Digitalization and network capability as enablers of business model innovation and sustainability performance: The moderating effect of environmental dynamism

Dynamic capabilities theory

The resource-based view of firms regards organizations as collections of valuable, rare, imperfectly imitable, and irreplaceable resources that support the achievement of competitive advantage (Barney, 1991). Scholars have argued that this view should be extended to explain how such resources contribute to long-term and sustainable competitive advantages in a dynamic environment (Huang et al., 2012; Teece and Pisano, 1994). DCT was proposed primarily to address how firms can survive under rapidly changing business conditions (Li and Liu, 2014; Teece et al., 1997). Teece and Pisano (1994) defined dynamic capabilities as an organization’s abilities to adjust, integrate, and reconfigure its internal and external resources and its abilities to adapt to ever-changing environments. Subsequent studies have divided dynamic capabilities into three dimensions: sensing, seizing, and transformation (Teece, 2007). The sensing capability helps organizations capture environmental signals. The seizing capability enables timely decision-making to seize emerging opportunities and neutralize threats. The transformation capability concerns the adjustment of organizational structures and processes to adapt to environmental changes and maintain competitiveness.

DCT explains that when superior dynamic capabilities are appropriately combined and integrated, they help organizations obtain sustainable competitive advantage (Teece and Pisano, 1994; Teece, 2007; Teece et al., 1997). Dealing with changes and achieving competitive advantage are major challenges in the context of manufacturing in emerging markets. Such conditions characterize the highly dynamic context familiar to Chinese manufacturers because they deal with fierce competition—both domestic and global—and struggle to achieve the upgrades and transformations necessary for innovation-intensive and high-quality development (Jiao et al., 2019; Long and Liao, 2021; Yuan et al., 2021). In this context, Ludwig and Pemberton (2011) used DCT to explain why it is difficult to change traditional manufacturing processes on short notice and that a firm’s dynamic capabilities can help it change manufacturing processes by adapting current routines and determining how processes can be permanently changed in the future. The study by Ludwig and Pemberton (2011) focused on the Russian steel industry, but other studies have demonstrated that DCT can provide apt explanations of the efficacy Chinese manufacturers’ dynamic capabilities in dealing with fierce competition and rapidly changing business conditions. ²

We indicate that in Chinese dynamic manufacturing context, digitalization, network capability, and BMI represent three critical dynamic capabilities that support digital transformation at the process level, organization level, and business domain level, respectively. Digitalization and network capability help organizations acquire resources and information to sense opportunities and integrate capabilities of their functional departments and business partners to seize opportunities, which enable BMI to transform resources and capabilities, leading to improved economic and environmental performance. Next, we explain the foundational concepts we leverage with respect to digitalization, network capability, and BMI.

Digitization, digitalization, and digital transformation

Digitization, digitalization, and digital transformation are regarded as the three phases of digital transformation (Verhoef et al., 2021). Digitization is the conversion of information from analog form to a digital representation based on bits and bytes (Teubner and Stockhinger, 2020). Digitalization refers to the increasing use and general institutionalization of digitized data and digital technologies (Buer et al., 2021; Ghobakhloo, 2020). That is, digitization emphasizes technological elements, whereas digitalization is the interaction of digital technologies with social and institutional processes (Teubner and Stockhinger, 2020). Driven by advanced technologies (Ghobakhloo, 2020), digitalization enables the transformation of existing business processes. Soluk and Kammerlander (2021) defined digital transformation as an organization’s use of digital technologies to improve its strategies, processes, capabilities, products, services, and even stakeholder relations. ³

Although scholars have recognized firms’ adoption of digital technologies to transform business models (Li et al., 2020; Sjödin et al., 2021; Weking et al., 2020), the relationship between digitalization and BMI has not been demonstrated empirically. ⁴ Most studies have focused on broad economic performance, and only a few have explored the environmental performance of digitalization. ⁵ Overall, empirical evidence of the effect of digitalization on environmental performance remains limited. Our study partially addresses the need to validate how digitalization leads to improved economic and environmental performance through BMI.

Network capability

The sustainable development of an organization is supported not only by its internal resources but also by its external resources, that is, those obtained from various relationships in its partnering network (Yang et al., 2018). Chi et al. (2010) described this network as a system in which organizations interact with one another, making it a platform for the collaborative development of resources and capabilities. Mu and Benedetto (2012) defined network capability as an organization’s ability to leverage existing external connections and establish new ones—an ability that enables the flexible allocation of resources that support the development of competitive advantage in a dynamic environment. ⁶ Network capability involves formal or informal relationships in which organizations exchange resources and capabilities with external partners (Yang et al., 2018).

Whereas some studies have emphasized the ability of organizations to strengthen relationships with external partners, others have explained network capability from a more comprehensive perspective. Walter et al. (2006) argued that network capability should comprise relational skills, coordination, partner knowledge, and internal communication from both internal and external perspectives: The external information, knowledge, and resources organizations acquire need to be transferred to internal departments and employees (Walter et al., 2006). Partanen et al. (2020) maintained that internal communication and partner knowledge, proposed by Walter et al. (2006), are innovative, distinguishing network capability from alliance management capability and relational capability. Our study follows Walter’s conceptualization of network capability.

Previous studies have demonstrated the positive influence of network capability on economic performance. Chi et al. (2010) suggested that firms rely on their partnering networks to access external resources and enable competitive actions, thus improving competitive performance. Parida et al. (2016) found that network capability can positively influence sales growth. Scholars have also proposed that network capability affects innovation performance and new-product performance (Mu, 2014; Najafi-Tavani et al., 2018). Walter et al. (2006) emphasized that organizations with excellent network capabilities are more likely to stimulate innovation than those with poor capabilities. Firms’ network capabilities can help them rapidly identify opportunities and paths for launching new services and achieving efficient commercialization. Parida et al. (2017) analyzed how network capability contributes to organizational innovativeness and influences customer, sales, and innovation performance. In one of the few studies that have addressed sustainability performance as an outcome of network capability, Amara and Chen (2020) found that network capability fosters eco-innovation capability and sustainability performance.

Business model innovation (BMI)

Teece (2010) defined a business model as an architecture composed of the mechanisms of value creation, delivery, and capture. Saebi et al. (2017) further described a business model as an organization’s market segmentation and value proposition, the structure of the value chain, value capture mechanisms, and the organic connections between these elements. BMI refers to the discovery of new value propositions to generate new sources of profit (Zott and Amit, 2007). Scholars have argued that BMI consists of an organization’s new methods of creating and obtaining value for stakeholders by modifying or improving at least one element of its business model (Casadesus-Masanell and Zhu, 2013). Efficiency, novelty, lock-in, and complementarity are the four types of BMI (Amit and Zott, 2001). Following Zott and Amit (2007), we investigate novelty- and efficiency-centered designs, which are a firm’s fundamental alternatives for creating value and which facilitate the building and testing of a parsimonious theory. Novelty-centered BMI aims to create new transaction types using a new method of communication among transaction participants and fostering product differentiation, whereas efficiency-centered BMI seeks to help organizations achieve transaction efficiency and reduce transaction costs (Zott and Amit, 2007). Scholars have also emphasized that novelty- and efficiency-centered BMI serve as the firm’s business strategies for achieving sustainable competitive advantage through product differentiation and price leadership, respectively (Teece, 2010; Zhang et al., 2021; Zott and Amit, 2007).

