Sustainable Competitive Advantage Driven by Big Data Analytics and Innovation

: Big data analytics (BDA) is one of the main pillars of Industry 4.0. It has become a promising tool for supporting the competitive advantages of firms by enhancing data-driven performance. Meanwhile, the scarcity of resources on a worldwide level has forced firms to consider sustainable-based performance as a critical issue. Additionally, the literature confirms that BDA and innovation can enhance firms’ performance, leading to competitive advantage. However, there is a lack of studies that examine whether or not BDA and innovation alone can sustain a firm’s competitive advantage. Drawing on previous studies and dynamic capability theory, this study proposes that big data analytics capabilities (BDAC), supported by a high level of data availability (DA), can improve innovation capabilities (IC) and, hence, lead to the development of a sustainable competitive advantage (SCA). This study examines the proposed hypotheses by surveying 117 manufacturing firms and analyzing responses via partial least squares–structural equation modeling (PLS-SEM) statistical software. Findings reveal that BDAC relies significantly on the degree of DA and has a significant role in increasing IC. Furthermore, the analysis confirms that IC has a significant and direct effect on a firm’s SCA, while BDAC has no direct effect on SCA. This study provides valuable insights for manufacturing firms to improve their sustainable business performance and theoretical and practical insights into BDA implementation issues in attaining sustainability in processes.


Introduction
Recently, modern manufacturing systems and advanced technologies have been generating data in greater volumes from a variety of sources in real time, a process referred to as "big data" [1,2]. Businesses today are increasingly utilizing advanced technologies like the Internet of Things (IoT), cyber-physical systems and human-robot collaboration that generate massive amounts of data [3][4][5][6][7][8][9][10][11][12].
However, traditional software is incapable of handling such data properly [13]. This is the role of big data analytics (BDA), extracting the value from these data to enhance a firm's competitive advantage [14]. Simply, BDA is the utilization of diverse analytical approaches to handle a variety of large-scale and complex data to extract usable results that serve a firm's performance [15,16].
Despite many studies having validated the significant effect of BDA on firm performance [17][18][19][20], many firms are still not certain regarding whether or not to adopt BDA for a variety of reasons. They might lack a sufficient understanding of or capabilities with which to realize its prerequisites or to integrate BDA with their existing processes and systems to extract value [13]. Firms also overlook organizational Drawing on these identified research gaps, this study aims to propose a strategic conceptual model that considers the effect of DA on BDA and the effect of BDA and IC on SCA. This study also attempts to benefit practitioners by providing an explanation and statistical validation for the proposed model. This study makes many contributions to the field. Firstly, it considers different topics and investigates their roles in attaining SCA; secondly, it provides empirical evidence supporting the extant literature; and thirdly, it considers three diverse aspects of sustainability in one model. Consequently, this study aims to address the following research questions: (1) How does DA affect BDAC? (2) How does BDAC contribute to a firm's IC? (3) How does BDAC enhance a firm's SCA? (4) What is the relationship between IC and a firm's SCA? Figure 1 portrays this study's research model, which consists of four reflective constructs: DA, BDAC, IC and SCA. The model is theorized based on the dynamic capabilities view, in which firms must adapt and innovate their strategic capabilities to cope with the rapid advances and changing trends in the marketplace [47]. This model represents the dynamic capabilities provided by BDAC and IC and is based on the proposition that DA is capable of enhancing firms' sustainable performance, empowered by BDA and innovative processes. Finally, the aforementioned capabilities are proposed to enhance firms' SCA. Table 1 clarifies the constructs' items and their referents.  BDAC is a dynamic capability since it possesses the characteristics of dynamic capabilities, e.g., repeatability, which create an experience that is difficult to imitate and thus create valuable assets that enhance competitive advantage [56]. However, to extract value from BDA, data must be available; a lack of data can be critical to overall performance and, hence, competitive advantage [57]. In addition, lost data cannot be recovered after being identified [49]. Therefore, firms aim to maximize the availability of their data to enhance their BDAC [58] by addressing related issues, such as data source, size and collection method [57].

