Abstract
Brands are increasingly using the metaverse to promote sales of products and services, but the social influence of avatars in that context is largely unknown. Grounded in social impact theory and impression management literature, this research investigates the impact of retail store employee avatar presence on consumer behavior. Two between-subjects studies from laboratory and online settings uncover consumers’ aversive responses to employee avatar presence—living (human) and nonliving (robot)—on several key outcomes. This negative effect only emerges in embarrassing shopping settings, such as the purchase of condoms and adult diapers, but not in ordinary (nonembarrassing) settings.
Management Slant
This research shows that both living (human) and nonliving (robotic) employee avatars decrease consumers’ purchase behavior, including time and money spent on a purchase, in embarrassing metaverse contexts.
Consumers are also less likely to recall the brand of an embarrassing product but more likely to recall that the embarrassing product was on sale in the presence (versus absence) of avatar employees.
The negative impact of avatar employee presence emerges in embarrassing virtual shopping settings but not in more mundane (nonembarrassing) settings.
Marketers, advertisers, and brand owners operating in the metaverse should avoid using avatar employees (both human and robotic) to promote embarrassing products.
INTRODUCTION
As the metaverse becomes increasingly popular, brands are exploring ways to offer consumers memorable experiences while promoting awareness, reach, and sales of both virtual and physical products (Yoo, Welden, Hewett, and Haenlein, 2023). Promotions in the metaverse are unique, as the metaverse utilizes highly immersive techniques that are enabled through frictionless virtual reality technologies to make promoted content more memorable (Cowan, Ketron, Kostyk, and Kristofferson, 2023; Giang Barrera and Shah, 2023). Nike became one of the first big brands to enter the metaverse when it created “Nikeland.” It launched a dedicated virtual world where users can customize their avatars with Nike products; attend virtual events, including concerts and fashion shows; and, most important, shop for Nike products (Reuters, 2021). A core pillar of the foundation of the metaverse, however, which facilitates social interaction and enables users to immerse themselves in a vast virtual realm (Hennig-Thurau, Aliman, Herting, et al., 2023), remains unknown: the effects of social presence through avatars. This raises the question as to whether brands should promote their products and services using employee avatars or abstain from relying on such social cues in promotional activities.
Researchers have extensively documented the importance of social presence effects on several key customer outcomes in physical consumption contexts (e.g., Argo and Dahl, 2020; Argo, Dahl, and Manchanda, 2005; Otterbring and Lu, 2018; Söderlund, 2016). Few studies, however, have tested whether the social presence of avatars in virtual shopping environments, such as the metaverse, plays a role in shaping consumer responses (Chang, Chu, and Kao, 2021). Notably, scholars find that virtual reality environments lead to higher levels of presence than physical store environments (Pizzi, Vannucci, and Aiello, 2020). Hence, one cannot take for granted that results from physical and online retail settings automatically replicate in virtual shopping settings, particularly considering that (nonliving) avatars, rather than actual human employees, constitute one of the most distinctive characteristics of the metaverse. This also implies that virtual consumption contexts might have a competitive edge relative to other offline and online retail formats, as the metaverse is a virtual rather than the real world, making it particularly promising for the promotion of certain personal or embarrassing products that consumers might not want to purchase in more traditional retail settings.
Building on social impact theory (Argo, 2020; Latané, 1981; Otterbring, 2021) and findings from the impression management literature (Argo, White, and Dahl, 2006; Kurt, Inman, and Argo, 2011; Leary and Kowalski, 1990), the present research tests the thesis that, when shopping for embarrassing products in the metaverse, employee avatar presence results in negative consumer responses. The authors examine whether avatar social presence influences the time consumers devote to purchasing embarrassing products, the amount of money spent, the number of products purchased, and the likelihood of making impulse purchases, as well as consumers’ willingness to buy a given product. They also explore whether the effects of social presence extend to consumers’ recall rates of brands within the embarrassing-product category along with their likelihood of recalling whether embarrassing products are promoted in the store. To this end, the authors used a carefully designed virtual reality study with conditions that include the promotion of embarrassing products (condoms) through human and robotic virtual employee avatars (Study 1). They then complemented this behavioral virtual reality study, high in realism and ecological validity, with a controlled online experiment involving both embarrassing and nonembarrassing shopping settings to address some limitations in an improved design with enhanced rigor, control, and internal validity (Study 2).
The findings suggest that consumers respond negatively to embarrassing products that are actively promoted by employee avatars, regardless of whether they represent human employees or robots. Consumers, therefore, spend less time before making their purchase, purchase fewer products, spend less money, and make fewer impulse purchases when shopping for an embarrassing product in the presence of employee avatars than they do when shopping alone (Study 1). The consumption context, however, moderates the focal employee avatar presence effect, such that the aforementioned results only apply when shopping for embarrassing products but not when shopping for more ordinary (nonembarrassing) products (Study 2). Together, these findings have implications for how to effectively advertise and sell embarrassing products in the metaverse. The results suggest that marketers, advertisers, and brand owners need to be cautious in using avatars to promote embarrassing products, even in virtual settings.
CONCEPTUAL BACKGROUND
Social Presence in Extended Reality Environments
Social impact theory posits that the real, implied, or imagined presence of others, including their actions, has the capacity to influence an individual’s emotions, cognitions, and behaviors (Latané, 1981). Some studies have shown that social presence in extended reality environments can positively influence consumer attitudes and intentions (Animesh, Pinsonneault, Yang, and Oh, 2011). “Social presence,” in this context, is defined as the degree to which a person is perceived as being “real” and “present” in a mediated communication. It is a psychological state in which virtual or artificial social actors are experienced as real (Lee, 2004).
Social presence is influenced by a myriad of factors, including the realism of the medium, the way the social actor is represented, and user expectations (Kim, Park, and Sundar, 2013). For example, virtual reality experiences with higher levels of social presence can lead to increased brand engagement and advocacy (De Regt, Plangger, and Barnes, 2021). Similarly, augmented reality experiences with increased sociability can lead to greater user–user engagement (Scholz and Smith, 2016). In massive virtual environments such as the metaverse, social presence can be designed to foster or hinder social interaction, which can have a significant impact on sharing, collaboration, and co-creation (Bonsu and Darmody, 2008). In the context of user-generated content, social functions can motivate consumers to perform activities that are important and relevant to others in online communities (Daugherty, Eastin, and Bright, 2008). Co-creation and its associated social connotation has been proposed, furthermore, to be particularly influential among highly involved consumers (Cowan and Ketron, 2019a), who might be more prone to co-create content if they feel a sense of social presence in the metaverse. This is because social presence can help to boost perceptions of shared ownership and responsibility, which can motivate consumers to contribute to the metaverse community.
