The performance implications of digital transformation (DT) have been extensively scrutinized; however, few studies have focused on the mechanisms that boost performance. This study investigates the potential benefits of digital transformation (DT) from a value creation perspective within the value creation-appropriation (VCA) framework. Using panel data from listed firms in China, we show that DT improves firm performance by simultaneously enhancing innovation capability and reducing cost stickiness, two critical determinants of efficient value creation. We also identify customer concentration and relative capital intensity—key structural bases of value creation activities—as moderating these value creation mechanisms. Specifically, customer concentration weakens the positive effect of DT on innovation capability, while relative capital intensity strengthens both the innovation-enhancing and cost-reducing effects of DT. These findings contribute to existing literature on DT by revealing the underlying mechanisms linking digital initiatives to economic outcomes and identifying critical contingencies that shape this relationship.
Digital transformation (DT) is a multifaceted, continuous process through which firms leverage and integrate advanced digital technologies, such as artificial intelligence (AI), blockchain, cloud computing, and big data (often called “ABCD”), into their business operations. Engaging in DT, firms aim to make significant changes to their business processes and models, and the overall strategic direction (Verhoef et al., 2021; Vial, 2019). These changes can reshape a firm’s value creation paradigms (Hess et al., 2016; Li et al., 2018), in some cases, introducing entirely new business models to achieve superior value creation and value appropriation (Bharadwaj et al., 2013; Fitzgerald et al., 2014). Thus, DT is not just a technological trend but a fundamental shift in how firms create and capture value in an increasingly digital economy.
Research in DT is growing, with a substantial focus on its diverse impacts. Broadly, DT has been found to support innovation by fostering new skills, competencies, and knowledge acquisition (e.g., Cheng et al., 2023), and drive operational efficiency by enhancing business processes, supply chain integration, and organizational agility (Bresciani et al., 2022). Studies also find that DT increases customer engagement by enabling service personalization (Wielgos et al., 2021). Beyond these operational and strategic shifts, empirical evidence shows a positive association between DT and firm performance, including gains in productivity, profitability, and market share (Ferreira et al., 2019; Li, 2022; Lu et al., 2023). Most of these studies apply a resource-based view (RBV) and the dynamic capabilities perspective to find that DT can reconfigure resources and capabilities (Bresciani et al., 2022; Vial, 2019). More recently, scholars have examined contingency factors that influence DT’s performance outcomes, specifically addressing the so-called “performance puzzle” in DT implementation and offering a more nuanced understanding of its performance implications (e.g., Guo et al., 2023; Li et al., 2022b; Wang et al., 2025).
Despite the growing body of research on DT, three gaps remain. First, while prior DT research draws on the RBV or dynamic capability perspective to explain why performance differences occur (Hess et al., 2016; Li et al., 2018), they largely treat the value creation process as a “black box”, that does not shed light on how DT shapes the specific drivers that generate competitive advantage. The literature on competitive advantage, particularly within a value creation-appropriation (VCA) framework, emphasizes value creation as the fundamental process through which firms generate and capture economic rents in market exchanges and formalizes a clear value creation equation (Lepak et al., 2007). Here, an integrated perspective rooted in value creation is applied to unlock the “black box” of value creation. Focusing on the specific mechanisms through which DT enhances firm performance provides a profound understanding of how firms benefit from leveraging DT.
Second, prior research on value creation has largely focused on customer-side dynamics, such as personalized services, engagement, and co-creation (Matarazzo et al., 2021), with relatively limited attention to value creation at the focal firm level. Even among the few studies that consider the focal firm, the emphasis is mainly on competitive isolating mechanisms, such as innovation protection and entry barriers (Sharapov & MacAulay, 2022), rather than on internal efficiency-enhancing mechanisms, such as agile resource allocation and flexible cost structures. This fragmented approach fails to consider the fact that sustainable value creation relies on both product innovation and operational efficiency—innovation enhances revenue by creating offerings that customers are willing to pay for, while operational efficiency reduces costs and supports revenue generation by ensuring these offerings are effectively delivered (Bowman & Ambrosini, 2000; Lepak et al., 2007). Therefore, examining how DT simultaneously shapes these two dimensions to reconfigure firms’ internal value creation processes provides a realistic and comprehensive understanding of the underlying mechanisms through which DT drives firm performance.
