Next Article in Journal
Diverse Counterfactual Explanations (DiCE) Role in Improving Sales and e-Commerce Strategies
Previous Article in Journal
Instagram Mega-Influencers’ Effect on Generation Z’s Intention to Purchase: A Technology Acceptance Model and Source Credibility Model Perspective
 
 
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Configuration Path Analysis of the Virtual Influencer’s Marketing Effectiveness

by
Min Tian
,
Haiqiang Hu
and
Meimei Chen
*
Glorious Sun School of Business and Management, Donghua University, Shanghai 200051, China
*
Author to whom correspondence should be addressed.
These authors contributed equally to this work and have co-first authorship.
J. Theor. Appl. Electron. Commer. Res. 2025, 20(2), 95; https://doi.org/10.3390/jtaer20020095
Submission received: 21 March 2025 / Revised: 27 April 2025 / Accepted: 29 April 2025 / Published: 8 May 2025

Abstract

:
As an emerging marketing tool, virtual influencers (VIs) have attracted increasing scholarly attention. However, existing research predominantly adopts linear causal analysis models, which fail to capture the complex, nonlinear interaction between consumers and VIs. Grounded in the 5W communication model and utilizing a fuzzy-set qualitative comparative analysis (fsQCA), this study systematically explores how different configurational paths influence consumer engagement, drawing on empirical data from 205 participants. The findings reveal that (1) the synergy of entertainment, information, and credibility is a core prerequisite for achieving high engagement; (2) two equivalent paths—namely, the technology-driven path (media richness + content synergy) and the cognition-driven path (technology acceptance + content synergy)—lead to high engagement, both with a solution consistency of 0.98; and (3) the joint absence of content, media richness, and audience cognition results in low engagement. Theoretically, this study challenges traditional linear approaches by validating causal asymmetry and revealing configurational interdependencies among communication elements. It also extends the Media Richness Theory (MRT) and the Technology Acceptance Model (TAM) into the context of virtual influencer (VI) marketing. Practically, the proposed dynamic configuration model offers marketers a novel framework for optimizing VI campaigns through resource-adaptive strategies.

1. Introduction

With the rapid advancement of generative AI and virtual reality technologies, virtual influencers (VIs) have evolved beyond traditional marketing tools. They are reshaping the brand communication paradigm in the digital era [1]. Headline companies, such as Meta and L’Oreal, are allocating up to 30% of their marketing budgets to virtual influencers. Luxury brands, such as Prada and Porsche, have reported a 175% increase in young user reach through the use of virtual spokespersons [2]. Despite this industry boom, academic research on the marketing effectiveness of VIs remains limited.
On the one hand, compared to human influencers (HIs), virtual influencers are computer-generated, possess customizable personas, and operate independently of real-life constraints [3]. While human influencers rely on authentic self-representation and personal charisma, VIs offer consistent, always-on engagement [4]. However, despite these technological advantages, VIs are often perceived as less trustworthy due to their artificial nature [5]. Existing studies tend to generalize findings from HIs to VIs without adequately addressing these structural differences. But recent studies have shown that VIs and HIs elicit markedly different consumer responses, particularly in terms of trust formation, perceived autonomy, and emotional attachment [6]. These findings underscore the need to treat VIs as a distinct category of communicators, as applying conclusions from HI studies may misrepresent the underlying communication dynamics. To address this gap, the present study examines VIs as unique communicators whose effectiveness is shaped by the interplay of information content, media, and audience cognition.
On the other hand, most existing literature adopts a linear analytical paradigm and overlooks the nonlinear synergies among key communication elements. This has led to inconsistent and paradoxical findings in practice. For example, human-like virtual influencers perform well in beauty-related promotions [7] but generate trust issues in public service contexts [5]; similarly, while high anthropomorphism enhances emotional connection [8], it can also trigger discomfort due to the “uncanny valley” effect when overused [9].
These paradoxes highlight the limitations of the traditional isolation research. This is despite Shao’s call for attention to the combinatorial relationship of factors that influence the marketing effectiveness of virtual influencers [10]. The current literature remains fragmented: it is either limited to the single dimension of content–emotion [11] or stops at the static analysis of characteristics such as technology acceptance [12] and fails to systematically unpack the dynamic network of interactions among the communicator, content, media, and audience. Furthermore, the multifactor framework proposed by Sorosrungruang et al., based on the 5W communication model, has yet to be empirically tested [2]. Traditional regression models are inherently limited in capturing the combinatorial effects of multiple interdependent conditions [13].
To address these gaps, this study adopts a fuzzy-set qualitative comparative analysis (fsQCA), which transcends the cognitive boundaries of linear thinking, to conduct a configuration path analysis of the marketing effectiveness of virtual influencers. The research focuses on two core questions: (1) what are the necessary conditions for effective virtual influencer marketing? (2) What equivalent configurations lead to high or low user engagement? By developing a synergistic fit model based on key communication elements, this paper contributes to theory in three ways. First, it identifies entertainment, information, and credibility as core content conditions and demonstrates their dynamic interplay with media channels and audience characteristics. Second, it validates causal asymmetry and functional substitutability within the 5W communication model, particularly between media richness and technology acceptance. Third, it shows that the simultaneous absence of content, media richness, and audience cognition results in minimal engagement, revealing a threshold effect.
The remainder of this paper is structured as follows: Section 2 integrates the 5W communication model with usage satisfaction theory to construct a configurational research framework. Section 3 presents the calibration procedures, fsQCA methodology, and resulting configuration paths. Section 4 discusses the theoretical contributions, practical insights, and future research directions for virtual influencer marketing.

2. Literature Review and Research Framework

2.1. Marketing Effectiveness of Virtual Influencers

As a core metric for measuring the effectiveness of social media marketing, user engagement reflects the depth and continuity of the interaction between consumers and brands [14]. In the context of virtual influencer (VI) marketing, engagement is typically expressed through behaviors such as liking, commenting, and sharing—actions that embody both emotional involvement and behavioral responses from consumers [15]. Previous studies have shown that high engagement not only fosters brand loyalty [16] but also indirectly enhances purchase intentions [17] through trust accumulation [18] and reputation reinforcement [19]. For example, Brazilian virtual influencer Lu do Magalu reportedly earned up to $33,000 from a single post due to high-frequency user interaction [20]. This example demonstrates the direct correlation between high engagement and business value conversion. Therefore, increasing user engagement has become a common goal of social media marketing [17].
From a communication perspective, user engagement is viewed as the foundation of communication activities [21]. Key elements in this process—including the communicator, content, channel, and audience—interact dynamically to shape how users respond to virtual influencers [22]. Therefore, this study aims to explore how these communication elements work in combination to influence user engagement with VIs, thereby offering both theoretical insight and practical guidance for virtual influencer marketing strategies.

