A comparison of consumers’ brand attitude and brand image between virtual worlds and websites - differences in high and low brand familiarity levels

Abstract Numerous companies have engaged in managing their brands in virtual worlds. This study intends to explore whether the strategy of utilizing virtual worlds has a significantly larger effect on consumers’ brand attitude and brand image than the traditional brand promotion strategy of using an official website. An experiment was conducted by randomly assigning 80 voluntary participants to one of four scenarios for a comparative analysis followed by a content analysis and observation of the companies’ official websites and virtual business locations. The experimental results showed significant differences between participants’ brand attitudes towards companies with which they have different levels of familiarity before the study. Through a questionnaire, this study found that virtual world approaches are more effective than official websites in enhancing the participants’ brand attitude toward an unfamiliar brand. In addition, their perceived experiential and symbolic brand images are significantly different between the two brand promotional strategies (virtual world > website) for an unfamiliar brand. These results imply that when consumers are unfamiliar with a brand, perceived brand image can serve as an indicator of any potential increases in their recognition of the brand, especially when there are more experiential or symbolic objects in the virtual business locations. The results of content analysis and observations through the affordance view also showed that vivid and interactive features can enhance users’ overall satisfaction, provide them with new experiences, and update their brand knowledge when viewing the companies’ websites or virtual worlds. More findings and discussions were provided in this study.

toward an unfamiliar brand.In addition, their perceived experiential and symbolic brand images are significantly different between the two brand promotional strategies (virtual world > website) for an unfamiliar brand.These results imply that when consumers are unfamiliar with a brand, perceived brand image can serve as an indicator of any potential increases in their recognition of the brand, especially when there are more experiential or symbolic objects in the virtual business locations.The results of content analysis and observations through the affordance view also showed that vivid and interactive features can enhance users' overall satisfaction, provide them with new experiences, and update their brand knowledge when viewing the companies' websites or virtual worlds.More findings and discussions were provided in this study.

Introduction
Nowadays, companies most often use the Internet in their marketing.The companies own official websites where they can promote their brands by conducting various marketing activities, such as showcasing new inventions on an announcement page, displaying products with images or photos, and participating in social welfare events.With the official websites, consumers are able to receive product brand messages while purchasing products online (Grudzewski et al., 2018;Holzwarth et al., 2006).In addition, numerous types of three-dimensional virtual worlds have been used to help companies conduct various marketing activities (Loureiro et al., 2019;Messinger et al., 2009;Srivastava & Chandra, 2018;Wedel et al., 2020).For example, some of them have entered Linden Lab's Second Life (SL) to create virtual business locations or "Islands" and to hold online meetings or training in SL.On these islands, the companies can build virtual buildings, provide product information, and offer technical support to their consumers.In brief, establishing virtual business locations is an alternative strategy for marketing and better engaging consumers and changing their perceptions of the company's brand recognition (Nah et al., 2011;S. R. Park et al., 2008;Yoo et al., 2015).
In prior studies, brand management has been examined to understand its impact on consumer brand recognition (Keller, 1993;Verhellen et al., 2013).Brand recognition can be regarded as the status of a company's brand equity (Jansen et al., 2011;Keller, 1993Keller, , 2010;;Nah et al., 2011).Consumers easily generate different brand associations which can further influence their brand recognition, including brand attitude and brand image, due to companies' different brand management approaches (Belén del Río et al., 2001;Keller, 1993;Ng & Houston, 2008;Park et al., 1986).As the companies provide consumers with good shopping experiences, their brand recognition can be increased (Baker et al., 2019;Kocak & Ruzgar, 2017).However, it is uncertain if the consumers' brand recognition would differ depending on the approaches that companies adopt to manage their brands, especially when the consumers have different levels of familiarity with the company brands.It is neither unknown how well these brand management approaches perform from the design perspective.
Many companies not only own an official website but build virtual business locations to promote service ideas and carry out marketing campaigns.These companies can provide content related to product information and technical support both on their official websites and in their virtual locations.More interactive features can be provided in their virtual locations than their official websites.Well-known global companies, e.g., IBM and Dell, have created virtual objects, including virtual products, virtual offices, and virtual buildings in their virtual locations to help consumers better know the products and to increase their purchase intention, which would enable the companies to reach their marketing goals (Baker et al., 2019).Based on the literature, to further compare the differences between different brand promotion approaches in influencing consumers' brand attitude and brand image, three characteristics, i.e., control, content, and process, can be used to examine the high/low performance levels of marketing activities held by the companies on their official websites and in their virtual worlds (Part et al., 2008).Here, three research questions are raised: (1) Is there a significant difference in consumers' perceived brand attitude before and after experiencing companies' different brand management approaches (official websites vs. virtual worlds)?(2) Is there a significant difference in the three dimensions of consumerperceived brand images between the different brand management approaches?(3) How well do these approaches perform in the three design characteristics (i.e., control, content, and process), especially when the consumers' familiarity with well-known global brands is uncertain?To answer these questions, a mixed method (using more than one method for collecting different types of data from different data sources) was considered an effective approach for the research purpose and was adopted in this research (Creswell, 2003).The value of adopting a mixed method to achieve cross-validation has been confirmed in prior studies (Belén del Río et al., 2001;Grudzewski et al., 2018;Holzwarth et al., 2006;Nah et al., 2011), such as avoiding bias, further clarifying research findings, increasing the level of robustness of the research findings, and obtaining different facets of study findings for the same research topic (Creswell & Clark, 2007).
Therefore, this research started out proposing the hypotheses by assuming that the strategy of utilizing virtual worlds has a significantly larger effect on consumers' brand recognition than the traditional strategy of using an official website.In a scenario-based experiment, the consumers' level of brand familiarity was first confirmed as high or low, followed by exploring the difference in consumers' brand recognition between SL virtual business locations and official websites.Then, this study carried out the content analysis and observation of select SL virtual business locations and official websites to analyze the performance of brand management business operations in SL virtual business locations and official websites.Accordingly, the research purpose is to explore which brand management approach (official website vs. virtual worlds) has a significantly larger effect on consumers' brand recognition (including brand attitude and three dimensions of brand images) and which design characteristics affect consumers' brand recognition.Overall, this study is expected to contribute to the literature in the fields of brand management, consumer behavior, and application of virtual contents.

