Critical Success Factors Evaluation for Blockchain’s Adoption and Implementing

Abstract

Blockchain has completely changed how business is performed today, thus making it one of the most disruptive technologies in recent times. However, it is a challenging task to adopt and implement blockchain technologies in different services and industries. Therefore, this study introduces a framework for investigating critical factors influencing the successful adoption of blockchain technologies in different applications and prioritizes them using the hierarchical Decision-Making Trial and Evaluation Laboratory (DEMATEL) technique. First, it provides fourteen critical success factors with the help of the extant literature and further classifies them into three categories: technological, organizational, and environmental. In addition, a set of sixteen key performance indicators (KPI) of successful blockchain adoption is introduced and classified into five categories: overall performance, system robustness, data robustness, accessibility, and overall cost. Then, the fourteen success factors are ranked based on their degree of prominence and relationships. It is concluded that environmental factors are the most critical factors for successful blockchain adoption, and law and policies and competitive pressure are the top two factors needed for blockchain adoption. In the technological context, only blockchain scalability is ranked among the top significant factors for blockchain adoption. On the other hand, adequate resources, top management support, and financial constraints are highly ranked in the organizational context.

1. Introduction

A blockchain is a sequential chain of blocks, each of which contains complete information about all network activities and is linked together to form a distributed ledger [1]. Blockchain is ranked first for transferring transactions because it is an Internet protocol that does not require the assistance of a third party [2], saving time and money because it does not require payment to a third party, such as banks, to transfer transactions [3]. Still, blockchain adoption is in its nascent stage and, thus, requires an assessment of factors influencing its adoption [4]. Since a critical success factor is defined as the cause of success, such as what you need to do to be successful, identifying such factors could accelerate the spreading of blockchain and its integration into organizations. This research aims to answer three research questions.

• What are the CSFs of blockchain technology from a broad perspective and not for specific applications?
• What are the types of relationships among factors, and what is the significance of these relations?
• What measures can be used to quantify the success of a blockchain?

Blockchain technology is still in its early phase, so limited literature is available on it. Therefore, little is known about the adoption of blockchain in organizations. Therefore, this paper makes several important contributions. First, it contributes to identifying and prioritizing the factors that can potentially influence the adaption of blockchain technology (BCT) in applications and to identify blockchain-adoption key performance indicators (BCAKPIS). Second, this research paper presents a systematic literature review to identify the critical success factors (CSFs) of BCT adoption in applications. Later, based on the literature, a hierarchical model was established based on the identified 14 CSFs and sixteen key performance indicators (KPI) of successful blockchain adoption. A final contribution of this paper is that we used the hierarchical DEMATEL technique to prioritize critical success factors and determine their relations.

The remainder of this paper is organized as follows: after a literature review of the works addressing the critical success factors of blockchain’s implementation in Section 2, Section 3 highlights the research model and methodology of the study. Then, the factors’ and indicators’ validation is addressed in Section 4, and the results are discussed in Section 5. Finally, the study is concluded in Section 6.

2. Related Work

The critical success factor (CSF) of adopting new technologies is an exciting research topic that attracts many researchers, e.g., the success factors of business intelligence [5,6,7], enterprise resource planning [8,9], and artificial intelligence [10,11]. Regarding blockchain, Post et al. identified the factors influencing blockchain-technology diffusion using grounded theory-based data collection and analysis [12]. Other studies discussed the key challenges and critical success factors of blockchain implementation in a specific application. For example, Zhou et al. discussed the challenges and the success factors of blockchain technology in the Maritime industry in Singapore using the analytic hierarchical process (AHP); a fishbone diagram; and a political, economic, socio-cultural, technological, environment and legal (PESTEL) analysis [13]. The analytic network process (ANP) was used by Juliet et al. to investigat the factors that affect the successful adoption of blockchain technologies in the freight logistics industry [14]. Prasad et al. identified the CSFs of blockchain-based cloud services [15], and Hartmann et al. conducted a literature review on success factors for conventional and blockchain-based crowdfunding [16]. In addition, Shardeo et al. identified different crucial success factors for the adoption of BT in the freight transportation [4]. Shoaib et al. proposed a framework to identify the CSFs that can positively influence implementing a blockchain-based supply chain using AHP [3]. Parmentola et al. investigated the complex relationship between blockchain adoption and environmental sustainability. Therefore, we employed a systematic literature review and the preferred reporting items for systematic review and meta-analyses (PRISMA) protocol [17].

Even though numerous studies addressed the identification of CSFs of blockchain adoption in a specific application, very few studies addressed the relationships among those factors. For example, using the Decision-Making Trial and Evaluation Laboratory (DEMATEL), Kouhizadeh et al. provided a comprehensive overview of the barriers to adopting blockchain in sustainable supply chains [18]. Furthermore, Yadav and Singh proposed an incorporated principal component analysis and Fuzzy-DEMATEL to study the use of blockchain technology in sustainable supply-chain management [19]. In addition, Kayikci et al. examined CSFs for blockchain-based circular supply chains (CSCs) through a short systematic literature review and using the integrated fuzzy cognitive mapping and fuzzy best-worst method (FCM-FBWM) [20]. Finally, Elhidaoui et al. proposed a conceptual framework of a green supply-chain management (GSCM) model to investigate the critical success factors (CSFs) of blockchain-technology (BCT) adoption in green supply-chain management [21].