Third, while prior studies examine boundary conditions in the DT–performance relationship, most focus on managerial and environmental factors (Guo et al., 2023; Li, 2022), assuming that value creation structures are homogeneous. This assumption overlooks that the foundations of firms’ value creation activities vary widely, especially in relation to customer base (identifying target customers) and resource base (allocating resources to serve target customers). The requirements for these foundations vary depending on the value creation model a firm follows. As a result, the effectiveness of DT-driven value creation may vary widely across firms with different combinations of customer and capital base. As customer concentration and relative capital intensity are key metrics that reflect these structures, investigating their moderating roles in DT’s value creation mechanisms offers a nuanced and comprehensive understanding of how DT affects firm performance.
In seeking to uncover DT’s performance implications through the lens of value creation, we address two key questions. First, what are the mechanisms through which DT affects firm performance from the value creation perspective? Second, how do value creation contingencies—customer concentration and relative capital intensity—shape these effect mechanisms? Drawing on the VCA framework, we conceptualize the efficiency stemming from DT along two dimensions: enhancing innovation capability, which expands and improves firms’ product offerings; and reducing cost stickiness, which constrains production and supply efficiency. We argue that DT contributes to firm performance by simultaneously increasing innovation capability and mitigating cost stickiness. Moreover, we posit that the structure of a firm’s customer base and capital base defines important boundaries for its value creation activities. Specifically, customer concentration is expected to weaken, while relative capital intensity is expected to strengthen DT’s value creation mechanisms. Our predictions are strongly supported by empirical evidence from a panel dataset of Chinese manufacturing firms listed on the Shanghai and Shenzhen Stock Exchanges during the period from 2013 to 2020.
Our study makes three contributions. First, it advances DT research by shifting focus from resources and capabilities—associated with the RBV and the dynamic capability perspective—to provide a deeper analysis of why DT drives positive outcomes. Drawing on the value creation perspective within the VCA framework, we unlock DT’s value creation “black box” by examining both the factors that determine efficiency and the underlying contingencies. Doing so offers a comprehensive theoretical perspective of how DT influences firm performance. Second, we integrate two isolated but critical value creation efficiency drivers, namely innovation capability and cost stickiness, into a unified analytical framework and examine their mediating roles in how DT affects firm performance. The findings broaden understanding of DT’s performance effects, and provide deeper insights into the internal value creation mechanisms within the VCA framework. Third, we enrich the contingent view of firm strategy implementation by incorporating customer concentration and relative capital intensity as the key structural factors, proposing and testing their moderating effects on DT’s value creation mechanisms. The findings offer novel insights into how firms can align their DT strategies with their specific customer and resource foundations for improved performance.
THEORETICAL foundation AND hypothese developmentFirm digital transformation from the value creation perspectiveDigital transformation (DT) refers to the continuous process by which organizations leverage and integrate advanced digital technologies into business operations to achieve significant organizational changes and improvements (Vial, 2019). Prior studies applied the RBV and the dynamic capability perspectives to explain DT as a value-based strategy to support firms’ competitive advantage and improve firm performance. Research shows that DT strengthens firms’ market-sensing and decision-making capabilities by enabling them to capture, process, and analyze fast-changing market information (Ghosh et al., 2022; Gupta et al., 2020; Heredia et al., 2022). Moreover, DT also supports the reconfiguration and orchestration of strategic resources and organizational capabilities, leading to process optimization, greater agility, and timely responses to emerging market opportunities (Bresciani et al., 2022; Fitzgerald et al., 2014; Li, 2022; Li et al., 2022a). These studies provide valuable insights into why DT strategically matters for value creation and firm performance, but not how DT translates into superior performance. That is, the underlying mechanisms through which DT influences the efficiency and effectiveness of value creation remain underexplored.