Brand management
A brand is the combination of consumers' psychological feelings, recognitions, and experiences about a company and its products or services (Keller, 1993(Keller, , 2010)).Brand equity is used to describe and measure consumers' perceived brand value (Christodoulides & Chernaton, 2004;Fine et al., 2016;Zheng et al., 2018).Brand value can be promoted significantly by companies as high-value product attributes are created for consumers, such as pleasing features, good quality, impressive innovative product brand design (Wu & Chen, 2019).Marketing personnel thinks carefully from the consumers' perspectives; they also need to establish brand knowledge in the consumers' minds by implementing various marketing strategies (Heuer et al., 2015;Loureiro et al., 2019).Numerous companies have entered the virtual world to hold different kinds of marketing activities other than maintained their official websites, especially during the pandemic period, to continue enhancing consumers' brand awareness or recognition and to provide them with good shopping experiences (Baker et al., 2019;Bleize & Antheunis, 2019;Keller, 1993).
Companies establish deep and vivid consumption memories, mainly to encourage consumers to recognize the company's brand and hopefully generating positive attitudes (Aaker & Jacobson, 2001;Baker et al., 2019).Consumers' brand image of a company can be positive if their memories associated with the brand are positive, and it can be expected to further influence the consumers' purchase willingness or behavior (Jansen et al., 2011;C. Park et al., 1986;Zenetti et al., 2014).In the current research, consumers' brand recognition can be identified by exploring their brand attitudes and brand images under different brand management approaches.Brand attitude refers to an individual's overall positive or negative feelings toward a brand; brand image is the image formed by the consumers' associations with their consumption experiences and memories about the brand (Aaker & Jacobson, 2001;Kocak & Ruzgar, 2017).

Three-dimensional virtual world
Virtual worlds designed in three-dimensional space have the characteristics of a simulation, which gives the users the feeling of being inside the space, allowing them to interact with others in the virtual world through an avatar (Bleize & Antheunis, 2019;Loureiro et al., 2019;Wedel et al., 2020;Zheng et al., 2018).The avatar representing the actual consumer plays an important role in marketing since it is capable of influencing the consumer's willingness to purchase real products, and it also provides an opportunity to bring him/her into potential markets (Bleize & Antheunis, 2019;Grudzewski et al., 2018;Wedel et al., 2020;Zheng et al., 2018).Numerous activities can thus be conducted between the virtual world and things in the real world through a user-created SL avatar (Bleize & Antheunis, 2019).
The SL virtual world, launched by Linden Lab in 2003, is considered an important channel for marketing and brand building (Barnes & Mattsson, 2011).The companies are able to establish their virtual business locations and create virtual objects in SL.Some global companies are still in SL where consumers/avatars can understand the companies' products and services (Maliaka, 2022).They can deliver multimedia messages in the real world; they can also provide simulated product functions in SL.Both approaches help maintain or enhance companies' brand equity This means that consumers' recognition of companies' brands is influenced by different marketing approaches, i.e., official websites and SL virtual business locations.Overall, SL virtual worlds can be explored to identify consumer' brand recognition in terms of brand attitudes and brand image.

Consumer behaviors in real and virtual worlds
A deep understanding of consumer behaviors in a real market, on the Internet, or in a virtual world is critical to all companies when it comes to marketing, product sales, and brand establishment (Baker et al., 2019;Bleize & Antheunis, 2019;Heuer et al., 2015;Manis & Choi, 2019;Zheng et al., 2018).Consumer behaviors are affected by their intentions at a given moment, and these intentions are produced by their attitudes toward the brands, according to the Theory of Reasoned Action (TRA) (Fishbein & Ajzen, 1975).Consumers' attitudes are affected by their beliefs, which are related to the level of acceptance influenced by the experiences they have with the company, such as joyful emotion, curiosity about new technologies, and personal avatar style (Bleize & Antheunis, 2019;Kocak & Ruzgar, 2017;Loureiro et al., 2019;Wedel et al., 2020;Wu & Chen, 2019).If a company delivers messages to consumers effectively, then they will be able to obtain positive experiences of the company, recognize the brand (i.e., brand recognition), and their brand attitude toward and brand image of the company will both be positive (Verhellen et al., 2013;Wu & Chen, 2019).For example, effective brand companies can quickly attract customers' attention and accurately predict their needs and preferences.
In addition, for unfamiliar company brands, even though consumers have no experience with it, they can make an overall judgment by receiving a variety of information from the company's different marketing channels, such as official websites, online comments on well-known social networks (e.g., Facebook and Twitter) and shopping websites (e.g., Amazon and eBay) or friends' opinions (i.e., word of mouth) in a virtual world (Baker et al., 2019;Bleize & Antheunis, 2019).For example, consumers can utilize the inquiry function provided by company websites or read online review comments from reliable sources before making a purchase decision.However, for unfamiliar company brands, the literature has no extensive discussions about whether consumers' brand attitudes and brand images could also be enhanced by offering them positive experiences using different brand management approaches.Therefore, the current research is expected to close a gap in the literature by carefully taking the consumers' high and low levels of brand familiarity into consideration while exploring how the different brand management approaches might lead to different levels of consumer brand recognition, including brand attitude and brand image.