Social presence can help to boost perceptions of shared ownership and responsibility, which can motivate consumers to contribute to the metaverse community.
Social presence is an important dimension to consider in the context of sales promotions targeting customers in the metaverse. The literature suggests that social factors influence merchandise and service quality (Baker, Grewal, and Parasuraman, 1994). The same is true for online store atmospherics (Eroglu, Machleit, and Davis, 2001). In the context of sales promotions in the metaverse, consumers might react differently as a function of whether people, in the form of avatars, are present or absent in the virtual store environment. For example, if consumers make their purchases in the presence of others, they may purchase certain products, such as more expensive items, that they do not really want because of impression management concerns and the desire to convey a positive image of themselves to others (Argo et al., 2005; Gasiorowska, Folwarczny, Tan, and Otterbring, 2023; Otterbring, 2021). Some studies, however, suggest that social presence has adverse effects on certain consumption responses (e.g., Esmark Jones, Stvens, Noble, and Brezeale, 2020; Luck and Benkenstein, 2015; Rosenbaum, Ramirez, El-Manstrly, and Sit, 2021), particularly in embarrassing consumption contexts (Dahl, Manchandra, and Argo, 2001; Esmark Jones et al., 2017). First, social presence can increase the likelihood that a consumer will be evaluated by others. This can lead to feelings of anxiety and self-consciousness, which can also make the consumer embarrassed. Second, social presence may increase the likelihood that the consumer will become aware of the social norms surrounding a particular purchase, which can further emphasize embarrassment.
The Role of Consumer Embarrassment in the Metaverse
“Consumer embarrassment” is the feeling of shame or discomfort that a person experiences when buying a product or service that, in some sense, is perceived as undesirable (Krishna, Herd, and Aydınoğlu, 2019). This feeling can be caused by several factors, such as the price of the product, the brand of the product, or the social context in which the purchase takes place (Fisk and Grove, 1996). In practice, consumers tend to develop their own strategies to cope with embarrassment. For example, a consumer may avoid buying condoms when they anticipate feeling embarrassed (Brackett, 2004); but empathetic employees can provide customers with a sense of comfort. If employees approach potentially embarrassing situations with professionalism and sensitivity, they may be able to assist customers and alleviate any discomfort, thus resulting in positive customer responses.
The role of consumer embarrassment in the context of metaverse shopping experiences is expected to play a significant role, as consumers are likely to feel more self-conscious when visiting virtual shopping worlds through avatars than they would when merely browsing an online store (Penz and Hogg, 2011). Consumers who are shopping for a certain outfit in the metaverse, for example, are more likely to feel embarrassed if they know that other people will be able to see their product selection. Likewise, when employee avatars promote embarrassing products, consumers may be less likely to purchase them if they think that another customer or employee in the store can see them (Esmark, Noble, and Brezeale, 2017).
Under certain circumstances, consumers may be more likely to make purchases if a metaverse shopping environment is designed to be social and interactive. This could be the result of consumers feeling more comfortable buying products in a virtual environment when they can interact with others, such as seeking help from employee avatars. When shopping for embarrassing products, however, the opposite might apply, considering that embarrassing shopping settings tend to result in negative consumer responses in the mere presence of others (Dahl et al., 2001; Krishna et al., 2019; Otterbring and Lu, 2018).
It is important to note that, when shopping in the metaverse, employee presence can only be established through virtual avatars rather than real physical persons. Some scholars suggest that the perceived presence of such virtual stimuli cannot be compared with the real world, similar to how a phone call cannot compensate for the presence of a face-to-face conversation (Hennig-Thurau et al., 2023). It is important to acknowledge the possibility that social presence in the metaverse may not necessarily lead to the same effects as those established in physical commercial settings (e.g., Argo et al., 2005; Söderlund, 2016; Tombs and McColl-Kennedy, 2010), and that employee presence does not automatically affect customers’ purchase behavior in relation to virtually promoted embarrassing products.
The literature is inconsistent as to whether replacing human employees with robotic counterparts may be desirable in embarrassing consumption contexts. Some studies show positive effects on consumer responses in embarrassing shopping settings, regardless of whether customers are (or are not) acknowledged by a human or a robotic employee (Frank and Otterbring, 2023b), whereas other studies have indicated that consumers feel less judged by a robot, compared with a human employee, when having to engage in an embarrassing service encounter, such as when acquiring medication to treat a sexually transmitted disease or being confronted with one’s own mistakes by a frontline employee (Holthöwer and van Doorn, 2023; Pitardi, Wirtz, Paluch, and Kunz, 2022).
HYPOTHESES
The influence of employee presence and salespeople in physical retail stores is well documented, but the same cannot be said for sales promotions in the metaverse. This study, therefore, aimed to empirically test propositions made regarding sales promotions through human and robotic employee avatars in the virtual realm and, more broadly, the impact of social presence effects on consumer responses in the metaverse (See Figure 1). The bulk of prior literature from physical retail settings in general (e.g., Argo et al., 2005; Luck and Benkenstein, 2015; Rosenbaum et al., 2021), and regarding embarrassing consumption contexts in particular (e.g., Dahl et al., 2001; Esmark et al., 2017; Otterbring and Lu, 2018), has shown social presence to generate a wide array of negative consumer responses. For instance, a retail field study showed that the presence of other individuals in the immediate vicinity of consumers reduced their inclination to interact with a battery-testing station situated close to a battery display from which consumers were to select batteries, suggesting that social presence decreased consumers’ decision time (Argo et al., 2005). Likewise, research has documented that the presence of another unknown individual standing next to target consumers reduced the number of alternatives they considered in front of a supermarket shelf, indicating that their decision time decreased because of social presence (Luck and Benkenstein, 2015). The latter study further revealed that such social presence resulted in a reduced willingness to buy, lower levels of satisfaction, amplified attempts to leave and get out of the shopping situation, decreased exploratory behaviors, and active attempts to avoid other people.
