Exploration of the Critical Factors Influencing the Development of the Metaverse Industry Based on Linguistic Variables

Abstract

Recently, the development of the Metaverse has emerged as a pivotal concern within both industrial and academic realms. The future development of the Metaverse industry is shrouded in uncertainty, complexity, and a dearth of technical and economic information. To address these challenges, this paper integrates the fuzzy Delphi method and fuzzy DEMATEL based on linguistic variables to explore the critical factors of the Metaverse industry. In accordance with the proposed methodology, a case study is presented to explore the critical factors of the Metaverse industry in Taiwan. The results of the empirical analysis demonstrated that the order of importance for the three principal dimensions is as follows: “infrastructure”, “consumer behavior”, and “user experience”. From the perspective of causality, “infrastructure” can be considered a driving dimension, whereas “user experience” can be regarded as a passive dimension. Regarding the critical factors, it can be observed that “virtual and real integration”, “equipment lightweight”, and “network communication” act as driving factors, exhibiting a high degree of correlation with the advancement of the Metaverse industry. Therefore, the proposed method not only possesses a robust theoretical foundation but also offers tangible practical value in the real world.

1. Introduction

In recent years, there has been a notable progression and evolution in information technology, leading to the maturation of Internet-based technologies and applications [1]. The advent of the smartphone has enabled individuals to swiftly access the Internet, engage in real-time online communication, participate in virtual meetings, conduct searches, access multimedia content, and participate in gaming activities at any time and from any location [2].

With the rapid development of network technology, network applications are often not cross-platform, and their interfaces remain limited to computers and mobile devices, despite the powerful functions they offer. Recently, there has been a rapid development of software and hardware technologies, including 5G networks, cloud computing, virtual reality/augmented reality, the Internet of Things, blockchain technology, virtual currency NFTs, artificial intelligence, and digital twinning. The aspiration of establishing an integrated platform that bridges the virtual and tangible realms is poised to be actualized. Consequently, the concept of the Metaverse has been developed by integrating the virtual and the real worlds [3,4]. The overarching objective of the Metaverse is the integration of the virtual and the real worlds. This can be achieved through the utilization of virtual reality (VR), augmented reality (AR), and mixed reality (MR) technologies.

Despite the existence of VR/AR technologies for a considerable period, the applications of virtual reality remain largely confined to the consumer, healthcare, educational, and entertainment sectors [4,5,6]. As indicated by data from the International Data Corporation (IDC), the global market for augmented reality and virtual reality (AR/VR) headsets are expected to decline 8.3% in 2023. However, the IDC predicted that the shipments of AR/VR headsets will grow in the future [7]. Despite the introduction of new generations of VR headsets, consumers remain reluctant to purchase them due to their bulky size, inconvenience in transportation, and high cost [8].

It can be reasonably asserted that the factors affecting the development of the Metaverse industry include the advancement of software and hardware technologies. The accuracy and update rate of sensors may be insufficient to prevent eye fatigue, dizziness, and other issues that could impair the user’s immersive experience. In the future, regardless of the platform ecology, hardware requirements, infrastructure, content presentation, and other factors, there is a need for constant innovation. The Metaverse industry will continue to expand in line with improvements in technological standards [9,10]. In order to understand the development of the Metaverse industry, it is essential to investigate the pivotal factors influencing its evolution. The technologies and products of the Metaverse industry are still in the initial stage of development [11]. There are numerous factors affecting the development of the Metaverse industry, which are characterized by high market uncertainty, a high complexity of product development and use, and a paucity of technical and economic information [12]. Considering these circumstances, it is challenging for managers to effectively identify the key factors influencing the evolution of the Metaverse industry and the interrelationships among these factors.

In the initial stage of industry development, a group of experts can provide a valuable and pragmatic assessment of emerging technologies and market trends, offering insights that can help navigate the inherent uncertainties in this early stage of growth. Therefore, this study assesses the pivotal factors influencing the evolution of the Metaverse industry through the insights of experts. However, the opinions of experts are frequently subjective and not readily quantifiable, rendering conventional crisp value expressions unsuitable. In such cases, linguistic variables are used in this study to offer an effective means of expressing the subjective opinions of experts. Furthermore, this study employs the fuzzy Delphi method [13,14] to integrate the fuzzy opinions of experts and utilizes the fuzzy DEMATEL method [12,14,15,16] to analyze the relationships of the key factors of Metaverse industry development. Ultimately, this study proposes a model for analyzing the key influencing factors for the development of the Metaverse industry, which can assist enterprises in conducting a systematic evaluation and analysis of the influencing factors and providing guidance on the future direction of the Metaverse industry.

2. Literature Review

2.1. Metaverse

The concept of the Metaverse has attracted considerable attention in recent years. López-Belmonte et al. [17] conceptualized the Metaverse as the next generation of Internet applications, a social form and a virtual world. It employs novel technologies to construct a virtual living environment that can be developed and modified by users. The Metaverse can be defined as a virtual world of technologies and applications that transcend the boundaries of reality. The accelerated advancement of virtual reality (VR), augmented reality (AR), mixed reality (MR), and extended reality (XR) technologies has prompted individuals to venture into a virtual domain that exists beyond the confines of a tangible reality [4,9].

Hwang [18] pointed out that it is feasible to engage in communication and interaction with others through digital avatars as a representation of each user within the Metaverse, thereby facilitating shared immersion and synchronous 3D experiences. In virtual worlds, people’s digital avatars engage in real-time and dynamic interactions and socio-cultural exchanges, thereby engaging in economic behavior. Consequently, the Metaverse is a vast and complex concept, and no single business or organization can exert control over it in isolation. It is a virtual platform that employs the spatially transcendent interactivity of digital avatars to facilitate connections with the real world for economic activities.