The value creation perspective within the VCA model offers a lens to bridge this gap. The VCA framework, grounded in the value-based strategy view (Brandenburger & Stuart, 1996), conceptualizes competitive advantage as arising from a firm’s efficiency in two distinct but interrelated processes, namely value creation and value appropriation (Lepak et al., 2007). In this view, value creation represents the initial stage, where firms generate economic value by widening the gap between customers’ willingness to pay (WTP) and the opportunity costs of a new offering. To maximize profit, firms need to simultaneously raise customer WTP and minimize offering costs by efficiently leveraging resources from diverse stakeholders (Barney, 2018). Compared with RBV, which focuses on the role of valuable strategic resources and capabilities, the VCA framework captures how those resources are turned into customer-valued offerings and how efficiently they are delivered. This value creation perspective offers a more comprehensive understanding of DT’s impact on firm performance by connecting and reconfiguring resources (RBV/dynamic capabilities) with processes that transform them into value creation and firm performance.
Innovation capability and cost stickiness as determinants of value creation efficiencyFrom the value creation perspective within the VCA framework, explaining the performance implications of DT requires explicit attention to the two fundamental dimensions that determine value creation efficiency: (1) the price customers are willing to pay for offerings (revenue); and (2) the costs incurred to produce and deliver these offerings (cost) (Lepak et al., 2007). To operationalize this framework, we identify two critical determinants of value creation efficiency: innovation capability and cost stickiness.
Innovation capability, which can be defined as a firm’s capability to imagine, invent, and develop new solutions with the goal of creating value (Owen et al., 2013, p.32), is closely tied to customers’ WTP and, in turn, to value creation efficiency. Drawing on Schumpeter’s (1942) theoretical foundations of economic development and creative destruction, innovation goes beyond incremental improvements, transforming markets and redefining customer expectations. Firms achieve temporary monopolistic advantages by introducing novel products or processes that competitors cannot replicate, thereby disrupting existing market equilibria and shifting demand curves upward. Thus, as firms innovate to introduce superior products or services, they not only redefine customer expectations but also establish pricing power that drives up customers’ WTP, thereby enhancing value creation efficiency (Lepak et al., 2007). Additionally, from the RBV perspective, innovation capability is a valuable and inimitable resource that enables firms to generate Schumpeterian rents and sustain competitive advantage over time (Barney, 1991). Accordingly, firms that continuously build their innovation capability can establish stronger pricing power through technologically advanced and hard-to-imitate value propositions and offerings, thereby increasing customers’ WTP and amplifying the revenue side of the value creation equation.
Cost stickiness, by contrast, constitutes a critical determinant of value creation efficiency on the cost side of the equation. It refers to asymmetric cost behavior, where costs decrease less when sales fall than increase when sales rise (Anderson et al., 2003). This rigidity often originates from inefficient resource allocation or organizational inertia, which causes costs to persist even during sales downturns driven by rapidly changing market demands. Such stickiness undermines a firm’s value creation efficiency by constraining its ability to adjust operations flexibly (Banker et al., 2014). Empirical evidence supports this view; for instance, Zhang et al. (2022) show that the stickiness of selling, general and administrative (SG&A) expenditures is negatively associated with firm performance, particularly in highly competitive contexts. Similarly, Jang and Yehuda (2021) find that acquirers’ cost stickiness is negatively related to their value creation efficiency in mergers and acquisitions (M&As). These findings suggest that firms with flexible cost structures can dynamically realign resource allocation with sales fluctuations, reducing production and delivery costs and preserving value creation.
Thus, innovation capability and cost stickiness can be considered two critical and complementary determinants of firm value creation efficiency. Innovation capability drives revenue by increasing customers’ WTP, while lower cost stickiness safeguards costs by enabling efficient resource and cost management. In the following sections, we examine how DT shapes firm performance through these two determinants, revealing the mechanisms through which DT translates into improved value creation efficiency and financial outcomes.
Digital transformation and firm performance from the value creation perspectiveDigital transformation, innovation capability, and firm performanceDigital transformation (DT) involves the deep integration of digital technologies into organizational processes (Verhoef et al., 2021; Vial, 2019). It strengthens firms’ technological expertise in identifying, acquiring, recombining, and leveraging internal and external knowledge to drive innovation (Zhou & Wu, 2010). For instance, cloud-based platforms enable firms to integrate third-party application programming interfaces (APIs) or open-source tools to refine existing products or prototype new ones, thereby accelerating the innovation cycle (Nambisan et al., 2017). More importantly, DT allows firms to anticipate and strategically navigate emerging technology fields, enabling agile reconfiguration of technical resources and exploration of opportunities transcending existing boundaries (Karimi & Walter, 2015). As such, DT fosters both exploitative and explorative innovation, strengthening innovation capability and value creation efficiency (Zhou & Wu, 2010).