Conceptual Model
On the basis of the aforementioned and related findings, the authors posit that the social presence of employee avatars, relative to their absence, should result in negative consumer responses in embarrassing shopping settings, regardless of whether the employee avatars represent humans or robots. They predicted, therefore, that such social presence should minimize the time consumers spend on deciding which embarrassing product to purchase in an attempt to leave the virtual store as quickly as possible (Argo et al., 2005; Tversky and Shafir, 1992; Verbeke and Bagozzi, 2002). They also predict that social presence should lead to fewer purchases in embarrassing shopping settings, both generally and regarding impulse purchases, as the increased layer of complexity because of social presence should make consumers more prone to getting their task completed rapidly, leading to fewer products purchased and a lower likelihood of making impulse purchases (Andersson, Wästlund, and Kristensson, 2016; Luck and Benkenstein, 2015; Mattila and Wirtz, 2008). This should also be reflected in spending patterns, such that consumers in embarrassing consumption contexts spend less money when shopping in the presence, rather than the absence, of human or robotic employee avatars (Dahl et al., 2001; Krishna et al., 2019; Li, Yang, and Zhou, 2018; Van Ittersum, Wansink, Pennings, et al., 2013). As both human and robotic employees are represented by virtual avatars, the authors did not anticipate differences between these avatar types in how consumers will respond (cf. von der Pütten, Krämer, Gratch, and Kang, 2010). Thus:
H1: The presence (versus absence) of employee avatars generates more negative consumer responses in embarrassing shopping settings, such that they:
(a) make more rapid purchases,
(b) purchase fewer products,
(c) spend less money, and
(d) make fewer impulse purchases.
Previous research has shown that the presence of (avatar) employees influences how consumers perceive and evaluate offline and online retail environments (Amorin and Bashashi Saghezchi, 2014; Holzwarth, Janiszewski, and Neumann, 2006). This relationship has often been discussed in the context of impression management (Diestre, Montauti, and Pinto De Sousa, 2023; Young, Gardner, and Gilbert, 1994). Impression management in social contexts is defined as “the goal-directed activity of controlling or regulating information in order to influence the impressions formed” (Schlenker, 1980, p. 6). In the present research, the authors assume that employee avatars constitute an important social element from an impression management perspective, as employees contribute to the overall customer experience in retail environments. For example, in an ordinary shopping context (e.g., weekly grocery shopping), friendly and attentive staff members can create a welcoming environment, making customers feel more comfortable and positively influencing their perception of the shopping situation, which can strengthen the success of promotional activities (Bluvstein Netter and Raghubir, 2020; O’Cass and Grace, 2008). The findings of earlier studies, however, imply that the effects of employee avatar presence on consumer responses are likely to depend on the consumption context (Demoulin and Willems, 2019; Seiders, Voss, Grewel, and Godfrey, 2005). For example, in embarrassing consumption contexts, such as when customers are purchasing sensitive or private items (e.g., condoms, lubricant), employee avatar presence is likely to make customers feel uncomfortable and self-conscious, with negative downstream effects on aspects such as impressions of the retail space, consumer spending, and impulse buying (e.g., Grace, 2007; Verbeke and Bagozzi, 2003). The authors propose a moderating impact of the consumption context (embarrassing versus ordinary) on the link between employee avatar presence and consumer responses, when compared with employee avatar absence. Thus:
H2: The impact of employee avatar presence (versus absence) on consumer responses is moderated by the consumption context (embarrassing versus ordinary), such that negative employee avatar presence effects only occur in embarrassing shopping settings but not in more ordinary settings.
Prior research has also documented that consumers perform worse on cognitive tasks and have greater difficulties with focusing in the presence of others (Donthu, Cherian, and Bhargava, 1993; Esmark Jones et al., 2020; Rosenbaum et al., 2021), while simultaneously being more prone to rely on heuristics in their decision-making under such social circumstances (Otterbring, 2021; Skalski and Tamborini, 2007). Consumers who shop in the presence of avatar employees, thus, might be less likely to recall the brand of an embarrassing product on sale, given that the additional social element may interfere with such memory-based functions. In the specific case of recalling whether an embarrassing product was promoted in the virtual store, however, consumers shopping in the presence of avatar employees might potentially outperform their counterparts who shop without such additional social elements, considering the established tendency of consumers to increasingly rely on heuristics (e.g., purchasing promoted products and perceiving such products as bargains) under conditions of social presence (Otterbring, 2021; Wästlund, Otterbring, Gustafsson, and Shams, 2015). Because of the lack of explicit findings related to the aforementioned possibilities, however, the authors refrain from specifying formal hypotheses. Instead, they sought to explore the following bipartite research question:
RQ: Does the presence of avatar employees influence consumers’ recall rates in embarrassing shopping settings? If so, will the effect be different for consumers’ ability to recall (a) the brand of an embarrassing product on sale relative to (b) the embarrassing product that was on sale?
STUDY 1: VIRTUAL SHOPPING IN AN EMBARRASSING SETTING
Method
Participants. A total of 111 participants (age: M = 26.1 years, SD = 8.3; age range = 18–67 years; 54.1 percent female and 45.9 percent male), recruited through the behavior lab panel at a large university in Northern Europe, completed all parts of the study. Participants were prescreened under the conditions of speaking English fluently, being legally allowed to work, being at least 18 years old, and having reported shopping for groceries at least occasionally.
Design and Stimuli. The study was laboratory based, with a total of three between-subjects conditions: employee presence (human avatar versus robotic avatar) and employee absence (control). The two employee presence conditions, featuring either a human avatar or a robotic avatar, actively promoted an embarrassing product (condoms) at a promotional stand at the entrance of the virtual store included in the study. The depiction of employee avatars was manipulated through visual presence of animated nonplayer characters in conjunction with an audible cue in the form of an artificial intelligence (AI)-generated recording of a human voice or robotic voice (See the accompanying Web Appendix for more detailed information). The two employee avatars were used to test social presence effects related to both human and robotic employees, given that human employees are increasingly replaced by service robots (Frank and Otterbring, 2023b).
The virtual store, used in a series of consumer behavior experiments (Frank, Peschel, Otterbring, et al., 2024; Peschel, Frank, Blumenkranz, et al., 2024), was designed to be as naturalistic as possible, fully equipped with shelves, freezers, and a wide range of grocery products (foods, drinks, snacks, sweets, dairy, meat, alcohol, etc.). The store was populated with a total of four additional nonplayer character avatars representing other customers, animated with looped movements to enhance overall realism and ecological validity (cf. Morales, Amir, and Lee, 2017; Otterbring, 2023b).
Procedure. At the beginning of each session, participants were given the shopping scenario, which instructed them to shop in the virtual store in preparation for a romantic surprise dinner for their significant other. Specifically, they were instructed to buy two missing items: an item from the intimates section, “to enhance the passionate after-dinner activities,” and a bar of dark chocolate. Participants received 200 Denmark kroner (200 DKK; approximately US$30) to purchase these items. (See Table A1 for a full description of the scenario.) After confirming that the shopping task was understood, participants were then introduced to the virtual reality equipment (controller, headset, etc.) to ensure that they had an equal understanding of the placement of the thumbstick, buttons, and triggers before obscuring their vision with the headset. Participants had assistance when putting the device on to ensure a consistent fit on their face. If they wore glasses, they were instructed, on the basis of their own comfort level, to make a determination as to what gave them the clearest level of detail. No participant chose to put their glasses on or take them off once committing to their initial decision.