2.2. Factors Influencing the Development of the Metaverse

In recent years, the impact of the global pandemic and the increased demand for remote immersive technology services have contributed to the accelerated development of a range of Metaverse applications across diverse industries. The Metaverse has the potential to be widely applied in education, healthcare, and industry, with the support of the Internet, 5G, VR, and other technologies [3,4,9,11]. In the field of education, the Metaverse has the potential to offer students an immersive educational experience [6]. In the industrial sector, the Metaverse can facilitate the creation of a virtual design environment, which can be employed to plan and optimize product lifecycles. This has the potential to address the issues associated with lengthy trial periods and unstable manufacturing processes [11]. Bale et al. [19] highlighted that users in the Metaverse interact with virtual environments through a range of somatosensory devices, creating a surreal experience that engenders the sense of actuality. Furthermore, the advent of 5G network infrastructure enables real-time information transfer, thereby enhancing users’ autonomy.

In a recent publication, Park and Kim [8] highlighted the enhanced economic stability of the Metaverse, enabled by decentralized exchanges facilitating in-game item trading. The Metaverse is a decentralized platform that facilitates seamless interconnectivity between the physical and virtual realms, enabling interactions between individuals and entities in both domains. Mourtzis et al. [20] conceptualized the Metaverse as a permanent and persistent multi-user environment that integrates physical reality and digital virtualization. Furthermore, the Metaverse facilitates connections between social interactive networks and networked immersive environments. The Metaverse represents a new era of Internet connectivity, characterized by key features such as interactivity, simulation, and a decentralized environment.

López-Belmonte et al. [17] identified the core areas of the Metaverse as follows: interoperability, scalability, confidentiality, immersive experiences, and ubiquity of identity. Nevertheless, several issues pertaining to the application of the Metaverse remain to be addressed, including those related to privacy, health, personal data protection, access inequality, the necessity for specific virtual universe legislation, and the threat of hacking attacks [10]. Arpaci et al. [21] posited that one of the key motivating factors influencing the future utilization of the Metaverse is “performance expectations”. The probability of the acceptance and utilization of a system is enhanced when it is straightforward for individuals to comprehend and learn from. In other words, the likelihood of people avoiding the Metaverse is contingent upon their perception of the ease and speed with which they can derive benefits from it.

Dwivedi et al. [22] pointed out that the Metaverse has the potential to expand the physical world through the utilization of AR and VR technologies, which will facilitate user interaction in both real and virtual domains using avatars and holographic projections. Mystakidis [23] highlighted the considerable weight of the VR headset as a significant limitation to prolonged use, citing the potential for head and neck fatigue. Yoo et al. [24] asserted that all elements of the Metaverse are based on sufficiently realistic sensory effects. The fundamental elements of virtual worlds are motion capture, digital augmentation, digital avatars, and simulated conversation modes. It can be reasonably deduced that the advancement of associated intelligent technologies will have an impact on the prospective evolution of the Metaverse industry. Wu and Ho [25] identified medical education and training as a significant potential application area for augmented reality (AR) and virtual reality (VR). The field of emergency medicine requires practitioners to possess the skills and abilities to adapt to a variety of clinical emergencies. Nevertheless, healthcare providers must ensure that personal data are used in an ethical manner and with due consideration for data privacy protection, security, and governance when utilizing the technologies of the Metaverse.

The advantages of the Metaverse are manifold. Primarily, it allows users to interact with others through digital avatar representations, thus providing shared immersion and synchronized 3D experiences [18]. The Metaverse employs virtual reality and augmented reality technologies to afford users a shared and real-time three-dimensional virtual environment. This virtual environment is interactive in that it transcends the limitations of time and distance, thereby facilitating interaction and socialization at any given moment [17]. The rapid development of various infrastructures provides further opportunities for the realization of the Metaverse. As devices and individuals become increasingly interconnected and more data are collected, concerns about the privacy, safety, and security of Metaverse applications will undoubtedly arise. In terms of user experience, the current devices are too bulky to be worn for extended periods of time, and their use is confined to specific locations, which precludes their integration into everyday life [12,20]. This paper presents a summary of the factors influencing the growth of the Metaverse industry in

Table 1. Factors affecting the development of the Metaverse.

	Source	A	B	C	D	E	F	G	H	I	J	K	L
Factors
Immersive Experience		√	√		√		√	√		√	√	√	√
Network Communication		√	√					√			√
Content Diversity			√								√	√	√
Decentralization				√				√				√
Digital Assets				√			√				√	√
Cross-Platform Operation				√				√
Sensory Devices
Social Interactivity		√	√	√	√		√	√		√	√	√
Real-Time					√		√	√				√
Scalability		√	√		√						√	√	√
Virtual–Real Integration			√					√			√
Device Lightweight						√		√
System Usability									√
Product Price									√
Security and Privacy					√	√				√			√
Consumer Habits									√

Notes: A: Dionisio et al. [26], B: Ning et al. [3], C: Park & Kim [8], D: López-Belmonte et al. [17], E: Mystakidis [23], F: Bale et al. [19], G: Mourtzis et al. [20], H: Arpaci et al. [21], I: Dwivedi et al. [22], J: Yoo et al. [24], K: Hwang [18], and L: Wu & Ho [25]. The symbol √ indicates that this literature contains this factor.

.

2.3. The Fuzzy Delphi Method

The Delphi method is a technique for the collection of subjective judgments based on the experience and knowledge of multiple participants, as well as their intuitive and value-based insights. To prevent the participants from influencing each other, an anonymous questionnaire is typically employed to ascertain the perspectives of everyone [27]. Following the processing of the surveys, the opinions of the participants exhibited a gradual convergence, with the scope of the gap gradually narrowing to reach the consensus opinion [27]. Nevertheless, the traditional Delphi method is associated with several disadvantages, including a lengthy timeframe, high costs, and a low response rate.