It is noteworthy that both of the above studies, as shown in

Table 1. Systematic review studies.

Author	Research Area	Tools Used	Critical Factors Identified	Successful Implementation
Post et al. [12]	Blockchain	Grounded theory-based data collection and analysis	Strategic (sector pressure, organizational size) Tactical (knowledge deficit, implementation method) Operational (technical shortcomings, process maturity)	x
Zhou et al. [13]	Maritime industry	AHP, a fishbone diagram and PESTEL analysis	Lack of experience, lack of blockchain knowledge, and scalability	x
Juliet et al. [14]	Freight-logistics industry	ANP	Technological (infrastructural facility, complexity, compatibility) Organizational (training facilities, top management support) External environmental context (law and policy, competitive pressure)	x
Prasad et al. [15]	Cloud services	Total interpretive structural modeling (TISM)	Regulatory clarity, and law of experiences	x
Shardeo et al. [3]	Supply chain	AHP	x	System strength (transparency, disintermediation, immutability) Overall efficiency (effectiveness and efficiency, automation) Reliability and ecoreconciliation (reliability, immutability, decentralization) overall cost
Kouhizadeh et al. [18]	Supply chains	DEMATEL	Technological (infrastructural facility, immaturity) Organizational (financial constraints, top management support) Environmental (lack of governmental policies)	x

, addressed sustainable supply-chain management. On the other hand, Biswas and Gupta used the DEMATEL technique to identify relationships among ten broad barriers, not success factors, of blockchains adoption [22]. To the best of our knowledge, no studies have addressed the relationships of CSFs of blockchain in a broad perspective or even for applications other than sustainable supply chains. One of the possible reasons for such a research gap is that DEMATEL is not suitable to address such complex multi-level problems. Therefore, it may be concluded that using the recent hierarchical DEMATEL technique [23] to identify the relationships among the CSFs of blockchain adoption from a broad perspective, other than sustainable supply chains, is a clear research gap that needs to be addressed.

Therefore, the research objective is realized by conducting an extensive review of relevant literature to identify the critical factors and evaluate these factors using the opinions of experts. The opinions of these experts were sourced using hierarchical DEMATEL to prioritize critical success factors and extract the interlinked relationships among those factors [23].

3. Research Model

The identification and classification of blockchain adaption success factors (BCASF) is the first research question of this study; moreover, it is a prerequisite for addressing the second research question. On the other hand, the third research question quantifies blockchain adoption’s success by determining key performance indicators (BCAKPIS). Therefore, the research methodology consists of seven steps, as shown in Figure 1, to address these questions. The first step involves data collection, elaborated in the following, Section 3.1 and Section 3.2. The second step involves selecting critical success factors and classifying them into various categories with the help of a literature review and identifying BCAKPIS. The third step involves the distribution of questionnaires among blockchain experts. The fourth step consists of collecting responses from experts and generating the final direct-relationship matrix based on the collected responses. The fifth step determines the causality and prominence of the success factors using the hierarchical DEMATEL, as described in Section 4.2. The sixth step involves the verification of the obtained results from Step 5 with the help of experts and revisiting extant literature. If there are significant deviations, the experts are communicated, and their responses are collected again. The seventh and final step involves deriving results and discussions, as described in Section 5.

The literature data were gathered from search engines and digital libraries using the method suggested by [24,25,26]. Well-defined key terms from research questions are used to create the search strings. Based on published articles about blockchains and their critical success factors, alternatives to search strings and synonyms were developed. The following search strings were used with Boolean operations (OR and AND) [27]: (“Factors” OR “critical success factors” OR “Successful implementation factors”) AND (“Blockchain” OR “smart contract” OR “Distributed” OR “Decentralized”). The search strings were applied in search engines such as Google Scholar, Science Direct, IEEE, and Springer to gather research papers between 2018 and 2022; since blockchain is the newest perspective technology in the modern economy [28], where the word blockchain was first used by Nakamoto (2008), we focus on research papers from the last five years.

To filter the formal studies, Afzal et al. proposed a five-phase technique known as the “tollgate approach” [29]. The phases of the tollgate approach are given below.

• Phase 1: “Search using search terms”.
• Phase 2: “Exclusion based on title and abstract”.
• Phase 3: “Exclusion based on introduction and conclusions”.
• Phase 4: “Exclusion based on the full text”.
• Phase 5: “Final selection of primary studies”.

Using this method resulted in the selection of eleven studies from roughly two hundred and forty-four publications.

Two hundred and twenty-four articles were chosen from the mentioned data sources based on the inclusion and exclusion criteria shown in

Table 2. Key inclusion and exclusion criteria.

Inclusion Criteria	Exclusion Criteria
Published from 2018 to 2022	Published not in the English language
In the field of block chain	Not relevant to challenges or success factors
Presented empirical data	Review papers or non-technical papers
Peer-reviewed	Grey literature (white papers, editorial comments, book reviews)

, before applying the tollgate approach. Thus, as shown in Figure 2, eleven publications were finally selected after applying the tollgate approach’s five phases, representing 4.9% of the total publications.