Before the task commenced, every participant engaged in an untimed, interactive tutorial featuring a sequence of guided steps. This tutorial aimed to acquaint them with the fundamental controls, including teleportation within the environment, handling and evaluating various items, and utilizing the virtual shopping basket. The purpose was to ensure that participants gained familiarity with the task of shopping in the metaverse retail store (Meißner, Pfeiffer, Peukert, et al., 2020; Roberts and Grassi, 2021) and to reduce the novelty that might be experienced when exposed to the timed aspect of the experience. Only successful completion of these three functions allowed them to conclude the tutorial.
After completion, and before “entering” the virtual store, participants were asked to restate the shopping task that they had been given. In case they were unsure or stated the task incorrectly, they were reminded and refamiliarized with the instructions of their task. Because participants verbally repeated the task back, the authors ensured that they understood the instructions even after the feeling of immersion had been reached.
Next, participants began their shopping task by selecting “Start,” which allowed freedom of movement within the shopping scenario. Participants could teleport around the store, interact with products on the shelves, evaluate what they wanted to purchase, and make decisions according to their preferences within the scenario framework. Product information (brand, logo, package size, price, etc.) for each item was clearly viewed and easily identifiable, alongside the price and promotional communication that was needed to complete the task accurately within the budget constraints of the scenario. When participants were finished, they had to teleport themselves to a large “exit” sign, and they would then leave the virtual store, concluding the virtual shopping part of the study. The duration of the study was approximately 20 minutes, of which participants spent approximately 2.5 minutes in the virtual retail store.
Measures. The study captured both behavioral and self-reported measures, described in further detail as follows.
Time of Purchase Decision. This behavior-dependent variable measured the total number of seconds it took until participants purchased the embarrassing product during their shopping trip (as captured at the precision of milliseconds). The timer started once participants entered the store and stopped when the embarrassing product was placed in their virtual shopping cart.
Number of Products Purchased. The construct for number of products purchased measured the total number of products that were purchased during the entire virtual shopping trip. This number reflects the count of all items that participants left the virtual store with in their shopping cart after confirming through a virtual interface prompt that they would like to check out and leave the store (the alternative being to continue shopping).
Spending. The spending construct measured the total basket price, which was calculated by summarizing the price of all products participants left the store with, while factoring in whether said products were on a sales promotion or not.
Impulse Purchases. The impulse purchases construct conceptualized as the propensity of participants purchasing additional products beyond the two they were instructed to buy (i.e., condoms and dark chocolate), such that two purchased products reflected no impulse purchases, whereas three or more purchased products reflected impulse purchases. As each participant was exposed twice to the instructions—once at the beginning and once after the feeling of immersion had been reached—novelty was limited, and a sense of familiarity had been obtained at the time of performing the shopping task in the virtual store. For this reason, impulse purchases and spending results are less likely a departure from the instructions but, rather, serve as a facet of consumer behavior related to spontaneity and impulsiveness.
Recall Rates. Two recall measures were captured as part of the study. First, brand recall measured the ability of consumers to remember a brand after being presented with a cue (Alba and Chattopadhyay, 1986). As a way of capturing how well the promoted brand was memorized and recognized by consumers, brand recall was measured by asking participants to list any names of the brand of the promoted product into an empty text field. Recall was coded as 1 if participants listed the promoted brand as one of the brands and 0 otherwise (Besharat, Kumar, Lax, and Rydzik, 2013). Second, to capture whether participants correctly identified that the embarrassing product was promoted, participants responded to a follow-up question, indicating whether the recalled brand was promoted (or not).
Because of the high fidelity of the virtual reality headset that participants used, details related to the task could be easily obtained.
Perceived Embarrassment. This construct measured participants’ perceived embarrassment during their virtual shopping trip to the metaverse store through a postexperience questionnaire on a self-report scale ranging from 1 (not at all) to 7 (extremely) (Frank and Otterbring, 2023b), as further elaborated in the Discussion section of Study 1.
Analysis. The data obtained from the researchers’ session notes, the behavioral virtual reality data, and the postexperience questionnaires completed by participants were combined before analysis. The merged dataset was then cleaned to eliminate any missing values, ensuring the data were accurate and complete. Owing to unforeseen software-related issues during the virtual store deployment, the sample size was reduced to 85 complete responses for the behavioral outcomes. (See the accompanying Web Appendix for details regarding this unforeseen issue, as well as sample size justifications and statistical power considerations across studies.)
As the human (n = 40) and robotic (n = 21) employee conditions did not differ significantly on any of the dependent variables (all ps > .20), these conditions were combined into a joint employee avatar presence condition and then compared with the employee avatar absence condition (n = 24 for the behavioral part, and n = 50 for the self-report part) to facilitate parsimonious analyses and boost statistical power (Frank and Otterbring, 2023a). The nature and significance of all tested variables remained unchanged when the data were analyzed using planned contrasts, in which both the employee presence conditions were compared with the employee absence condition as the first contrast, and the two employee presence conditions were separately compared against one another as the second contrast.
Results
A series of independent samples t tests and a Pearson’s chi-square test were conducted to examine whether employee avatar presence influenced the time (in seconds) that participants took to decide which embarrassing product to purchase, as well as the number of products purchased, the amount of money they spent in the virtual store, and the likelihood of making impulse purchases (H1a–H1d).
First, in partial support of H1a, the authors found a marginally significant and moderately sized difference regarding the time to participants’ product purchase decision in the predicted direction, t(83) = 1.67, p = .099; d = 0.40. Consumers in the employee avatar presence condition (M = 52.34 seconds, SD = 36.27; n = 61) thus made their initial purchase decision related to the embarrassing product more rapidly than their counterparts in the employee avatar absence condition (M = 66.92 seconds, SD = 36.33; n = 24).
Second, supporting H1b, the authors found a large and statistically significant difference in the number of products purchased, t(83) = 3.00, p = .004; d = 0.72. Consumers in the employee avatar presence condition (M = 2.34, SD = 0.60; n = 61) purchased fewer products than those in the employee avatar absence condition (M = 2.79, SD = 0.66; n = 24).
Third, in line with H1c, there was a large and highly significant effect on money spent, t(83) = 3.53, p < .001; d = 0.85. Of their assigned budget of 200DKK (approximately US$30), consumers in the employee avatar presence condition (M = 120.61DKK, SD = 38.42; n = 61) spent less money in the store than those in the employee avatar absence condition (M = 153.66DKK, SD = 40.06; n = 24).