Digital transformation (DT) also improves customer-orientation innovation. Digital driven interactive customer databases provide real-time insights into the customer value proposition, demand, product perception, and usage experience across the entire purchase journey (Verhoef et al., 2021). These data-driven insights help firms better align their innovation efforts with customer expectations and enhance their innovation capability and value creation efficiency (Peltier et al., 2020). In addition, digital tools such as social media and search engines allow firms to monitor and respond dynamically to customer preferences, as customers evolve into informed and active participants (Lamberton & Stephen, 2016).
Furthermore, DT enables firms to establish collaborative innovation networks, platforms, and even ecosystems where they can integrate partners, suppliers, and customers into their innovation initiatives (Bonina et al., 2021). These collaborations generate network effects that further enhance innovation capability. As firms deepen their DT efforts, their digital strategies shift from optimizing business processes to scaling new business models based on digital supply chains and platform interactions (Nambisan et al., 2017). In these models, firms co-create value with network participants who contribute complementary resources and learn from each other, accelerating knowledge exchange and innovation and enhancing value creation (Grönroos & Voima, 2013).
In sum, DT strengthens innovation capability through enhancing exploitative and explorative innovation, enabling customer interactive innovation, and fostering collaborative innovation. These mechanisms increase customer WTP and value creation efficiency and, ultimately, firm performance. We propose:
H1 DT improves firm innovation capability, which serves as a key determinant of firm value creation efficiency, thereby leading to improved firm performance.
We further argue that DT plays an important role in mitigating cost stickiness and, thereby, enhancing value creation efficiency. By integrating digital technologies such as big data, artificial intelligence, blockchain, and cloud computing, firms gain operational flexibility through real-time coordination across internal functions and external stakeholders (McIntyre & Srinivasan, 2017). This enables dynamic production inventory adjustments to demand fluctuations, reducing inefficiencies triggered by rigid cost structures (Bresciani et al., 2022). Thus, DT enables firms to optimize resource utilization both during expansion, through rapid scalability, and downturns, by preventing excess cost accumulation.
Digital transformation (DT) also mitigates cost stickiness arising from managerial over-optimism. Managers are often overly optimistic in their forecasts, leading to overinvestment and high fixed costs that are difficult to adjust during downturns (Banker et al., 2014). In fact, DT reduces this bias by enabling data-driven decision-making. For example, AI-powered predictive analytics and machine learning systems accurately forecasts market trends, decreasing reliance on subjective judgment and lowering the risk of overinvestment (Lu & Ramamurthy, 2011).
In addition, DT alleviates cost stickiness linked to agency problems. Information asymmetry between managers and stakeholders can lead to inefficient resource allocation (Banker et al., 2014). DT improves transparency and coordination through real-time analytics, digital collaboration tools, and cloud-based dashboards that provide timely access to financial and operational data (Chen & Zhang, 2024). Therefore, enhanced data-sharing across departments, suppliers, and partners aligns incentives and reinforces coordination within the value network, further reducing agency-driven cost rigidity. In sum, DT reduces cost stickiness by improving operational flexibility, correcting managerial biases, and lowering agency costs. This reduction enhances value creation efficiency and ultimately strengthens firm performance. We propose:
H2 DT mitigates overall cost stickiness, which serves as a key determinant of firm value creation efficiency, thereby leading to improved firm performance.
The above arguments are framed within the general foundation of firms’ value creation, which depends on two key dimensions: the customer base, defining for whom value is created; and the capital base, determining the resources with which value is generated (Brandenburger & Stuart, 1996). The customer base directly shapes a firm’s ability to serve and capture value, while the capital base reflects the labor, technology, and capital assets that support and sustain operational and strategic initiatives. As firms advance DT implementation, both dimensions evolve; therefore, it is important to investigate how firms condition the link between DT and value creation efficiency, and thereby firm performance. Specifically, customer concentration, the extent to which a firm’s revenue depends on a few key customers, directly influences how firms pursue innovation and adaptability within their value creation paradigm (Krolikowski &Yuan, 2017). Thus, this feature of customer base is central to understanding DT value creation effects. Similarly, relative capital intensity, the extent to which a firm relies more on capital than labor (Acemoglu & Restrepo, 2019), affects operational flexibility, investment in digital infrastructure, and response to emerging digital opportunities, thereby influencing the firm’s DT value creation process.