The fuzzy Delphi method was developed by combining fuzzy set theory with the traditional Delphi method [28] to address the limitations of the latter. The application of the fuzzy Delphi method has the potential to reduce the time required for the integration of consensus opinions in the traditional Delphi method [13,28]. The application of the fuzzy Delphi method in group decision-making can address the issue of ambiguity in the degree of consensus among experts’ opinions while also reducing the number of rounds of questionnaires, thereby enhancing the efficiency and quality of the questionnaires [29].

The following section outlines the execution steps of the fuzzy Delphi method [13,14] as follows.

Step 1: A summary of the m influencing factors was compiled from the study to facilitate the assessment of their relative importance.

Step 2: Each expert provides an assessment for each factor using linguistic variables as illustrated in

Table 2. The triangular fuzzy numbers that correspond to the linguistic variables [33].

Level	Linguistic Variables	Triangular Fuzzy Numbers
1	Very Unimportant (VUI)	(0, 0, 0.25)
2	Unimportant (UI)	(0, 0.25, 0.5)
3	Ordinary (OD)	(0.25, 0.5, 0.75)
4	Important (IM)	(0.5, 0.75, 1)
5	Very important (VIM)	(0.75, 1, 1)

. These linguistic variables can be represented by triangular fuzzy numbers [30,31,32]. Let us assume that there are K experts. The importance assessment for the jth factor by the kth expert (${\tilde{w}}_{kj}$) is shown below:

$${\tilde{w}}_{kj}=\left(a_{kj},b_{kj},c_{kj}\right),k=\mathrm{1,2},\dots ,K,j=\mathrm{1,2},\dots ,m$$

(1)

Step 3: It is assumed that the kth expert assigns a relative importance value to the jth factor as ${\tilde{w}}_{kj}=\left(a_{kj},b_{kj},c_{kj}\right)$, where k = 1, 2, …, K. The resulting fuzzy weight of the jth factor, denoted as ${\tilde{w}}_j=\left(a_j,b_j,c_j\right)$, is calculated as follows:

$$a_j=Min\left\{a_{kj}\right\},b_j=\sqrt[m]{\prod _k^m{b}_{kj}},c_j=Max\left\{c_{kj}\right\}$$

(2)

Step 4: The fuzzy weight (${\tilde{w}}_j$) of each factor can be defuzzied to the crisp value DF(${\tilde{w}}_j$) using the area method, which is calculated as follows [34,35]:

$$DF\left({\tilde{w}}_j\right)={\displaystyle \frac{1}{4}}\left(a_j+2b_j+c_j\right)$$

(3)

Step 5: When filtering factors, it is essential to set an appropriate threshold value. If the threshold is set too high, the number of factors to be filtered will be insufficiently representative. Conversely, if the threshold is set too low, the number of factors to be filtered will be excessive. Assume that β represents a threshold value for the importance of the factors; then, the jth factor is a critical factor if $DF\left({\tilde{w}}_j\right)\ge \mathsf{\beta }$, and conversely, if $DF\left({\tilde{w}}_j\right)<\mathsf{\beta }$, it is deleted. Consequently, the critical factors can be identified based on the input from the experts.

3. Fuzzy DEMATEL Method

In the process of decision analysis, many influencing factors are frequently interrelated and exert a mutual influence upon one another. When experts evaluate the extent of mutual influence between influencing factors, the qualitative nature of the factors and the ambiguity of the subjective judgment of the experts often render it challenging to ascertain the degree of mutual influence. Accordingly, the fuzzy DEMATEL method represents an effective approach for addressing these issues and elucidating the interactions between factors [14,15,36,37,38,39]. The following steps constitute the fuzzy DEMATEL method [15]:

Step 1: The initial step is to assume that there are n influencing factors to be evaluated.

Step 2: Using the linguistic variables as illustrated in

Table 3. The linguistic variables of the influenced degree [15].

Linguistic Variables	Triangular Fuzzy Numbers
No influence	(0, 0, 0)
Low influence	(0, 1, 2)
Medium influence	(1, 2, 3)
High influence	(2, 3, 4)
Very high influence	(3, 4, 4)

, each expert provides a judgment on the degree of influence between two factors, thereby constructing a fuzzy direct-relation matrix (${\tilde{X}}^k$), as demonstrated below:

$${\tilde{X}}^k={\left[{\tilde{x}}_{ij}^k\right]}_{n\times n}$$

(4)

In this context,${\tilde{x}}_{ij}^k$ represents the linguistic evaluation of the extent to which factor i affects factor j as determined by the kth expert (with k = 1, 2, …, K). These linguistic variables can be represented by triangular fuzzy numbers [30,31,32].

Step 3: Incorporate the linguistic evaluations of experts.

The intersection of the fuzzy valuation of each expert with at least one other expert’s fuzzy valuation indicates the existence of a consensus among all experts. Once it has been established that the opinions of the experts have reached a consensus, the fuzzy Delphi method is employed for integrating the fuzzy direct-relation matrix of each expert. The fuzzy direct-relation matrix of K experts is integrated as $\tilde{X}={\left[{\tilde{x}}_{ij}\right]}_{n\times n}$, where ${\tilde{x}}_{ij}=(a_{ij},b_{ij},c_{ij})$.

Step 4: The area method is employed to defuzzify ${\tilde{x}}_{ij}$ as demonstrated below [27,28]:

$$DF({\tilde{x}}_{ij})={\displaystyle \frac{1}{4}}\left(a_{ij}+2b_{ij}+c_{ij}\right)$$

(5)

Step 5: Construct a direct-relation matrix as shown below:

$$A={\left[DF\left({\tilde{x}}_{ij}\right)\right]}_{n\times n},i=\mathrm{1,2},\dots ,n,j=\mathrm{1,2},\dots ,n$$

(6)

In this context, $DF\left({\tilde{x}}_{ij}\right)$ represents the unambiguous value of the degree of influence exerted by factor i on factor j.