BMI lays the foundation for sustainable business success (Casadesus-Masanell and Zhu, 2013). Scholars have demonstrated that BMI has a positive influence on firm growth (Wei et al., 2014) and firm performance (Latifi et al., 2021; Zott and Amit, 2007). Both novelty- and efficiency-centered BMI have positive effects on economic performance, but there is limited empirical evidence concerning the relationship between BMI and environmental performance. Klein et al. (2021) noted that businesses prefer to integrate sustainability concerns into their products, services, and processes, and they investigated how a commitment to sustainability influences firms’ BMI efforts. However, whether BMI can result in improved environmental performance requires further investigation.

Extant research has explored both internal and external factors affecting BMI. Internal factors include company size, entrepreneurship, culture, strategy, human resources, and organizational capabilities, whereas external factors include technologies, markets, competition, and policies (Klein et al., 2021; Saebi et al., 2017; Teece, 2010). Scholars have considered the effects of new technologies on BMI in the digital economy. Digital technologies (e.g., AI, cloud computing, and big data) can give rise to a variety of novel business models (Guo et al., 2020). Weking et al. (2020) conducted case studies to explain how firms foster BMI by leveraging cyber-physical systems, IoT, and smart factories. Sjödin et al. (2021) showed how AI capabilities enable and Li (2020) identified emerging trends in the use of digital technologies to facilitate BMI in creative industries. Although extensive research has pointed out the key role of digitalization in BMI, there is still a lack of empirical validation concerning whether and how digitalization leads to novelty- and efficiency-centered BMI.

Digitalization and network capability as enablers of BMI

A firm that embraces digitalization relies on the use of digital technologies to connect its business processes across intra- and interorganizational boundaries (Li et al., 2020; Yang et al., 2021), whereas a firm’s network capability reflects how it manages the boundaries within its business networks (Amara and Chen, 2020; Parida et al., 2017). Firms with excellent digitalization capability can sense opportunities to collaborate with potential business partners (Cenamor et al., 2019). Our study argues that digitalization provides a technological infrastructure to connect these business partners to establish firms’ networks (Cenamor et al., 2019; Chi et al., 2010; Pagani and Pardo, 2017). The functional departments and external partners can be connected efficiently through the application of digital technologies. Such digital connections can increase information transparency within the network (Battleson et al., 2016), which improve the efficiency of communication and coordination and promote mutual understanding among business partners (Pagani and Pardo, 2017). Cenamor et al. (2019) proposed that digital-technology-enabled platforms facilitate intra- and interorganizational interactions and help firms manage their networks. In addition, digitalization enables firms to integrate new, heterogeneous resources across organizational boundaries and strengthen network connections (Cenamor et al., 2019). Lyytinen et al. (2016) proposed that digital technology adoption accelerates network integration. Thus, we hypothesize that:

H1

Digitalization has a positive influence on a firm’s network capability.

Digitalization improves firms’ resource acquisition and integration capabilities (Sousa-Zomer et al., 2020; Von Briel et al., 2019; Yu et al., 2019). Digital technologies provide organizations with better access to external information and resources that can improve their capabilities to sense market changes and uncertainties (Li et al., 2020), and both the integration of diverse and heterogeneous resources and the establishment of data assets through digitalization strengthen the organizations’ seizing capability to promote new-product development (Durmusoglu and Kawakami, 2021). Moreover, firms can easily involve consumers in product development processes through digitalization when creating new products and services in line with market demands (Guo et al., 2020; Lyytinen et al., 2016). Francesco et al. (2022) demonstrated that big-data-analytics capabilities support effective marketing actions, such as new-product launches, by capturing and analyzing customers’ needs and behavior patterns and can thus help optimize product customization. With digital technologies, firms can upgrade production lines to provide new combinations of products and services (Buer et al., 2021). Our study assumes that digitalization will help firms overcome geographical barriers and physical constraints, thereby facilitating connections with new business partners to generate collaborative and innovative solutions. For example, cloud-based technologies make it easy to share knowledge among decentralized supply chain partners to enable innovation and continuous change (Gupta et al., 2020). Thus, we predict:

H2a

Digitalization has a positive influence on a firm’s novelty-centered BMI.

Digitalization influences efficiency-centered BMI in several respects. First, digitalization, such as adopting automation technologies and robotics, enables standardized and large-scale production processes that will improve production efficiency (Buer et al., 2021; Ghobakhloo, 2020). The transformation of production processes in modern manufacturing firms relies mostly on the application of digital technologies (Buer et al., 2021). Second, digitalization can improve communication efficiency with business partners (Li et al., 2020). The application of digital technologies facilitates real-time information exchange and thus increases transparency while reducing transaction costs (Yang et al., 2021). Improved information-gathering and information-processing capabilities also promote quick decision-making and transaction efficiency (Li et al., 2020). Third, digital technologies open boundaries so firms can interact with more business partners and promote the free mobility and efficient integration of heterogeneous resources (Pagani and Pardo, 2017). Broekhuizen et al. (2021) pointed out that digitalization enables firms to establish digital-related dynamic capabilities and to integrate internal and external resources to support efficiency-centered innovation. In summary, we hypothesize:

H2b

Digitalization has a positive influence on a firm’s efficiency-centered BMI.

We also propose that network capability can effectively promote novelty-centered BMI by integrating and leveraging resources and capabilities from business partners. First, an extensive partnering network and a strong network capability can broaden a firm’s sources of information and resources, which help the firm rapidly identify market opportunities and introduce new businesses (Zhu et al., 2019). Firms obtain information and knowledge through their networks to deepen their understanding of the business environment (Guo et al., 2021) and to innovate their business models to adjust to ever-changing conditions (Zhang et al., 2021). Second, network capability enables companies to obtain heterogeneous external capabilities, develop new resource combinations and cooperative relationships, drive new-product development, and provide resources and support for novelty-centered BMI (Chen et al., 2021; Mu, 2014; Najafi-Tavani et al., 2018); Zhang et al. (2021). Zhang et al. (2021) demonstrated that exploration across organizational boundaries contributes to knowledge integration and generation and promotes the overall evolution of a product class and technology. Third, firms with strong network capabilities can reduce the uncertainty in the process of novelty-centered BMI through sufficient information and knowledge sharing (Von Delft et al., 2019). Therefore, we predict:

H3a

Network capability has a positive influence on a firm’s novelty-centered BMI.

Similarly, our study explains the positive effect of network capability on efficiency-centered BMI. Networks provide flexible communication channels for all types of information and resources (To et al., 2019). Companies in networks can exchange and share resources more easily by establishing external connections, and they can obtain at lower costs the knowledge and capabilities required to promote efficiency-centered BMI (Zhang et al., 2021). The establishment of partnering networks also reduces firms’ dependence on certain scarce proprietary resources, thus mitigating opportunistic behaviors, promoting cost reduction, and improving transactional efficiency (Chi et al., 2010). Yu et al. (2019) showed that communication and mutual understanding between partners reduce the risks of information asymmetry. Moreover, good network relationships can provide the focal firm with a cushion (i.e., alternative sources) that reduces the risk of vendor lock-in, subsequently preventing unexpected supply chain disruptions and lowering the average costs of production (Riccardo et al., 2021). Therefore, firms with strong network capabilities can reduce transaction costs and improve transaction efficiency, providing favorable conditions for efficiency-centered BMI (Zott and Amit, 2007). Thus, we hypothesize the following:

H3b

Network capability has a positive influence on a firm’s efficiency-centered BMI.