Accordingly, investigating how customer concentration and relative capital intensity interact with DT’s value creation mechanisms is essential for a more nuanced understanding of its implications for firm performance.
Moderating effects of customer concentrationCustomer concentration reflects the structure of a firm’s customer base, formed through the strategic choices made when prioritizing key customers for cost reduction and economies of scale (Patatoukas, 2012). While this approach provides efficiency in traditional settings (Patatoukas, 2012), the digital economy—driven by customer-centric, demand-driven value creation—enables firms to adopt a more interactive and adaptive value creation model (Verhoef et al., 2021). In this context, a dispersed customer base exposes firms to a broader knowledge pool, enhancing innovation aligned with personalized demand patterns, supporting greater WTP and value creation efficiency. By contrast, high customer concentration limits knowledge recombination and experimentation, constraining firms’ ability to leverage DT for dynamic adaptation.
Moreover, DT transforms traditional supply chains into interconnected digital value networks, where value emerges from interactions between firms, customers, and suppliers (McIntyre & Srinivasan, 2017). Within these networks, strong customer concentration reduces exposure to diverse market insights, hampering co-innovation and digital resource orchestration (Cenamor et al., 2019). Conversely, a dispersed customer base fosters knowledge spillovers and access to external resource, thus strengthening digital innovation and enhancing value creation efficiency. We propose:
H3a Customer concentration weakens the positive effect of DT on innovation capability, which, in turn, enhances firm performance.
We argue that customer concentration moderates the relationship between DT and cost stickiness by constraining the flexibility DT provides in cost management. A concentrated customer base compels firms to tailor operations and resources around a few dominant customers, increasing cost rigidity and reducing adaptability to market fluctuations (Chang et al., 2021). This lack of demand diversification further impedes firms’ responsiveness to broader market shifts in the digital business environment. While DT enhances data analytics, real-time monitoring, and resource optimization, firms with high customer concentration may fail to fully leverage these tools to streamline cost adjustments, thus diminishing DT’s ability to mitigate cost stickiness.
Moreover, while DT improves decision making by aligning it with real-time market conditions, firms with a concentrated customer base receive limited external market signals, and therefore fewer decision-relevant insights (Zhong et al., 2021). As a result, even with enhanced transparency and decision-support systems, managers may continue to rely on narrow information sets, diminishing DT’s effectiveness in reducing agency costs and cost stickiness. In sum, customer concentration reinforces structural rigidities and limits information diversity, constraining firms’ ability to use DT for cost flexibility and weakening its capacity to mitigate cost stickiness. We propose:
H3b Customer concentration weakens the negative effect of DT on cost stickiness, which, in turn, decreases firm performance.
In summary, customer concentration moderates the relationships between DT and both innovation capability and cost stickiness, serving as a key contingency that weakens DT’s positive effect on firm performance through enhanced value creation efficiency. Therefore, we propose:
H3 Customer concentration weakens the value creation efficiency mechanisms of DT, thereby weakening the positive relationship between DT and firm performance.
Capital intensity refers to a firm’s reliance on physical assets, such as machinery, equipment, and infrastructure (Kaldor, 1939), whereas relative capital-labor intensity indicates the degree to which a firm relies on capital versus labor in production (Acemoglu & Restrepo, 2019). Kaldor (1939) emphasized that changes in capital intensity reflect shifts in production methods, making capital structure a key determinant of value creation. In the context of DT, where technologies are inherently standardized and can be easily replicated by competitors, firms must integrate these technologies with their unique resource and capabilities to fully realize the value creation potential of their DT (Warner & Wäger, 2019).
To our knowledge, firms with higher relative capital-labor intensity are better positioned to leverage digital technologies, which foster innovation by increasing efficiency and enabling rapid adaptation to market dynamics. These firms can invest substantially in technology upgrades to unlock new opportunities, fostering rapid innovation cycles and strengthening innovation capability. In contrast, labor-intensive firms, where innovation relies on human capital, tend to have slower technology adoption, constraining the extent to which DT can enhance innovation. While DT can still enhance data-driven decision-making and operational efficiency, its impact is likely to be pronounced in capital-intensive firms. Therefore, we propose:
H4a Relative capital-labor intensity strengthens the positive effect of DT on innovation capability, which, in turn, enhances firm performance.