Step 6: The normalized direct-relation matrix $D$ is constructed in accordance with the direct-relation matrix as follows:

$$D={\displaystyle \frac{A}{max\left(\underset{1\le i\le n}{\mathit{max}}\sum _{i=1}^{n}DF\left({\tilde{x}}_{ij}\right),\underset{1\le \mathrm{j}\le n}{\mathit{max}}\sum _{j=1}^{n}DF\left({\tilde{x}}_{ij}\right)\right)}}$$

(7)

Step 7: The total influence-relation matrix $T$ is computed according to the normalized direct-relation matrix, as follows:

$$T=\mathrm{D}{(I-\mathrm{D})}^{-1}$$

(8)

where $I$ represents the identity matrix.

Step 8: It is assumed that $T$ is represented by the matrix $\left[t_{ij}\right] (i=\mathrm{1,2},\dots ,n,j=\mathrm{1,2},\dots ,n)$. The influence value $\left(d_i\right)$ and the influenced value $\left(r_i\right)$ of factor i are calculated as follows:

$$d_i=\sum _{j=1}^n{t}_{ij}$$

(9)

$$r_i=\sum _{j=1}^n{t}_{ji}$$

(10)

The value $d_i$ represents the degree to which factor i affects the other factors. The value $r_i$ signifies the extent to which factor i is influenced by the other factors. The value $\left(d_i+r_i\right)$ represents the degree of centrality of factor i among all factors. The value $\left(d_i-r_i\right)$ indicates the net influencing degree of factor i among all factors. A positive result indicates that factor i exerts influence over other factors, whereas a negative result suggests that factor i is the recipient of influence from other factors.

Step 9: The importance value of factor i (${\overline{w}}_i$) is calculated as follows, based on the values $d_i$ and $r_i$ [40].

$${\overline{w}}_i=\sqrt{{\left(d_i+r_i\right)}^2+{\left(d_i-r_i\right)}^2}$$

(11)

In accordance with the importance value of factor i (${\overline{w}}_i$), the normalized weight value of factor i $\left(w_i\right)$ is calculated as follows:

$$w_i={\displaystyle \frac{{\overline{w}}_i}{\sum _{i=1}^n{\overline{w}}_i}}$$

(12)

Step 10: The construction of a causal relationship between factors.

The threshold value (α) for the degree of impact is calculated based on the average influence degree among the factors. According to the total influence-relation matrix (T), the threshold value (α) can be defined as the following equation:

$$\alpha ={\displaystyle \frac{\sum _{i=1}^n\sum _{j=1}^n{t}_{ij}}{n\times n}}$$

(13)

The construction of the cause-effect relation matrix (C) is based on the total influence-relation matrix (T) and the threshold value of the degree of influence (α). The following steps are required:

$$C={\left[c_{ij}\right]}_{n\times n},i=\mathrm{1,2},\dots ,n,j=\mathrm{1,2},\dots ,n$$

(14)

where

$$c_{ij}=\left\{\begin{array}{c}1, t_{ij}\ge \alpha \\ 0, t_{ij}<\alpha \end{array}\right.$$

A value of $c_{ij}$ equal to 1 indicates that factor i exerts an influence on factor j. Conversely, a value of $c_{ij}$ equal to 0 signifies that factor i has no impact on factor j.

4. Empirical Analysis

In order to gain insights into the pivotal elements driving the advancement of the Metaverse industry in Taiwan, a committee comprising six expert groups was established to assess the significance of the underlying factors. The background information of each expert is presented in

Table 4. Background information of experts.

Gender	Age	Industry	Years of Experience	Company Size
Male	40~49	Information Technology	20~30 years	More than 1000 people
Male	40~49	Information Technology	10~20 years	More than 1000 people
Male	40~49	Information Technology	10~20 years	100~500 people
Male	40~49	Education	10~20 years	100~500 people
Male	40~49	Technology R&D	10~20 years	More than 1000 people
Male	40~49	Information Technology	10~20 years	100~500 people

.

4.1. Filtering Critical Factors

4.1.1. Collecting the Influence Factors

Following a comprehensive analysis and collation of the relevant literature (as illustrated in Table 1), this study identified 16 key factors that are likely to influence the development of the Metaverse industry. The factors influencing the development of the Metaverse industry have been identified as follows: immersive experience (F₁), network communication (F₂), content diversity (F₃), decentralization (F₄), digital assets (F₅), cross-platform operation (F₆), sensory devices (F₇), social interactivity (F₈), real-time (F₉), scalability (F₁₀), virtual–real integration (F₁₁), device lightweight (F₁₂), system usability (F₁₃), product price (F₁₄), security and privacy (F₁₅), and consumer habits (F₁₆). The interpretation of each factor is presented in

Table 5. Explanation of the significance of the factors.