Development of the research model and hypotheses
The idea of maintaining or enhancing consumers' brand recognition can be discussed from two specific perspectives, i.e., brand attitude and brand image.According to the TRA, if consumers believe a brand offers quality products, their attitudes toward the brand can significantly influence their purchase willingness or behavior (Jansen et al., 2011;Zenetti et al., 2014).Brand attitude is a core reference for consumers before purchasing products or during the purchase decisionmaking process (Keller, 1993).Consumers' brand attitudes toward products can be regarded as positive if they like purchasing certain brands of products and as negative if they dislike doing so (Baker et al., 2019).Most companies try to ensure that the marketing platform (official website or a virtual world) where messages are delivered to the consumers is effective in triggering positive beliefs about the consequence of a behavior and in generating positive attitudes toward the company brands that will later lead to actual purchase behavior.
Brand image is usually generated from consumers' memories associated with a certain brand (Keller, 1993).Consumers then establish a strong link in their memories to the company's brand products and services (Keller, 1993).Brand image is included in the larger concept of brand equity (Aaker & Jacobson, 2001).Different types of brand images (i.e., functional, experiential, and symbolic) determine different types of brand equity managed by the company (C.Park et al., 1986).Functional brand image is related to intangible product characteristics when consumers analyze the value of the brand.Functional brand image represents the benefits that consumers obtain to satisfy their fundamental need for variety and innovation (Belén del Río et al., 2001;Kocak & Ruzgar, 2017).For example, certain brand products are useful and safe to use, and the quality of the brand's products is high, with a full warranty.The experiential brand image not only is related to the product characteristics but also emphasizes the satisfactory level perceived by consumers as they experience the products (Bleize & Antheunis, 2019;Messinger et al., 2009).Consumers' emotions may fluctuate over time as they experience different products (Manis & Choi, 2019;Wu & Chen, 2019).For example, consumers can find abundant product information by viewing the many webpages on the company's official website; they can also have interesting and enjoyable experiences in the virtual world as they interact with virtual objects, i.e., the products.Symbolic brand image is unrelated to the characteristics of the products themselves.Instead, it reflects the intangible features provided by the products, such as being fashionable, free, modern, and unique.Those features can satisfy consumers' psychological needs, such as personal taste or style, personal expectations and value systems, interpersonal relationships, social identity or status, and above all, self-concept (Kocak & Ruzgar, 2017).For example, consumers wear a certain brand of sports shoes (e.g., Nike) to publicly present themselves as runners.
As mentioned above, consumers may form different levels of brand recognition, including brand attitude and brand image, when encountering different brand management approaches, i.e., SL virtual business locations or websites (Fine et al., 2016;Heuer et al., 2015;Verhellen et al., 2013).Accordingly, the current research proposes the following hypotheses and the research model in Figure 1.Hypothesis 1 and Hypothesis 2 can be further divided into two sub-hypotheses each for differentiating the participants' levels of brand familiarity (high vs. low).

Hypothesis 1 (H1):
There are significant differences in consumers' brand attitudes toward companies that use different brand management approaches, such as websites and virtual worlds.

Hypothesis 2 (H2):
There are significant differences in consumers' brand image in terms of three dimensions (i.e., functional, experiential, and symbolic) for companies that use different brand management approaches, such as websites and virtual worlds.

Research design
This study selected two multinational enterprises, i.e., IBM and Dell, in the computer industry.Both companies owned SL virtual business locations other than official websites at the time of the research.For a fair comparison of analysis results, this study adopted only the English version of the SL virtual business locations and the English version of the official websites.

Research participants
The participants were recruited immediately after the production of the experimental scenarios, which were recorded videos described in detail below.At the time of conducting this scenariobased experimental study, it was required that all participants be legal adults aged over 20, or obtain parental consent in order to participate.This precaution was implemented to safeguard the protection of participants' rights in the study.Also, the selected participants were predominantly junior and senior college students in computer-related majors, chosen for their expected experience with computers and interactive features, serving as a control variable.It is important to note that this study did not aim to compare participants' different demographic backgrounds.As the virtual world scenarios involved a higher level of interactivity, non-computer majors might find these features more exciting, potentially influencing their responses to the brand attitude and image questionnaires.
The participants had satisfactory English comprehension abilities, verified in their freshmen or sophomore years, enabling them to process English content while viewing the scenarios or freely browsing the websites or virtual worlds.In addition, when recruiting the participants, Although the participants were familiar with the official websites of IBM and Dell, their knowledge of the companies' SL virtual business locations was uncertain.It was neither unknown if they were consumers of IBM or Dell.Therefore, their brand recognition was assessed through a questionnaire that captured their brand attitude and brand image.It is worth mentioning that the gender ratio was not deliberately controlled, as gender difference had minimal influence on the overall study results and was not the focus of this study.

Reliability and validity of the questionnaire
The content validity of the questionnaire was established by reviewing the literature related to brand management and revisiting this study's purposes.For each construct, the measurement items were developed by referring one article and then further revision was made after reviewing more articles.The final revision was made through an internal peer-examination process carried out voluntarily by two scholars with expertise in brand communication and consumer behavior.Their examination was to further enhance the content validity of the questionnaire developed for this study by offering concrete suggestions to the authors.There were six items in the construct of brand attitude, and 18 items in the construct of brand image.All 24 items adopted a seven-point Likert-type scale, from 1-strongly disagree to 7-strongly agree (Table 1).This final version of the questionnaire was administered to a group of 28 college students to participate in a pilot study.The accuracy and comprehensiveness of the questionnaire contents were confirmed by adopting Bartlett's Sphericity Test and the Kaiser-Meyer-Olkin (KMO) Test.For the validity of the questionnaires, the pilot study results indicated a significant relationship between three dimensions (i.e., functional, experiential, and symbolic) of brand image.Thus, the three dimensions were analyzed together as the participants' overall brand image, which was found its Bartlett's Sphericity Test to reach a significant standard of 0.00 (<0.05) with a KMO measure of 0.449 (not above the threshold 0.60) (Hair et al., 2006).Although the KMO measure did not reach a satisfactory value, increasing the number of participants could improve this.The reliability of the questionnaires was measured using Cronbach's α (>0.60) to confirm the stability of the scores obtained from the participants (Hair et al., 2006).As a result, Cronbach's α for brand attitude was 0.805, while Cronbach's α for the brand image from the functional, experiential, and symbolic dimension were 0.631, 0.723, and 0.839, in that order.Overall, the measurement items developed in the questionnaire in this study had reached an acceptable level of validity and reliability.