We further contend that relative capital-labor intensity moderates the relationship between DT and cost stickiness. Specifically, in capital-intensive industries, firms typically face significant fixed costs associated with machinery, equipment, and infrastructure (Harris, 1988), which can be rapidly optimized during DT implementation. Inherently, digital technologies, such as automation and predictive analytics, are better suited to reconfiguring capital assets than labor inputs, which are often constrained by organizational rigidity (Acemoglu & Restrepo, 2019). This allows firms to dynamically adjust fixed capital investments, reducing over-optimistic forecasting and minimizing agency costs, thereby mitigating cost stickiness.
Additionally, the predominance of capital resources in capital-intensive firms creates an operational environment that is amenable to the rapid, technology-enabled adjustments afforded by DT. These firms can leverage digital tools to streamline production and reallocate resources swiftly in response to demand fluctuations. By contrast, labor-intensive firms are hampered by the inflexibility of human capital adjustments. Thus, the ability of DT to restrict cost stickiness is strengthened in capital-intensive firms, leading to agile cost management and improved overall firm performance. Therefore, we propose:
H4b Relative capital-labor intensity strengthens the negative effect of DT on cost stickiness, which, in turn, reduces firm performance.
In summary, relative capital intensity moderates the relationships between DT and innovation capability and cost stickiness, serving as a contingency that strengthens DT’s positive effect on firm performance through enhanced value creation efficiency. Therefore, we propose:
H4 Relative capital-labor intensity strengthens the value creation efficiency mechanisms of DT, thereby strengthening the positive relationship between digital transformation and firm performance.
Based on the above arguments, we build a conceptual framework, as outlined in Fig. 1.
METHODSample and data collection
Our sample comprised Chinese manufacturing firms listed on the Shanghai and Shenzhen Stock Exchanges during the period from 2013 to 2020. We focused on this setting for three reasons. First, China’s digital economy has undergone rapid expansion, with the DT of manufacturing firms serving as a key driver of industrial upgrading. Investigating this context offers valuable insights for firms in other countries seeking value creation in their DT implementations (e.g., Li et al., 2022a). Second, manufacturing, given its substantial contribution to economic development, deep integration of digital technologies, and complex production and supply chain structures, provides an ideal setting for investigating DT’s value creation mechanisms in the digital era. Third, the diversity of China’s manufacturing subsectors results in varying degrees of DT adoption in procurement, production, manufacturing, and sales processes (Zhang, 2025), thus enabling a comprehensive analysis of how DT strategies shape performance outcomes.
To address potential endogeneity and establish causal links, we lagged the independent, moderating, and all control variables by one year, with independent variables beginning in 2013 and the dependent variable in 2014. The base year 2013 was chosen because it marks a critical turning point in China’s national strategy to embed digital technologies into traditional industrial sectors. Data were drawn from the CSMAR database, covering all eligible A-share manufacturing firms, excluding modern manufacturing firms to avoid confounding effects from varying DT paradigms. To mitigate irregularities associated with newly listed firms, we removed firms with significant operational or financial anomalies during the study period and those listed for less than one year (Cheng et al., 2022). After eliminating observations with missing or abnormal data, the final unbalanced panel included 1770 firms and 8225 firm-year observations. Data sources included annual reports for extracting DT-related information and the CSMAR and WIND databases, from which firm-level characteristics such as innovation capability, cost stickiness, customer concentration, relative capital intensity, and firm performance were sourced. Both databases are authoritative and extensively employed in strategy and management research (Cheng et al., 2022).
MeasuresDependent variablesFirm Performance. Consistent with prior research (Cheng et al., 2022), we measured firm performance using return on assets (ROA), a widely adopted financial metric that captures a firm’s overall ability to generate profits relative to its total assets.