Factors	Explanation of Meaning
Immersive Experience	In a virtual environment, all the senses are engaged, including sight, sound, touch, smell, and taste. It is recommended that the experience be made as immersive as possible.
Network Communication	The speed of a network directly correlates with the amount of information that can be transferred. The reduction in latency results in an enhanced user experience.
Content Diversity	In addition to providing a platform for gaming, virtual worlds also facilitate the delivery of a range of services, including education, healthcare, and business. The virtual world provides more freedom and diversity than the real world.
Decentralization	The absence of a centralized authority allows each member to exercise autonomy, enabling them to engage in peer-to-peer transactions and operations without the necessity for third-party involvement.
Digital Assets	Furthermore, the virtual items and creations of players in the Metaverse can be transformed into digital assets. Individuals and businesses are afforded the opportunity to create, own, invest in, and sell.
Cross-Platform Operation	It can accommodate a considerable number of second parties, including those representing various types of film and television, sports, performing arts, clothing, tourist attractions, and other intellectual properties. Additionally, it facilitates collaboration between manufacturers. Users and businesses alike can explore, create, socialize, and participate in a plethora of experiences across a multitude of platforms.
Sensory Devices	The incorporation of sensory devices, such as motion tracking, tactile feedback, and eye tracking, expands the scope of user experience beyond the limitations of traditional display and audio modalities.
Social Interactivity	The term “interactive behavior” is used to describe the way users engage with one another in a digital environment. The Metaverse represents a significant expansion of the boundaries of physical space, offering a highly interactive, shared, and engaging social experience.
Real-Time	It facilitates the rapid transmission of data and the swift generation of interactive responses, thereby achieving synchronization between the dissemination of information and operational processes.
Scalability	The Metaverse offers users the capacity to innovatively alter map structures, construct assets, and upload an array of digital content, thereby enabling the creation of bespoke virtual environments.
Virtual–Real Integration	The question thus arises as to whether the virtual space is sufficiently realistic to enable accurate and seamless connection with the real world. It is possible to integrate the physical world with the Metaverse.
Device Lightweight	The dimensions and mass of a wearable device have the potential to influence the user’s long-term experience. In this regard, a lightweight device may foster a greater willingness to utilize it.
System Usability	The utilization of the Metaverse necessitates a degree of comprehension, proficiency, and expertise. The ease with which a user can join the Metaverse is contingent upon their ability to utilize or learn the requisite technical equipment.
Product Price	The cost of the disparate devices and products that employ the Metaverse will influence the extent to which individuals are willing to utilize them.
Security and Privacy	The question thus arises as to whether the Metaverse can guarantee the privacy of personal information and the security protection of various data.
Consumer Habits	The potential for a given technology to be used in the future is contingent upon its historical usage. To illustrate, users are already inclined to utilize cell phones for communication and entertainment purposes. Consequently, they have established a firm preference for this technology, which can readily impact the uptake of the nascent Metaverse technology.

.

4.1.2. Assessment of the Relative Importance of the Identified Factors

The experts employed linguistic variables (presented in Table 2) to express their assessments of the significance of the factors influencing the advancement of the Metaverse industry. Each expert provided an importance rating of each factor, and the results are presented in

Table 6. Triangular fuzzy numbers that correspond to semantic evaluations.

Factors	Expert 1	Expert 2	Expert 3	Expert 4	Expert 5	Expert 6
F1	(0.5, 0.75, 1)	(0.75, 1, 1)	(0.75, 1, 1)	(0.5, 0.75, 1)	(0.5, 0.75, 1)	(0.75, 1, 1)
F2	(0.75, 1, 1)	(0.75, 1, 1)	(0.75, 1, 1)	(0.75, 1, 1)	(0.5, 0.75, 1)	(0.75, 1, 1)
F3	(0.5, 0.75, 1)	(0.5, 0.75, 1)	(0.75, 1, 1)	(0.75, 1, 1)	(0.5, 0.75, 1)	(0.5, 0.75, 1)
F4	(0.5, 0.75, 1)	(0.5, 0.75, 1)	(0.5, 0.75, 1)	(0.5, 0.75, 1)	(0.75, 1, 1)	(0.5, 0.75, 1)
F5	(0.5, 0.75, 1)	(0.75, 1, 1)	(0.75, 1, 1)	(0.75, 1, 1)	(0.5, 0.75, 1)	(0.75, 1, 1)
F6	(0.75, 1, 1)	(0.75, 1, 1)	(0.75, 1, 1)	(0.5, 0.75, 1)	(0.5, 0.75, 1)	(0.75, 1, 1)
F7	(0.5, 0.75, 1)	(0.75, 1, 1)	(0.5, 0.75, 1)	(0.5, 0.75, 1)	(0.5, 0.75, 1)	(0.75, 1, 1)
F8	(0.5, 0.75, 1)	(0.75, 1, 1)	(0.75, 1, 1)	(0.75, 1, 1)	(0.5, 0.75, 1)	(0.75, 1, 1)
F9	(0.75, 1, 1)	(0.75, 1, 1)	(0.25, 0.5, 0.75)	(0.75, 1, 1)	(0.75, 1, 1)	(0.75, 1, 1)
F10	(0.5, 0.75, 1)	(0.5, 0.75, 1)	(0.5, 0.75, 1)	(0.75, 1, 1)	(0.75, 1, 1)	(0.75, 1, 1)
F11	(0.75, 1, 1)	(0.5, 0.75, 1)	(0.5, 0.75, 1)	(0.75, 1, 1)	(0.5, 0.75, 1)	(0.75, 1, 1)
F12	(0.5, 0.75, 1)	(0.5, 0.75, 1)	(0.75, 1, 1)	(0.75, 1, 1)	(0.75, 1, 1)	(0.75, 1, 1)
F13	(0.75, 1, 1)	(0.5, 0.75, 1)	(0.75, 1, 1)	(0.75, 1, 1)	(0.75, 1, 1)	(0.75, 1, 1)
F14	(0.5, 0.75, 1)	(0.5, 0.75, 1)	(0.75, 1, 1)	(0.75, 1, 1)	(0.5, 0.75, 1)	(0.75, 1, 1)
F15	(0.75, 1, 1)	(0.75, 1, 1)	(0.75, 1, 1)	(0.75, 1, 1)	(0.75, 1, 1)	(0.75, 1, 1)
F16	(0.75, 1, 1)	(0.5, 0.75, 1)	(0.25, 0.5, 0.75)	(0.75, 1, 1)	(0.75, 1, 1)	(0.5, 0.75, 1)

.