Theoretical Model
Advanced Industry 4.0 based manufacturing technologies like cyber-physical systems and human-robot collaboration with sensing devices, e.g., depth camera, proximity sensors, radio frequency identification (RFID), etc., can connect to the IoT to generate big data [4][5][6][7][8][9][10][11][12]29,31] that can be connected to the cloud to ensure DA [26]. In addition, management information systems, such as enterprise resource planning (ERP), can be integrated with BDA [28,30]. In this regard, advanced wireless networks are supposed to support real-time DA, e.g., 5G [32,33]. However, DA is an essential factor that contributes to the effectiveness of BDA [59]. Therefore, since higher DA promotes better data analytics, we propose that DA is an enabler of BDAC. Hereafter, we propose the following hypothesis:

Research Hypothesis 2: BDAC as an Enabler of IC
BDA has high potential in promoting innovation in products, processes and markets since it provides advanced decision-making prospects [60,61]. BDAC triggers firms to leverage big data, boosting innovation and firm development [41], via enabling novel solutions and rapid adaptation to changes and challenges on a pseudo-real-time basis [62]. In this regard, Erevelles et al. affirmed that BDAC assists firms in creating innovative opportunities [37]. For instance, firms can utilize their customers' big data to innovate their products and services [63]. Moreover, Niebel et al. confirmed that the use of BDAC stimulates firms' tendencies to innovate their products, processes and solutions [60].
Firms can utilize BDA in different data sources, e.g., customer analysis, processes and equipment monitoring, etc. [14,64], to enhance their IC in products, processes and markets [51,52]. In addition, BDAC can facilitate the routine processes and efforts exerted by an innovation team, thus enhancing IC [41]. Moreover, Mikalef et al. [65] proved that BDAC has a positive influence on IC. Therefore, we propose the following hypothesis: Hypothesis 2 (H2). BDAC have a positive effect on IC.

Research Hypotheses 3 and 4: The Role of BDAC and IC in Enhancing SCA
Both capabilities-BDA and innovation-allow firms to renovate and reconfigure their resources [56]. Breznik et al. [66] also concluded that IC can be considered as both a part and a result of the dynamic capabilities that a firm possesses and that both capabilities contribute to the development of a firm's SCA. BDA permits data-driven decision-making that creates novel ways to innovate and learn how to create value, thus improving a firm's sustainable performance [67]. Consequently, we aim to examine the effects of both capabilities on a firm's SCA.
Firstly, BDAC represents a key source of knowledge creation [64]. This is attained by exploiting different data sources, e.g., production, processes and customers, to support firms in making better decisions, enhancing performance and, thus, sustaining competitive advantage [24]. For instance, Dubey et al. found a significant effect of BDA on a firm's environmental and social sustainability [39]. Additionally, Côrte-Real et al. examined the significant impact of BDA on firms' financial performance. Secondly, innovation as a dynamic capability plays a crucial role in increasing sales, profits and performance excellence through product and process development [68]. Moreover, innovativeness that creates faster adaption to market changes encourages the creation of competitive advantage [42].
Thus, IC contributes significantly to firm performance and competitiveness [69,70]. Both dynamic capabilities, i.e., BDA and innovation, assist firms in overcoming external challenges and competencies by enhancing and sustaining their competitive advantage and overall performance [38,42]. Consequently, we propose the following hypotheses: Hypothesis 3 (H3). As a dynamic capability, BDAC has positive effect on firms' SCA.
Hypothesis 4 (H4). As a dynamic capability, IC has positive effect on firms' SCA.

Sustainable Competitive Advantage (SCA)
Competitive advantage is attained when a firm becomes superior in its success as compared to its current or potential competitors [14]. Moreover, advanced technologies, like BDA, are enablers of sustainable performance by promoting data-driven performance [36]. However, a firm's performance index, compared to its competitors, is a widespread way to indicate competitive advantage [14,44,71]. Thus, to be consistent with this conceptualization, we operationalize the SCA construct to function as a measure for superior performance compared with competitors [71] in terms of social, environmental and economic performance. Therefore, based on previous studies that provide the competitive advantage construct, this construct is the first that constitutes the three aspects of sustainability on the firm level.

