Big data analytics in supply chain management and its impact on business performance

Resource-Based Theory (RBT) originated in the field of economics and was used more in the field of strategic management, and then gradually evolved into the fields of operations management, marketing, and human resource management and entrepreneurship management [24]. There are three main perspectives included in RBT, the resource-based perspective, the knowledge-based perspective, and the relationship-based perspective.

Enterprise resources are the basis for the formation of capabilities, through which capabilities are established and further developed into core capabilities. RBT theory summarizes the formation of core capabilities of an enterprise into three stages, and promotes the transformation of the three stages under the role of the environment, as shown in the specific model in Figure 1. In the first stage, the enterprise employees transform the enterprise resources into a kind of personal professional skills through learning, for example, the employees with IT technology background are converted to learn the big data analysis technology and become more and more professional in the field of big data analysis through work practice. In the second stage, each department of the enterprise learns to develop the work practices of the employees in the organization into an efficient operational practice. In the third stage, the organization learns how to respond to the external environment and develops its own core competencies.

Figure 1.

Formation process of core competence

The primary goal of enterprises to carry out big data analysis is to establish big data analysis capability, and what resources need to be available to enhance big data analysis capability is a key issue that enterprises need to think about, and resource-based theory provides a theoretical basis for the sources of big data analysis capability formation.

While RBT emphasizes the selection of resources, dynamic capabilities emphasize the development, allocation and renewal of resources. Having basic resources does not directly provide a sustainable competitive advantage for a firm, whereas dynamic capabilities help firms to solve problems systematically, allow them to make timely, market-oriented decisions, and introduce innovative changes in their operations that are more likely to produce superior performance [25].

The evolution of capabilities in a dynamic environment is proposed from an IT perspective as shown in Figure 2, where IT architectural capabilities provide access to information, increase information flow and collaboration opportunities, and are further elevated to the level of business capabilities of an organization that include dynamic capabilities, rapid response capabilities, and operational capabilities to bring about strategic advantages. Dynamic capabilities are also hierarchical, and the higher the hierarchy, the more abstract and complex they become. Dynamic capabilities are categorized into lower-order and higher-order dynamic capabilities, with higher-order dynamic capabilities representing higher levels of strategic insight and complex capabilities that are more difficult to observe and decode. Big data analytics capabilities face the status quo of rapidly changing technology and high investment in realizing the value of the enterprise, so the integrated hierarchical and contextual perspectives contribute to a better understanding of the value of big data analytics in this paper.

Figure 2.

Hierarchical evolution process of capability

Value theory originated as an advancement of resource-based theory to help firms identify valuable resources. For enterprises, internal resources create use value and exchange value is realized through interaction with the outside world. At the same time they categorized value activities into 3 parts: value creation, value capture and value destruction [26].

Innovation is the source of enterprise survival and development, the key to the success or failure of enterprises lies in innovation, and the key to the success or failure of innovation lies in management [27]. The innovation choice space model is shown in Figure 3 to analyze the sources of innovation activities. The vertical coordinate of the model represents the two dimensions of innovation, “incremental and breakthrough innovation”, and the horizontal coordinate represents “the degree of environmental uncertainty (the higher the degree represents the greater the number of environmental elements and the closer the potential interaction between the elements)”. Regions 1 and 2 are two types of innovation activities of firms when the level of environmental uncertainty is low, when firms still innovate according to the established framework due to the relative stability of environmental elements. Region 1 Exploitation is the most basic category of innovation activity, where a stable shared framework is assumed to exist and firms realize improvements in the tools and methods of technological research and market research through adaptive and incremental exploitation. Area 2 limited search is an option when entering new areas that have stretched the known boundaries of the firm, and the firm’s innovative activities are usually realized by deepening its connections with its stakeholders.

Figure 3.

Innovation choice space model

Regions 3 and 4 are two types of innovation activities for firms when the level of environmental uncertainty is high, when firms need to seek new frameworks to sustain their growth or ensure competitive advantage due to the close interaction between environmental factors and firms. Area 3 reconfiguration is an attempt to explore different permutations of environmental factors and reorganize the business model to match the environmental factors. This phase is riskier for the firm and may lead to failure of the transformation or to the emergence of a strong business model and the formation of a virtuous cycle. Area 4 synergistic evolution is an unprecedented innovation for firms and represents a passive innovation activity with blurred boundaries.