3.1. Blockchain Success Factors

However, because blockchain adoption is still in its early stages, it is necessary to evaluate the factors that influence its adoption. Therefore, the success factors that affect the spreading of blockchain were identified through a literature review. In this sub-section, BCASFs are identified from the extant literature summarized into three significant categories: technological, organizational, and environmental [14,18].

Table 3. BCASF.

Categories	Factors
Technology	Scalability [13,30,31]
	Infrastructural facility [14,18,22,32,33,34,35,36]
	Complexity [14,32,33]
	Compatibility [14,32,33,34,35]
	Immaturity of Technology [18,22,37]
	Distributed design [22]
Organization	Lack of experience and knowledge [13,15,32,38,39]
	Training facilities [13,27,30,31,32,36,37,38,39,40,41]
	Top management support [14,18,30,33,34,35,37,39,40,41,42,43,44,45,46]
	Organizational culture [13,14,18,30,33,34,37,39,40,43,44,45,46]
	Financial Constraints [18,22,33,37,47,48,49]
	Adequate resource [37]
Environment	Laws and Policy [14,19,22,30,31,32,33,35,36,39,45,46,47,50,51]
	Competitive pressure [14,30,35,36,39,45,46,50]

summarizes the factors of each context and indicates the research papers that supported the inclusion of each factor as a crucial success factor of blockchain adoption.

3.1.1. Technological Context

The first category focuses on how technological features can affect the adoption of blockchain technologies in different applications. Therefore, the technological context includes five factors: scalability, infrastructural facility, complexity, compatibility, immaturity of the technology, and distributed design. Juliet et al. classified into four categories—scalability, infrastructural facility, complexity, and compatibility—technological factors [14]. Moreover, Yadav and Singh suggested the immaturity of the technology and distributed design as important technological factors for the success of blockchain adoption [19].

Theoretically, a massive block containing each transaction in the header could be constructed. One of the key problems with Ethereum, for instance, is that each node must process every transaction and keep a complete record of every account’s balance, contract code, storage, etc. Even though this approach offers high security, it severely restricts Ethereum’s ability to handle more transactions under a single capability. For a peer-to-peer network (P2P), scalability is defined as the non-zero growing rate of the vertex sets that the network can accommodate [52]. In other words, scalability refers to how quickly the blockchain system can handle data as it is loaded [13].

The infrastructural facility can affect the use of blockchains in different applications [33]. The infrastructural facilities are the hardware and software components, including networks as essential resources for blockchain adoption [18,53]. Another technology-related factor is the complexity that represents a attribute of blockchains: that it need specific skills to ease their adoption [54]. In addition, compatibility is essential in this context, and can be defined as the straightforwardness of integrating blockchain technologies into related platforms in different sectors [36]. On the other hand, incompatibility can result in time-consuming and expensive processes [55].

Furthermore, technology’s immaturity is considered a challenge of blockchain technology, which still suffers from latency and throughput issues [18]. Distributed design is another crucial factor, meaning where blockchain is an unalterable distributed ledger [56]. A distributed ledger is a database across several locations or multiple participants, forming a peer-to-peer network through which the servers communicate to maintain the most accurate and up-to-date transaction records [57].

Juliet et al. [14] and Kouhizadeh et al. [18] considered the security and privacy of information and the immutability and accessibility of a blockchain as success factors in the technological context. Still, this research views them as the success indicators of blockchain implementation. These indicators ensure that the information shared is secure and unchanged in the ledgers.

3.1.2. Organizational Context

The organizational context describes the adoption organization’s attributes, characteristics, and resources that can either ease or hinder the adoption of blockchains; it includes six sub-factors: lack of experience and knowledge, training facilities, top management support, organizational culture, financial constraints, and adequate resources. For example, Juliet et al. [14] classified organizational context into three sub-factors: the presence of training facilities, top management support, and organizational culture, but Yadav and Singh mentioned only two factors: organizational culture and financial constraints [19]. Saberi et al. added two factors: lack of experience and knowledge and adequate resources, to the distinct organizational culture [37].

For example, the availability of suitable training facilities enhances the team adaption of blockchain technologies and handles the lack of experience and knowledge of emerging blockchain technologies [30]. Top management support is defined as the capability of senior managers to offer guidance, resources, and necessities throughout and after the firms’ acquisitions of blockchain technologies [35]. On the other hand, organizational culture is considered a management philosophy to manage and improve work performance and influence thoughts, emotions, and communication. Organizational culture affects how companies react to external stresses and make strategic business decisions [58]. The leadership’s role is to create and manage a culture to achieve superior business results [45].

Additionally, the availability of adequate funds during the development of blockchain constitutes a crucial factor for the success of its adoption [18]. Furthermore, other resources, e.g., human resources, are vital to guarantee the success of blockchain technologies [14,45]. Therefore, the decisions of the firms considering adopting blockchain are typically based on a certain set of their organizational characteristics [59].