Our study further proposes that network capability mediates the effects of digitalization on novelty- and efficiency-centered BMI. Generally, the use of digital technologies can help firms establish collaborative platforms (Li et al., 2020), which serve as resource pools that enable network participants to innovate (Wang and Hu, 2020). However, not all external resources are equally accessible to all firms (Nwankpa and Datta, 2017). The firm needs to absorb and integrate valuable heterogeneous organizational resources from its business partners by developing network capability (Chen et al., 2021). Network capability can promote cooperation among business partners and the “wisdom of the crowd” (Chen et al., 2021) and enhance the quality of resources gathered through digital technologies, which subsequently promote innovation activities (Chen et al., 2021). Digitalization provides a technological architecture that enables network partners to exchange information and discover new business opportunities and network capability can help partners promote product innovation and improve transaction efficiency (Cenamor et al., 2019; Chi et al., 2010; Mikalef et al., 2021a). Firms without adequate network capabilities cannot fully use external resources of their business partners to conduct novelty- and efficiency-centered BMI (Chen et al., 2021; Mu, 2014). Our study emphasizes the role of network capability in high-quality and high-efficiency decision-making, which cannot be replaced by digitalization (Cenamor et al., 2019; Chi et al., 2010). In summary, digitalization should influence BMI activities through network capability. Thus, we hypothesize the following:

H4

A firm’s network capability positively mediates the effects of digitalization on (a) novelty-centered BMI and (b) efficiency-centered BMI.

Influence of BMI on sustainability performance

This study claims that novelty-centered BMI can help firms achieve unique forms of competitive advantage, thereby leading to improved economic performance. Novelty-centered BMI aims to connect new trading partners or reconnect existing trading partners in innovative ways (Amit and Zott, 2001). Firms focusing on novelty-centered BMI can reconfigure heterogeneous knowledge and resources from partners to create new values or profit (Pati et al., 2018). Guo et al. (2020) pointed out that by supporting the development of novel products, services, or both, innovative business model designs can help companies leap into new markets. Novelty-centered BMI also motivates firms to continuously learn and accumulate specialized knowledge from their partners (Pati et al., 2018), rapidly replace scarce assets and resources, establish high switching costs for buyers, and set up invisible entry barriers for competitors (Zott and Amit, 2007). If in fact novelty-centered BMI has these positive aspects, they should enable firms to gain potential first-mover advantages, thus improving their economic performance. Scholars have also found that successful novel business models tend to entice imitations (Teece, 2010). In this case, firms conducting novelty-centered BMI can develop an excellent reputation and corporate/brand image in the market, which will help them expand their market share and increase their pricing power. In summary, novelty-centered BMI can strengthen a firm’s unique competitive advantage, consequently enhancing the firm’s economic performance. Thus, we hypothesize:

H5a

Novelty-centered BMI is positively related to a firm’s economic performance.

We also propose that novelty-centered BMI can improve environmental performance. Again, in the last few decades, environmental sustainability has become increasingly important to both organizations and consumers (Elliot, 2011). This concern has led to a downstream demand for responsible production and environmentally friendly products and services (Loeser et al., 2017). Firms that engage in novelty-centered BMI can better understand downstream customers’ demand for such products and services and develop innovative eco-friendly products and services to bolster a green image (Klein et al., 2021) and improve their environmental performance. Extensive collaboration with various partners, such as universities and research institutions, can be beneficial to sustainability-related innovation (Sudusinghe and Seuring, 2022). Furthermore, as firms get better at leveraging knowledge partners and delivering novelty-centered BMI for environmental performance, they can better introduce, encourage, and integrate such environmental thinking when they connect with new trading partners (Amit and Zott, 2001). In summary, with the help of knowledgeable partners, firms tend to obtain and absorb environmental management knowledge to improve green production and engage in green-product development. Therefore, the following hypothesis is consistent with prior research:

H5b

Novelty-centered BMI is positively related to a firm’s environmental performance.

Efficiency-centered BMI can improve transaction efficiency and reduce transaction costs for firms and their partners. Low cost is always associated with pricing advantages, and it not only encourages existing customers to increase their purchase quantities but also attracts new customers and trading partners (Zott and Amit, 2007). In our context, firms implementing efficiency-centered BMI can continuously expand transaction scales and increase their market share and sales, ultimately improving their economic performance (Latifi et al., 2021; Wei et al., 2014). Previous studies have shown that efficiency-centered BMI enables firms to simplify coordination and communication procedures and mechanisms with their business partners (Zhang et al., 2021; Zott and Amit, 2007). Thus, a smooth trading experience will strengthen a firm’s customer retention and loyalty, and the firm’s partners, such as suppliers and logistics providers, will in turn become more willing to collaborate for the purpose of value co-creation (Sudusinghe and Seuring, 2022). Moreover, efficiency-centered BMI improves transparency and reduces uncertainty and information asymmetry (Wei et al., 2014). Firms that cannot respond to uncertainty and risks effectively tend to suffer from higher risks of recovery costs and loss of profit (Syed et al., 2020). Thus, we predict:

H6a

Efficiency-centered BMI is positively related to a firm’s economic performance.

Environmental performance involves resource saving, energy consumption reduction, and environmental protection (Zhu and Sarkis, 2004). Efficiency-centered BMI can promote environmental performance by improving resource utilization efficiency and simplifying trading procedures (Zott and Amit, 2007). When implementing efficiency-centered BMI, firms generally reallocate or reconfigure available resources to avoid redundancy, eliminate unnecessary trading procedures to save resources and energy, and improve trading transparency to reduce waste of resources, rework, and errors (Zhang et al., 2021), thereby improving environmental performance. Efficiency-centered BMI also enables firms to rapidly gather and process information that helps them deal with the increasing uncertainty and complexity of environmental management activities (Klein et al., 2021). In summary, to comply with sustainable-development strategies, firms introduce efficiency-centered BMI to implement energy-saving and environmentally friendly processes and reduce waste and pollution, hence improving their environmental performance. Thus, we hypothesize the following:

H6b

Efficiency-centered BMI is positively related to a firm’s environmental performance.

Mediating effect of BMI

Based on DCT, digitalization and network capability enable organizations to gather and aggregate information, knowledge, competencies, and resources for identifying opportunities and challenges by connecting organizations with business partners and seizing the opportunities or neutralize threats through investment in digital resources and organizational resources, whereas novelty- and efficiency-centered BMI enable organizations to recombine and reconfigure internal and external resources that enable them to transform and adjust to changes (Teece, 2018). As a result, the effects of digitalization and network capability on the two dimensions of sustainability performance are mediated by novelty- and efficiency-centered BMI.