2.2. 5W Communication Model for Virtual Influencers

Similar to other communication models, firms deliver value to consumers through virtual influencer marketing to maximize firm profits [23]. It has been shown that the 5W communication model still has applicability in virtual influencer marketing [21]. It clearly describes the communication process by which information flows from the virtual influencer to the audience [22]. However, most studies have been limited to the impact of virtual influencers’ characteristics and content on marketing effectiveness, ignoring the role of audience characteristics and channels [4]. As shown in Table 1, most of the existing studies analyze the effects based on only a single or dual element of the 5W communication elements, failing to capture the synergistic interactions among all elements. In this regard, Shao emphasizes that the relationship between consumers and VIs is not linear but rather the result of complex interactions [10]. Therefore, based on the 5W communication model, this paper comprehensively examines the elements of the communicator, content, channel, and audience and explores their combined effects on the marketing effectiveness of virtual influencers to fill the theoretical gap in existing research. Specifically, virtual influencers are categorized as human-like, anime-like, and non-human (communicators); entertainment, information, and credibility (content); high or low media richness (channels); and high or low technology acceptance (audiences), as shown in Figure 1.

2.2.1. Communicator Characteristics of Virtual Influencers

Past research on virtual influencer characteristics has primarily focused on three dimensions: sensory, social, and functional attributes. Specifically, sensory features refer to the ability of virtual influencers to interact with users through visual, auditory, etc. Among these, appearance anthropomorphism [7,37,38,39] and aesthetic features [36] are the focus areas of research. Social features focus on the social distance between consumers and virtual influencers [40,41,42], social contexts [3,38,43] and patterns of social interaction [4,11,44]. Functional features studies cover product category suitability [27,45,46], algorithmic transparency [29], technical implementation, and cost-effectiveness [28,47].
Among these dimensions, visual salience cues are a significant driver of marketing effectiveness [48]. Accordingly, this study focuses on the visual aspect of sensory features, particularly the degree of anthropomorphism in virtual influencers. Although the effects of anthropomorphism on consumer cognition, emotion, and behavior have been widely discussed, the findings remain inconsistent. For example, Yan found that anime-like and non-human virtual influencers elicited greater emotional attachment and perceptions of authenticity than human-like virtual influencers [31]; however, other studies have demonstrated that human-like virtual influencers are more advantageous in enhancing purchase intentions and user engagement [4,7,8]. These contradictory results suggest that the effects of different VI types (human-like, anime-like, and non-human) have not been fully explained. Therefore, this paper attempts to reconcile the divergence of existing studies by analyzing the combination effects of different types of virtual influencers and revealing their optimal marketing combination strategies in various contexts.

2.2.2. Content Characteristics of Virtual Influencers

Prior studies have identified various content characteristics of virtual influencers, including sponsorship disclosure [49], the content form [3], the image composition, and taglines [21]. However, these studies primarily focus on the objective features of the content itself, while largely overlooking the subjective user experience and perceived attributes of such content. According to the uses and gratifications theory, consumer experience and perceived satisfaction are crucial [50]. Therefore, this paper explores the psychological antecedents of users’ engagement with virtual influencers, verifying the existence of these antecedents and their differential impact on different types of virtual influencers.
Specifically, the uses and gratifications theory suggests that people choose external information according to their needs and desires [51]. Within this framework, entertainment, information, and credibility have been consistently recognized as three core dimensions influencing user engagement [50,52]. Entertainment refers to the extent to which a virtual influencer’s content can satisfy consumers’ needs for aesthetic enjoyment or emotional release. Information pertains to the completeness and richness of information about the product or service provided by the virtual influencer. Credibility reflects to the authenticity and trustworthiness of the virtual influencer’s content [53,54]. Based on this theoretical foundation, this paper focuses on users’ perceived entertainment, information, and credibility of the virtual influencer content to assess their combined effects on user engagement.

2.2.3. Media Characteristics of Virtual Influencers

It has been shown that there is a significant positive relationship between virtual influencers’ media richness and user engagement. This finding aligns with media richness theory, which states that media richness significantly enhances users’ cognitive, affective, and behavioral responses to media content [8]. However, other studies suggest the opposite: low media richness may result in virtual influencers with low morphological authenticity, but according to expectancy confirmation theory, consumers automatically lower their expectations of the behavioral competence of virtual influencers with low morphological authenticity. In such cases, if virtual influencers have high behavioral authenticity (e.g., through story settings or human-like expressions), they can instead enhance users’ emotional attachment [34]. This paradox suggests that the impact of media richness on marketing effectiveness may depend on the type of virtual influencer.
To address this complexity, the present study categorizes media richness into two levels—high and low—and employs configuration analysis to explore their differential effects in various VI contexts. Specifically, high media richness platforms can provide multi-dimensional sensory stimuli (e.g., visual, auditory, etc.), enhancing users’ expectations of human-like VIs and their marketing effectiveness. Virtual influencers on low-media-richness platforms have lower morphological authenticity (e.g., non-human virtual influencers) but can still effectively enhance user engagement through high behavioral authenticity (e.g., narrativized content or human-like interactions). Through this framework, we attempt to reveal the fitness relationship between media richness and virtual influencer types, providing a theoretical basis for virtual influencer marketing strategies in different media environments.

2.2.4. Audience Characteristics of Virtual Influencers

Prior research has shown that audience characteristics such as age [21], cultural background [29,55], racial consistency [32], and personality traits [30] all influence user engagement with virtual influencers. For example, Generation Z has higher engagement with virtual influencers [21], which may be related to Generation Z’s high acceptance of emerging technologies [56]. Based on this insight, this paper hypothesizes that Gen Z’s high engagement with virtual influencers may stem partly from their high technology acceptance. To test this hypothesis, this paper categorizes technology acceptance into high and low categories to explore its impact on user engagement. Our study not only helps to reveal the role of technology acceptance in virtual influencer marketing but also provides further explanations and additions to previous studies.

3. Methodology

3.1. Data Collection

In this study, the designed questionnaire was distributed via the Questionnaire Star platform, where participants filled it out online to complete the data collection. The survey was conducted from January to February 2025. The questionnaire utilized an audience recall method [57]. Filter questions were set at the beginning of the questionnaire, with criteria focusing on whether participants had heard of virtual influencers and whether they had engaged in interactions with them. Eligible respondents were invited to answer questions related to this study, such as engagement with virtual influencers, information, entertainment, credibility, media richness, and technology acceptance. To further ensure the authenticity and reliability of the data, a time limit was imposed for answering the questions. Based on the pre-simulated time spent, questionnaires with a response time of less than 100 s were classified as invalid. A total of 240 questionnaires were collected, of which 28 were excluded for non-compliance with the screening criteria (based on the filter questions and time spent), and 7 were incomplete. Consequently, 205 questionnaires were considered valid and used for subsequent analysis. The demographic information of the participants is presented in Table 2. All measurements were derived from established research scales, with specific references and questions detailed in Table 3.
It is worth noting that the sample primarily consisted of Chinese Gen Z consumers, who tend to exhibit high digital fluency and openness to virtual technologies. While this demographic is relevant to virtual influencer research, it may not fully represent broader consumer populations across different age groups or cultural contexts. Future studies should consider expanding the sampling scope to enhance generalizability.