4.1.3. Integration of Expert Opinions

In order to integrate the fuzzy evaluation of the importance of each factor by six experts, the fuzzy Delphi method was employed, as illustrated in

Table 7. The integrated fuzzy and crisp importance values of each factor.

Factor Code	Factor Name	Integration of TFN	Crisp Values
F1	Immersive Experience	(0.5, 0.87, 1)	0.81
F2	Network Communication	(0.5, 0.95, 1)	0.85
F3	Content Diversity	(0.5, 0.83, 1)	0.79
F4	Decentralization	(0.5, 0.79, 1)	0.77
F5	Digital Assets	(0.5, 0.91, 1)	0.83
F6	Cross-Platform Operation	(0.5, 0.91, 1)	0.83
F7	Sensory Devices	(0.5, 0.83, 1)	0.79
F8	Social Interactivity	(0.5, 0.91, 1)	0.83
F9	Real-Time	(0.25, 0.89, 1)	0.76
F10	Scalability	(0.5, 0.87, 1)	0.81
F11	Virtual–Real Integration	(0.5, 0.87, 1)	0.81
F12	Device Lightweight	(0.5, 0.91, 1)	0.83
F13	System Usability	(0.5, 0.95, 1)	0.85
F14	Product Price	(0.5, 0.87, 1)	0.81
F15	Security and Privacy	(0.75, 1, 1)	0.94
F16	Consumer Habits	(0.25, 0.81, 1)	0.72

.

4.1.4. Defuzzification

The integrated triangular fuzzy numbers of each factor are defuzzied using the area method, as illustrated in Table 7.

In this study, the threshold of importance was set at 0.8. In accordance with the crisp importance of factors, this study identified 11 critical influencing factors and classified them into three principal dimensions, namely, “user experience”, “consumer behavior”, and “infrastructure”, based on the characteristics of the factors in question. Finally, the framework of the key influencing factors for the development of the Metaverse industry is presented in Figure 1.

4.2. Calculation of the Weights of Critical Factors

In accordance with the critical factors framework for the development of the Metaverse industry, this study employed the fuzzy DEMATEL (Fuzzy Decision-Making Laboratory Analysis) method to determine the weights and identify the causal relationships between the three dimensions. Furthermore, the fuzzy DEMATEL approach was employed to determine the relative importance of indicators within each dimension and to identify the underlying causal relationships. In the second stage of the study, five experts were invited to conduct the evaluation. These experts were senior engineers and executives from the Information Industry Council and the Science Park. The background information of each expert is presented in

Table 8. Background information of experts in the second phase.

No.	Gender	Age	Years of Experience	Title
1	Male	30~39	10~20 years	Senior Engineer
2	Male	40~49	10~20 years	Head of Department
3	Male	40~49	10~20 years	Senior Engineer
4	Male	30~39	Under 10 years	Senior Engineer
5	Male	30~39	10~20 years	Senior Engineer

. The experts employed linguistic variables (as illustrated in Table 3) to ascertain the extent of mutual influence between the dimensions and between the factors under each dimension.

4.2.1. The Construction of the Direct-Relation Matrix of Dimensions

The experts employ linguistic variables (as illustrated in Table 3) to assess the extent of the interrelationship between the two dimensions. By utilizing the fuzzy Delphi method and integrating the fuzzy valuation of the experts, a fuzzy direct-relation matrix can be derived (as presented in

Table 9. Fuzzy direct-relation matrix of dimensions.

	$C_1$	$C_2$	$C_3$
$C_1$	(0, 0, 0)	(1, 3.104, 4)	(0, 2.352, 4)
$C_2$	(1, 2.702, 4)	(0, 0, 0)	(1, 3.104, 4)
$C_3$	(2, 3.565, 4)	(1, 2.930, 4)	(0, 0, 0)

). The crisp direct-relation matrix is presented in

Table 10. Crisp direct-relation matrix of dimensions.

	$C_1$	$C_2$	$C_3$
$C_1$	0	2.802	2.176
$C_2$	2.601	0	2.802
$C_3$	3.282	2.715	0

.

4.2.2. Establishment of Total Influence-Relation Matrix

The normalized crisp direct-relation matrix is obtained by normalizing the direct-relation matrix, as illustrated in

Table 11. Normalized direct-relation matrix (D).

	$C_1$	$C_2$	$C_3$
$C_1$	0.000	0.467	0.363
$C_2$	0.434	0.000	0.467
$C_3$	0.547	0.453	0.000

. Accordingly, the total influence-relation matrix, T can be obtained as shown in

Table 12. Total influence-relation matrix (T).

	$C_1$	$C_2$	$C_3$
$C_1$	3.011	3.213	2.957
$C_2$	3.507	3.078	3.178
$C_3$	3.783	3.605	3.057

.

4.2.3. Calculate the Weights of the Dimensions

The total influence-relation matrix allows the relative importance (${\overline{w}}_i$) and the normalized weight value ($w_i$) of each dimension to be calculated. The results are presented in

Table 13. Normalized weights of dimensions.

	$d_i$	$r_i$	$d_i+r_i$	$d_i-r_i$	${\overline{w}}_i$	$w_i$	Rank
$\mathrm{User} \mathrm{experience} (C_1$)	9.181	10.301	19.482	−1.120	19.488	0.331	3
$\mathrm{Consumer} \mathrm{behavior} (C_2$)	9.763	9.896	19.659	−0.133	19.659	0.334	2
$\mathrm{Infrastructure} (C_3$)	10.445	9.192	19.637	1.253	19.677	0.335	1

.

4.2.4. Creating a Cause-Effect Relation Matrix

The threshold value (α) is calculated as 3.265 in accordance with the total influence-relation matrix T. The causal relation matrix of the dimensions can be obtained based on the total influence-relation matrix and the threshold values, as demonstrated in

Table 14. Cause-effect relation matrix of the dimensions.