Digitalization allows firms to access external information and resources from their business partners and to establish links with their external business environments by leveraging digital technologies (Li et al., 2020). Similarly, network capability enables firms to acquire resources from their network partners through social or relational mechanisms (Chi et al., 2010). Adequate and diverse information and resources enhance firms’ sensing capabilities, which enable them to rapidly pick up signals from the environment and identify business opportunities and challenges (Mikalef et al., 2021). For example, firms can obtain the best insights into customer demand through big-data-analytics techniques (Shamim et al., 2019). The fourth Industrial Revolution (i.e., Industry 4.0) has introduced technologies such as virtual reality, augmented reality, and digital twins that can detect or monitor bottlenecks in the production process, enable predictive maintenance, and minimize production-line downtime (Frank et al., 2019; Mourtzis, 2020). External partners in a firm’s network can also inspire the firm with new knowledge and ideas about the business (Chi et al., 2010).

However, identifying opportunities and gathering resources offer limited capabilities that directly support the continuous improvement of sustainability performance. To deal with potential business changes, firms should implement novelty- and efficiency-centered BMI to dynamically reconfigure resources acquired through digitalization and network capability. BMI requires firms’ transformation capabilities to periodically reposition themselves for the purpose of addressing new opportunities and threats (Teece, 2018). Mezger (2014) proposed a capability-based conceptualization of BMI and showed that new business models are designed based on bundles of valuable competencies and resources. Teece (2018) argued that the refinement and transformation of business models are high-order capabilities that support the adjustment and recombination of low-order capabilities and resources. BMI serves to orchestrate and connect organizational elements based on a core value proposition to provide sustainable competitive advantage (Teece, 2010; Zott and Amit, 2007). Novelty-centered BMI enables firms to innovate for new value propositions, new ways of conducting transactions, and new links with business partners, whereas efficiency-centered BMI focuses on improving transaction efficiency and reducing transaction costs (Zott and Amit, 2007). In summary, with the support of distinct BMI types, firms maximize the sustained values of resources obtained using digital technologies and network-partner collaborations. Thus,

H7

Novelty-centered BMI positively mediates the effect of digitalization on a firm’s (a) economic performance and (b) environmental performance.

H8

Efficiency-centered BMI positively mediates the effect of digitalization on a firm’s (a) economic performance and (b) environmental performance.

H9

Novelty-centered BMI positively mediates the effect of network capability on a firm’s (a) economic performance and (b) environmental performance.

H10

Efficiency-centered BMI positively mediates the effect of network capability on a firm’s (a) economic performance and (b) environmental performance.

Moderating effect of environmental dynamism

Environmental dynamism can be defined as a high rate and large volume of changes in a firm’s business environment (Azadegan et al., 2013; Dess and Beard, 1984). In a highly unstable and unpredictable environment, product life cycles tend to be shortened, and existing production processes can easily compromise a firm’s competitiveness (Azadegan et al., 2013; Chan et al., 2016). Therefore, a dynamic environment drives firms to constantly introduce new products, services, or processes to respond to changing market demand and intensified market competition (Chan et al., 2016). Our study proposes that the implementation of novelty-centered BMI to cope with the ever-changing environment involves two basic requirements: the speed and heterogeneity of resource acquisition. The effect of digitalization on novelty-centered BMI can be strengthened by rapid information-gathering and processing and timely acquisition of heterogeneous resources (Guo et al., 2021; Li, 2020). Digital technologies enable real-time connectivity to support quick decision-making (Gillani et al., 2020) and standardize digitalized resources to gather heterogeneous resources (Thomas and Carsten Lund, 2020). Hence, the firm can efficiently and rapidly utilize heterogeneous resources to generate novel ideas and respond to high environmental uncertainties. Moreover, the new markets and business opportunities in a dynamic environment are fleeting if not promptly grasped (Yuan et al., 2021). Firms have to enhance the perception of rapid changes (Mikalef et al., 2020) in customer needs and technological breakthroughs (Li and Liu, 2014). We extend this connection to propose that novelty-centered BMI relies strongly on digitalization to quickly sense and seize opportunities in the presence of environmental dynamism.

On the other hand, environmental dynamism is highly related to organizational pressure, causal ambiguity, and exchange hazards (Yuan et al., 2021), which lead to high transaction costs and low transaction efficiency. The exploratory use of emerging digital technologies motivates firms to introduce novel solutions (Wei et al., 2014), but it can also increase the complexity of business operations (Paolucci et al., 2021), particularly when technologies are premature in their development or have not been well integrated in the operations of the firm. In a dynamic environment, the efficiency of business operations and transactions can be seriously harmed by the inappropriate use of digital technologies (Rana et al., 2021; Son et al., 2021). In contrast to novelty-centered BMI, efficiency-centered BMI favors a more stable organizational environment and disciplined decision-making (Yuan et al., 2021). Rana et al. (2021) emphasized the technology, privacy, and security risks associated with digitalization. The presence of these risks implies that the positive effects of digitalization on efficiency-centered BMI can be weakened by the increased complexity and risks caused by environmental dynamism. Thus, we hypothesize:

H11a

Environmental dynamism positively moderates the relationship between a firm’s digitalization and novelty-centered BMI.

H11b

Environmental dynamism negatively moderates the relationship between a firm’s digitalization and efficiency-centered BMI.

Environmental dynamism requires firms to make decisions and take actions in a timely and flexible manner. Both novelty- and efficiency-centered BMI drive firms to transform their methods of conducting transactions with business partners (Zott and Amit, 2007). When firms engage in novelty- and efficiency-centered BMI in highly unstable environments, business partners are required to accept and adjust to frequent changes in the business models (Pati et al., 2018). Firms also need to respond to a variety of new or existing partners from different domains (Mikalef et al., 2021). Network capability can increase the flexibility of relationships with such partners (Parida et al., 2016). Firms with strong network capabilities can efficiently coordinate and communicate with business partners to reconfigure competencies and resources and escape unfavorable path dependencies collaboratively (Cenamor et al., 2019; Chen et al., 2021). In addition, strong network capabilities can help firms ensure the quality of resources acquired from business partners, contributing to both novelty- and efficiency-centered BMI in an unpredictable environment (Chi et al., 2010). Despite the potential risks related to environmental dynamism, firms can leverage communication and coordination mechanisms, relational skills, and partner knowledge to govern business partners and prevent the negative effects on BMI (Chen et al., 2021). Our study thus proposes that environmental dynamism can strengthen the effects of network capability on novelty- and efficiency-centered BMI. Thus, we predict:

H12a

Environmental dynamism positively moderates the relationship between a firm’s network capability and novelty-centered BMI.

H12b

Environmental dynamism positively moderates the relationship between a firm’s network capability and efficiency-centered BMI.

Sampling and data collection

Our study focused on large manufacturing firms in China, a burgeoning economy marked by swift economic expansion and robust productivity (Chen et al., 2014; Yang et al., 2021). In alignment with national initiatives like “Made in China 2025,” the Chinese government has been actively encouraging digital transformation as a pathway to high-quality development (Wang and Hu, 2020). Specifically within the manufacturing industry, firms are increasingly adopting digital solutions as a strategic response to intense global competition and the disruptions precipitated by the COVID-19 pandemic (Pan and Zhang, 2020). Additionally, driven by mounting concerns over rapid resource depletion and environmental degradation, these Chinese manufacturing entities are taking deliberate steps toward environmental conservation and resource efficiency to achieve sustainable development (Chan et al., 2016; Li et al., 2020).