Fourth, consistent with H1d, the chi-square analysis using a 2 (Condition: employee avatar presence versus absence) × 2 (Impulse Purchases: yes versus no) cross-tabulation revealed a large and significant effect, χ2(1, N = 85) = 9.91, p = .002; V = .33. Thus, the proportion of consumers who made impulse purchases was significantly lower in the employee avatar presence condition (29.5 percent; n = 61) than in the absence condition (66.7 percent; n = 24).
As the data were not normally distributed regarding the authors’ time, quantity, and amount variables (i.e., time to product decision, number of products purchased, and money spent), they supplemented the aforementioned analyses with their nonparametric counterparts and ran a series of Mann-Whitney U tests (See the Web Appendix for details). It is important that the findings remained largely unchanged. If anything, there was even stronger empirical evidence for all hypotheses using nonparametric tests (as the time effect became statistically significant rather than only marginally significant, whereas the quantity and amount metrics remained significant).
Regarding the recall measures used to address Research Question 1, a Pearson’s chi-square analysis using a 2 (Condition: employee avatar presence versus absence) × 2 (Condom Brand Recalled: yes versus no) cross-tabulation found a significant and moderately sized effect, χ2(1, N = 111) = 7.24, p = .007; V = .26. The proportion of consumers who correctly recalled the brand of the embarrassing product was significantly lower in the employee avatar presence condition (34.4 percent; n = 61) than in the absence condition (60.0 percent; n = 50). It is notable that a similar analysis on the likelihood of correctly recalling that the condom brand was on sale (yes versus no) yielded a significant and moderately sized effect in the opposite direction, χ2(1, N = 111) = 6.38, p = .012; V = .24. Accordingly, the proportion of consumers who correctly recalled that the brand of the embarrassing product was on sale was significantly higher in the employee avatar presence condition (16.4 percent; n = 61) than in the absence condition (2.0 percent; n = 50).
Discussion
Study 1 provides robust evidence for the notion that employee avatar presence negatively influences a series of key customer outcomes within a metaverse shopping context centered on purchasing an embarrassing product. Shopping for condoms in this study, however, yielded substantially lower embarrassment ratings (M = 2.11 on the 7-point scale, SD = 1.55) compared with the embarrassment ratings for condoms from previous research focusing on physical retail stores (e.g., Blair and Roese, 2018; Dahl et al., 2001; Moore, Dahl, Gorn, and Weiberg, 2006; Otterbring and Lu, 2018). This discrepancy underscores the divergence between findings in physical and virtual environments. To investigate whether the shopping context (as embarrassing or not) might moderate the study’s employee avatar presence results, and to further amplify embarrassment, the authors conducted a follow-up experiment (Study 2) utilizing a considerably more embarrassing product (adult diapers). This subsequent study also enabled them to rule out the possibility that their initial results regarding time and money spent, impulse purchases, and the number of products purchased could be attributed to participants’ disregard for instructions (as participants in Study 1 were solely instructed to purchase two products, although a large proportion of them ended up purchasing three or more products). Using a different embarrassing product also means that the authors tested the generalizability and external validity of their findings, as called for by several scholars (Borau, Otterbring, Laprote, and Fosso Wamba, 2021; Loebnitz, Fank, and Otterbring, 2022; Winer, 1999).
STUDY 2: REPLICATION AND MODERATION
In Study 2, the authors sought to replicate and extend the focal social presence effect of employee avatars on consumers’ purchase responses while simultaneously testing whether this effect is moderated by the consumption context as embarrassing or not. Specifically, they examined whether the negative social presence effect of employee avatars only occurs in embarrassing shopping settings, but not in more “mundane” (nonembarrassing) shopping settings. Study 2 was designed, furthermore, to address a series of potential limitations from Study 1.
Study 1 provides robust evidence for the notion that employee avatar presence negatively influences a series of key customer outcomes within a metaverse shopping context centered on purchasing an embarrassing product.
First, Study 1 did not use a formal manipulation check for social presence. Although many researchers perceive such manipulation checks as redundant when manipulating objectively different information across conditions (e.g., O’Keefe, 2003; Perdue and Summers, 1986), which was the case in the initial study, a critic might still claim that the absence of a manipulation check means that one cannot ascertain that their manipulation worked as intended. Second, the authors have yet to demonstrate whether their focal employee avatar presence effect generalizes across consumption contexts or only emerges in embarrassing shopping settings. Third, although the authors deem it unlikely that all their obtained effects on time spent, number of products purchased, money spent, and the likelihood of making impulse purchases from Study 1 solely emerged as a function of participants not following instructions (as many of them purchased more products than they were instructed to buy), they cannot explicitly rule out this potential confound. Study 2, therefore, used an improved design that effectively mitigated these issues, thus ensuring rigor, control, and high internal validity.
Method
Participants. A total of 169 U.S. consumers (age: M = 38.1 years, SD = 10.6, age range = 19–78 years; 50.3 percent female and 49.7 percent male) completed the experiment in full. Participants were recruited through a well-established online panel and prescreened such that they had a high command of spoken English, were allowed to legally work, were at least 18 years old, and had reported being primary shoppers for groceries.
Design and Stimuli. The study was an online experiment with a total of six between-subjects conditions: social presence (none versus human avatar versus robot avatar) and embarrassment (yes versus no).
The authors pretested 10 product categories in terms of how embarrassing U.S. consumers (N = 100; age: M = 36.7 years, SD = 11.4; 50 percent female) found purchasing these products when shopping in a supermarket (See Table A2 in the Web Appendix). Participants rated each product on a scale ranging from 1 (“not at all embarrassing”) to 7 (“extremely embarrassing”). The product category that was the most embarrassing was adult incontinence products—specifically, adult diapers (M = 4.43, SD = 2.02)—which scored significantly above the scale midpoint of 4, t(99) = 2.13, p = .035. The product category that was the least embarrassing was ingredients—specifically, pizza sauce—which scored way below the scale midpoint (p < .001), approaching the minimum value (M = 1.02, SD = 0.14).
In-store advertisements were designed with the two selected products (adult diapers and pizza sauce) placed into a scene where all branding and recognizable features to known products were digitally edited (See Figure A3 in the Appendix). The scene was constructed to include a familiar yet unrecognizable aisle that one would typically find in any supermarket throughout the United States. A two-dimensional promotional display, which was a close visual replica of one used in Study 1, was placed in the aisle. The selected promotional products were placed under a promotion sign to call attention to the products. The human social presence condition was designed with the help of an AI-generated female avatar. For the robotic presence condition, a robot stimulus was used from the established literature (Frank and Otterbring, 2023b). The avatars were placed in the scene behind the promotional display to manipulate presence.