3.1.3. Environmental Context

Environmental factors are related to the business operating field, such as competitive pressure and government laws and policies. Juliet et al. pointed out two sub-factors in the environmental context: competitive pressure and government laws and policies [14]. Competitive pressure is related to the outer environment of the industry. Therefore, it is described as the continuous desire of firms to show their competence to stakeholders or investors by looking for methods to grow and improve their competitive advantage [50]. In addition, government laws and policies constitute the capability of related government agencies to deliver aid and establish the rubrics and regulations to encourage blockchain adoption in the application industry [46].

3.2. Blockchain Success Indicators

The key performance indicators (KPI) of successful blockchain adoption can be classified into five categories: overall performance, system robustness, data robustness, accessibility, and overall cost [19], as shown in

Table 4. BCAKPIS.

Measures	Factors
Overall performance	Efficiency [19,38,60]
	Effectiveness [19,38,60]
	Speed [3,22,31,61,62]
	Quality [3,16]
System robustness	Transparency [3,4,19,31,38,60,61,62]
	Security [3,4,19,31,38,60,61,62]
	Disintermediation [3,19]
	Trust [4,19,60]
Data	Immutability [3,19,31,60,63]
	Reliability [3,4,60]
	Decentralization [3,19]
	Data accuracy [31]
Accessibility	Traceability [3,4,19,37,38,60,63,64]
	Integrity [3]
Overall cost	Cost reduction [3,4,19,31,32,60]
	Save energy [3]

.

3.2.1. Overall Performance

With the successful deployment of blockchain technology, the overall performance of an organization as a whole should be improved in terms of efficiency, effectiveness, and the speed and quality of operations. [19]. Yadav and Singh [19] pointed out that the overall performance impact of a successful blockchain adoption can be measured through efficiency, effectiveness, and operation speed. Shoaib et al. [4] added quality control and fairness to effectiveness, efficiency, and automation (to speed up the operations). Successful blockchain adoption results in a high speed of information flow, removing the intermediary, traceable smart contracts, and streamlined processes [19]. Better quality can be achieved through the verification and validation of blockchain transactions conducted by blockchain nodes.

3.2.2. System Robustness

System robustness is a system’s ability to withstand a specific degree of failure and hostile inputs during execution and remain functioning under disturbances [65]. It includes four indicators: transparency, security, disintermediation, and trust [4,19]. As a significant indicator of robustness, transparency makes peer-to-peer transactions at the minor end verifiable, and updated data cannot be modified or hacked. Security is another crucial indicator to achieve through blockchain adoption because it is difficult to erase or modify the block once it has been created [19]. Finally, both disintermediation, which means peer-to-peer transactions without encompassing a third party, and trust are significant indicators of successful blockchain adoption because blockchain’s main objective is to remove the core of the assessment of the credibility of the network members [4].

3.2.3. Data Robustness

Data robustness can be identified through its immutability, reliability, decentralization, and accuracy. Data immutability means it is impossible to modify data or the information on the blockchain [19]. Data reliability is related to the elimination of the risk of system failure, loss of data, and malicious attacks. Decentralization means that transactions are authenticated and protected without the involvement of a central authority [66] and are stored at several nodes in the network [67]. Garg et al. [31] related data immutability and accuracy to the successful adoption of blockchain. Yadav and Singh [19] mentioned decentralization and immutability. Shoaib et al. [3] mentioned reliability, immutability, and decentralization.

3.2.4. Accessibility

Accessibility is identified through data traceability and integrity [3]. Traceability focuses on the transparency of the transaction history [68]. Recently, blockchain technology has also been introduced to support the enhancement of data traceability. Successful implementation of blockchain allows data traceability through a direct link, which is checked for quality control and safety and is available at all phases and at any time requisite [69]. On the other hand, successful blockchain adoption provides a high level of data integrity by ensuring it is modified only as intended and designed through nonrepudiation [70].

3.2.5. Overall Cost

The overall cost reduction is an intuitive success indicator of any system and can be measured through cost reduction and energy saving [3]. Cost reduction means reducing paper, other consumable items, and the time needed to conduct operations. It was reported that successful blockchain adoption might result in a 15% reduction in overhead costs by removing intermediaries and paperwork and reducing tracking, audits, transport costs, and energy [19].

Consequently, a blockchain-adoption model is proposed to address these questions as shown. Therefore, the proposed model is divided into two primary parts: blockchain adaption success factors (BCASF), as shown in Figure 3, and key performance indicators (blockchain adaption key BCAKPIS), as shown in Figure 4.

4. The Relationships among the Factors

In a practical situation, handling a large set of interrelated system success factors challenges decision-makers to achieve an organization’s mission. Therefore, engaging advanced multi-criteria decision-making (MCDM) techniques is imperative to prioritize critical success factors and determine their relations. In situations where there are several conflicting criteria, MCDM can also assist the decision maker depending on their preferences [71].

Amongst the related MCDM methods available, the interpretive structural modeling (ISM) and analytical hierarchy process (AHP) are being widely applied by researchers. Recently, various studies have applied the Decision Making Trial and Evaluation Laboratory (DEMATEL) technique [22,72,73]. When examining mutually dependent factors, scholars consider that ISM and DEMATEL are preferable to AHP. The DEMATEL technique can also identify each factor’s total degree of effect. Finally, dependent relationships can be established by classifying these factors into causal and receiver groups using the DEMATEL technique [22].