3.2. Data Calibration and Reliability Test

The calibration of the variables of the scales in this paper was performed using the direct calibration method [58]. The measured values were converted to fuzzy scores of 0–1 based on the “Calibrate (0.95, 0.5, 0.05)” method of the fsQCA3.0 software and further analyzed. Specifically, the scale itself provides information on the calibration of these variables, so that “full affiliation points” (7), “intersection points” (4), and “full non-affiliation points” (1) were set [59]. The virtual influencer’s appearance characteristics are categorized into 3 classes, and a 3-valued fuzzy set is constructed for calibration in this paper. None of the three are calibrated to 0, non-human VIs are calibrated to 0.33, anime-like VIs are calibrated to 0.67, and human-like VIs are calibrated to 1. For media richness and technology acceptance, dichotomization was applied by coding values ≥ 5 as “high” and < 5 as “low”. This cut-off was determined based on the empirical distribution (median split) and the theoretical threshold above which respondents express clear agreement. This approach aligns with prior fsQCA applications [60] and balances parsimony with interpretability.
In addition, this paper used SPSS19.0 software to analyze the reliability of the scale. The Cronbach reliability coefficient (α), combined reliability (CR), and average extracted variance (AVE) indicators are shown in Table 3. The α values of the variables were all greater than 0.7, indicating that this scale has high reliability. The AVE values of the variables are all higher than 0.5, and the CR values are all greater than 0.7, so this scale has high convergent validity. Meanwhile, the square root values of AVE exceeded the correlation coefficients between the corresponding variables, so this scale also has good discriminant validity.
Table 3. Scale items, reliability, and validity analysis.
Table 3. Scale items, reliability, and validity analysis.
Measurement ItemSpecific ItemLoadCronbach’s αC.RAVE
VI content characteristics [50,52]EntertainmentI think the virtual influencer is fun and interesting0.7840.7810.7040.877
I found the virtual Influencer to be enjoyable0.803
I found the virtual influencer interesting0.924
InformationI think the virtual influencer is a good source0.8800.8150.7330.891
I think the virtual influencer is a good channel0.828
I think the virtual influencer can provide information0.859
CredibilityI trust the influencer0.8590.8510.7890.918
I think the influencer is reliable0.882
I find the virtual influencer’s information convincing0.924
VI media characteristics (high/low richness) [8]I can provide and receive timely feedback through the virtual influencer0.9060.7730.8210.902
I can interact with the virtual influencer via text, audio, video, etc.0.906
VI audience characteristics (high/low technology acceptance) [61]I have a favorable attitude towards the virtual influencer0.9030.7240.8150.901
I am receptive to virtual influencers0.903
Engagement [8]I am interested in virtual influencers0.9300.8250.7680.908
I will pay attention to virtual influencers0.875
I can communicate and interact with virtual influencers0.820

3.3. Necessary Condition Analysis

In this paper, we use a fuzzy set qualitative comparative analysis (fsQCA) to explore the marketing effectiveness of virtual influencers based on a configuration perspective. fsQCA is a configurational approach that emphasizes multiple factors combined to influence results [60]. For example, different combinations of conditions can achieve the same outcome, and a variable may be effective in one combination of conditions but ineffective in another. In this study, we selected the communication features of virtual influencers and mapped them to the corresponding 5W communication model. Specifically, virtual influencers are categorized as human-like, anime-like, and non-human (who); entertainment, information, and credibility of content (says what); high or low media richness (in which channel); high or low technology acceptance (to whom); and engagement (what effect).
Before proceeding with the analysis of the condition configurations, the researcher must assess the “necessity” of each condition individually. In this paper, we evaluate whether a single condition (along with its nonsets) serves as a necessary condition for the marketing effectiveness of a virtual influencer. Consistency is a crucial test of a necessary condition; when consistency exceeds 0.9, the condition is deemed necessary for the outcome. As shown in Table 4, the consistency levels of entertainment, information, and credibility of the virtual influencer content all exceed 0.9, indicating that they are necessary conditions for generating high levels of engagement.

3.4. Conditional Configuration Analysis

Unlike the analysis of necessary conditions described above, a configuration analysis attempts to reveal the adequacy of the results of different configurations consisting of multiple conditions. The consistency level of adequacy should not be less than 0.75 [60]. Depending on the specific research context, different consistency thresholds have been used, such as 0.76 by scholar Zhang Ming and 0.8 by Cheng Cong [62,63]. The frequency threshold should be determined according to the sample size, with a frequency threshold of 1 for small and medium samples and greater than 1 for large samples [60]. In this study, considering the minimum number of cases required for the solution and the sample size of this paper, the consistency threshold was finally determined to be 0.80 and the frequency threshold to be 2, which meets the general criteria for a qualitative comparative analysis. Based on the above settings, simple and intermediate solutions were calculated using fsQCA3.0 software. The results of high- and low-engagement configuration paths were plotted and obtained as shown in Table 5 and Table 6.
Table 5 presents the different configuration paths for the marketing effects of high engagement, where the solution consistency is 0.98, which means that 98% of all cases of virtual influencer marketing effects that satisfy this type of conditional configuration show high engagement. The solution coverage is 0.84, which means that this type of conditional configuration explains 84% of the cases of high engagement marketing. The consistency and coverage of the solutions are higher than the critical value, indicating that the empirical analysis is valid.
Specifically, conditional configuration 1 suggests that virtual influencers will have higher engagement when their marketing content is entertaining, informative, and credible and when media richness is high. This path explains about 81% of the high engagement. In addition, about 5% of the high-engagement cases can only be explained by this path. Conditional configuration 2 suggests that virtual influencers will have high engagement when their content is entertaining, informative, and credible, as well as when audience technology acceptance is high. This path explains about 79% of the high-engagement marketing effectiveness cases. In addition, about 2% of high-engagement marketing effectiveness cases can only be explained by this path.
Table 6 presents the different configuration paths for the low engagement, where the solution consistency is 0.99, which means that 99% of all cases of virtual influencers that satisfy this type of conditional configuration show low engagement. The solution coverage is 0.66, which means that this type of conditional configuration explains 66% of the cases of low engagement. The consistency of the solutions and the coverage of the solutions are both higher than the critical value, indicating the validity of the empirical analysis.
Specifically, conditional configuration 1 suggests that virtual influencers will have lower levels of engagement when their marketing content is less entertaining, informative, and credible, when media richness and when audience technology acceptance are low. This path explains about 66% of the low-engagement cases. In addition, about 66% of the low-engagement cases can only be explained by this path.