To enhance perceived presence (cf. Pizzi et al., 2020) and to stay as close as possible to the manipulation used in Study 1, the authors added AI-generated promotional voices to complement the visual scenes. The female (or robotic) voice greeted participants with the line: “Hey, welcome to the store. Would you like to purchase some adult diapers (pizza sauce)? They’re on sale today.” The audible cue accompanied was played once, on participants entering the respective two-dimensional scene in the experiment on their personal devices. Participants were reminded to turn up the volume on their audio-capable devices before entering the scene and were allowed to manually play the voice if they wanted to hear it again.
Procedure. At the outset of the session, participants received instructions prompting them to envision themselves in a shopping scenario, emphasizing the necessity of having their audio set to a comfortable level for study completion. Their assigned task involved imagining a situation after a minor medical procedure, wherein their doctor recommended either wearing adult diapers (embarrassing condition) or staying home (nonembarrassing condition) for the subsequent two weeks (See the Web Appendix for a full scenario description). On the basis of the assigned condition, participants continued the scenario by envisioning shopping for adult diapers (embarrassing condition) or procuring a missing ingredient (pizza sauce) for cooking one of their favorite dishes (nonembarrassing condition). After the scenario presentation, participants navigated to a page with the two-dimensional representation of the store aisle featuring the promoted product together with the audible cue. Subsequently, each participant completed a questionnaire where the following self-reported measures and demographic details were collected.
Measures. The primary dependent variable was a single-item measure: “How likely is it that you would purchase the promotional item in this situation?” Participants indicated their response on a 7-point scale ranging from 1 (“extremely unlikely”) to 7 (“extremely likely”). Such single-item measures are valid if they are clear and unambiguous (Bergkvist and Rossiter, 2007; Otterbring, 2020), as in the present case.
In addition, the authors captured perceived embarrassment through a three-item, 7-point bipolar rating scale adapted from Dahl et al. (2001), encompassing the dimensions of embarrassment, ranging from 1 (“not embarrassed at all”) to 7 (“very embarrassed”); discomfort, ranging from 1 (“not uncomfortable at all”) to 7 (“very uncomfortable”), and awkwardness, ranging from 1 (“not awkward at all”) to 7 (“very awkward”). These items demonstrated high reliability (Cronbach’s α = 0.94) and were merged into a composite score representing perceived embarrassment, which served as the authors’ manipulation check of the consumption context (as embarrassing or not).
The authors also assessed perceived social presence through a single-item measure: “Please indicate the extent to which you perceived yourself to be in the presence of others (e.g., humans or robots) at the promotional stand where you found the promoted product.” Participants rated this perception on a 5-point scale ranging from 1 (“not at all”) to 5 (“a great deal”). The authors used participants’ responses on this item as a manipulation check of avatar social presence.
Analyses. There were no discernible differences between human (n = 55) and robotic (n = 58) employee avatars across any of the authors’ focal measures, including the manipulation checks (See the Web Appendix for details). They, therefore, merged these conditions into a unified social presence condition to facilitate parsimonious analyses, consistent with the approach used in Study 1. To test for moderation, the authors performed a 2 (Condition: employee avatar presence versus absence) × 2 (Consumption Context: embarrassing versus nonembarrassing) between-subjects analysis of variance (ANOVA) on participants’ willingness to buy their assigned product. Both their manipulations were effective and were verified by their manipulation checks (See the Web Appendix for the manipulation check analyses).
Results
A 2 (Employee Avatar: presence versus absence) × 2 (Embarrassment: yes versus no) between-subjects ANOVA on willingness to buy found a significant and negative effect of employee avatar presence when compared with employee avatar absence (MPresence = 4.24, SD = 1.15, versus MAbsence = 4.88, SD = 1.74), F(1, 165) = 10.28, p = .002; ·2p = .06. Of central importance for the present investigation, although the main effect of embarrassment was nonsignificant, F(1, 165) = 0.35, p = .553; η2p < .01; the two-way interaction emerged, F(1, 165) = 3.95, p = .048; η2p = .02 (See Figure 2).
Willingness-to-Buy Scores across Conditions
Follow-up pairwise comparisons revealed that there was no statistically significant difference in participants’ willingness to buy the product between the absence and presence conditions in the nonembarrassing shopping setting (MPresence.No = 4.15, SD = 1.94, n = 53, versus MAbsence.No = 4.52, SD = 1.65, n = 27), t(78) = 0.84, p = .402; d = 0.19. Conversely, in the case of the embarrassing shopping setting, the presence of an employee avatar had a strong and significant negative effect, resulting in a reduction of more than 1.5 scale points in willingness to buy the product (MPresence.Yes = 3.53, SD = 2.30, n = 60, versus MAbsence.Yes = 5.21, SD = 1.78, n = 29), t(87) = 3.45, p < .001; d = 0.74.
Last, to enhance statistical robustness of their focal comparison (i.e., employee avatar presence in the embarrassing shopping setting compared with all other conditions combined), the authors supplemented their pairwise comparisons (shown earlier) with a planned contrast analysis, as contrast analysis on the entire dataset has considerably higher statistical power than simple pairwise comparisons on a subset of the data (Otterbring and Folwarczny, 2024; Rosnow and Rosenthal, 1996). In further support of their theorizing, participants’ willingness to buy the product in the presence of an employee avatar in the embarrassing condition (MYes.Presence = 3.53, SD = 2.30; n = 60) was significantly lower than in all other conditions (MYes.Absence No.Presence No.Absence = 4.52, SD = 1.86; n = 109), t(167) = 3.04, p = .003; d = 0.49.
Finally, the authors ascertained that the results pertaining to the pairwise comparisons and the planned contrast analysis were also robust to nonparametric analyses using Mann-Whitney’s U test, which was consistently found to be the case (See the Web Appendix for details).
Discussion
The results of Study 2 conceptually replicate the focal employee avatar presence effect on consumers’ purchase responses found in Study 1. The authors also find that this effect is moderated by the shopping setting, such that the negative employee avatar presence effect only emerges under embarrassing conditions but not in more regular (nonembarrassing) shopping settings.