Experimental scenarios
There were four experimental scenarios (i.e., IBM website, IBM virtual business location, Dell website, Dell virtual business location), which were all presented in recorded videos instead of the actual website browsing or SL interactions since it was necessary to control what design elements the participants saw in these two approaches to ensure that any difference was due to the approach and not the participants' random navigation.The recording tool, "Debut Video Capture Software v1.49," was utilized to record videos of virtual business locations and official Notes: The main references used for the development of the items are underlined.
websites.Then, "VideoPad Video Editor Professional v2.20" was used to edit videos and to add Chinese captions.
Since the companies' official websites were publicly available to gain access, here Figures 2 and 3 only showed the screenshots of the videos recorded from visiting the companies' virtual business locations by one researcher's registered SL account.The recorded videos were as complete and clear as possible in introducing different products and services offered by the company.In addition, each video was kept for about ten minutes to control the impact of the video's length on the participants (IBM: 9'44 minutes; Dell: 9'54 mins).
Priority was specifically given to the virtual locations and website contents that were directly related to customers or were visited more frequently by users in general.The videos about the official websites also included the "About" webpage containing the company history, the webpages about the company vision, goals, and value declaration, and the webpages about recent and historical issues and topics promoted by the company.The videos about the virtual business locations included the introduction of the first welcome or orientation island, where SL users could use maps or some tools offered by the company to get familiar with the company in the virtual space.Also introduced in the video were the islands where the company products were displayed and the islands where SL users could engage in interactive learning activities or use virtual objects.

Figure 2. IBM's SL virtual business location. Figure 3. Dell's SL virtual business location.
To be noted, the official website content created by the companies at the time when we collected the data (the second quarter of 2017) was different from that of today (IBM: https:// www.ibm.com/us-en;Dell: https://www.dell.com/en-us).The SL virtual worlds created by the companies and recorded by this study are currently inaccessible due to being private invitation-only groups or managed as separate virtual communities.To access these virtual worlds, a registered SL account is still required.While some groups are open for public joining, others may require a small payment in Linden dollars or L$ (e.g., membership fee: L$100; currency exchange rate L$1: USD$0.00313).

Research procedures
The recruitment plan involved sending electronic invitation letters to lecturers or teaching assistants who could then refer their students to participate in the experiment.The snowball method was also utilized, whereby participants who had completed the experiment were asked to invite or recommend a friend to participate in the study.Moreover, the study topic and a contact window were announced on a public bulletin board of the university's course management system.All participants voluntarily participated in the study by receiving an implied consent form and completing a short anonymous survey to verify if they shared similar demographic backgrounds.
Before the experiment, no stimulus was given to help link the participants' memories or update their brand knowledge, so it was difficult for them to give a fair overall judgment about their perceived brand image.However, based on the participants' previous beliefs, they were able to respond to their perceived brand attitude by responding to the questions regarding their perceived brand attitude toward the company.Then, they watched the instruction slides created to produce stable experimental results and to ensure a smooth experimental process.In the instructional slide, the participants pretended to help a good friend asking for their opinions on purchasing an IBM or Dell server computer.The pretending was to enhance participants' engagement in the experimental scenarios later on presented to them in the videos.The slides also helped them follow the correct procedures for taking part in this study.
In the scenario-based experiment study, the participants were randomly assigned to one of the four scenarios (i.e., IBM website, Dell website, IBM virtual business location, Dell virtual business location).Importantly, all scenarios were presented consistently in this study to carefully control for unrelated variables, such as language and duration of each scenario (Podsakoff et al., 2012).During the experiment, depending on the participants' previous beliefs, they were able to establish new associations with the brands while updating their brand knowledge.All the participants were asked to play the videos by using the same software, i.e., Windows Media Player 12.The experimental scenario intended to offer the participants first-person experiences while browsing a website or exploring a virtual world.Therefore, while the participants were watching the video clips, they were free to browse the companies' official websites or to explore the companies' virtual business locations by logging into the test account set up by researchers in a computer lab with Internet access.They were also allowed to freely watch videos back and forth.These flexibilities were considered to have minimal influence on the overall study results, so they were not intentionally controlled during the experiment.
After the experiment, the participants re-took the same questions regarding their perceived brand attitude.They also responded to questions regarding their perceived brand image including three dimensions.It could be further determined how their brand recognition was affected by one of the experimental scenarios.Finally, the researchers utilized a photographic method to conduct content analysis along with observation, based on S. R. Park et al. (2008) affordance views, to compare the performance status of IBM's and Dell's official websites and virtual business locations.
This study adopted a mixed-method approach which was similar to what Holzwarth et al. (2006) utilized in their study to combine quantitative with qualitative data.They carried out two experimental studies (quantitative data) combined with the content analysis (qualitative data) of a retail website to examine the influential role of different types of avatars on the website.Grudzewski et al. (2018) study also combine two data types (experimental study plus observation) for their research purpose, i.e., to explore which marketing communication tool is significantly more effective than others in presenting marketing messages and delivering advertising information for individuals involving in using new technologies.

Data analysis methods
The participants' demographic information, as well as their responses to the questionnaire, were analyzed using descriptive statistical analyses.An independent sample t-test was then carried out using SPSS Statistics 17.0 to determine if any significant differences were found in the responses to the measurement items regarding their perceived brand attitude between IBM and Dell before the experiment.If so, it meant that the participants had different baselines in their perceived brand attitude toward IBM and Dell, and their brand attitudes might be further influenced when watching the video clips presenting different brand management approaches.Thus, their responses to all the measurement questions after the experiment were separately analyzed for the two companies.The Least Significant Difference post hoc test was also conducted to verify the differences among the 2 × 2 scenarios.In addition, given the fact obtained from the pilot study that the three dimensions of the brand image were significantly related to one another, a MANOVA was thus appropriate to determine if there were any significant differences in the participant's responses to the questions regarding their overall perceived brand image after watching the video clips.Furthermore, to verify hypotheses H1 and H2, it was necessary to separate the data and analyze one company at a time: H1-IBM and H1-Dell, H2-IBM, and H2-Dell.Finally, by giving them another questionnaire after the experiment, their brand recognition affected by one of the experimental scenarios could be determined.