Given this intricate, volatile, and dynamic environment, the Chinese manufacturing landscape offers an ideal research context for examining the interplay between digital transformation and sustainability performance through the lens of DCT (Yuan et al., 2021). To empirically validate our research model, which posits that digitalization enhances sustainability performance, we gathered data from Chinese manufacturing firms using an online survey methodology.

The potential sampling pool was determined based on the Yellow Pages of China Telecom (Jacobs et al., 2016), which contains information on Chinese manufacturing companies’ names and contacts. To improve the quality and efficiency of data collection, we collaborated with a national information technology company to contact the manufacturing companies directly. Data collection was conducted from October 2021 to December 2021. An email linked to the survey was sent to potential respondents, inviting them to participate (Dillman, 2007). We explained that this study had received Institutional Review Board approval from the lead author’s institution and that the data collected would be used with absolute confidentiality and only for academic research purposes (Yang et al., 2021). We also offered to share a report of the findings at the end of the study. The questionnaire contained two attention-check questions. Followed-up messages and calls were used to remind participants to respond. These measures helped ensure the quality and reliability of the data. A total of 1681 questionnaires were distributed, and 255 valid responses were returned.

Table 1(a)–(d) provides a comprehensive summary of the respondent companies, categorizing them based on industry, sales volume, workforce size, and geographical regions. Our survey targeted middle- and top-level managers, such as CEOs, general managers, and departmental managers in areas like production, product management, marketing, and procurement. These individuals possess in-depth knowledge of their companies’ business operations, product lines, core production processes, and key performance indicators (Jacobs et al., 2016).

OPEN IN VIEWER

Table 1.

Demographic information of responding companies

Industry	Frequency	Percentage (%)
Publishing and printing	3	1.2
Electronics and electrical	67	26.3
Textiles and apparel	12	4.7
Arts and crafts	1	0.4
Chemicals and petrochemicals	10	3.9
Building materials	9	3.5
Metal, mechanical, and engineering	52	20.4
Wood and furniture	2	0.8
Food, beverage, and alcohol	14	5.5
Toys	1	0.4
Rubber and plastics	9	3.5
Pharmaceutical and medical	21	8.2
Equipment	54	21.2

Firm sales (million CNY)	Frequency	Percentage (%)
<10	23	9.0
10–50	61	23.9
50–100	66	25.9
100–400	59	23.1
>400	46	18.1

Number of employees	Frequency	Percentage (%)
<100	28	11.0
101–300	57	22.3
300–500	65	25.5
501–1000	52	20.4
>1000	53	20.8

Regions	Frequency	Percentage (%)
Circum-Bohai Sea	51	20.0
Pan-Pearl River Delta	96	37.7
Yangtze River Delta	72	28.2
Other areas in China	36	14.1

The survey demographics reveal that 36.5% (93 respondents) held top-management positions, whereas the remaining 63.5% (162 respondents) were from middle management. As for their tenure, 8.7% had been in their roles for 1–3 years, 52.9% for 4–6 years, and 38.4% for 7 years or longer.

Regarding industry subsectors, the participating firms were predominantly situated in electronics and electrical manufacturing (26.3%), equipment manufacturing (21.2%), and metal, mechanical, and engineering (20.4%). Only a minimal proportion (0.8%) came from the toy, arts, and crafts subsector. This distribution closely mirrors the concentration of manufacturing sectors in China (Flynn et al., 2010; Yang et al., 2021).

In terms of company size, based on the number of employees, the sample covered a spectrum of small (<300 employees), medium (300–1000 employees), and large (>1000 employees) enterprises. Specifically, the proportions were 33.3, 45.9, and 20.8%, respectively, aligning well with the overall composition of manufacturing firms in China (Wei et al., 2014).

Geographically, the firms were located across major Chinese economic zones, including the Circum-Bohai Sea Economic Zone (encompassing cities such as Beijing, Tianjin, Hebei, Liaoning, Shandong, and Shanxi), the Pan-Pearl River Delta Economic Zone (e.g., Guangdong, Guangxi, Fujian, Jiangxi, Hunan, Sichuan, Yunnan, and Guizhou), and the Yangtze River Delta Economic Zone (e.g., Shanghai, Zhejiang, Jiangsu, and Anhui). Some were also situated in other areas, including Gansu and Chongqing. These zones are the primary industrial hubs of China, covering a broad range of eastern, central, and western cities. Consequently, as corroborated by Table 1(d), our sample of respondent companies aptly represents the distribution of China’s key economic zones (Yang et al., 2021).

Construct measures

The measurement scales for digitalization, network capability, novelty-centered BMI, efficiency-centered BMI, economic performance, and environmental performance were adapted from well-established scales, as detailed in Appendix A. We first developed an English version of the questions with a 7-point Likert-type scale. A back-translation procedure was carried out by two bilingual researchers to ensure conceptual equivalence between the versions (Douglas and Craig, 2007). Subsequently, a pilot study was conducted among 15 practitioners. It gave the researchers the opportunity to further improve the wording and readability of the items, which can enhance content validity (MacKenzie et al., 2011). We summarize the measurements of the key constructs in the following sections.

Establishing the factorial validity and reliability of reflective constructs

SmartPLS 3.2.8 software was used to analyze the survey data. The measurement model was first evaluated to determine the factorial validity and reliability of the survey instruments. The hypothesized relationships in the structural model were then examined through a bias-corrected and accelerated bootstrapping procedure with 255 cases and 5000 subsamples.

Appendix B, Table B1 shows the measurement-assessment results of all the first-order reflective constructs. Reliability was confirmed using factor loadings, Cronbach’s α, and composite reliability (CR). As Appendix A shows, the values of all the factor loadings for each construct were greater than 0.50 and significant at the 0.001 level. The Cronbach’s α values were between 0.65 and 0.85, and internal consistency reliability was thus ensured (Nunnally, 1978). Moreover, the CR values for all the constructs were larger than 0.800, indicating satisfactory construct reliability. The average variance extracted (AVE) values were used to measure each construct’s convergent validity. AVE values higher than 0.50 indicated that more than 50% of the variance in measurement items was explained by the corresponding construct (Hair et al., 2019).

Discriminant validity was evaluated based on the heterotrait–monotrait (HTMT) ratio of the correlations. Table B1 shows that all the HTMT values were lower than the 0.85 threshold suggested by Henseler et al. (2015). We also assessed discriminant validity by determining whether the Fornell–Larcker criterion (Fornell and Larcker, 1981) was satisfied. The AVE values of all the constructs were greater than their squared correlations with other constructs. Appendix B, Table B2 summarizes all the items’ cross loadings. Each item’s loading on its corresponding construct was higher than its cross loadings on other constructs. The discriminant validity of the reflectively measured constructs was therefore confirmed.

Establishing the factorial validity of second-order formative constructs

Network capability was modeled as a second-order formative construct with four reflective first-order constructs (Walter et al., 2006). Following Becker et al. (2012), we estimated the parameters in the reflective–formative-type model by combining the repeated indicator approach and the two-stage approach. First, the first-order constructs—including coordination, internal communication, relational skills, and partner knowledge—were reflectively measured and linked to network capability, which was formatively measured using all 12 first-order construct indicators. Because the reliability and validity of the first-order constructs were confirmed, we then assessed the formative second-order construct in terms of indicator weights, significance of weights, and low multicollinearity of indicators, which was assessed with variance inflation factors (VIFs) lower than 3 (Becker et al., 2012; Hair et al., 2019). The results presented in Table 2 indicated that the reflective–formative model for network capability passed all the thresholds of acceptability and thus represented a valid second-order construct.