GENERAL DISCUSSION
This research investigated the impact of employee avatar presence on consumers’ shopping behavior and recall rates in the metaverse, with the empirical work situated around sales promotions taking place in virtual shopping settings. Building on social impact theory and impression management literature, the authors hypothesized and found that employee avatar presence, relative to absence, negatively affects consumers’ shopping responses in terms of time and money spent, the number of products purchased, and the likelihood of making impulse purchases, as well as their willingness to buy in embarrassing shopping settings but not in more mundane (nonembarrassing) shopping settings. As such, their focal employee avatar presence effect was moderated by the consumption context, as embarrassing or not. In embarrassing shopping settings, furthermore, they also explored and documented that employee avatar presence differentially affected consumers’ recall rates depending on whether the recall measure focused on branding or promotional aspects. Although employee avatar presence generally reduced the recall rates of an embarrassing brand promoted in the virtual store, consistent with prior literature (Esmark Jones et al., 2020; Luck and Benkenstein, 2015; Rosenbaum et al., 2021), such presence increased consumers’ inclination of recalling that the embarrassing product was promoted. This latter finding supports recent studies, which indicate that consumers have a stronger tendency to rely on heuristics, such as perceiving promotions as persuasive, in the mere presence of others (Otterbring, 2021).
Theoretical Contributions
This research is the first of its kind, to our knowledge, to test the impact of employee avatar presence on behavioral and recall-based outcomes related to sales promotions of embarrassing products in a metaverse shopping setting. As such, the present work makes a novel contribution to the emerging stream of literature on consumer behavior in the metaverse (Giang Barrera and Shah, 2023; Jacobsen, Krogsgaard-Jensen, and Peschel, 2022; Meißner et al., 2020; Peschel, Jacobsen, Frank, and Steinmann., 2022) regarding the use and role of avatars in shaping consumer responses (Holzwarth et al., 2006; Miao, Kolzenkova, Wang, et al., 2022).
In addition, the authors extend the existing knowledge on social presence, as previously primarily researched in physical retail stores (Baker et al., 1994; Eroglu et al., 2001), particularly with respect to active employee sales promotions (Argo and Dahl, 2020; Söderlund and Berg, 2019), including those administered by nonliving (robotic) entities (Holthöwer and van Doorn, 2023; Frank and Otterbring, 2023b).
At a more general level, the results contribute to retail-relevant advertising research in that most of the authors’ tested variables are behavioral rather than self-reported. This is important, given that most empirical investigations in advertising, retailing, and marketing are restricted to self-reported responses, such as consumers’ stated attitudes, intentions, and preferences, with only a small minority of studies capturing actual behavioral outcomes (Gneezy, 2017; Otterbring, Sundie, Li, and Hill, 2020; Pham, 2013; Simester, 2017). By contrast, the present work combined objective behavioral outcomes, such as time spent on purchase decisions, with a complex research environment by means of highly immersive virtual reality technology (Burke, Harlam, Kahn, and Lodish, 1992; Cowan and Ketron, 2019b; Pizzi, Scarpi, Pichierri, and Vannucci, 2019), which is unusual in this stream of literature. As a result, the authors’ findings should be more ecologically valid than those from studies solely using self-report measures, because of the realistic virtual settings and the behavioral data used to support their theorizing (Jacobsen et al., 2022; Morales et al., 2017; Otterbring et al., 2023a).
When it comes to metaverse shopping behavior, the results of the present research suggest that virtual worlds are truly immersive and social (Giang Barrera and Shah, 2023). More precisely, the findings indicate that avatars can, indeed, induce the same social presence effects in a metaverse sales promotion context as prior studies have found in physical commercial settings (Argo and Dahl, 2020), which could not be taken for granted before the present studies were conducted, for reasons delineated in the study’s conceptual background. It is important to note that the social presence effects are not restricted to living entities in the form of human employee avatars but generalize to the mere virtual presence of nonliving entities, analogous to the service robots that are increasingly replacing human workforce in the retail space (Frank and Otterbring, 2023b). This finding goes against recent academic work, which finds robot presence to induce different effects than human presence in embarrassing consumption contexts (Holthöwer and Van Doorn, 2022; Pitardi et al., 2022).
One explanation for the discrepancy between the findings from the present research and those from former studies might be the immersiveness of the virtual world, which puts users into a state where they become more accepting of fiction (Huang and Klippel, 2020). This allows them to perceive employee avatars, whether living or nonliving, as being more social and present. This explanation also resonates with research that has shown the metaverse to be a space where users expect to connect, communicate, and engage in a diverse range of activities resembling human presence (Giang Barrera and Shah, 2023; Hennig-Thurau et al., 2023).
A final theoretical contribution of the present research is the finding that participants in the employee avatar presence condition got better at noticing that the embarrassing product was on sale while getting worse in their ability to recall the actual name of the brand of the promoted embarrassing product. The improvement in recalling that the promoted product was, indeed, on sale can be attributed to consumers’ increased reliance on heuristics when shopping in the presence of others (e.g., Skalski and Tamborini, 2007) combined with the possibility that it might be easier to justify their potential purchase of an embarrassing product if it is on sale. The finding, however, that consumers became worse in recalling the promoted brand supports the notion that shoppers are generally less attentive during their shopping trip when unknown others are around and can watch their purchases (Esmark Jones et al., 2020; Rosenbaum et al., 2021).
Practical Implications
Implementation of advertisements in a new medium implies many challenges. As society’s understanding is developed, however, with respect to what the metaverse is and how it can be utilized from a consumer standpoint, companies have greater opportunities for advertising pioneers to craft and develop strategies to perfection to yield the best return on investment. The present work indicates that the presence of employee avatars with any degree of humanization can significantly affect consumers’ purchase behavior and recall rates in the metaverse. When being faced with designing metaverse shopping experiences, therefore, marketers must decide whether the inclusion of virtually represented avatar employees or even anthropomorphic embodiments produce the intended consumer responses. The present research calls for caution when it comes to including such social cues in virtual consumption contexts that risk resulting in embarrassment, such as when shopping for condoms or buying adult diapers.
As the studied employee avatar presence effects extend from purchase behavior to recall rates linked to promoted embarrassing products, the authors’ findings suggest that the presence of employee avatars (both human depictions and robotic illustrations) in the metaverse negatively influences consumers’ purchase behavior and brand recall. Accordingly, advertisers who seek to make brands more memorable in the metaverse should consider strategies that minimize the presence of avatar employees during consumers’ decision-making process in embarrassing shopping settings. This could involve changing store layouts to reduce employee visibility, providing self-service interface options, or using alternative methods to persuasively position brands.
Consumers in the employee avatar presence condition made their initial purchase decision of condoms more rapidly compared with those in the employee avatar absence condition. When viewed in light of the embarrassment shopping setting, these “speedy” decisions might indicate uninformed purchases, possibly resulting in regret and postpurchase dissatisfaction. Advertisers can leverage this insight by carefully selecting how to communicate embarrassing product information that educates consumers in the metaverse (e.g., by minimizing social cues in connection to products perceived as embarrassing).