Content analysis and observation
After employing the above statistical analyses to test the two hypotheses, additional analyses were conducted to further understand the difference between the two brand management approaches.For that effort, the study adopted the view of affordance in the virtual world, proposed by S. R. Park et al. (2008).A review of the literature related to website design and virtual world pointed to content analysis and observation as appropriate methods for examining IBM and Dell's official websites and SL virtual business locations.The participants' experiences when they entered the virtual world scenarios were expected to be clearly described by the affordance view.There are three characteristics, i.e., control, content, and process, in virtual worlds (S.R. Park et al., 2008).These characteristics are influenced by 21 affordances (Table 2).
In this study, the upper-case letters, such as A1 and B2, were used to label the affordances that were well presented either on the official websites or the virtual business locations; while poorly presented ones were labeled with lower-case letters, such as a1 and b2.The symbol "-" (minus) indicates the affordances that were missing from either the official websites or the virtual business locations.Overall, the designs of IBM's and Dell's websites were analyzed by referring to these characteristics to examine the high/low performance levels of marketing activities held by the companies on their official websites and in their virtual worlds.Each affordance was carefully analyzed by observing the official websites and SL virtual business locations with inquiry strategies, i.e., asking questions to determine if the design of the official websites or the virtual business locations reflects that specific affordance.
The content characteristics of the virtual world are affected by ten kinds of affordances across two dimensions.The interactivity dimension of the content characteristics is affected by five affordances: social networks (B1 or b1), sports events (B2 or b2), event hosting (B3 or b3), C2 W: This affordance can be considered as missing unless the website presents a spatial 3D environment for achieving some marketing purposes.V: The extent to which the company uses a 3D environment to build its virtual business location.
The process characteristics are affected by six affordances.Three of them influence an individual's extrinsic motivation: monetary incentives (D1 or d1), free trials and services (D2 or d2), and competitive events (D3 or d3).The other three influence an individual's intrinsic motivation for the process characteristics of the virtual world: exclusive memberships (E1 or e1), exclusive ownership (E2 or e2), and games (E3 or e3).

Results
A total of 90 participants who had not taken part in the pilot study were recruited, but ten of the incomplete questionnaire responses to all the measurement items were found and removed.The final data thus had 80 questionnaire responses from 52 males (65%) and 28 females (35%).Each scenario contained 20 valid responses.Based on the descriptive analysis results of the participants' demographic information, it was confirmed that most participants had the basic capabilities needed to use computers and had the experience of using multimedia interfaces as shown in Table 3.All the participants were thus qualified to participate in this study's experimental scenarios.Several Harman's one-factor tests were initially conducted to assess whether any common method bias was potentially introduced due to factors such as the flexible timing of participation, variations in instructional slides presented by different research assistants, or differences in the level of engagement exhibited by participants during the experimental scenarios (Podsakoff et al., 2003).All the test results confirmed the absence of common method bias, as the percentage of variance accounted for by a single factor was below the threshold value of 50% (32.99%).This was observed in the pre-measurement of brand attitude (39.59%), post-measurement of brand attitude (42.22%), and post-measurement of three-dimensional brand images (functional: 44.53%, experiential: 47.47%, symbolic: 42.95%).Based on these results, it can be asserted that the experimental study demonstrated reliability and validity without bias.
As per the result of the participants' responses to the questionnaires, this study obtained a Cronbach's α of 0.779 for the reliability of the pre-experiment measure of brand attitude and 0.812 for the post-experiment measure of brand attitude.The Cronbach's α of the post-experiment measure of the functional, experiential, and symbolic brand images were 0.821, 0.839, and 0.812, respectively.The overall Cronbach's α of the post-experiment measure was 0.917.Table 4 showed the item-to-total statistics including the corrected item-to-total correlation coefficient values and "Cronbach's α values if item deleted" for each item in its construct.Every "Cronbach's α value if item deleted" was lower than its corresponding coefficient, indicating that no item should be removed, and the reliability of the questionnaire was supported.Furthermore, the validity of the post-experiment measure of all the responses to the questionnaire reached the significant standard of 0.00 after being measured by the Bartlett's Sphericity Test, and the KMO measure had reached 0.829.Therefore, all the responses collected from the questionnaires developed in this study were both reliable and valid.Subsequent analysis results were obtained by only conducting comparisons between the official websites and virtual business locations of the same company since there were significant differences (t = 2.743, p = 0.008) between the participants' brand attitude towards IBM (M = 4.729, SD = 0.717) and Dell (M = 4.283, SD = 0.737) before they participated in any of the scenarios.
Since the participants' brand attitudes towards IBM and Dell had different baselines, it was necessary to analyze the participants' responses to the post-questionnaires of the same company.The hypotheses H1-IBM and H1-Dell were tested to determine if there was a significant difference in consumers' brand attitudes between the two brand management approaches (the website and the virtual world).Separate verification was also carried out for each company (H2-IBM and H2- Dell) to determine if there was a significant difference in consumers' brand image between those two approaches.Tables 5 and 6 showed the results of the descriptive analysis of the pre-and postquestionnaire responses for IBM and Dell in the different scenarios.The differences between the official websites and virtual business locations for the two companies in the participants' brand attitude or brand image were reported in the following sections.