3.5. Robustness Test

In QCA research, the robustness test of the analyzed results is necessary. Given that QCA is an ensemble theory method, this paper refers to the method of the QCA robustness test by prior scholars, changing the consistency threshold to conduct the robustness test [60]. In the robustness test, a stricter threshold is used to analyze the results, and the consistency threshold is increased from 0.80 to 0.85. The results show that there is a completely consistent combination of conditions, coverage, and consistency level with the original model, which indicates that the results of the study are robust.

4. Discussion

4.1. Research Conclusions

Using a fuzzy-set qualitative comparative analysis (fsQCA), this study identifies distinct configurational paths that drive the marketing effectiveness of virtual influencers. Between the two core configurations associated with high engagement (Table 5), the synergy between content attributes (entertainment, information, and credibility) and either high media richness (configuration 1) or high audience technology acceptance (configuration 2) emerges as a critical determinant. These configurations exhibit a solution consistency of 0.98 and a coverage of 0.84, indicating strong explanatory power. This finding complements previous linear studies: it validates the boundary effect of “product sensory attributes and virtual spokesperson suitability” proposed by Zhou et al. [1]. And it also echoes Volles et al.’s emphasis that virtual influencer uniqueness (e.g., novelty and customization) needs to be amplified by the channel [64]. The low-engagement configuration (Table 6) suggests that when both content attributes and media technology conditions are at a low level, it is difficult to increase the engagement, even if they possess anthropomorphic features. This explains the phenomenon found by You and Liu that the “autonomy” advantage of virtual influencers in social advocacy may be transformed into a credibility disadvantage if they lack real interactions [6].

4.2. Theoretical Contributions

This study offers several important theoretical contributions. First, our findings align with existing theories and reconcile previous anthropomorphism research. On the one hand, the technology-driven path resonates with Media Richness Theory, which posits that richer media enhance information processing [8]. The cognition-driven path mirrors the Technology Acceptance Model, where perceived usefulness and ease of use foster trust in AI-driven interactions [56]. On the other hand, the two equivalent paths explain prior contradictory findings: for instance, while human-like VIs thrive in high-media-richness contexts [8], another type of VI succeeds when paired with high-quality content [21], resolving the anthropomorphism debate through configurational logic.
Second, we challenge the traditional linear causal paradigm and reveal the mechanism of elemental synergy. Most prior studies have been based on the isolated effects of variables (e.g., exploring the service failure response mechanism [24]); this study finds that the effect of entertaining content needs to be realized by combining credibility and media conditions. Thereby, we respond to Sorosrungruang et al. call for “frameworks need to integrate the technology, content, and relationship dimensions” [2]. Furthermore, our configuration analysis validates the interaction logic of the 5W communication model. For instance, our findings show that low media richness can be effectively compensated by high audience technology acceptance, providing new insights into functional substitutability across elements. These results lay a foundation for the future development and the empirical refinement of virtual influencer research.
Third, these findings demonstrate the theoretical utility of fsQCA in reconciling fragmented empirical outcomes across contexts, thereby advancing scenario-based theorization in VI research. Previous research has indicated that virtual influencers excel in certain contexts [1], such as high-tech product demonstrations, but are less effective in scenarios requiring emotional resonance, like service marketing [65]. These seemingly contradictory findings stem from different success conditions across contexts: some campaigns prioritize technological immersion, while others rely more on content quality and emotional appeal. By identifying multiple equifinal configurations that lead to high engagement, our study offers a clear framework for selecting optimal strategies tailored to specific scenarios.

4.3. Practical Implications

The business strategy based on the configuration path can be designed at two levels. The first is the content optimization layer: building a “trinity” communication matrix. Businesses need to prioritize the entertainment-information-credibility synergy (a necessary core requirement). For example, virtual influencer’s narratives should be embedded with fun interactions (e.g., gamified tasks) to enhance entertainment [64], third-party authenticated sources to enhance credibility [35], and real-time scenario simulations to provide in-depth information [1].
The second is the technology adaptation layer: choose the path of configuration based on resource endowment. Enterprises with sufficient resources can adopt technology-driven strategies: investing in augmented reality (AR) or virtual image simulation technology to improve media richness [65]. For example, Gucci launched AR glasses that allow consumers to virtually try on shoes. This technology-driven strategy has dramatically increased engagement. Technology-constrained enterprises can adopt cognition-driven strategies: increase technology acceptance through progressive audience education. For example, Samsung has managed to win consumers’ trust in its technology by releasing an “AI Technology Transparency White Paper” that explains the fairness of its algorithms. All in all, in choosing between the technology-driven and cognition-driven strategies, marketers should weigh their resource availability, campaign goals, and audience characteristics.

4.4. Limitations and Prospects

While this study provides critical insights into the configuration paths of virtual influencer marketing, certain limitations must be acknowledged. First, the study sample consisted exclusively of Gen Z consumers in China. While this group is highly relevant to the virtual influencer phenomenon due to their digital nativeness and high technology acceptance, the findings may not generalize to older cohorts or international markets with different cultural perceptions of AI and virtual entities. Specifically, older age groups may have more conservative attitudes toward virtual technologies and their trust mechanisms and media use behaviors differ [21]. In addition, cultural values (such as collectivism in China versus individualism in the West) may largely influence consumer responses to virtual influencers. For example, perceptions of anthropomorphism or media richness may be interpreted differently depending on social norms and technological familiarity [10]. Therefore, future research is encouraged to replicate and extend this study using cross-cultural and multi-generational samples to enhance the external validity and robustness of the configuration paths identified here.
Additionally, the reliance on self-reported engagement and a single platform (online surveys) introduces potential biases, such as response bias or sampling bias, which may limit the generalizability of the findings. Future studies could incorporate multiple data sources or objective measures to enhance validity.
Finally, the current study adopts a dichotomous calibration for constructs such as media richness and technology acceptance, in line with the conventions of fsQCA to ensure model simplicity and consistency. While this approach is methodologically sound for our sample size and configurational logic, future research could explore multi-value or continuous calibrations to capture more granular variations in user perceptions and platform characteristics. Such refinement may further enhance the explanatory power and flexibility of fsQCA in analyzing virtual influencer effectiveness.

Author Contributions

Conceptualization, M.T. and H.H.; methodology, M.T. and H.H.; validation, M.C.; formal analysis, H.H.; investigation, M.T.; data curation, M.T.; writing—original draft preparation, M.T.; writing—review and editing, M.C.; visualization, M.T.; supervision, M.C. All authors have read and agreed to the published version of the manuscript.