OPEN IN VIEWER

Table 2.

Assessment of second-order constructs for network capability.

Dimension	Weight	SE	t-values	p-values	VIF
Coordination	0.307	0.008	36.177***	0.000	1.697
Relational skills	0.311	0.008	37.014***	0.000	1.879
Partner knowledge	0.310	0.009	35.569***	0.000	1.748
Internal communication	0.304	0.009	32.318***	0.000	1.708

Testing for nonresponse bias

Following a widely applied approach (Armstrong and Overton, 1977), we tested for potential nonresponse bias by comparing early and late responses. The 255 responses were categorized into two groups according to response time (127 early responses collected before November 13, 2021; 128 late responses collected after this time). We performed a t-test between the groups on firm size (measured by the number of employees) and firm ownership type. The results indicated no significant differences in firm size (t = −0.036, p = 301) or type of ownership (t = 1.003, p = .317) between the two groups. Thus, nonresponse bias was not an issue of concern.

Controlling for and testing common method bias

Common method bias was controlled through a priori procedural methods and tested using ex post statistical methods. For the procedural methods, the measurement scale of each construct was separated from the others in the questionnaire. We selected knowledgeable top- and middle-level managers with rich management experience and ensured complete confidentiality to encourage honest responses (Podsakoff et al., 2003). We conducted Harman’s one-factor test with the nonrotation and extraction method based on an eigenvalue greater than 1. The results indicated that the extracted first factor accounted for 35.845% of the total variance. We also created a latent common method factor to estimate all the items in the PLS analysis and, at the same time, estimated each item on its theoretical construct (Lindell and Whitney, 2001; Mikalef et al., 2021a, 2021b). The results indicated no notable differences between the estimated path model relationships with the common method factor and those without it. Therefore, common method bias was not a critical concern for our PLS analysis.

Structural model analysis of the main effects

The direct effects were first evaluated based on the PLS model depicted in Figure 2, and Table 3 summarizes the details. The coefficients for DT→NC, DT→NBMI, and DT→EBMI were 0.568 (t = 9.678, p = .000), 0.436 (t = 5.216, p = .000), and 0.264 (t = 3.631, p = .000), respectively. Therefore, the effects of digitalization on network capability, novelty-centered BMI, and efficiency-centered BMI were positive and significant, supporting H1, H2a, and H2b. In support of H3a and H3b, network capability was positively and significantly related to novelty-centered BMI (β = 0.423, t = 5.676, p = .000) and efficiency-centered BMI (β = 5.595, t = 9.684, p = .000). Novelty-centered BMI exhibited a positive and significant relationship with economic performance (β = 0.444, t = 5.936, p = .000) and environmental performance (β = 0.327, t = 3.992, p = .000), supporting H5a and H5b. The results also indicated that efficiency-centered BMI positively and significantly influenced economic performance (β = 0.330, t = 4.701, p = .000) and environmental performance (β = 0.445, t = 5.643, p = .000), supporting H6a and H6b. Additionally, the effect of firm nature on novelty-centered BMI was positive (β = 0.124, t = 2.684, p = .007), which indicated that privately owned Chinese companies facing fierce competition pursued novelty-centered BMI (Zhu et al., 2019) more actively than did their non-privately owned counterparts.OPEN IN VIEWER

OPEN IN VIEWER

Table 3.

Estimated direct relationships of the structural model.

Structural path	β	SE	t-values	p-values	f 2	Support?
H1. DT → NC	0.568***	0.059	9.678	0.000	0.476	Supported
H2a. DT → NBMI	0.436***	0.084	5.216	0.000	0.308	Supported
H2b. DT → EBMI	0.264***	0.073	3.631	0.000	0.115	Supported
H3a. NC → NBMI	0.423***	0.074	5.676	0.000	0.296	Supported
H3b. NC → EBMI	0.595***	0.061	9.684	0.000	0.593	Supported
H5a. NBMI → ECOP	0.444***	0.075	5.936	0.000	0.221	Supported
H5b. NBMI → ENVP	0.327***	0.082	3.992	0.000	0.105	Supported
H6a. EBMI → ECOP	0.330***	0.070	4.701	0.000	0.125	Supported
H6b. EBMI → ENVP	0.445***	0.079	5.643	0.000	0.199	Supported
Control variables
FSIZE → NBMI	0.035 n/s	0.050	0.689	0.491	0.003	Not supported
FSIZE → EBMI	−0.015 n/s	0.046	0.328	0.743	0.000	Not supported
FNATURE → NBMI	0.124**	0.046	2.684	0.007	0.031	Supported
FNATURE → EBMI	−0.028 n/s	0.054	0.521	0.603	0.001	Not supported
FSIZE → ECOP	−0.016 n/s	0.040	0.399	0.690	0.001	Not supported
FSIZE → ENVP	0.038 n/s	0.041	0.926	0.355	0.003	Not supported
FNATURE → ECOP	−0.040 n/s	0.041	0.985	0.325	0.003	Not supported
FNATURE → ENVP	−0.025 n/s	0.045	0.547	0.584	0.001	Not supported

Note. *** = p significant at 0.001; ** = p significant at 0.01; n/s = not significant; DT = digital transformation; EBMI = efficiency-centered BMI; ECOP = economic performance; ENVP = environmental performance; FNATURE = firm nature; FSIZE = firm size; NBMI = novelty-centered BMI; NC = network capability; SE = standard error.

We also evaluated the structural model’s explanatory power using the coefficient of determination (R ² ). R ² values of 0.25, 0.50, and 0.75 indicated weak, moderate, and strong in-sample predictive power, respectively (Hair et al., 2019). As Figure 2 shows, the R ² value of network capability, 0.323, implied that the model explained 32.3% of the variance for network capability, representing weak-to-moderate predictive power. The model explained 59.2% of the variance for novelty-centered BMI, 59.7% for efficiency-centered BMI, 56.7% for economic performance, and 50.7% for environmental performance. Therefore, the model has moderate-to-strong predictive power for BMI and sustainability performance.

We further evaluated the model using the f ² effect size. The effect size ranged from small to medium to large when the f ² values were higher than 0.02, 0.15, and 0.35, respectively. Table 3 indicates that the effect size for each hypothesized path ranged from 0.105 to 0.593, indicating medium-to-high effect sizes.

Finally, we calculated the cross-validated redundancy measure Q² for each endogenous latent variable based on a blindfolding procedure. The value of Q² for network capability was 0.193, which was higher than 0 but lower than 0.25, indicating small-to-medium predictive relevance (Hair et al., 2019). The values of Q² for novelty-centered BMI (0.277), efficiency-centered BMI (0.280), economic performance (0.318), and environmental performance (0.281) were higher than 0.25 but lower than 0.50, demonstrating medium-to-large predictive accuracy (Hair et al., 2019).