The innovation theory sheds light on this paper in the following ways:

First, the innovation theory explains that the innovation activities of enterprises need to experience the transformation of factor combination from progressive to breakthrough. Strategic flexibility emphasizes the flexible adjustment of the resource portfolio, and strategic innovation emphasizes the breakthrough, reorganization and innovation of the resource portfolio model, and it is obvious that strategic flexibility is the foundation and prerequisite for enterprises to realize strategic innovation, which provides the basis for the research hypothesis that strategic flexibility promotes strategic innovation.

Secondly, the innovation choice space model illustrates that along with the enhancement of the enterprise’s recognition of environmental factors, the enterprise is able to realize the transformation from the existing framework to the new framework and achieve sustainable growth.

Specifically, with the enhancement of analyzing power of consumer market, enterprises can satisfy consumers’ changing consumer needs and further creatively reshape consumer needs, so as to seize the consumer market and establish competitive advantages, which provides the basis for this paper to explore the chain relationship of “big data analyzing power - strategic flexibility - strategic innovation - enterprise performance” based on the theory of innovation. This provides the basis for this paper to investigate the chain relationship of “big data analytic capability--strategic flexibility--strategic innovation--enterprise performance” based on innovation theory.

Supply chain management plays a pivotal role for organizations, especially in the face of unexpected crises, timely supply/delivery and responding to customer demand often tends to plague organizations. Taking a retail enterprise as an example, in understanding the impact of big data analytics capability on supply agility, it is important to reflect not only the results but also the process. First understand the enterprise’s big data analytics in the supply side of the application mechanism as shown in Figure 4, the diversification of application scenarios to enhance the depth of cooperation with suppliers, mainly reflected in product development, inventory replenishment, product traceability, advertising and marketing, supply logistics and supply selection, a total of six aspects.

Figure 4.

Big data analysis on the supply end

In this paper, we have sorted out the relevant theories of big data capability, strategic flexibility, strategic innovation, environmental uncertainty, enterprise performance, and Internet enterprise. Further, based on the resource base theory, dynamic capability theory, innovation theory, and value theory, and taking supply chain management enterprises as the research object, the impact of big data analytic capability on enterprise performance is discussed in depth, the important mediators and weighting factors between the two are analyzed, and the theoretical model of the mechanism of the role of big data analytic capability on enterprise performance is constructed as shown in Fig. 5.

Figure 5.

Theoretical model of big data analysis ability on enterprise performance

The theoretical model contains the following main elements:

The use of big data technology in the supply chain can enhance the flexibility of the supply chain, reduce the whiplash effect among supply chain members, strengthen the integration of supply chain resources and accelerate the transmission of information, so that supply chain-related enterprises can better cope with the increasingly competitive market environment and enhance their market competitiveness.

Based on big data technology and closed-loop supply chain, a competitive intelligence system has been established to provide effective support for realizing supply chain flexibilization. Based on the dynamic capability theory to study the realization mechanism of big data on enterprise supply chain agility, it is found that the reasonable use of big data analysis technology is of great significance to improve the flexibility of supply chain.

Therefore, this paper makes the following hypotheses:

H1, big data analytics capability positively affects supply chain flexibility.

With the rapid development of big data technology, the competition among enterprises has developed from the competition of products to the competition of comprehensive strength of enterprises. Big data can indirectly improve the overall performance of enterprises by improving tangible and intangible benefits. The relationship model between data empowerment, strategic flexibility and enterprise performance is constructed with strategic flexibility as the mediating variable, and it is found that data empowerment has a significant effect on enterprise performance. Enterprises apply big data technology to transform data into various information assets, from which they can analyze the needs of customers to realize precise marketing, increase the order rate of customers, expand market share, and promote the growth of enterprise performance.

Therefore, this paper makes the following hypotheses:

H4, Big data analytics capability positively affects able firm performance.

The development of supply chain flexibilization allows firms to engage in real-time sharing of information in order to provide the most rapid response to changing market and customer demands, and to coordinate inventory quantities, reduce unnecessary cost expenditures, and promote growth in performance among supply chain firms.