Funding

This research was not given any external financial support.

Institutional Review Board Statement

The study was conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee of Donghua University (protocol code SRSY202406150027 and 15 June 2024).

Informed Consent Statement

Informed consent was obtained from all subjects involved in the study.

Data Availability Statement

The data presented in this study are available on reasonable request from the corresponding author.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Zhou, X.; Yan, X.; Jiang, Y. Making Sense? The Sensory-Specific Nature of Virtual Influencer Effectiveness. J. Mark. 2023, 88, 84–106. [Google Scholar] [CrossRef]
  2. Sorosrungruang, T.; Ameen, N.; Hackley, C. How Real Is Real Enough? Unveiling the Diverse Power of Generative AI-Enabled Virtual Influencers and the Dynamics of Human Responses. Psychol. Mark. 2024, 41, 3124–3143. [Google Scholar] [CrossRef]
  3. Zhou, Q.; Li, B.; Li, H.; Lei, Y. Mere Copycat? The Effects of Human versus Human-like Virtual Influencers on Brand Endorsement Effectiveness: A Moderated Serial-Mediation Model. J. Retail. Consum. Serv. 2024, 76, 103610. [Google Scholar] [CrossRef]
  4. Xie-Carson, L.; Benckendorff, P.; Hughes, K. Not so Different after All? A Netnographic Exploration of User Engagement with Non-Human Influencers on Social Media. J. Bus. Res. 2023, 167, 114194. [Google Scholar] [CrossRef]
  5. Mo, Z.; Zhou, M. Don’t Like Them but Take What They Said: The Effectiveness of Virtual Influencers in Public Service Announcements. J. Theor. Appl. Electron. Commer. Res. 2024, 19, 2269–2288. [Google Scholar] [CrossRef]
  6. You, L.; Liu, F. From Virtual Voices to Real Impact: Authenticity, Altruism, and Egoism in Social Advocacy by Human and Virtual Influencers. Technol. Forecast. Soc. Change 2024, 207, 123650. [Google Scholar] [CrossRef]
  7. Wan, C.; Lee, D.; Ng, P. The Role of Anthropomorphism and Racial Homophily of Virtual Influencers in Encouraging Low- versus High-Cost pro-Environmental Behaviors. Psychol. Mark. 2024, 41, 1833–1853. [Google Scholar] [CrossRef]
  8. Sun, Y.; Zhong, Y.; Zhang, Z.; Wang, Y.; Zhu, M. How Technical Features of Virtual Live Shopping Platforms Affect Purchase Intention: Based on the Theory of Interactive Media Effects. Decis. Support Syst. 2024, 180, 114189. [Google Scholar] [CrossRef]
  9. Gutuleac, R.; Baima, G.; Rizzo, C.; Bresciani, S. Will Virtual Influencers Overcome the Uncanny Valley? The Moderating Role of Social Cues. Psychol. Mark. 2024, 41, 1419–1431. [Google Scholar] [CrossRef]
  10. Shao, Z. Understanding the Switching Intention to Virtual Streamers in Live Streaming Commerce: Innovation Resistances, Shopping Motivations and Personalities. J. Res. Interact. Mark. 2024, 19, 333–357. [Google Scholar] [CrossRef]
  11. Yao, R.; Qi, G.; Wu, Z.; Sun, H.; Sheng, D. Digital Human Calls You Dear: How Do Customers Respond to Virtual Streamers’ Social-Oriented Language in e-Commerce Livestreaming? A Stereotyping Perspective. J. Retail. Consum. Serv. 2024, 79, 103872. [Google Scholar] [CrossRef]
  12. Akhtar, N.; Hameed, Z.; Islam, T.; Pant, M.K.; Sharma, A.; Rather, R.A.; Kuzior, A. Avatars of Influence: Understanding How Virtual Influencers Trigger Consumer Engagement on Online Booking Platforms. J. Retail. Consum. Serv. 2024, 78, 103742. [Google Scholar] [CrossRef]
  13. Sukhov, A.; Friman, M.; Olsson, L.E. Unlocking Potential: An Integrated Approach Using PLS-SEM, NCA, and fsQCA for Informed Decision Making. J. Retail. Consum. Serv. 2023, 74, 103424. [Google Scholar] [CrossRef]
  14. Jones, A.R.; Lee, S.-E. Factors Influencing Engagement in Fashion Brands’ Instagram Posts. Fash. Pract. 2022, 14, 99–123. [Google Scholar] [CrossRef]
  15. Yu, J.; Egger, R. Color and Engagement in Touristic Instagram Pictures: A Machine Learning Approach. Ann. Tour. Res. 2021, 89, 103204. [Google Scholar] [CrossRef]
  16. Labrecque, L.I. Fostering Consumer–Brand Relationships in Social Media Environments: The Role of Parasocial Interaction. J. Interact. Mark. 2014, 28, 134–148. [Google Scholar] [CrossRef]
  17. Kumar, V.; Aksoy, L.; Donkers, B.; Venkatesan, R.; Wiesel, T.; Tillmanns, S. Undervalued or Overvalued Customers: Capturing Total Customer Engagement Value. J. Serv. Res. 2010, 13, 297–310. [Google Scholar] [CrossRef]
  18. Oliveira, A.B.d.S.; Chimenti, P. “Humanized Robots”: A Proposition of Categories to Understand Virtual Influencers. Australas. J. Inf. Syst. 2021, 25, 3223. [Google Scholar] [CrossRef]
  19. Song, S.; Park, S.; Park, K. Thematic Analysis of Destination Images for Social Media Engagement Marketing. Ind. Manag. Data Syst. 2020, 121, 1375–1397. [Google Scholar] [CrossRef]
  20. de Brito Silva, M.J.; de Oliveira Ramos Delfino, L.; Alves Cerqueira, K.; de Oliveira Campos, P. Avatar Marketing: A Study on the Engagement and Authenticity of Virtual Influencers on Instagram. Soc. Netw. Anal. Min. 2022, 12, 130. [Google Scholar] [CrossRef]
  21. Xie-Carson, L.; Magor, T.; Benckendorff, P.; Hughes, K. All Hype or the Real Deal? Investigating User Engagement with Virtual Influencers in Tourism. Tour. Manag. 2023, 99, 104779. [Google Scholar] [CrossRef]
  22. Lasswell, H.D. The Structure and Function of Communication in Society. Commun. Ideas 1948, 37, 136–139. [Google Scholar]
  23. Leung, F.F.; Gu, F.F.; Palmatier, R.W. Online Influencer Marketing. J. Acad. Mark. Sci. 2022, 50, 226–251. [Google Scholar] [CrossRef]
  24. Zheng, X.; Miao, M.; Zhang, L.; Huang, L. Standing out or Fitting in? How Perceived Autonomy Affects Virtual Influencer Marketing Outcomes. J. Bus. Res. 2024, 185, 114917. [Google Scholar] [CrossRef]
  25. Zhao, T.; Ran, Y.; Wu, B.; Wang, V.L.; Zhou, L.; Wang, C.L. Virtual versus Human: Unraveling Consumer Reactions to Service Failures through Influencer Types. J. Bus. Res. 2024, 178, 114657. [Google Scholar] [CrossRef]