Consumers' brand attitude and brand image-IBM
The independent-sample t-test results showed a significant difference in the IBM participants' overall responses before and after the experiments (t = 3.034, p = 0.003; before: M = 4.729, SD = 0.717; after: M = 5.188, SD = 0.632), and their brand attitude was significantly enhanced (↑).Specifically, after separating the participants' brand attitude responses about the IBM official websites from those about its virtual business locations, a significant difference was only found in their pre-vs.post-experiment responses to the official website (website: t = 2.591, p = 0.014; virtual world: t = 1.633, p = 0.111).However, no significant difference was found in their responses to the post-experiment questionnaire of brand attitude toward IBM between its official websites and its virtual business location (t=-1.130,p = 0.265; website: M = 5.300 > virtual world: M = 5.075; mean difference: 0.225).Thus, H1-IBM was not supported (website↑ > virtual world↑).
According to the results of the Pearson correlation analysis, significant correlations were found among functional (F), experiential (E), and symbolic (S) brand images (F&E = 0.672***, F&S = 0.647***, E&S = 0.635***, *** p < 0.000).A MANOVA was then conducted to determine if any differences were found in the participants' responses to the post-experiment Brand Image Questionnaire between the two experiment scenarios, i.e., the IBM official website and virtual business location.No significant difference was found in their responses between the two scenarios regarding the IBM brand image (F = 1.404, p = 0.257, website almost equaled to the virtual world).There was also no significant difference in their responses between the two brand management approaches, regardless of which types of brand image they perceived when completing the post-experiment questionnaire (Functional: F = 0.065, p = 0.800; Experiential: F = 0.808, p = 0.374; Symbolic: F = 0.571, p = 0.455).Thus, H2-IBM was not supported (functional and symbolic: website > virtual world; experiential: virtual world > website).

Consumers' brand attitude and brand image-Dell
The participants' overall responses to the pre-and post-experiment questionnaires showed a significant difference in their brand attitudes toward Dell (t = 2.839, p = 0.006; before: M = 4.283, SD = 0.737; after: M = 4.804, SD = 0.896), and their brand attitudes toward the company were significantly enhanced (↑).Specifically, there was a significant difference in the participants' brand attitude toward Dell's official website before and after the experiment (t = 2.327, p = 0.025).However, the change in their brand attitude toward Dell's virtual business location was merely approaching significant (t = 1.864, p = 0.070).Besides, no significant difference was found in their brand attitude in the post-experiment questionnaire between the official website and the virtual business location (t = 1.653, p = 0.107; website: M = 4.575 < virtual world: M = 5.033; mean difference: 0.458).Thus, H1-Dell was not supported (virtual world↑> website↑).
The group of participants assigned to the Dell official website was compared with the group assigned to the virtual business location to find the differences in their responses to the postexperiment questionnaires regarding the company's brand image.The Pearson correlational analysis results showed significant relationships among the functional, experiential, and symbolic brand images (F&E = 0.640***, F&S = 0.559***, E&S = 0.715***, *** p < 0.000).The MANOVA showed that there was a significant difference in the participants' brand image of Dell between the two brand management approaches (F = 12.105, p = 0.000, website < virtual world).Specifically, the experiential and symbolic brand image perceived by the participants had significant differences in the post-experiment responses (Functional: F = 1.746, p = 0.194; Experiential: F = 28.167,p = 0.000; Symbolic: F = 10.670,p = 0.002).Thus, H2-Dell was partially supported (functional, experiential, and symbolic: virtual world > website).

Verification of hypotheses
No significant difference was found in the participants' responses to the post-experiment brand attitude questionnaire between the website and the virtual world scenarios for IBM or Dell.However, significant differences and certain levels of enhancement were found for both the IBM and the Dell scenarios in the participants' overall or separate responses (official website vs. virtual world) to the preand post-experiment questionnaires.These mean that the integrated brand management approach helps increase the participants' brand attitudes toward both companies.Simply launching an official website is insufficient for a company to maintain its brand in the virtual world economy nowadays.
A website integrating a user-oriented technology and interactive features is an effective approach for a company to continuously maintain its brand and to make the brand durable (Grudzewski et al., 2018;Holzwarth et al., 2006;Manis & Choi, 2019;Wedel et al., 2020).Overall, H1 was not supported.The virtual approach was significantly more effective than managing official websites for enhancing brand attitude toward unfamiliar brands.However, there was no difference between the two approaches for enhancing brand attitude toward brands that were already familiar.Eventually, consumers could possess an equivalent level of brand attitude toward both companies.
The participants' brand image responses were not significantly different after watching videos about IBM's official website and virtual business location.However, the participants' responses to the experiential brand image of the IBM virtual business location were higher than the ones to the IBM official website.In addition, compared with the participants' responses to all three dimensions of the brand images regarding IBM's official websites, the mean values were all higher than the ones responded by those assigned to Dell's official website.Again, the reason may be due to the participants' different levels of familiarity with IBM and Dell at the beginning of the study.On the contrary, the mean values of the experiential and symbolic dimensions of brand images to Dell's virtual business locations are higher than the ones for IBM's virtual business location.These results highlighted the effectiveness of developing a virtual world to enhance consumers' abstract thought about a brand, and it was an important platform to offer an interactive environment for consumers who may need different personalized experiences with the products or services (Baker et al., 2019;Keller, 2010).Overall, H2 was partially supported.It does not matter whether the participants are more familiar with one brand or the other, the experiential dimension of brand image could be easily enhanced in both the companies' websites and virtual worlds.The participants' different levels of familiarity with the brand differentiate only their functional and symbolic brand images.

Content analysis and observation results of IBM and Dell official websites and virtual business locations
The results of the comparison based on the concept of affordance by S. R. Park et al. (2008) were summarized in Table 7 for IBM's and Dell's SL virtual business locations and official websites.Both IBM's and Dell's official websites and virtual business locations contained texts, pictures, and videos that constituted the rich content.
The IBM official website placed more emphasis on social groups and event hosting, especially more training lessons, so its overall design presented a set of diversified and interactive items in a varied and functional way.On the other hand, Dell's official website's main design direction was the ease of use and browsing.The website focused on social groups and event hosting, offered users service consultation, and used e-coupons to attract more users to browse the website.The control features of IBM's virtual business location emphasized convenient explanations for the users, while the design of Dell's virtual business location encouraged users to obtain as many training sessions as possible, so that they could quickly become familiar with the virtual business location.For the interactivity dimension of the content, even though IBM's virtual business location focused more on reactive objects than Dell, both contained the same number and a similar level of affordances.For the vividness dimension of the content characteristics, Dell's virtual business location had many stronger affordances, such as virtual object simulation and three-dimensional space design.For the process characteristics, both virtual business locations presented a similar level of affordances with a focus on users' extrinsic and intrinsic motivations, such as offering interesting landscape designs.

