment value, Web site information value, Web site credibility, Web site graphic design, Web site conceptual design, user satisfaction, attitude towards the product, and buying intention.

RESULTS AND DISCUSSION

The American data was compared to the German data using SPSS data analysis software. A univariate analysis of variance was run to test the relationship between the experimental factor treatment with three factor levels (attractive avatar, credible avatar, no avatar) and each of the eight dependent variables.

As previously stated, the first hypothesis was that the use of a credible avatar has a significantly more positive effect on Germans’ Web site perceptions than on Americans’ Web site perceptions. To test this hypothesis, for each of the eight dependent variables, the data from Germans who received a credible avatar was compared to the data from Americans who received a credible avatar. Table 1 displays the means and standard deviations of the results from each of the eight dependent variables tested using credible avatars. The conceptual design of the Web site, however, was the only dependent variable that yielded a highly significant result (p<.001 at _<.05). Therefore, Germans who received a credible avatar generally had a much more positive impression of the Web site’s conceptual design than did Americans who received a credible avatar. This is a particularly valuable finding because the conceptual design is an indicator of the ease of navigation of the Web site. Consequently, global firms catering to a German consumer audience could find it useful to use credible avatars to improve the navigation of their Web sites.

The second hypothesis was that the use of an attractive avatar has a significantly more positive effect on Americans’ Web site perceptions than on Germans’ Web site perceptions. To test this hypothesis, for each of the eight dependent variables, the data from Americans who received an attractive avatar was compared to the data from Germans who received an attractive avatar. The means and standard deviations of the results from each of the eight dependent variables tested using attractive avatars are displayed in Table 1. These analyses did not yield any significant results, although it is interesting to note that across all three treatments (attractive avatar, credible avatar, and no avatar), the information value of the Web site received higher ratings from Americans than Germans. This means that on average, Americans perceived the Web site as more informative than did Germans, although not statistically significantly.

This result supports the unanimous view of the cross-cultural literature that Germans require detailed information and extensive background research before making decisions. Hall (2002) repeatedly contends that “Germans require much more information than most from other cultures” (p. 35) and that “Germans demand facts, facts, and more facts” (p. 50). Accordingly, taking this cultural characteristic of Germans into consideration could explain why Germans were not as easy to please with the information content of the Web site in the experiment. It could therefore be crucial for firms to be aware of this requirement of extensive information when targeting their Web sites at German consumers.

Although most of the results were not of statistical significance, the majority of the analyses showed a positive trend in consumer attitudes influenced by the use of avatars. The graphic design of the Web site, for example, was perceived almost exactly the same by both Germans and Americans. Such identical responses show that the avatars were comparably effective for both groups.

The lack of significant results from this experiment could possibly be attributed to the weak manipulation of the avatars’ physical features. In order to obtain a more accurate view of the cultural differences between perceptions of avatars, more research should be conducted using avatars with stronger feature manipulations that are therefore easily distinguishable as either attractive or credible. The similarity in the German and American data results, however, could be an indication that avatars have positive effects in both cultures and therefore can be successfully implemented for numerous target cultures globally.
CONCLUSION
Although the ways that avatars affect German and American consumers remain to be adequately tested, it is nevertheless of obvious importance for global firms to take the cultural differences of their online audiences into consideration when designing a Web site. Avatars are becoming ubiquitous on the Web, and therefore more research should be conducted to explore the effects of using avatars as virtual assistants on commercial Web sites with culturally different consumers. With deeper insights into avatar perception in various cultures, firms will have powerful knowledge to go beyond borders in reaching their target cultures and maximizing customer satisfaction worldwide.

---

REFERENCES

1. Baack, D. & Singh, N. (2004), “Web Site Adaptation: A Cross-Cultural Comparison of U.S. and Mexican Web Sites,” in Journal of Computer-Mediated Communication, Vol. 6. No. 1.
2. Barber, W. & Badre, A. (1998), “Culturability: the merging of culture and usability,” in Human Factors and the Web. www.research.att.com/conf/hfweb/proceedings/barder/index.htm
3. Bernstein, E. (2004), Culture and Customs of Germany, West Port, CT: Greenwood Press, p. 123.
4. Chau, P., Cole, M., Massey, A., Montoya-Weiss, M., & O’Keefe, R. (2002), “Cultural differences in the online behavior of consumers,” in Journal of the Association for Computing Machinery, Vol. 45, No. 10.
5. Eimeren, B. van, Gerhard, H., & Frees, B. (2003), ARD/ZDF-Online Study 2003, http://www.daserste.de/service/ardon103.pdf
6. Hall, E. & Hall, M. (1990), Understanding Cultural Differences, Yarmouth, MA: Intercultural Press, p. 35, 50.
7. Hofstede, G. (1998), “Cultural Constraints in Personnel Management, ” in Journal of International Business, Special Issue 2/98, Management International Review, p. 8-9.
8. Internet World Usage Statistics. Retrieved December, 2004, from http://www.internetworldstats.com/stats2.htm
9. Kilbourne, J. (1999), Deadly Persuasion, New York: Free Press.
10. Ohanian, R. (1991), “The impact of celebrity spokespersons’ perceived image on consumers’ intention of purchase,” in Journal of Advertising Research, Vol. 31, No. 1, p. 46-54.
    

---

--top--

Back to the Journal of Undergraduate Research