	$C_1$	$C_2$	$C_3$
$C_1$	0	0	0
$C_2$	1	0	0
$C_3$	1	1	0

.

In accordance with the cause-and-effect relation matrix of the dimensions (as illustrated in Table 14), the cause-and-effect relations of the three dimensions can be plotted as demonstrated in Figure 2. In accordance with the weights illustrated in Table 13, the order of importance of the three dimensions is as follows: infrastructure ($C_3$), consumer behavior ($C_2$), and user experience ($C_1$). The net influenced degree of the “infrastructure ($C_3$)” dimension is positive, and it is therefore considered an “influential dimension”. The net influenced degree of the dimensions of “user experience ($C_1$)” and “consumer behavior ($C_2$)” are negative, indicating that they are “influenced dimensions”. The “infrastructure ($C_3$)” exerts a direct influence on the other two dimensions while being largely independent of their respective influences. Consequently, it can be classified as a “driving dimension”. The “user experience ($C_1$)” can be defined as a “passive dimension”, in that it is only affected by the other two dimensions and does not affect them in return.

4.3. Calculate the Weight of Each Factor Under Each Dimension

4.3.1. User Experience Dimension

The influence relationship and normalized weights of the factors under the dimension of “user experience” are presented in

Table 15. Weights of factors under the dimension of “user experience ${(C}_1)$”.

	$d_i$	$r_i$	$d_i+r_i$	$d_i-r_i$	${\overline{w}}_i$	$w_i$	Rank
immersive experience $\left(C_{11}\right)$	35.671	35.674	71.345	−0.003	71.345	0.255	1
virtual–reality integration ${(C}_{12})$	35.102	34.817	69.919	0.285	69.920	0.250	2
social interaction $\left(C_{13}\right)$	34.289	34.792	69.081	−0.503	69.083	0.247	4
scalability $\left(C_{14}\right)$	34.731	34.51	69.241	0.221	69.241	0.248	3

, and the cause-effect relation is illustrated in Figure 3. As illustrated in Figure 3, the immersive experience ($C_{11}$) is the most significant factor within the dimension of “user experience”. The net influence degrees of virtual reality integration ($C_{12}$) and scalability ($C_{14}$) are positive and are thus classified as “influential factors”. The net influenced degrees of the immersive experience ($C_{11}$) and social interactivity ($C_{13}$) are negative, and thus they are classified as “influenced factors”.

4.3.2. Consumer Behavior Dimension

The influence relationships and weights of the factors pertaining to the dimension of “consumer behavior” are presented in

Table 16. Weights of factors under the dimension of “consumer behavior ${(C}_2)$”.

	$d_i$	$r_i$	$d_i+r_i$	$d_i-r_i$	${\overline{w}}_i$	$w_i$	Rank
Device Lightweight ($C_{21})$	12.543	11.606	24.149	0.937	24.167	0.339	1
System Usability $\left(C_{22}\right)$	11.484	12.430	23.914	−0.946	23.933	0.336	2
Product Price $\left(C_{23}\right)$	11.615	11.606	23.207	0.009	23.221	0.325	3

, and the cause-and-effect relations are illustrated in Figure 4. As illustrated in Figure 4, device lightweight ($C_{21}$) is the most significant factor within the domain of consumer behavior. The net influence degrees of device light-weight ($C_{21}$) and product price ($C_{23}$) are positive; thus, they are classified as “influential factors”. The net influenced degree of system usability ($C_{22}$) is negative, indicating that it is an “influenced factor”. The “device lightweight ($C_{21}$)” exerts a direct influence on the other two factors and is not affected by the other factors. Consequently, it can be classified as a “driving factor”. System usability ($C_{22}$) is classified as a passive factor because it does not affect the other factors.

4.3.3. Infrastructure Dimension

The relative importance of the factors under the dimension of “infrastructure” and the associated cause-and-effect relationships are illustrated in

Table 17. Weights of factors under the dimension of “infrastructure ${(C}_3)$”.

	$d_i$	$r_i$	$d_i+r_i$	$d_i-r_i$	${\overline{w}}_i$	$w_i$	Rank
Network communication ($C_{31}$)	8.862	7.618	16.480	1.244	16.527	0.254	2
Cross-platform operation $\left(C_{32}\right)$	7.929	7.635	15.564	0.294	15.568	0.240	3
Security and privacy $\left(C_{33}\right)$	8.724	8.780	17.504	−0.056	17.504	0.270	1
Digital assets $\left(C_{34}\right)$	6.882	8.364	15.246	−1.482	15.318	0.236	4

and Figure 5. As illustrated in Figure 5, the factor designated “security and privacy ($C_{33}$)” is identified as the most significant element within the infrastructure dimension. The net influence degrees of network communication ($C_{31}$) and cross-platform operation ($C_{32}$) are positive and thus may be considered as “influential factors”. The net influenced degrees of “security and privacy ($C_{33}$)” and “digital assets ($C_{33}$)” are negative, indicating that they are considered “influenced factors”. The “digital assets ($C_{34}$)” is classified as a “passive factor” because it does not affect any other factors.

4.4. Analysis and Discussion

4.4.1. Key Factors in the Development of the Metaverse Industry

This study presents a summary of 16 influencing factors that were identified as affecting the development of the Metaverse industry. The factors were selected based on a comprehensive review of the relevant literature and a detailed analysis of their impact. The 11 key factors affecting the development of the Metaverse industry were identified through a process of expert assessment and integration using the fuzzy Delphi method.