Discussions and implications
The results of this study indicated that the participants had significantly better brand attitudes towards IBM than Dell before the experiments.Regardless of the brand management approach (virtual business location or official website) adopted by IBM or Dell, the participants' brand recognition of the company influenced their brand attitude after the experiment (Baker et al., 2019;Verhellen et al., 2013).Therefore, it was more appropriate to have a separate discussion of the experimental results for each company.
For the unfamiliar brand (Dell), consumers' brand recognition is more strongly associated with their perceived brand image than brand attitude.Based on the hypothesis verification results, the company's virtual business location has proved to be more effective than the official website in enhancing brand recognition, specifically the brand image, in consumers unfamiliar with the brand.The reason is that virtual experiences offer these consumers a special emotion and interaction with avatars and virtual objects and encourage the consumer's mind to make a brand association (cognitive perspectives) with enjoyment (psychological aspects), producing a positive perceived brand image (Barnes & Mattsson, 2011;Keller, 2010).The consumers' perceived brand image can be significantly enhanced with vivid and interactive design contents, such as providing product characteristics (functional brand image) on the website, caring more about the consumers' perceived satisfactory level of the products (experiential brand image), and considering the intangible features provided by the products (symbolic brand image).However, instead of the functional dimension of brand image, consumers' perceived experiential and symbolic dimensions are worthy of exploration in positioning brand personality with social telepresence and flow features for different types of consumers with different technology confidence levels (Keller, 2010;Nah et al., 2011).The official websites usually contain textual descriptions and static images, while the high-technology functions of virtual business locations are expected to have a high-quality virtual presentation of products and services.Yet both the website and the virtual business location must consider the basics of introducing the brand to consumers who are unfamiliar with it (Wu et al., 2019).Over time, the unfamiliar brand will eventually become well-known as consumers can immerse themselves in telepresence (higher experiential brand image) and find meaningful sensemaking interactions with the quality products and services (symbolic brand image) each time they enter the virtual business locations (Nah et al., 2011).
Based on the overall statistical analysis results, consumers might have a higher expectation of the familiar brand's (IBM) virtual world in terms of brand recognition, including brand attitude and brand image.For a well-known brand entering the virtual interactive platform, this is a reminder of how the company can consistently position its brand by offering effective and interactive virtual objects to produce a full sensory experience for their consumers as in the real world while presenting its brand on its official website (Barnes & Mattsoon, 2011;C. Park et al., 1986C. Park et al., , 1991)).As more and more consumers can afford to purchase virtual technologies to familiarize themselves with and engage in the virtual world to conduct novel e-commerce transactions, their virtual experiences obtained from a familiar brand can directly influence their brand attitude and brand image, which then further influence their purchasing behaviors (Baker et al., 2019).In the world of virtual business locations, a familiar brand needs to maintain its brand concept through interactive functions for its existing consumers to hold the same or higher level of brand recognition.It also needs to continuously create virtual events or objects to enrich the virtual consumption experiences and make consumers recall positive memories to deepen their loyalty and "stickiness" to the brand.
In addition, the content analysis and observation results show that the IBM official website and virtual business location designs both focused more on control and content characteristics, such as A3-online help, B5-design template, and C4-multimedia than on the process ones (Part et al., 2008).Therefore, even though users could obtain rich information-including multimedia resources-from the right place, the new experiences and memories associated with the participants' previous brand knowledge were not obviously available for them to enhance their brand image.Differently, Dell's content analysis and observation results show that even though their official website contained more textual descriptions and static images, more emphasis was given to process characteristics, like D1-monetary incentives, while its virtual business location focused on A4-training program and C2-spatial.Consequently, users could find interesting ways to participate in various interactive activities and could obtain satisfactory new experiences with events or virtual objects, and this could then influence the establishment of the brand image.Based on the above findings, we can conclude that if a company designs its official website and virtual business location from the perspectives of control and content characteristics, then the design will have good functionality and can better present vivid and interactive features in its affordances (McMillan & Hwang, 2002;S. R. Park et al., 2008).As highly interactive and simulation technologies emerge, the consumers' total experience with the company product or brand needs to be satisfactory for not only the physical stores but also the virtual ones (Messinger et al., 2009).
Moreover, this study found that certain content designs on websites and in virtual business locations are capable of providing multiple affordances at the same time.An interdependent relationship (as it is referred to in contingency theory) thus exists among these design affordances (Zheng et al., 2018).For example, the relationship between A5-design templates and A4-training programs shows that if users are offered personalized or membership services on companies' official websites, they are allowed to edit their own webpage viewing layout and select customized functions.Certainly, sufficient training and explanation must be provided in videos or other interesting ways to assist the users in the utilization of the websites.Moreover, the relationships among B1-social networks, E1-exclusive memberships, and E2-exclusive ownership show that if the users want to use most of the functions on company social websites, they must register for membership first.After registration, members can own their blog webpages and identify each other via membership accounts/numbers.Placing membership registration restrictions makes it possible for social networks to identify members versus nonmembers.The relationship between C4-multimedia and C5-haptic feedback shows the importance of increasing the number of visual and auditory effects on official websites.As more video or animation elements are added, more haptic feedback contents are required so that users can easily control the multimedia contents, such as playing videos, dragging animated selection menus, scrolling, browsing pictures, and so on.
In virtual business locations, the relationships among A1-navigation aids, A3-online help, and A4training programs show that users need to be familiarized with the three-dimensional spaces before interacting with the virtual objects, and this needs to be taught to them by the companies.A learning zone in a virtual world is helpful to provide explanation cards, training courses, and online navigation assistance, including the basic functions of road signs and moving to other spaces.Also, the relationship between B1-social networks and B3-event hosting shows that social events, such as international conferences, can be held in real-time for numerous users in the virtual world.Therefore, for the purpose of conducting smooth social events in the virtual world, an event host must be present to control the process and sequence of events, just as in the real world.The relationships among B5-reactive objects, C1-realistic objects, C2-spatial 3D, and C5-haptic feedback show that virtual objects in the virtual world can be designed to have the same appearance and size as those in the real world.These virtual objects can further be simulated to interact with users.For example, users can plant a seed that can grow into a tree following users' programming in Dell's virtual business locations.In IBM's virtual business location, the server model can be turned on to allow users to install parts in their computers.In addition, the relationships among B5-reactive objects, C5-haptic feedback, and E3-games show that some virtual objects with a recreational or game nature are expected to engage in interactions with users, and as such, the design of these objects needs some haptic feedback to bring more amusement to the users.
Overall, the more vivid the interactions are in the three-dimensional spaces, the higher the consumers' willingness to browse product information and the better their understanding of the product features will be, and consumers will then gain a deeper brand recognition of the product (Christodoulides et al., 2004).The current study suggests that companies should consider redesigning their official websites to offer more personalized services, so that consumers can easily find and understand how to browse the system and use the functionalities of the webpages.Consumers can then effectively obtain the most appropriate and richest product/service contents from the websites, and their attitudes toward the brand can be enhanced.The same suggestion is also given to companies about re-designing their virtual business locations because it is more effective to enhance consumers' brand recognition in the virtual world than on official websites.Companies are thus suggested to provide experiential and interactive activities, including vivid and interactive video contents, both on official websites and in virtual business locations, so that the company brands can be promoted in varied and interesting ways.Sample activities using symbolic and decorative virtual objects can be designed in the companies' virtual business locations, so that consumers can continuously find special features in different brand management approaches.When consumers feel satisfied with the companies' brand promotion activities, their brand image will then be enhanced.