Survey Design
A web-based questionnaire was designed to test the research model and hypotheses. The questionnaire included 15 questions, all rated by respondents on a five-point Likert scale from "1: strongly disagree" to "5: strongly agree". The questionnaire was developed based on the previous literature. The questions were used to establish multi-item reflective measures for the four constructs of the research model. The minimum sample size, determined by G*Power 3.1 software, was 117 samples. Given this, the values for significance level, power and effect size were set to 5%, 90% and 0.3, respectively.

Data Collection and Analysis Method
This study targeted manufacturing companies from different industrial sectors who were familiar with and had some experience in Industry 4.0 technologies to test the hypotheses. The study did not include the service sectors, since some questions were specifically developed for manufacturing sectors. In addition, the questionnaire aimed to collect data as a primary source to test the hypotheses of the model's constructs. Respondents, members of upper-level management, were targeted to fill out the questionnaire to increase the credibility of the gathered data. This study used partial least squares-structural equation modeling (PLS-SEM) as a statistical software tool for data analysis to test our hypotheses because of its powerful capability in analyzing small sample sizes and its efficiency in analyzing mediating variables and indirect relationships [72]. SmartPLS can be also directly used for the assessment of convergent validity, discriminant validity and reliability. It has a very user-friendly interface and it has very attractive graphical outputs. Its output is very flexible compared to other software packages.

Results
PLS-SEM software analyzes the conceptual model in two steps: first, by assessing the measurement model (see Figure A1); second, by assessing the structural model (see Figure A2). The results and explanations of both steps are detailed in the following two sub-sections.

The Measurement Model: Reliability and Validity Analysis
Reliability and validity tests of the constructs should be conducted to assess the reflective measurement models [73]. Both tests are prerequisites to structural model assessment [74]. These tests aim to assess the internal consistency of the values recorded by the respondents by examining each variable versus its related indicator's values and the validity of the questionnaire's questions. The internal consistency, i.e., reliability, was assessed via Cronbach's alpha (CA) and composite reliability (CR) tests, while the validity was assessed by convergent and discriminant validity tests [72].
Cronbach's alpha is used to measure the internal consistency of a construct's items [75]. The reliability of items/factors must be 0.60 or higher. In addition, the composite reliability, which is used to double-check the internal consistency, should have values above 0.7 [76]. The results are reported in Table 2, which indicates that all constructs are reliable since their corresponding Cronbach's alpha values are above 0.6 and their composite reliability values are higher than 0.7. Therefore, the results affirm that the internal consistency of all constructs is reliable and acceptable. Convergent and discriminant validity tests were used to assess the correlation between each construct and its related items [77]. Convergent validity is measured by the average variance extracted (AVE) for all constructs' items with a minimum acceptable value of 0.50. This value indicates that the construct explains 50% or more of the variance of its items [76]. The results reported in Table 2 indicate that all AVE values are above 0.5, thus confirming convergent validity. Discriminant validity is measured by calculating the square root of all constructs' AVE and comparing the resulting value for each construct with other constructs. Hence, when the correlations within each construct exceed the correlations with all other constructs, the discriminant validity is accepted [78]. The results in Table 3 indicate that the discriminant validity is acceptable. In addition, the Heterotrait-Monotrait ratio (HTMT) test was used to double-check the discriminant validity; it measures the degree of differences between overlapping constructs. Accepted values range between zero and one and values less than one indicate good reliability [79]. The HTMT values for this study are below 0.90, as shown in Table 4; therefore, discriminant validity is confirmed.  Table 5 depicts the PLS estimation of the factor outer loadings of the reflective indicators. Acceptable values are equal to 0.7 or higher for outer loadings [80] or are lower than 0.7 but above 0.4 if the indicator has been recently established [81]. It is essential to check the outer model by testing the t-statistics with values above 1.96 to confirm that the outer model loadings are highly significant. Thus, according to the outer loading and t-test values, the indicators are highly significant.