  26. Zhang, Y.; Wang, X. The Effect of Virtual Anchor Appearance on Purchase Intention: A Perceived Warmth and Competence Perspective. J. Prod. Brand Manag. 2024, 34, 84–103. [Google Scholar] [CrossRef]
  27. Yu, T.; Teoh, A.P.; Bian, Q.; Liao, J.; Wang, C. Can Virtual Influencers Affect Purchase Intentions in Tourism and Hospitality E-Commerce Live Streaming? An Empirical Study in China. Int. J. Contemp. Hosp. Manag. 2024, 37, 216–238. [Google Scholar] [CrossRef]
  28. Yu, J.; Dickinger, A.; So, K.K.F.; Egger, R. Artificial Intelligence-Generated Virtual Influencer: Examining the Effects of Emotional Display on User Engagement. J. Retail. Consum. Serv. 2024, 76, 103560. [Google Scholar] [CrossRef]
  29. Yang, D.; Zhang, J.; Sun, Y.; Huang, Z. Showing Usage Behavior or Not? The Effect of Virtual Influencers’ Product Usage Behavior on Consumers. J. Retail. Consum. Serv. 2024, 79, 13859. [Google Scholar] [CrossRef]
  30. Yan, R.; Tang, Z.; Liu, D. Can Virtual Streamers Replace Human Streamers? The Interactive Effect of Streamer Type and Product Type on Purchase Intention. Mark. Intell. Plan. 2024, 43, 0263–4503. [Google Scholar] [CrossRef]
  31. Yan, J. The Effect of Different Types of Virtual Influencers on Consumers’ Emotional Attachment. J. Bus. Res. 2024, 177, 114646. [Google Scholar] [CrossRef]
  32. Xu, H.; Wu, Y. Do Virtual Endorsers Have a Country-of-Origin Effect? From the Perspective of Congruent Explanations. Technol. Forecast. Soc. Change 2024, 206, 123530. [Google Scholar] [CrossRef]
  33. Xie, J.; Wu, H.; Liu, K.; Cui, Y.; Zhang, X. Is Virtual Streamer Useful? Effect of Streamer Type on Consumer Brand Forgiveness When Streamers Make Inappropriate Remarks. J. Retail. Consum. Serv. 2024, 79, 103899. [Google Scholar] [CrossRef]
  34. Sun, L.; Tang, Y. Avatar Effect of AI-Enabled Virtual Streamers on Consumer Purchase Intention in e-Commerce Livestreaming. J. Consum. Behav. 2024, 23, 2999–3010. [Google Scholar] [CrossRef]
  35. Shao, Z. From Human to Virtual: Unmasking Consumer Switching Intentions to Virtual Influencers by an Integrated fsQCA and NCA Method. J. Retail. Consum. Serv. 2024, 78, 103715. [Google Scholar] [CrossRef]
  36. Song, X.; Lu, Y.; Yang, Q. The Negative Effect of Virtual Endorsers on Brand Authenticity and Potential Remedies. J. Bus. Res. 2024, 185, 114898. [Google Scholar] [CrossRef]
  37. Quach, S.; Cheah, I.; Thaichon, P. The Power of Flattery: Enhancing Prosocial Behavior through Virtual Influencers. Psychol. Mark. 2024, 41, 1629–1648. [Google Scholar] [CrossRef]
  38. Franke, C.; Groeppel-Klein, A. The Role of Psychological Distance and Construal Level in Explaining the Effectiveness of Human-like vs. Cartoon-like Virtual Influencers. J. Bus. Res. 2024, 185, 114916. [Google Scholar] [CrossRef]
  39. Kim, I.; Ki, C.-W.; Lee, H.; Kim, Y.-K. Virtual Influencer Marketing: Evaluating the Influence of Virtual Influencers’ Form Realism and Behavioral Realism on Consumer Ambivalence and Marketing Performance. J. Bus. Res. 2024, 176, 114611. [Google Scholar] [CrossRef]
  40. Ferraro, C.; Sands, S.; Zubcevic-Basic, N.; Campbell, C. Diversity in the Digital Age: How Consumers Respond to Diverse Virtual Influencers. Int. J. Advert. 2024, 43, 1342–1365. [Google Scholar] [CrossRef]
  41. Mrad, M.; Ramadan, Z.; Toth, Z.; Nasr, L.; Karimi, S. Virtual Influencers Versus Real Connections: Exploring the Phenomenon of Virtual Influencers. J. Advert. 2024, 54, 1–19. [Google Scholar] [CrossRef]
  42. Muniz, F.; Stewart, K.; Magalhaes, L. Are They Humans or Are They Robots? The Effect of Virtual Influencer Disclosure on Brand Trust. J. Consum. Behav. 2024, 23, 1234–1250. [Google Scholar] [CrossRef]
  43. Wu, R.; Liu, J.; Chen, S.; Tong, X. The Effect of E-Commerce Virtual Live Streamer Socialness on Consumers’ Experiential Value: An Empirical Study Based on Chinese E-Commerce Live Streaming Studios. J. Res. Interact. Mark. 2023, 17, 714–733. [Google Scholar] [CrossRef]
  44. Hu, H.; Ma, F. Human-like Bots Are Not Humans: The Weakness of Sensory Language for Virtual Streamers in Livestream Commerce. J. Retail. Consum. Serv. 2023, 75, 103541. [Google Scholar] [CrossRef]
  45. Dabiran, E.; Farivar, S.; Wang, F.; Grant, G. Virtually Human: Anthropomorphism in Virtual Influencer Marketing. J. Retail. Consum. Serv. 2024, 79, 103797. [Google Scholar] [CrossRef]
  46. Kim, H.; Park, M. Virtual Influencers’ Attractiveness Effect on Purchase Intention: A Moderated Mediation Model of the Product–Endorser Fit with the Brand. Comput. Hum. Behav. 2023, 143, 107703. [Google Scholar] [CrossRef]
  47. Xin, B.; Hao, Y.; Xie, L. Virtual Influencers and Corporate Reputation: From Marketing Game to Empirical Analysis. J. Res. Interact. Mark. 2024, 18, 759–786. [Google Scholar] [CrossRef]
  48. Bruce, N.I.; Becker, M.; Reinartz, W. Communicating Brands in Television Advertising. J. Mark. Res. 2020, 57, 236–256. [Google Scholar] [CrossRef]
  49. Kim, E.; Shoenberger, H.; Kim, D.; Thorson, E.; Zihang, E. Novelty vs. Trust in Virtual Influencers: Exploring the Effectiveness of Human-like Virtual Influencers and Anime-like Virtual Influencers. Int. J. Advert. 2024, 44, 453–483. [Google Scholar] [CrossRef]
  50. Shao, G. Understanding the Appeal of User-generated Media: A Uses and Gratification Perspective. Internet Res. 2009, 19, 7–25. [Google Scholar] [CrossRef]
  51. Katz, E.; Blumler, J.G.; Gurevitch, M. Uses and Gratifications Research. Public Opin. Q. 1973, 37, 509–523. [Google Scholar] [CrossRef]
  52. Papacharissi, Z.; Rubin, A.M. Predictors of Internet Use. J. Broadcast. Electron. Media 2000, 44, 175–196. [Google Scholar] [CrossRef]
  53. Brackett, L.K.; Carr, B.N. Cyberspace Advertising vs. Other Media: Consumer vs. Mature Student Attitudes. J. Advert. Res. 2001, 41, 23–32. [Google Scholar] [CrossRef]