Structural model mediation effects of BMI

Next, we tested the mediation hypotheses based on a bootstrap test of the indirect effects (cf. Flynn et al., 2010; Vance et al., 2015; cf. Zhao et al., 2010) in the structural model shown in Figure 3. The results are summarized in Table 4 and Appendix B, Table B3. The direct effects of digitalization on network capability (β = 0.566, t = 9.606, p = .000), novelty-centered BMI (β = 0.436, t = 5.365, p = .000), and efficiency-centered BMI (β = 0.266, t = 3.744, p = .000) were all positive and significant. The indirect path coefficients for DT→NC→NBMI and DT→NC→EBMI were 0.240 with a t-value of 4.484 and 0.336 with a t-value of 6.128, which meant that network capability partially mediated the effects of digitalization on novelty- and efficiency-centered BMI. Thus, H4a and H4b were supported. The direct effect of digitalization on economic performance (β = −0.045, t = 0.593, p = .553) and environmental performance (β = 0.031, t = 0.448, p = .654) were not significant, whereas the coefficients for indirect paths, including DT→NBMI→ECOP (β = 0.189, t = 3.428, p = .001), DT→NBMI→ENVP (β = 0.113, t = 2.328, p = .020), DT→EBMI→ECOP (β = 0.075, t = 2.893, p = .004), and DT→EBMI→ENVP (β = 0.092, t = 2.748, p = .006), were all significantly positive. Therefore, novelty- and efficiency-centered BMI fully mediated the effect of digitalization on economic and environmental performance, supporting H7a, H7b, H8a, and H8b.OPEN IN VIEWER

OPEN IN VIEWER

Table 4.

Results for mediation effects testing.

Indirect structural path	β	SE	t-values	p-values	Support?
H4a. DT → NC → NBMI	0.240***	0.053	4.484	0.000	Supported
H4b. DT → NC → EBMI	0.336***	0.055	6.128	0.000	Supported
H7a. DT → NBMI → ECOP	0.189**	0.055	3.428	0.001	Supported
H7b. DT → NBMI → ENVP	0.113*	0.048	2.328	0.020	Supported
H8a. DT → EBMI → ECOP	0.075**	0.026	2.893	0.004	Supported
H8b. DT → EBMI → ENVP	0.092**	0.033	2.748	0.006	Supported
H9a. NC → NBMI → ECOP	0.183***	0.047	3.913	0.000	Supported
H9b. NC → NBMI → ENVP	0.109**	0.042	2.622	0.009	Supported
H10a. NC → EBMI → ECOP	0.167**	0.060	2.766	0.006	Supported
H10b. NC → EBMI → ENVP	0.205**	0.069	2.950	0.003	Supported
DT → NC → ECOP	0.063 n/s	0.058	1.086	0.277	Not supported
DT → NC → ENVP	0.101 n/s	0.057	1.772	0.077	Not supported
DT → NC → NBMI → ECOP	0.104***	0.030	3.483	0.000	Supported
DT → NC → NBMI → ENVP	0.062*	0.026	2.392	0.017	Supported
DT → NC → EBMI → ECOP	0.095*	0.038	2.495	0.013	Supported
DT → NC → EBMI → ENVP	0.116**	0.043	2.700	0.007	Supported

Note. *** = p significant at 0.001; ** = p significant at 0.01; * = p significant at 0.05; n/s = not significant;β = beta coefficient; SE = standard error; DT = digital transformation; EBMI = efficiency-centered BMI; ECOP = economic performance; ENVP = environmental performance; NBMI = novelty-centered BMI; NC = network capability; SE = standard error.

Similarly, the direct effect of network capability on economic performance (β = 0.112, t = 1.139, p = .255) and its direct effect on environmental performance (β = 0.179, t = 1.822, p = .068) were nonsignificant, and the indirect path coefficients for NC→NBMI→ECOP, NC→NBMI→ENVP, NC→EBMI→ECOP, and NC→EBMI→ENVP were 0.183 (t = 3.913, p = .000), 0.109 (t = 2.622, p = .009), 0.167 (t = 2.766, p = .006), and 0.205 (t = 2.950, p = .003), respectively. Thus, H9a, H9b, H10a, and H10b were supported, indicating that novelty- and efficiency-centered BMI fully mediated the effect of network capability on economic and environmental performance.

We analyzed other indirect paths, which are summarized in Table 4. The coefficients for indirect paths, including DT→NC→NBMI→ECOP (β = 0.104, t = 3.483, p = .000), DT→NC→NBMI→ENVP (β = 0.062, t = 2.392, p = .017), DT→NC→EBMI→ECOP (β = 0.095, t = 2.495, p = .013), and DT→NC→EBMI→ENVP (β = 0.116, t = 2.700, p = .007), were all significantly positive. Table B3 summarizes the direct effects, indirect effects, and total effects, providing a full picture of how digitalization leads to enhanced economic and environmental performance.

Structural model moderation effects of environmental dynamism

We then generated interaction terms to validate the moderating effects of environmental dynamism and added them to the structural model shown in Figure 2. Figure 4 summarizes the structural model results for testing moderating effects. The product indicator approach was used to build product terms between the indicators of digitalization and those of environmental dynamism. All the indicators were mean centered to avoid multicollinearity, as suggested by Henseler and Fassott (2010). The interaction term (DT × EDYN) was measured with the product terms and linked to NBMI and EBMI in the structural model. The coefficients for the path from the interaction term of digitalization and environmental dynamism to novelty- and efficiency-centered BMI were −0.125 (t = 1.553, p = .120) and −0.164 (t = 3.028, p = .002), supporting H11b only.OPEN IN VIEWER

Because network capability was modeled as a formative construct, we used a two-stage method to establish the interaction term, which we then linked to novelty- and efficiency-centered BMI. Following Chin et al. (2003) and Henseler and Chin (2010); Henseler and Fassott (2010), in the first stage, we ran the PLS path model to obtain the latent variable scores for network capability and environmental dynamism. In the second stage, we calculated the interaction term (NC × EDYN) using the product of the latent variable scores and then added it to the main effect model. The results indicated that the coefficient for the path linking the interaction term between environmental dynamism and network capability with novelty-centered BMI was nonsignificant (β = 0.047, t = 0.851, p = .395), thus H12a was unsupported. The interaction term had a significantly positive influence only on efficiency-centered BMI (β = 0.105, t = 2.568, p = .010), supporting H12b. Table 5 summarizes the results for the moderating effects.

OPEN IN VIEWER

Table 5.

Results for moderation effects testing.

Moderation effects	β	SE	t-values	p-values	f 2	Support?
H11a. DT x EDYN → NBMI	−0.125 n/s	0.080	1.553	0.120	0.025	Not supported
H11b. DT x EDYN → EBMI	−0.164**	0.054	3.028	0.002	0.046	Supported
H12a. NC x EDYN → NBMI	0.047 n/s	0.056	0.851	0.395	0.005	Not supported
H12b. NC x EDYN → EBMI	0.105**	0.041	2.568	0.010	0.023	Supported

Contributions to research and theory

Our study helps explain the phenomenon of digital transformation by investigating the interrelationships among digitalization, network capability, and BMI. Digital transformation is far more than the adoption of digital technologies in changing business processes (Wessel et al., 2021): it also includes the changes in the organization and business domain level (Hanelt et al., 2021; Parviainen et al., 2017; Verhoef et al., 2021). Based on DCT, our study proposes digitalization, network capabilities, and BMI as three critical dynamic capabilities to support digital transformation in the process, organization, and business domain level, respectively. Digitalization as the precursory phase of digital transformation triggers organizational changes in networking relationships and business model transformation. The results confirm the positive effects of digitalization on network capability and BMI and the mediating effect of network capability on the association between digitalization and BMI. Following Li et al. (2018), Töytäri et al. (2018), and Cenamor et al. (2019), our study suggests that digitalization should be complemented by network capability development. Digitalization enables the organization to sense and seizing the business opportunities by integrating digital resources, whereas network capability can help the organization acquire and integrate organizational resources from its business partners to promote transformation. The findings therefore provide theoretical and empirical evidence of how digitalization contributes to business model transformation (Li et al., 2020; Weking et al., 2020) and advance the understanding of how successful digital transformation occurs.