Structural Model
The coefficient of determination (R 2 ) is a statistical value that explains the amount of endogenous construct variance that is explained by the exogenous constructs, with acceptable values between zero and one [76]. The R 2 value for SCA is 0.440, depicted in Table 2, which implies that 44.0% of the variance in the SCA constructs is explained by the BDAC and IC constructs. In addition, PLS-SEM was used to calculate the bootstrapping procedure for assessing the structural model and to obtain the t-values [72], as shown in Table 6. Table 6. Results of the model fit. It is significant to verify the path coefficient values and statistical significance of the relationships between the latent variables in the structural model to evaluate the proposed hypotheses. This verification is performed by assessing the standardized path coefficients (B-value), with values equal to or above 0.1 [80], and the associated t-value, having values above 1.96 or below −1.96 at a 0.05 level of significance, i.e., p-value [82].

B-Value STDEV t-Value T Statistics p-Value
The results of the model fit, provided in Table 6, confirm that DA has a significant and positive impact on BDAC (B-value = 0.410, t-value = 5.256) and that BDAC has a significant impact on IC (B-value = 0.491, t-value = 6.062) and is not significant on SCA (B-value = 0.106, t-value = 1.178). IC has a positive and significant impact on SCA (B-value = 0.605, t-value = 8.897). However, according to the results, IC has the strongest effect on SCA (0.605). Moreover, the results given in Table 6 confirm that H1, H2 and H4 are supported. However, the relation between BDAC and SCA is not supported.

Discussion
The purpose of this study was to investigate the influence of BDA and innovation as dynamic capabilities on firms' SCA. The empirical results demonstrate that BDAC has a significant effect on IC and that IC enhances SCA but BDA capabilities do not. The outcome of this study makes several contributions to the existing body of knowledge in this field. Specifically, this study is the first to explore the effect of DA level on BDAC and the role of BDAC and IC in attaining SCA.
Our findings prove that an increase in DA level positively affects BDAC; the results of the conceptual model support H1. Indeed, when a firm's system is integrated with BDA-all components are connected to IoT, and backup system is effective-the availability of data will increase, which, in turn, will enhance BDAC. However, the BDAC will not be enhanced if the DA in the manufacturing companies is not supported with IoT technology enablers.
In addition, an increase in BDAC has a positive effect on IC, so the results of the model support H2. Therefore, firms implementing BDA are more innovative than others that do not [60]. This result is consistent with the findings of Mikalef [66] and Lozada [83]. Moreover, an increase in IC has a positive effect on SCA, so the results of this study support H4. This result confirms the findings of many similar studies [14,42,84]. In practice, these results are obtained when a firm continuously utilizes new and up-to-date technologies in processes, introduces new products, uses the latest technologies in its products or adopts innovativeness in marketing and promotion. Additionally, the results emphasize that BDAC has a key role in enhancing IC by prompting a firm to adopt the latest technologies in products and processes, introduce new products and innovate in marketing and promotion. In this regard, if the manufacturing companies did not support the BDAC variables, the ability of the companies to cope with the rapid advancements in technologies and market changes will be reduced. Furthermore, whilst reviewing the related literature, we observed that very few studies investigated the BDAC effect on firm SCA. The only study found, to the extent of our review, was a study published in 2020 by Kamble et al. [36], which was conducted at the supply chain level in the agriculture industry and concluded that BDAC is an enabler of sustainable performance. Moreover, no other study has previously investigated the role of DA in reinforcing BDAC. Therefore, this study is the first in conceptualizing the DA construct and synthesizing the SCA construct at the firm level.
In practice, this model is applicable and very efficient in highly customized and technologically advanced production environments where massive amounts of data are generated. Therefore, this model is not useful when it is employed in traditional mass production systems where no highly customized products are needed and the big data will not benefit the producers compared with amount of investment in this filed. In addition to this, the model is not applicable when firms do not have an IT system or enough flexibility to adapt to a wide range of market requirements or to respond to all customer requirements. In such cases, the BD and BDAC may not impact the IC and thus the SCA will not be enhanced. The same can be considered for companies of low competition or market monopoly.
Finally, this study examines DA issues, e.g., network requirements, as a prerequisite for BDA. In addition, this study assists managers in deciding where to invest and allocate resources, focusing on dynamic capabilities as essential enablers for SCA at the firm level in the manufacturing industry.