  54. Wang, Y.; Sun, S. Examining the Role of Beliefs and Attitudes in Online Advertising. Int. Mark. Rev. 2010, 27, 87–107. [Google Scholar] [CrossRef]
  55. Zeng, W.; Ma, Z.; Chen, T. When and Why Do Consumers Resist Virtual Influencer Endorsement? The Role of Product Depth and Advertising Claim Types. Electron. Commer. Res. 2024, 24, 1–34. [Google Scholar] [CrossRef]
  56. King, W.R.; He, J. A Meta-Analysis of the Technology Acceptance Model. Inf. Manage. 2006, 43, 740–755. [Google Scholar] [CrossRef]
  57. Gioia S, Guiret M, Chen A, Tucker M, Gray BE, Vetter C, Garaulet M, Scheer FAJL, Saxena R, Dashti HS. How Accurately Can We Recall the Timing of Food Intake? A Comparison of Food Times from Recall-Based Survey Questions and Daily Food Records. Curr. Dev. Nutr. 2022, 6, nzac002. [CrossRef]
  58. Morgan, S.L. Redesigning Social Inquiry: Fuzzy Sets and Beyond. Soc. FORCES 2010, 88, 1934–1936. [Google Scholar] [CrossRef]
  59. Chi, M.M.; Zhao, J.; Li, Y.H.; Wang, W.J. The Achievement Mechanisms of the Platform Ambidexterity:A Fuzzy-set Qualitative Comparative Analysis. Nankai Bus. Rev. 2017, 20, 65–76. [Google Scholar] [CrossRef]
  60. Zhang, M.; Du, Y.Z. Qualitative Comparative Analysis (QCA) in Management and Organization Research: Position, Tactics, and Directions. Chin. J. Manag. 2019, 16, 1312–1323. [Google Scholar] [CrossRef]
  61. Zhu, X.F.; Lou, W.J. Risk Perception of the Public and Acceptance of Emerging Technologies: A Survey Experiment on Facial Recognition Technology. J. Manag. World 2024, 40, 118–132. [Google Scholar] [CrossRef]
  62. Zhang, M.; Chen, H.W.; Lan, H.L. Why Do Chinese Enterprises Completely Acquire Foreign High-Tech Enterprises—A Fuzzy Set Qualitative Comparative Analysis (fsQCA) Based on 94 Cases. China Ind. Econ. 2019, 4, 117–135. [Google Scholar] [CrossRef]
  63. Chen, C.; Jia, L.D. Research on the Driving Pattern of China’s Enterprise Cross-border M&As: A Qualitative Comparative Analysis. Nankai Bus. Rev. 2016, 19, 113–121. [Google Scholar] [CrossRef]
  64. Volles, B.K.; Park, J.; Van Kerckhove, A.; Geuens, M. How and When Do Virtual Influencers Positively Affect Consumer Responses to Endorsed Brands? J. Bus. Res. 2024, 183, 114863. [Google Scholar] [CrossRef]
  65. Meng, L.; Bie, Y.; Yang, M.; Wang, Y. The Effect of Human versus Virtual Influencers: The Roles of Destination Types and Self-Referencing Processes. Tour. Manag. 2025, 106, 104978. [Google Scholar] [CrossRef]
Figure 1. Research framework.
Figure 1. Research framework.
Jtaer 20 00095 g001
Table 1. Review of virtual influencer (VI) marketing literature (note: HI refers to human influencer).
Table 1. Review of virtual influencer (VI) marketing literature (note: HI refers to human influencer).
AuthorsResearch ContextVI FactorsContent FactorsMedia FactorsFollower FactorsOther FactorsOutcomesKey Findings
[3]Online survey with social media usersNot studiedImage vs. VideoNot studiedCultural differenceVI vs. HIEndorsement effectivenessEndorsements work better in video format and are influenced by different cultures.
[24]Online survey with social media usersPerceived autonomyNot studiedNot studiedNot studiedProduct type; VI vs. HIPurchase intentionPerceived autonomy and purchase intention of virtual influencers are negatively correlated but moderated by product type.
[25]Experiment with Instagram influencersNot studiedNot studiedNot studiedFamiliarityVI vs. HI Forgiveness propensity;
Punishment Intention
Consumers had a higher propensity to forgive virtual influencers, but familiarity had no significant effect.
[26]Online survey with social media usersAnimal-human-like or all-human-like virtual anchorNot studiedNot studiedCertainty of needsProduct typePurchase intentionAnimal–human mixing elicits higher purchase intentions, and high (low) certainty needs enhance purchase intentions through enhanced perceptual abilities (warmth).
[27]Survey 416 active viewers of VIs in THCLSNot studiedSource credibility ofvirtual influencersNot studiedNot studiedInfluencer–product
congruence
Purchase intentionSource credibility of virtual influencers positively affects purchase intentions, and influencer–product consistency strengthens the positive effect.
[28]Used 1028 pictures shared by Lil MiquelaFacial action unitNot studiedNot studiedNot studiedInfluencer–product
congruence
EngagementThe findings disclose the significance of happiness, sadness, disgust, and surprise in triggering user engagement when promoting diverse products with visually captivating content.
[6]Online survey with Instagram usersAutonomyNot studiedNot studiedNot studiedVI vs. HIDigital activism; Altruistic motivesThe advantages of virtual influencers in commercial marketing do not necessarily translate to, enhancing their role in, advocacy for social causes.
[11]Online surveys for college studentsHuman-like vs. animated Linguistic styleNot studiedNot studiedProduct typePurchase intentionThe positive impact of socially oriented language on purchase intentions is reinforced under both experience and search products when virtual anchors are human-like.
[29]Online surveys for college studentsHuman-like vs. animated VINot studiedNot studiedNot studiedProduct usage behaviorEngagementVirtual influencers demonstrating product use behaviors are more effective at increasing engagement, and human-like is more effective than anime-like.
[30]Online survey with social media usersNot studiedNot studiedNot studiedImplicit personalityVI vs. HI; product typePurchase intentionConsumers’ implicit personality variances also influence their willingness to accept virtual streamers.
[31]Online survey with social media usersMimic-human VI, Animated-human VI, and non-human VINot studiedNot studiedNot studiedSocial presenceEmotional attachment; benefit-seeking behaviorMimic-human VI has lower emotional attachment compared to the other two.