DEMATEL is also efficient in research including a small number of experts [74]. For example, past studies have used 15 experts [22], 3 experts [75], 19 experts [76], and 4 experts [77]. In the subsequent section, the DEMATEL technique is described in detail.

4.1. Hierarchical DEMATEL

The Battelle Memorial Institute Geneva Research Center first proposed the original DEMATEL technique [78]. DEMATEL is a structural modelling technique that examines interdependent relationships and determines the significant impacts of those important factors (in this study, success factors (SF)) in the form of a cause-and-effect diagram known as a diagraph [75,79]. It is conducted through the below steps:

• Generate the average direct-relation matrix “X”;
• Normalize the direct-relation matrix “N”;
• Obtain the total direct-relation matrix “T”;
• Compute prominence and relations between factors.

The hierarchical DEAMTEL is established for complex systems with numerous system factors and various sorts of impacts. Therefore, a complicated system can be broken down into many subsystems in accordance with specific rules. Each subsystem can be broken further into sub-sub systems, until it is eventually divided into factors. Three additional steps (H1–H3) need to be conducted for hierarchical DEMATEL models before continuing with the abovementioned five steps of the DEMATEL technique.

• Step H1. Hierarchical decomposition

The hierarchical decomposition is constructed, including vertical decomposition and horizontal decomposition. The goal of horizontal decomposition is to break complex systems’ critical-factor identification problem down into numerous simple problems. The goal of vertical decomposition is to break the complicated system down into multiple-level subsystems which follow a predetermined rule [23,80,81].

• Step H2. Direct influence analysis

After the hierarchical structure has been established, obtaining initial direct-relation (IDR) matrices in the original DEMATEL is necessary for examining the key factors. Then, the elements involved in the same IDR matrix can be contrasted with one another. The participants in various IDR matrices, in contrast, cannot be compared directly but, rather, through a specific conversion. The values on the main diagonal of the IDR matrix are defined as zero in the original DEMATEL, meaning that a system factor has no influence on itself; therefore, the values of the main diagonal in the IDR matrices involved in each subsystem are zero. In hierarchical DEMATEL, the main diagonal in the IDR matrices in the system may not be zero because the factors in the subsystem influence each other. The cluster’s self-feedback mechanism in the ANP can also confirm the same concept [23].

• Step H3. Construct the super initial direct-relation matrix

The term “super IDR matrix” refers to a matrix that contains all the direct influence degrees among the constituent factors of complex systems under a particular rule. In the super IDR matrix, the direct impact grades of each pair of elements in the system must be determined, taking into account not only the direct effects between the factors present in each subsystem but also those involved in the various subsystems. Experts are requested to make judgments and provide the IDR matrices X = [x_qq′]_Q×Q included in system F and X_q = [$x_{nn\prime }^{q }$]_Nq×Nq included in the subsystem Fq, q = 1,⋯, Q. A pair of subsystems’ direct interaction in the one-level subsystem structure can be described by Equations (1) and (2):

$$\mathrm{Fq}=\left\{f_i^q\right|\mathrm{i}=1,\cdots , \mathrm{Nq}\}$$

(1)

$${\mathrm{Fq}}^{\prime }=\left\{f_{\mathrm{j}}^{q\prime }\right|\mathrm{j}=1,\cdots , {\mathrm{Nq}}^{\prime }\}$$

(2)

where $f_i^q$ and $f_{\mathrm{j}}^{q\prime }$ are the factors present in the two subsystems q and q′ and $x_{ij}^{qq\prime }$ is the direct influence degree on $f_i^q$ →$f_{\mathrm{j}}^{q\prime }$.

When q = q′, F_q→F_q′ indicates the direct influences among the subsystem’s constituent factors $x_{ij}^{qq\prime }$ = $x_{nn\prime }^q$; (∀i, j) is obtained within the IDR matrix using Equation (3):

$${\mathrm{X}}_{\mathrm{q}}={\left[x_{nn\prime }^q\right]}_{N_q\times N_q}.$$

(3)

when q ≠ q′, F_q→F_q′ signifies the direct effects among the factors included in two diverse subsystems F_q and F_q′, whose degrees $x_{ij}^{qq\prime }$ are not obtained directly and need to be calculated.

It is clear that the straight effect grade on subsystems F_q→F_q′ is equivalent to the total of a factor’s direct influence degrees $f_i^q$ →$f_j^{q\prime }$, i.e., x_qq′ = δ_qq′Σ_iΣ_j $x_{ij}^{qq\prime }$, where δ_qq′ is a coefficient of conversion between the two neighboring levels produced by the relative 0–4 scales. To unite the units of the scales at the lower level (factor) with those at the upper level (subsystem), the conversion coefficient δqq′ is added. As already stated, when q = q′, xqq′ is the degree of direct influence on subsystems and $x_{ij}^{qq\prime }$ (I, j = 1,⋯, N_q); experts have provided information on factors. Consequently, Equation (4) can be used to determine the conversion coefficient.

$${\mathsf{\delta }}_{{\mathrm{qq}}^{\prime }}={\mathrm{x}}_{{\mathrm{qq}}^{\prime }}/{{\displaystyle \sum }}_i{{\displaystyle \sum }}_j x_{ij}^{qq\prime }$$

(4)