Limitations and Future Research Directions
This study provides valuable insights for manufacturing companies to improve sustainable competitive advantage as well as theoretical and practical insights into BDA implementation issues in attaining sustainability in processes. However, this study has its own limitations and weaknesses, which can be considered as future research opportunities. For example, the research questions of this study have been developed for different types of manufacturing companies in industrial sectors and it was not a sector-based study; Therefore, analyzing the differentiation between different industrial sectors would be inaccurate. Moreover, the service sector was out of the scope of this study and the findings of this study cannot be generalized to the service sectors. This is because the service sector has its own variables which cannot be measured through the same research questions of this study.
Although the sample size was enough to make decisions about the relationships between hypotheses using SmartPLS, for more precise results, it needs to be increased. Moreover, the size of the surveyed manufacturing companies was not taken into consideration in the study, which may impact the findings of the model. As BDA is one of the major pillars of Industry 4.0, the study lacks a real-world case study that explains the implementation of BDA within the Industry 4.0 system. These aspects should be considered as future research directions.

Conclusions
In answering the research questions introduced by this study, the results confirm that DA has a significant influence on BDAC, which, in turn, has a significant role in enhancing the relationship with IC. On the other hand, BDAC does not have a significant role in enhancing a firm's SCA. In addition, the results indicate that IC is a significant mediator in the BDAC and SCA relationship. Thus, managers can enhance a firm's SCA by investing in innovation; this might put the firm at the top in rivalries. In addition, they also should be aware of the prerequisites of and issues with BDA implementation.
A firm can make decisions based on the findings of each hypothesis. For example, in H1, since DA has a significant and positive effect on BDAC, firms must digitize all firm levels, i.e., enterprise, operational and manufacturing, to enhance data availability for BDA. At the manufacturing level, firms can digitalize their manufacturing systems by connecting smart entities such as machines, products, tools and devices to the IoT system, which in turn provides a stream of data to BDA. Moreover, firms will lose the benefits of horizontal and vertical connectivity if they do not employ a proper backup system for their generated BD.
H2 confirms that BDAC have a significant and positive role in enhancing firms' IC since continuous examination of the generated BD allows firms to enhance the main pillars of the innovation process, which are product, processes, market and organization innovation. Thus, BDAC maintains the firms' awareness of new technologies to which they must adapt in order to enhance their processes and products and thus be able to distinguish themselves among other competitors in the market. It is concluded that BDAC will enhance innovation management capabilities in manufacturing companies.
In H3, it is found that BDAC as a dynamic capability has positive effect on firms' SCA. Our results have not shown a significant effect of BDAC on firms' SCA. This result is justifiable; BDAC do not ensure that firms will implement the analyzed data and the available reports in the three aspects of sustainability.
As a dynamic capability, IC in H4 has a positive and significant effect on firms' SCA. This encourages the decision-makers to invest in technologies where the innovation capabilities are enhanced. This will reinforce the competitive edge of the firm.
Clearly, using BDA as an advanced analytical tool would be supportive of firms' managerial decision-making, as posited by [85]. The adoption of BDAC in a firm offers several opportunities for improving the IC of firms in the manufacturing industry.
This study proposes a beneficial framework that handles many topics such as DA, BDAC, IC and their roles in reinforcing the SCA of manufacturing firms. However, there is still room for further improvement; for example, the approach of this study limits the generalization of findings to the service sector. Therefore, a possible future work could be a sector-based study approach (i.e., either industrial or service sectors), where more specific variables can be added to the research questions to address specific variables in each sector and thus more accurate results could be obtained and sector differentiation could be precisely analyzed. Funding: This study received funding from King Saud University through researchers supporting project number (RSP-2020/145). The APC was funded by King Saud University through researchers supporting project number (RSP-2020/145).

Conflicts of Interest:
The authors declare no conflict of interest.