[32]Survey with campus networks, street stops, etcCountry-of-origin; anthropomorphismNot studiedNot studiedNot studiedVE image perception; product value perceptionWillingness to payWillingness to pay increases significantly when the product and VE country of origin are the same.
[33]Secondary data and situational experiments InfluencerNot studiedNot studiedNot studiedHPVS vs. RHS; brand reputationBrand forgivenessWhen influence is high, virtual anchors receive higher brand forgiveness.
[8]Online survey with VLSP usersAnthropomorphismNot studiedMedia richnessNot studiedNot studiedPurchase intentionDegree of anthropomorphism and media richness positively affect purchase intention.
[34]Online survey Form realism; behavioral realismNot studiedNot studiedNot studiedRelationship norm orientationPurchase intentionMorphological authenticity and behavioral authenticity interact to influence consumer purchase intention.
[10]Online survey Not studiedNot studiedNot studiedInnovation resistance; motivations; personalitiesNot studiedSwitching intentionUnveiled six configurations of arrangements, each characterized by a unique combination of causation.
[35]Online survey AI technology-like; human-like; social attributesNot studiedNot studiedPersonalitiesNot studiedSwitching intentionUnveiled six configurations of arrangements.
[36]Online survey Aesthetic imperfectionNot studiedNot studiedNot studiedmultiple brand endorsements; VI vs. HI Brand authenticityWhen endorsers are designed to be aesthetically imperfect, the negative effect of virtual endorsers on brand authenticity is attenuated.
[37]Online surveyAnthropomorphismFlatteryNot studiedNot studiedNot studiedProsocial behaviorWhen virtual influencers have a highly humanoid appearance, flattery enhances users’ perceptions of their authenticity, which in turn promotes prosocial behavior.
[21]Online survey with Instagram usersSource realnessImage composition and caption discourseNot studiedNot studiedNot studiedEngagementHumanlike VIs are preferred over 3D animated VIs and the least preferred influencers are 2D animated VIs. Pictures of scenes where the influencer does not exist are most popular. Users preferred rational discourse that provided a travel scenario.
[4]1112 user comments collected from 52 Instagram postsSource factorsContent factorsNot studiedNot studiedSource–content factorsEngagementUsers engage with non-human influencers for various reasons, including entertainment value, emotional connection, and educational content.
This paper205 questionnaires on online platformsHuman-like VI; anime-like VI non-human VIEntertainment information credibilityMedia richness (high/low)Technology acceptance (high/low)Different configuration pathsEngagementInformation synergy media richness and technology acceptance influence user participation
Table 2. Sample characteristics.
Table 2. Sample characteristics.
NameOptionsFrequencyPercentage
GenderFemale9354.63
Male11245.37
AgeLess than 18 years old125.85
19–24 years old9747.32
25–30 years old7938.53
31–40 years old157.32
Above 41 years old20.98
EducationHigh school and below2411.71
Specialized or undergraduate13565.85
Master’s degree and higher4622.44
IncomeLess than 2000 yuan6330.73
2001–5000 Yuan4421.46
5001–8000 Yuan4019.51
8001–11,000 Yuan3818.54
More than 11,000 Yuan209.76
Table 4. Consistency and coverage of single antecedent variables.
Table 4. Consistency and coverage of single antecedent variables.
Conditional VariableHigh EngagementLow Engagement
ConsistencyCoverageConsistencyCoverage
VI Communicator0.730.710.830.42
~VI Communicator0.400.820.420.44
Entertainment0.910.920.380.20
~Entertainment0.210.400.840.83
Information0.920.940.340.18
~Information0.190.360.880.86
Credibility0.910.930.350.18
~Credibility0.190.360.860.84
Media Richness0.920.920.360.19
~Media Richness0.190.370.840.84
Technology Acceptability0.900.910.410.21
~Technology Acceptability0.220.420.830.81
Table 5. Configuration of high engagement.
Table 5. Configuration of high engagement.
ConditionalConfiguration 1Configuration 2
VI Communicator
Content CharacteristicsEntertainment
Information
Credibility
Media Richness
Audience Technology Acceptance
Consistency0.980.99
Raw Coverage0.810.79
Unique Coverage0.050.02
Solution Coverage0.84
Solution Consistency0.98
Note: ⬤ or ● indicates that the condition exists, ⬤ indicates a core condition, and ● indicates a marginal condition. A blank indicates that the condition may or may not exist.
Table 6. Configuration of low engagement.
Table 6. Configuration of low engagement.
ConditionalConfiguration 1
VI Communicator
Content CharacteristicsEntertainment
Information
Credibility
Media Richness
Audience Technology Acceptance
Consistency0.97
Raw Coverage0.66
Unique Coverage0.66
Solution Coverage0.66
Solution Consistency0.99
Note: ⊗ indicates that the condition does not exist, ⊗ indicates a core condition.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Tian, M.; Hu, H.; Chen, M. Configuration Path Analysis of the Virtual Influencer’s Marketing Effectiveness. J. Theor. Appl. Electron. Commer. Res. 2025, 20, 95. https://doi.org/10.3390/jtaer20020095

AMA Style

Tian M, Hu H, Chen M. Configuration Path Analysis of the Virtual Influencer’s Marketing Effectiveness. Journal of Theoretical and Applied Electronic Commerce Research. 2025; 20(2):95. https://doi.org/10.3390/jtaer20020095

Chicago/Turabian Style

Tian, Min, Haiqiang Hu, and Meimei Chen. 2025. "Configuration Path Analysis of the Virtual Influencer’s Marketing Effectiveness" Journal of Theoretical and Applied Electronic Commerce Research 20, no. 2: 95. https://doi.org/10.3390/jtaer20020095

APA Style

Tian, M., Hu, H., & Chen, M. (2025). Configuration Path Analysis of the Virtual Influencer’s Marketing Effectiveness. Journal of Theoretical and Applied Electronic Commerce Research, 20(2), 95. https://doi.org/10.3390/jtaer20020095

Article Metrics

Back to TopTop