The IDR matrix involved in F_q can be changed using Equation (5):

$${\overline{X}}_q=[\overline{x}{ }_{{nn}^{\prime }}^q]\mathrm{Nq}\times \mathrm{Nq}=[\mathsf{\delta }{\mathrm{qq}}^{\prime } x_{{nn}^{\prime }}^q]\mathrm{Nq}\times \mathrm{Nq}.$$

(5)

As a result, the subsystem level with the IDR matrix Xq have unified scales. IDR matrices on a couple of subsystems F_q→F_q′ can be constructed as ${\overline{x}}_{q{q}^{\prime }}={\left[{\overline{x}}_{ij}^{qq\prime }\right]}_{N_q\times N_{q\prime }}$ [23]. As a result, Equation (6) can be used to characterize the direct influence degree among factors on the same subsystem and various subsystems.

$${\overline{x}}_{ij}^{qq\prime }=\left\{\begin{array}{c}\begin{array}{c}\frac{x_{q{q}^{\prime }}}{{{\displaystyle \sum }}_i{{\displaystyle \sum }}_j{x}_{ij}^{q{q}^{\prime }}}{x}_{ij}^{qq\prime }, q=q^{\prime }\end{array}\\ \frac{z_i^q{z}_j^{q\prime }}{{{\displaystyle \sum }}_i{{\displaystyle \sum }}_j{z}_i^q{z}_j^{q\prime }}{x}_{qq\prime }, q\ne q\prime \end{array}\right. for i=1 ,\cdots , N_q, j=1,\cdots ,N_q$$

(6)

The prominence $z_i^q$ corresponding to factor $f_i^q$ can be determined using the same steps as the original DEMATEL, i.e., $z_i^q$ = $r_i^q$ + $d_i^q$ = ${\displaystyle \sum }_{i\prime }{t}_{ii\prime }^q$ + ${\displaystyle \sum }_{i\prime }{t}_{ii\prime }^q$ where $t_{ii\prime }^q$ is the degree of the overall effect on $f_i^q$→$f_{i\prime }^q$ belonging to the total-relation matrix T_q = [$t_{ii\prime }^q$]_Nq×Nq in the subsystem F_q. The super IDR matrix can be created using the IDR matrices for all subsystem pairs using Equation (7):

$$\overline{x }={\left[{\overline{x}}_{ij}\right]}_{N\ast N}=\left[\begin{array}{ccc}\left[\begin{array}{ccc}{\overline{x}}_{11}^{11}& \cdots & {\overline{x}}_{1N_1}^{11}\\ ⋮& \ddots & ⋮\\ {\overline{x}}_{N_{1}1}^{11}& \cdots & {\overline{x}}_{N_1{N}_1}^{11}\end{array}\right]& \cdots & \left[\begin{array}{ccc}{\overline{x}}_{11}^{1Q}& \cdots & {\overline{x}}_{1N_Q}^{1Q}\\ ⋮& \ddots & ⋮\\ {\overline{x}}_{N_{1}1}^{1Q}& \cdots & {\overline{x}}_{N_1{N}_Q}^{1Q}\end{array}\right]\\ ⋮& \ddots & ⋮\\ \left[\begin{array}{ccc}{\overline{x}}_{11}^{Q1}& \cdots & {\overline{x}}_{1N_1}^{Q1}\\ ⋮& \ddots & ⋮\\ {\overline{x}}_{N_{Q1}}^{Q1}& \cdots & {\overline{x}}_{N_Q{N}_1}^{Q1}\end{array}\right]& \cdots & \left[\begin{array}{ccc}{\overline{x}}_{11}^{QQ}& \cdots & {\overline{x}}_{1N_Q}^{QQ}\\ ⋮& \ddots & ⋮\\ {\overline{x}}_{N_{Q1}}^{QQ}& \cdots & {\overline{x}}_{N_Q{N}_Q}^{QQ}\end{array}\right]\end{array}\right]$$

(7)

4.2. Quantifying the Relationship among the Factors

To respond to the second research question and identify the relation among the obtained success factors, a hierarchal DEMATEL questionnaire including these factors was distributed to blockchain experts. The hierarchical DEMATEL technique was used because the blockchain model is an intricate system with numerous success factors, and a variety of types of influences and hierarchies. The questionnaire was sent to ten experts because blockchain is a new technology which appeared in 2008; therefore, there exist few experts in this field. Three experts responded promptly to this questionnaire.

The following three sections made up the questionnaire: (i) information about the industry sector, job description, and experience; (ii) classifications and categories of the success factors; and (iii) suggestions on any new success factors for blockchain implementation. The experts were requested to assess the pair-wise effect among each success factor for blockchain implementation by providing five linguistic options: no influence (“0”), low influence (“1”), medium influence (“2”), high influence (“3”), and extreme influence (“4”). Experts were asked to recommend if there exist additional success factors other than in our proposed model.

After collecting the experts’ responses, the response data was transformed into the direct-relationship-matrix format ahead of applying the hierarchical DEMATEL technique. The factors that affect blockchain implementation are fourteen factors, which were clustered into three dimensions: technological (F_1⊃1), organizational (F_1⊃2), and environmental (F_1⊃3). The factor set F₁ = {f₁,⋯, f₁₄} and the two-level structure of blockchain implementation is shown in

Table 5. The factors of blockchain implementation.