4.4.2. The Cause-Effect Relation Between Dimensions

The results of the fuzzy DEMATEL method indicated that the centrality values of “infrastructure ($C_3$)” and “consumer behavior ($C_2$)” are greater than the average, suggesting that these two dimensions are more closely related to the development of the Metaverse industry. “Infrastructure ($C_3$)” belongs to the “influencing dimension”. In contrast, “consumer behavior ($C_2$)” and “user experience ($C_1$)” belong to the “influenced dimension”. “Infrastructure ($C_3$)” can be classified as a “driving dimension”, while “user experience ($C_1$)” falls within the “passive dimension”. These observations pointed out that enterprises should prioritize investments in “infrastructure” when entering the Metaverse industry, as this approach may enhance the possibility of success.

4.4.3. The Cause-Effect Relation of Key Factors

The cause-effect relation of the elements under the “user experience” dimension reveals that “virtual–real integration ($C_{12}$)” exhibits a high degree of correlation and belongs to the aspect of influence. Furthermore, “virtual–real integration ($C_{12}$)” exerts a direct impact on “immersive experience ($C_{11}$)” and “social interaction ($C_{13}$)”, and an indirect impact on “scalability ($C_{14}$)” through “immersive experience ($C_{11}$)”. Consequently, when enterprises invest in the Metaverse industry, it is recommended that they prioritize the integration of “virtual and real” to enhance the user experience.

In accordance with the cause-and-effect relationship between factors within the consumer behavior dimension, device lightweight ($C_{21}$) is identified as a highly correlated factor with a significant influence on consumer behavior. Consequently, enterprises should prioritize device lightweight as a key consideration in guiding consumer behavior.

According to the cause-effect relation of the elements under the “infrastructure” dimension, “network communication ($C_{31}$)” has a high degree of correlation and belongs to the influence factor. Therefore, enterprises should prioritize the improvement of “network communication ($C_{31}$)” to reinforce the infrastructure and enhance the probability of success for those investing in the development of the Metaverse industry.

4.5. Comparison Results with Different Methods

Based on the calculation results of both the fuzzy Delphi and fuzzy DEMATEL methods, the ranking of the weights of each factor in each dimension can be shown in

Table 18. Ranking order of factor weights in different methods.

	Factors	Fuzzy Delphi	Fuzzy DEMATEL
$C_1$	$\mathrm{Immersive} \mathrm{experience}\left(C_{11}\right)$	2	1
	$\mathrm{Virtual}–\mathrm{reality} \mathrm{integration}{(C}_{12})$	2	2
	$\mathrm{Social} \mathrm{interaction}\left(C_{13}\right)$	1	4
	$\mathrm{Scalability}\left(C_{14}\right)$	2	3
$C_2$	$\mathrm{Device} \mathrm{lightweight} \left(C_{21}\right)$	2	1
	$\mathrm{System} \mathrm{usability}\left(C_{22}\right)$	1	2
	$\mathrm{Product} \mathrm{price}\left(C_{23}\right)$	3	3
$C_3$	$\mathrm{Network} \mathrm{communication} (C_{31}$)	2	2
	$\mathrm{Cross}‐\mathrm{platform} \mathrm{operation}\left(C_{32}\right)$	3	3
	$\mathrm{Sec}\mathrm{urity} \mathrm{and} \mathrm{privacy}\left(C_{33}\right)$	1	1
	$\mathrm{Digital} \mathrm{assets} \left(C_{34}\right)$	3	4

. Since the fuzzy Delphi method only considers the subjective judgment of the experts but not the influence of the factors, it is not possible to obtain the ranking of the factor weights effectively. By using the fuzzy DEMATEL method, the ranking of factor weights can be effectively obtained by analyzing the influence relationship between factors. Therefore, the fuzzy DEMATEL method is a better way to calculate the weights of factors.

5. Conclusions

In recent years, the Metaverse industry, which emphasizes the integration of virtual and real, has emerged as a key area of interest for enterprises. Consequently, when investing in the Metaverse industry, it is imperative for enterprises to first grasp the pivotal influencing factors to optimize the probability of successful Metaverse product and market development. Many quantitative and qualitative factors should be considered for Metaverse product and market development. Because of the uncertainty and ambiguity of technology and market, it is possible to employ a combination of fuzzy set theory and multi-criteria decision analysis to analyze the pivotal factors influencing the development of the Metaverse industry.

This study employs the fuzzy Delphi method, a technique for integrating the evaluation of the importance of influencing factors by multiple experts, to identify 11 key impact factors affecting the development of the Metaverse industry. The method is based on linguistic variables. To facilitate analysis and effectively grasp the relationship between the factors, this study divided the eleven key impact factors into three major dimensions, in consideration of their characteristics. Accordingly, this study employs the fuzzy DEMATEL method to compute the weights and causal–effect relations pertaining to the influence dimensions and factors on the development of the Metaverse industry. By evaluating the causal–effect relation between the dimensions and key factors, it is possible to gain an understanding of the driving dimensions and the driving key factors, which will, in turn, improve the possibility of success in the Metaverse industry.

This study presents a systematic analysis model of the key factors influencing the development of the Metaverse industry. This study makes two principal contributions to the field. (i) The use of linguistic variables through the fuzzy Delphi method can effectively screen the key factors affecting the development of the Metaverse industry. (ii) The fuzzy DEMATEL method allows for the proposal of a systematic analysis model, which can effectively evaluate the importance of the key factors in the development of the Metaverse industry and its causal relationship. This approach enables enterprises to gain a more accurate understanding of the characteristics of key factors, thereby enhancing the feasibility of developing the Metaverse industry.

The following recommendations are made for future research on this study. (i) In future research, the utilization of alternative fuzzy multi-criteria decision-making methods for analysis may be considered to enhance the acceptability of results. (ii) It is acknowledged that the number of influencing factors will have an impact on the assessment of the development of the Metaverse industry. Therefore, in future studies, the potential merits of collecting influencing factors at a greater number of levels for the screening of key factors may be explored to enhance the value of the practical application.