Therefore, this paper makes the following hypotheses:

H5, supply chain flexibility positively affects firm performance.

To sum up, on the one hand, big data analytics can uncover useful information from a large amount of data, thus improving the flexibility of enterprise supply chain management. On the other hand, the enhancement of supply chain flexibility can also improve enterprise performance, in which the agile supply chain effect plays a partly mediating role in enterprise partnership and enterprise performance. This paper adopts process flexibility and logistics flexibility to measure supply chain flexibility, and analyzes the mediating effect of supply chain flexibility between big data and enterprise performance.

Therefore, this paper makes the following assumptions:

H8, Supply chain flexibility mediates between big data analytics capabilities and firm performance.

Since this study was conducted in the context of big data analysis in enterprise supply chain management, the Big Data Analytics Competency Measurement Scale was used to measure big data analytics competency, and the specific test items measured are shown in Table 1. Big data analysis ability is divided into 3 dimensions, namely, resource acquisition ability, integration and analysis ability, and insight and prediction ability, with a total of 12 items.

For the measurement of firm performance, the instrument used in this paper is shown in Table 2, which will measure firm performance in two dimensions, market performance and operational performance, with a total of eight question items.

In this study, supply chain flexibility refers to, through effective supply chain management, the ability to collaborate with supply chain members to economically and quickly produce products oriented to consumer demand. This paper combines the two divisions of supply chain flexibility structural dimensions, and based on this design enterprise supply chain flexibility measurement tool as shown in Table 3. The scale divides supply chain flexibility into new product flexibility, procurement flexibility, product flexibility, delivery flexibility and information system flexibility, totaling five dimensions. The scale adopts a Likert 5-level scale with a total of 22 items.

In order to ensure the reliability of the questionnaire, the study analyzed the reliability of the variable scales of the valid questionnaire using SPSS 26.0, where the Cronbach’s a value greater than 0.7 indicates good reliability, and the correlation coefficients of the question items with the total score, i.e., the CITC value is greater than 0.4 indicates a high level of reliability. The specific results are shown below.

For the scales in this paper, both supply chain flexibility and enterprise performance use mature scales, so they have good content validity. The scale of big data analyzing ability is also revised on the basis of the scales of previous studies, which also has good content validity. The measurement of construct validity was mainly conducted by factor analysis. Before conducting exploratory factor analysis, KMO test and Bartlett’s sphericity test were first performed to determine whether the data were suitable for factor analysis. If the KMO value is above 0.9, the data is very suitable for factor analysis, and above 0.8 is more suitable for factor analysis. Meanwhile, Bartlett’s spherical test result reaches the significant level, i.e. Sig.<0.05, the data has good correlation and is suitable for factor analysis.

In order to ensure the representativeness of the data, before the empirical analysis, the 264 valid samples recovered were subjected to general collation statistics, and the results are shown in Table 13. From the results, the position is mainly the manager of the company. The age of the company is mainly more than 5 years (87.5%). In terms of company ownership, private companies accounted for the largest share of 64.77%. In terms of industry distribution, manufacturing companies accounted for the largest proportion at 45.83%. The largest percentage of company capitalization is in not more than 10 million yuan, which is 38.26%. Therefore, it can be considered that the representation of the data collected in this paper is good to a certain extent.

Table 14 shows the descriptive statistical analysis of the variables, which demonstrates the descriptive statistical results of the independent variable (firms’ big data analytics capabilities), the mediator variable (supply chain flexibility), and the dependent variable (firms’ performance), with a total sample of 264, for the extremes, means, and standard deviations. The results show that the data collected in this study are good. In particular, the mean values of market performance and operational performance in enterprise performance are 5.7819 and 5.7858, respectively, which are in line with the actual situation, and the statistical results of the variables have a certain degree of credibility.

Before testing the path coefficients in the structural model, this paper first provides a discriminative description of the fitness and validity of the structural model.

Regarding the fitness of the structural model, the index provided by SmartPLS software is the SRMR value. When the SRMR value is less than 0.08, it indicates that the fitness of the model is good. In this paper, the software’s PLS algorithm is used to calculate the SRMR value, and the result shows that the SRMR value is 0.071 < 0.08, which indicates that the fitness of this model is good.