Dimension	Factor	Criteria
Technological (F1⊃1)	f1 f2 f3 f4 f5 f6	Scalability Infrastructural facility Complexity Compatibility Immaturity of technology Distributed design
Organizational (F1⊃2)	f7 f8 f9 f10 f11 f12	Lack of experience and knowledge Training facilities Top management support Organizational culture Financial constraints Adequate resource
Environmental (F1⊃3)	f13 f14	Laws and policy Competitive pressure

. This sub-section presents the results after applying the hierarchical DEMATEL technique.

The direct influences of the factors affecting blockchain implementation are all evaluated by experts using relative 0–4 scales, and all of the provided IDR matrices (X₁ and X_1⊃q2, q₂ = 1, ⋯, 3) which show the direct relationships between and within the subsystems F_1⊃q2. In addition, the values on the main diagonal of the IDR matrices implicated in the level-1 subsystems, as shown in Equation (8), may be greater than zero, indicating the factors involved in those subsystems may effect each other. However, the values on the primary diagonal of the IDR matrices implicated in the level-2 subsystems, as shown in Equations (9)–(11), are equal to zero, indicating that no direct influence exists between the factors and themselves.

$$x_1=\left[\begin{array}{ccc}3& 2& 1\\ 3& 3& 1\\ 1& 1& 1\end{array}\right]$$

(8)

$$x_{1\supset 1}=\left[\begin{array}{cccccc}0& 4& 4& 3& 4& 4\\ 4& 0& 3& 3& 2& 3\\ 3& 2& 0& 2& 3& 2\\ 3& 3& 2& 0& 3& 3\\ 3& 2& 3& 3& 0& 3\\ 3& 2& 3& 3& 2& 0\end{array}\right]$$

(9)

$$x_{1\supset 2}=\left[\begin{array}{cccccc}0& 1& 2& 3& 2& 2\\ 3& 0& 1& 3& 2& 3\\ 3& 4& 0& 4& 4& 4\\ 2& 2& 2& 0& 2& 2\\ 3& 4& 2& 2& 0& 4\\ 2& 4& 3& 2& 4& 0\end{array}\right]$$

(10)

$$x_{1\supset 3}=\left[\begin{array}{cc}0& 3\\ 2& 0\end{array}\right]$$

(11)

To create a super IDR matrix for all factors, the direct influences among the factors contained in the same or various subsystems should be altered or determined using Equation (5). Accordingly, the detailed direct-influence degrees among factors in subsystem F_1⊃2 described by the IDR matrix X_1⊃2 can be calculated as shown in Equation (12).

$${\overline{x}}_{1\supset 2}={\left[{\overline{x}}_{ij}^{1\supset 2}\right]}_{6\times 6}=\frac{x_{22}^1}{{{\displaystyle \sum }}_i{{\displaystyle \sum }}_j{x}_{ij}^{1\supset 2}} X_{1\supset 2}=\frac{3}{81}\times \left[\begin{array}{cccccc}0& 1& 2& 3& 2& 2\\ 3& 0& 1& 3& 2& 3\\ 3& 4& 0& 4& 4& 4\\ 2& 2& 2& 0& 2& 2\\ 3& 4& 2& 2& 0& 4\\ 2& 4& 3& 2& 4& 0\end{array}\right]$$

(12)

Meanwhile, ${\overline{x}}_{{qq}^{\prime }}$= ${\left[{\overline{x}}_{nn\prime }^{q{q}^{\prime } }\right]}_{Nq\times Nq}$ = ${\left[{\left(z_i^q{z}_j^{q\prime }/{\displaystyle \sum }_i{\displaystyle \sum }_j{z}_i^q z_j^{q\prime }\right)}_{x_{qq\prime }}{x}_{ij}^{qq\prime }\right]}_{Nq\times Nq}$ can be used to calculate the direct influence degrees among the factors included in two various subsystems. For q₂ = 3 and q′₂ = 2, the direct influence degrees on F_1⊃3→F_1⊃2 are the value 1 in the IDR matrix X₁; it means the sum of the direct-effect degrees that the elements included in subsystem F_1⊃3 influence those in F_1⊃2 is 1.

Table 6. The prominences of factors with total IDR matrices for F_1⊃2 and F_1⊃3.

Subsystem	Factor	${\mathit{x}}_1^{1\supset {\mathit{q}}_2}$	${\mathit{x}}_2^{1\supset {\mathit{q}}_2}$	${\mathit{x}}_3^{1\supset {\mathit{q}}_2}$	${\mathit{x}}_4^{1\supset {\mathit{q}}_2}$	${\mathit{x}}_5^{1\supset {\mathit{q}}_2}$	${\mathit{x}}_6^{1\supset {\mathit{q}}_2}$	ri	di	${\mathit{z}}_{\mathit{i}}^{1\supset {\mathit{q}}_2}$
F1⊃2	$x_1^{1\supset 2}$	0.215	0.299	0.270	0.366	0.326	0.339	1.817	2.300	4.118
	$x_2^{1\supset 2}$	0.382	0.276	0.251	0.395	0.355	0.411	2.072	2.612	4.684
	$x_3^{1\supset 2}$	0.521	0.615	0.306	0.580	0.584	0.611	3.218	1.828	5.047
	$x_4^{1\supset 2}$	0.314	0.344	0.269	0.231	0.327	0.342	1.829	2.428	4.257
	$x_5^{1\supset 2}$	0.448	0.531	0.343	0.422	0.331	0.528	2.605	2.441	5.047
	$x_6^{1\supset 2}$	0.418	0.545	0.388	0.432	0.516	0.366	2.666	2.598	5.264
F1⊃3	$x_1^{1\supset 3}$	2	3	—	—	—	—	5	4	9
	$x_2^{1\supset 3}$	2	2	—	—	—	—	4	5	9