Given that consumers in the presence of employee avatars made more rapid purchase decisions and purchased fewer products, both in general and on impulse, than those in the employee avatar absence condition, the presence of employee avatars might deter customers from exploring or considering additional products in embarrassing consumption contexts. Hence, marketers may face challenges in upselling or cross-selling when employee avatars are present, potentially resulting in lower sales and missed revenue opportunities. This creates a real burden on how advertisers can use computer-mediated experiences to attract customers (Spielmann and Orth, 2021). That is, if consumer exploration is stifled due to employee avatar presence in embarrassing shopping settings, this can lead to a reduction in impulse purchasing. Marketers who operate in domains where impulse purchases, such as hedonic products and services (e.g., snacks, soda, and candy brands) are particularly important may struggle to stimulate immediate sales in the presence of employee avatars, requiring different strategies to capture consumer interest and boost sales. Indeed, several hedonic products that involve sensory pleasure might also induce consumer embarrassment (e.g., sex toys), particularly so in the real or implied presence of others in the metaverse, implying that strategies for embarrassing products should focus on reducing social presence, even in these virtual shopping environments. This may be achieved by providing private session options to remove the friction caused by employee avatars in the context of embarrassing products in virtual stores.
Given that consumers in the presence of employee avatars made more rapid purchase decisions and purchased fewer products, both in general and on impulse, the presence of employee avatars might deter customers from exploring or considering additional products in embarrassing consumption contexts.
Limitations and Future Research
In this article, the authors document the impact of employee avatar presence in metaverse shopping settings as well as the moderating role of the consumption context (embarrassing or nonembarrassing). To shed further light on the employee avatar presence effects, it holds value to delve into the underlying psychological mechanisms that drive their observed effects. One potential mediator might relate to situational self-awareness (Govern and Marsch, 2001). This concept, which is based on three distinct dimensions—private self-activation, public self-activation, and surroundings self-activation—captures the capacity of consumers to shift their attentional focus between the environment and themselves. It is notable that, whereas consumer self-consciousness is typically considered a stable trait, private and public self-awareness are situation-specific states that are susceptible to manipulation (Carver and Glass, 1976). Further research on the dynamics of situational self-awareness could uncover why employee avatar presence in embarrassing shopping settings yields less favorable consumer responses and why such negative effects do not occur in more ordinary settings that do not induce embarrassment.
Another unexplored avenue for future research involves investigating the potential moderating effect of whether the employee avatar is controlled by a human or an automated system (e.g., generative artificial intelligence; Frank, Jacobsen, Søndergaard, and Otterbring, 2023). In the present work, despite an immersive appearance and use of voiceovers, no differences emerged in consumer responses as a function of whether the avatar was a human employee or a robot, leading the authors to combine these conditions into a joint employee avatar presence condition. Future research could explore whether the introduction of an additional layer of automation would alter consumers’ purchase behavior, potentially finding differences between these conditions depending on whether they feature a human-controlled avatar or an automated avatar.
One potential concern is that the second study was conducted as a slightly more simplistic scenario-based experiment, which was not performed in a virtual reality environment. It should be noted, however, that care was taken to boost perceived realism, even in this latter study, considering the auditory stimuli used (i.e., promotional voices) combined with the image-supported scenarios featuring a typical supermarket environment with a promotional display closely resembling the one used in Study 1. The effect size obtained for the “willingness to buy” comparison between the employee avatar presence and absence conditions in the embarrassing shopping setting from Study 2 does not differ significantly from the effect size for the “money spent” comparison between these conditions in Study 1 (Z = 0.33, p = .741). For this reason, and considering the realistic setup in Study 2, the authors are confident that their findings are generalizable to a more explicit virtual reality context. They call, nevertheless, for further metaverse studies to verify these tentative claims.
A final avenue for future research stems from the quantitative nature of the authors’ studies and their unknown generalizability to other consumption contexts. This research takes an important first step in uncovering the impact of human and robotic employee avatars on consumers’ purchase behavior and recall rates in virtual shopping settings, and future research should explore whether the negative effects of employee avatar presence on consumer behavior extend to different consumption contexts (e.g., watching a movie, shopping for services) and endure over extended periods of time. To this end, a longitudinal exploration would contribute to a comprehensive understanding of the lasting impact of such experiences on consumer responses. Using mixed-methods approaches that compare the dynamics across virtual, online, and offline shopping settings, preferably both qualitatively and quantitatively, could provide additional valuable insights (Pizzi et al., 2020).
ABOUT THE AUTHORS
Darius-Aurel Frank is an assistant professor of marketing at the Department of Management, Aarhus University, Denmark. His research focuses on life with technology in the future, and in particular, consumer behavior related to the adoption and use of AI, intelligent automation, service robots, virtual reality, and the metaverse. Frank’s work has appeared in journals such as Nature Human Behaviour, Technological Forecasting & Social Change, Psychology & Marketing, and the British Journal of Management.
Jason DiPalma is a PhD candidate at the Department of Management, Aarhus University, Denmark. His primary research interests encompass retail, the metaverse, and mixed reality. He aims to uncover valuable insights that shed light on emerging domains and the intersection of the physical and digital realms. DiPalma is part of the commercial and retail management group, which focuses on skills and knowledge in the consumer goods industry. He has worked on multiple research projects involving virtual reality.
Sascha Steinmann is professor of commerical and retailing management in the Department of Management, Aarhus University, Denmark. His research focuses on the causes and consequences of consumer behavior in omnichannel environments. Steinmann has published articles in international journals, including Journal of Business Research, Psychology & Marketing, Business and Society, Journal of Retailing & Consumer Services, and International Review of Retail, Distribution and Consumer Research.
Tobias Otterbring is professor of marketing in the Department of Management, University of Agder, Norway, and a member of the Young Academy of Norway. His research focuses on how the real, imagined, or implied presence of others influences consumers’ cognitions, emotions, and actual purchase or choice behaviors. Otterbring’s work has appeared in journals such as Nature Human Behaviour, Nature Communications, and PNAS, Journal of Marketing Research, Organizational Behavior and Human Decision Processes, British Journal of Management, Journal of Experimental Social Psychology, and Journal of Experimental Psychology: Applied.
- Received July 17, 2023.
- Received (in revised form) January 5, 2024.
- Accepted February 16, 2024.
- Copyright © 2024 ARF. All rights reserved.
REFERENCES
Vol 64 Issue 3