Regarding the validity of the structural model, the software provides R² For measuring the explanatory power of the structural model. Q² is used to measure the predictive accuracy of the structural model. Where, according to the discriminant criteria, the predictive accuracy of the structural model is adequate, moderate, and weak when R² is 0.75, 0.50, and 0.25, respectively. When Q² is 0.50, 0.25 and 0.00, respectively, the predictive accuracy of the structural model is large, medium and small.

First, this paper utilizes the software’s PLS algorithm to calculate the R² of the structural model, and the results show that the R² values corresponding to endogenous latent big data analysis capability, volume supply chain flexibility, and enterprise performance are 0.318, 0.445, and 0.351, respectively, which indicates that the explanatory power of this model is medium.

Subsequently, this paper utilizes the software’s Blindfolding algorithm to calculate Q². The results show that the Q² values corresponding to big data analytic capability, volume supply chain flexibility, and enterprise performance are 0.491, 0.371, and 0.408, respectively, which indicates that the predictive accuracy of this model is moderate.

Finally, in order to ensure the accuracy of the regression between the latent variables, this paper also calculated the covariance statistic (VIF) corresponding to each measurement question item using the PLS algorithm of the software. The results show that the VIF value corresponding to each question item is less than 3.5, which is within the suggested threshold (less than 5), indicating that there is no covariance problem in this model.

The structural equation model, i.e. PLS-SEM, constructed in this paper using SmartPLS software is shown in Fig. 6, which is composed of a measurement model and a structural model. After completing the discriminative description of the fitness and validity of the structural model, this paper utilizes the Bootstrapping function of the software to test the path coefficients of the structural model and its significance. Among them, the Bootstrapping function mainly uses the random sampling of the observations therein from the original dataset by pulling out and putting back to create a sub-sample, and the number of sampling times is set to 1000 in this study.

Figure 6.

Structural equation model

The results of the path coefficient test of the impact relationship are shown in Table 15. Regarding the test results of H1, between big data analytics capability and supply chain flexibility, it presents a significance at the 0.001 level (standardized coefficient β = 0.441, p=0.000<0.001), which indicates that big data analytics capability has a significant positive impact on supply chain flexibility, so hypothesis H1 is valid. Regarding the test results of H2 and H3, between big data analytics capability and process flexibility and logistics flexibility in supply chain flexibility, both of them present significance at 0.001 level, with standardized coefficients β = 0.448 and 0.637, respectively, which suggests that big data analytics capability has a significant positive impact on process flexibility and physical flexibility in supply chain flexibility, so hypotheses H1 and H3 are valid.

The standardized coefficient β between big data analytics capability and firm performance is 0.481, presenting significance at the 0.001 level (p=0.000<0.001), which indicates that big data analytics capability has a significant positive impact on firm performance, and hypothesis H4 is valid. Between Supply Chain Flexibility, Process Flexibility and Physical Flexibility and Firm Performance, presenting a significance at the 0.01 level, hypotheses H5, H6 and H7 hold.

This paper sequentially tests the mediating role of supply chain flexibility between big data analytics capabilities and firm performance based on the testing procedure for mediating effects. The test of mediating effect includes the following steps:

In the first step, model 1, i.e., the total effect c and its significance between big data analytic capability and firm performance, is tested.

In the second step, model 2, i.e., the regression coefficient a of big data analytic capability on supply chain flexibility and its significance, is tested.

In the last step, model 3 is tested, i.e., both the direct effect c′ of big data analytic capability on firm performance and its significance, and the regression coefficient b of supply chain flexibility on firm performance and its significance are considered.

The results of the mediating effect analysis are shown in Table 16, which shows that the mediating variable, supply chain flexibility, plays a partial mediating effect in the effect of big data analytic capability on enterprise performance, and its effect value is 0.237. Therefore, supply chain flexibility plays a mediating role between big data analytic capability and enterprise performance, and hypothesis H8 is valid. The effect values of process flexibility and logistics flexibility between big data analytic capability and firm performance are 0.041 (35.02%) and 0.154 (64.98%), respectively, and hypotheses H9 and H10 hold.