To provide a nuanced understanding of the differential influence of digital transformation from the sustainability perspective, our research model introduces economic and environmental performance as dependent variables. Previous studies have extensively discussed the influence of digital technologies on operational performance (Buer et al., 2021; Gillani et al., 2020), organizational performance (Nwankpa and Datta, 2017; Sousa-Zomer et al., 2020), and competitive performance (Mikalef et al., 2020)—all of which are reflected in the economic dimension of sustainability performance. Our study extends existing research on digital transformation by including the environmental dimension of sustainability performance (Dubey et al., 2019; Hanelt et al., 2017; Nayal et al., 2022). Following Vial’s (2019) suggestions, our study establishes links between digitalization and two dimensions of sustainability performance to elaborate the wider-ranging outcomes of digital transformation. We analyze and validate the mediating effects of network capability and BMI to reveals the paths through which digitalization leads to better economic and environmental performance (Soluk and Kammerlander, 2021; Wessel et al., 2021). Our results demonstrate that digitalization leads to novelty- and efficiency-centered BMI through network capability and that novelty- and efficiency-centered BMI mediate the effects of digitalization and network capability on economic and environmental performance. The results indicate that digital transformation has the potential not only to increase an organization’s economic profits but also to mitigate the negative influence of its manufacturing and operating activities on the natural environment. The paths of DT→NC→NBMI/EBMI→ECOP/ENVP provide unprecedented insights into the ways how digitalization can lead to desired sustainability-related outcomes on the firm level.

By investigating the moderating role of environmental dynamism, this study also specifies the boundary conditions under which digitalization and network capability lead to novelty- and efficiency-centered BMI. The results show that environmental dynamism negatively moderates the effect of digitalization on efficiency-centered BMI but positively moderates the relationship between network capability and efficiency-centered BMI. This means that in a highly dynamic environment, like that of Chinese manufacturing, the use of digital technologies can further increase the complexity of business operations and carry additional transaction risks (Paolucci et al., 2021; Rana et al., 2021; Yuan et al., 2021) that may impede efficiency-centered BMI—yet bolster novelty-centered BMI. However, network capability can enable firms to deal with a dynamic environment and improve transaction efficiency in a flexible and reliable manner through social and relational mechanisms, such as communication and coordination (Cenamor et al., 2019). Environmental dynamism does not strengthen the effects of digitalization and network capability on novelty-centered BMI, indicating that novelty-centered innovation in both stable and dynamic environments relies on digitalization and network capability. Thus, our study reveals the distinct effects of digitalization (a technology-driven approach) and of network capability (an organizational factor) on BMI in different environmental conditions. These findings are consistent with those of Gillani et al. (2020), which strongly urge future studies to take a more comprehensive perspective such as the Technology-Organization-Environment (TOE) framework when investigating the critical success factors of digital transformation. As a result, our study adds valuable evidence to the existing conceptualization of digital transformation as a multidisciplinary process.

Our study uses DCT to illuminate the role of digital transformation in Chinese manufacturing firms, which is characterized by high-speed development and rapidly changing environment. DCT highlights the organization can maintain sustainable competitiveness by improving dynamic capabilities of opportunity-sensing, opportunity-seizing, and transformation to manage the dynamic environment (Li and Liu, 2014; Teece, 2007). From the perspective of DCT, digitalization, network capability, and BMI can be regarded as manufacturers’ dynamic capabilities, which cannot be easily imitated, to enable digital transformation and achieve significant, sustained competitive advantage. Many scholars have used DCT to conceptualize digitalization (Soluk and Kammerlander, 2021), network capability (Chen et al., 2021), and BMI (Teece, 2018). Our study strengthens this research stream by identifying digitalization and network capability as keys to help the firm sense business opportunities and challenges and seize opportunities through integration of digital resources and organizational resources (Chi et al., 2010; Mikalef et al., 2021a, 2021b; Warner and Wäger, 2019). Novelty- and efficiency-centered BMI are dynamic capabilities that enable firms to reconfigure resources and competencies in ways that transform value creation and delivery (Mezger, 2014). Our study distinguishes the three dynamic capabilities and reveals the mediating role of network capability and BMI and the moderating role of environmental dynamism. These findings extend DCT by empirically validating the interplay among these dynamic capabilities and clarifying how the relationships vary with environmental conditions.

Implications for practice and management

In recent years, the COVID-19 outbreak and extreme weather events have had profoundly negative effects on the natural environment and have attracted widespread public attention (Pan and Zhang, 2020; Weilnhammer et al., 2021). China-based global manufacturing firms are under tremendous pressure to improve their economic and environmental performance and to maintain sustainable development using digital technologies (Pan et al., 2021). In highly dynamic environments, firms are encouraged to seize opportunities to promote sustainable business model transformation in the digital economy (Hanelt et al., 2017; Pan et al., 2021). Our study indicates that for these firms, digitalization leads to improved economic and environmental performance through novelty- and efficiency-centered BMI. Manufacturing firms need to leverage digital technologies to create new ways of conducting economic exchange as well as to reduce transaction costs and improve transaction efficiency, consequently promoting sustained returns. The findings thus offer practitioners insights into the integration of digital technology use and business model design in the pursuit of sustainable development.

Our study suggests that if a firm’s digital transformation relies only on technologies, it is likely to have a high chance of failure. Although our study is conducted based on survey data collected from the Chinese manufacturing industry, a recent McKinsey survey from “1793 participants representing the full range of regions, industries, company sizes, functional specialties, and tenures,” shows that the success rate of digital transformation is less than 30% (McKinsey, 2018). The result is a warning to all managers whose organizations aspire to engage in digital transformation. For the endeavor to succeed, firms need to consider all aspects of technological, organizational, and environmental factors. They must focus on intra- and interorganizational relationship management in the transformation process and take full advantage of internal and external network partnerships to establish reliable connections and ensure the acquisition of high-quality resources. Network capabilities, including effective communication and coordination, relational skills, and partner knowledge, play essential roles in a successful digital transformation.

In a highly dynamic global manufacturing environment, digital technology use has a dark side. Our findings suggest that environmental dynamism weakens the effect of digitalization on efficiency-centered BMI, at least for Chinese manufacturing firms. Such firms emphasize operational excellence and endeavor to become price leaders by lowering transactional costs (Zhang et al., 2021). They must effectively manage the complexity and risks related to digitalization, which impede transaction efficiency; moreover, to collaborate with business partners flexibly and reliably in view of efficiency-centered BMI, they must put more emphasis on network capability development.