Conclusion
Numerous companies have established virtual business locations in virtual worlds, such as the SL explored in this study.These three-dimensional spaces have become alternative brand marketing channels for providing diversified brand promotion activities.While a website contains abundant text and static images or photos, virtual business locations can be easier for consumers to surf.The virtual world usually contains a variety of rich content for consumers, generating a better brand image than the official website can, as more interactive virtual objects are provided to the consumers.Differences were also found between the companies and their different brand management approaches and were recorded in terms of the affordance characteristics, such as control, content, and process.
This study verified the hypotheses by carrying out a scenario-based experiment incorporating with the questionnaire approach to collect quantitative data.To collect qualitative data, this study employed a photographic method to conduct content analysis and observation for the websites and virtual business locations of IBM and Dell by adopting S. R. Park et al. (2008) view of affordance.Statistical results of the experiment and the content analysis and observation results are different facets of the research findings that the researchers were able to obtain, demonstrating the value of cross-validation by adopting a mixed method.This study may serve as a useful reference to brand management officers and virtual world platform developers to get a clear picture of brand management (regardless of how well-known their brands are) and to obtain a comprehensive examination of how different brand management approaches can affect consumers' brand recognition, including brand attitude and three dimensions of brand image.
This study had several limitations in its context.First, this work only selected SL to explore an alternative brand management approach, even though numerous companies have established virtual business locations in other three-dimensional spaces to simulate their real-world buildings.Moreover, due to the constraints of time and staff, the antecedents of brand attitude and other types of brand images perceived by consumers were not discussed in this study.Different perspectives and comments about companies' brand promotion activities or brand management approaches in virtual business locations may arise due to cultural and language-related factors.In addition, the distribution of the participants with regard to their basic demographic backgrounds needs more careful consideration in future studies.
Subsequent research can be conducted in three directions.Firstly, multiple virtual worlds can be selected, then comparisons can be conducted to verify if consumers' brand recognition is affected by the different features of different virtual worlds.Secondly, different types of companies can be selected to verify if the design affordances for websites and virtual business locations examined in this work can enhance consumers' brand recognition.Studies can also explore the extent to which consumers perceive varying levels of brand recognition when they visit different companies' official websites and virtual business locations with distinct designs.With the widespread popularity of social networking platforms such as Facebook, Instagram, LINE, WeChat, and WhatsApp, numerous companies and stores worldwide have established official accounts to leverage various social networking features for marketing activities, brand management, product promotion, and customer services.Consequently, it would be valuable to compare how consumers' brand recognition is influenced by different social networking platforms.
Finally, with the availability of increasingly advanced and convenient research devices (e.g., functional near-infrared spectroscopy, electroencephalogram, and electromyography), researchers currently have the capability to capture sophisticated and immediate consumer reactions during the presentation of experimental scenarios, while simultaneously obtaining accurate data on biological indicators and neurophysiological measurements (Wedel et al., 2020).In future studies, it is recommended that researchers focus on detecting consumers' momentary thoughts about a brand and identifying the neural correlates associated with specific brand attitudes and brand images.This will enable a deeper exploration of the interconnected relationship between different brand management approaches.

Table 1 . Brand attitude and brand image questionnaire items Brand Recognition Questionnaire Items
Park et al. (1986)d Chernaton (2004)ides and Chernaton (2004),Keller (1993), C.Park et al. (1986)I feel that when someone is browsing this brand's website/virtual world, they will ...B7.Feel happy.B8.Feel relaxed.B9.Be aware of their own mood.B10.Experience up and downs in their mood.B11.Feel that the website/virtual business location is very special.B12.Have feelings never experienced with other companies.Brand Image -SymbolicChristodoulides and Chernaton (2004),Keller (1993), C.Park et al. (1986)I think that browsing this brand's website/virtual world will . . .B13. Allow me to be trendy and fashionable.B14.Be a good match to my taste and style.B15.Help me notice what kind of people prefer the brand.B16.Help me find others who think of the brand the same way I do.B17.Make me feel special.B18.Highlight my individuality.