displays the outcomes of determining the prominences of the two subsystems’ total IDR matrices for the concerned factors using steps 2–3 in Section 4.1. Thus, the IDR matrix on F_1⊃3→F_1⊃2 can be determined as:

$${\overline{x}}_{1\supset 32 }={\left[{\overline{x}}_{ij}^{1\supset 32}\right]}_{2\times 6}=\left[\begin{array}{cccccc}0.0724& 0.0824& 0.0887& 0.0749& 0.0887& 0.0926\\ 0.0724& 0.0824& 0.0887& 0.0749& 0.0887& 0.0926\end{array}\right]$$

(13)

Repeating the above procedures for q₂, q′₂ = 1,2, 3, the determined IDR matrices ${\overline{x}}_{1\supset q_{2}q{\prime }_2}$ are integrated to create the super IDR matrix X1.

5. Results and Discussions

After constructing the super IDR matrix X₁, the DEMATEL is applied to generate the total relation matrix. Then, the prominence and ranking of every factor are then determined, as shown in

Table 7. The prominences of factors.

	Factor		${\overline{\mathit{r}}}_{\mathit{i}}-{\overline{\mathit{d}}}_{\mathit{i}}$	${\overline{\mathit{r}}}_{\mathit{i}}+{\overline{\mathit{d}}}_{\mathit{i}}$	Ranking
Technological	f1	Scalability	−0.23	5.73	3
	f2	Infrastructural facility	−0.22	4.78	8
	f3	Complexity	−0.57	4.64	12
	f4	Compatibility	−0.36	4.76	11
	f5	Immaturity of technology	−0.37	4.76	10
	f6	Distributed design	−0.51	4.77	9
Organizational	f7	Lack of experience and knowledge	0.10	4.23	14
	f8	Training facilities	0.11	4.84	7
	f9	Top management support	1.08	5.25	5
	f10	Organizational culture	0.05	4.38	13
	f11	Financial constraints	0.48	5.22	6
	f12	Adequate resource	0.44	5.43	4
Enviomental	f13	Laws and policy	0.45	7.17	1
	f14	Competitive pressure	−0.43	7.17	2

. As a result, when the value of (R − D) is positive, the success factor is a net causer. On the other hand, when the value of (R − D) is negative, the success factor behaves as a net receiver [74]. The experts ranked the environmental context as the most critical one for achieving successful blockchain adoption. Therefore, law and policies and competitive pressure are the top two factors needed for blockchain adoption. Among the technological context, only the scalability of blockchain is considered the top significant factor for blockchain adoption. Regarding the organizational context, adequate resources, top management support, and financial constraints are highly ranked.

On the other hand, the surveyed experts do not believe that technology represents a barrier to the successful adoption of blockchain. Therefore, infrastructure, complexity, compatibility, and the immaturity of blockchain technology are given lower ranks, as shown in Table 7. Moreover, the reason for not ranking distributed design highly may be that the experts considered the permissible blockchain applications in their minds.

The value of $\propto$, the average value of the total relationship matrix (T), is 0.187. Therefore, the elements of (T) with a higher value than 0.187 are considered to be significant relationships among the factors. The links in Figure 5 represent such significant relations among the six top prominences factors. Based on the value of (r-d) shown in Table 7, both scalability and the competitive pressures are net receivers. At the same time, management support, financial constraints, adequate resources, and law and policy are net causers.

6. Conclusions

Identifying blockchain critical success factors will help the industry and governments prioritize the strategic factors and achieve successful blockchain adoption. This research identifies those factors affecting blockchain-technology spreading and classifies them into three contexts: technology, organization, and environment. Moreover, measuring the success of blockchain adoption requires a tool to guide the implementation toward successful adoption. This research provides a set of sixteen indicators clustered into five groups that can signal this success: overall performance, system robustness, data robustness, accessibility, and overall cost.

Finally, the hierarchical DEMATEL was utilized to rank the success factors and determine the relationship among those factors. The top-two ranked factors are competitive pressures and laws and policies, which belong to the environmental context. The presented model could serve as a valuable framework for further and more in-depth research of the critical factors of other blockchain-technologies’ adoption, such as cryptocurrencies and smart contracts. Moreover, proposing tools to standardize the measuring of each of the sixteen KPIs may be an exciting research area.

Second, in the future, we suggest empirically testing our framework in some applications such as last-mile delivery, e-government, and public health to provide a greater validity of the critical factors influencing the successful adoption of blockchain technologies in different applications.