Independent variablesDigital Transformation (DT). To measure DT, we followed a methodology widely adopted in DT research (e.g., Leng & Zhang, 2024; Zhou & Li, 2023; Zhou et al., 2022), employing text analysis techniques to quantify the frequency of DT-related terms in firms’ annual reports. We first compiled a dictionary of DT expressions and categorized them into two dimensions: technical usage and practical application. The technical usage dimension included four categories: artificial intelligence, cloud computing, blockchain, and big data. The practical application dimension reflects the deployment of these technologies across business operations, and includes terms related to intelligence manufacturing, digital business model, and digital-enabled consumer experiences. Subsequently, DT is constructed as the logarithmic of one plus the total frequency of keywords across these two dimensions in each firm’s annual report.
Mediating variables and moderating variablesInnovation capability (IC). Consistent with previous literature (Saunila, 2020), we measured a firm’s innovation capability based on its R&D output, proxied by the number of patents independently filed in a given year. To normalize the distribution and reduce the impact of extreme values, we applied a natural logarithm transformation to this measure.
Cost stickiness (CS). To measure cost stickiness, we adopted the widely recognized Weiss model (Weiss, 2010). The specific formula is:
Where τ denotes 4, representing the four quarters in a year; τ‾ represents the most recent quarter in which sales declined, and τ¯ denotes the most recent quarter in which sales increased. ΔSALE and ΔCOSTcapture the quarterly changes in sales and costs, respectively. STICKY measures the difference in the slopes of the cost function between the most recent two quarters, specifically between sales decline and sales growth. Cost stickiness is evident when cost increases during sales growth exceeds cost reductions during sales declines. Therefore, a lower STICKY value indicates a higher degree of cost stickiness. For easier interpretation, we used its inverse as the firm’s CS measure.
Customer concentration (CC). Following Cohen and Li (2020), we measured customer concentration as the proportion of a firm’s total sales attributed to its top five customers. This metric reflects the degree of revenue dependence on a limited customer base, where higher concentration indicates greater reliance on a few key buyers.
Relative capital intensity (RCI). Consistent with Yao et al. (2023), we proxied relative capital-labor intensity using the logarithm of the ratio of fixed assets to the number of employees. A higher value denotes greater dependence on capital investment relative to labor in production processes.
Control variablesWe controlled for a series of variables in our analysis. First, we controlled for fundamental firm attributes, including Firm age, measured as the number of years since establishment, and Firm size, measured by the natural logarithm of total assets. To capture ownership heterogeneity, we included SOE status, a dummy variable coded as 1 if the firm’s ultimate controller is a government entity and 0 otherwise, along with State ownership, measured as the percentage of stakes held by government-affiliated entities. Second, we accounted for slack resources. Specifically, we controlled for absorbed slack and unabsorbed slack as used in previous studies (Cheng et al., 2022). Absorbed slack is measured by the ratio of SG&A to sales, representing resources embedded within existing operations. Unabsorbed slack is measured by the ratio of current assets to current liabilities, indicating managerial discretion in resource allocation. Given China’s institutional context, we also controlled for government subsidy, measured as the ratio of total government subsidies to total sales, capturing the role of state support. Third, we included top management team (TMT) characteristics, which shaped strategic choices and value creation activities. CEO duality was coded as 1 if the CEO also served as board chair and 0 otherwise, TMT ownership was measured as the percentage of shares held by the top executives. Finally, to account for market conditions, we incorporated regional market development, measured using the National Economic Research Institute (NERI) marketization index, widely applied in economics, finance, and strategy research. Additionally, all regression models included industry and year fixed effects to capture unobservable industry-specific and time-variant factors.
Estimation approachThe analysis leveraged the panel structure of the data, enabling us to capture both cross-sectional heterogeneity across firms and temporal dynamics within firms over time. The Hausman test favored a fixed-effects model over a random-effects specification, directing us to adopt fixed-effects estimation to test our hypotheses. Additionally, we controlled for a set of executive-, board-, firm-, and industry-level variables that may simultaneously affect a firm’s DT implementation and performance. Year and industry fixed effects were included to account for unobserved heterogeneity. Furthermore, to address potential endogeneity stemming from reverse causality, we employed lagged independent variables, mediators, moderators, and control variables (Cheng et al., 2022).
RESULTSMediation effects of innovation capability and cost stickinessTable 1 presents the descriptive statistics, including the mean, standard deviation, and correlation matrix of all variables. Correlation analysis indicates that multicollinearity is not a concern, as the correlation coefficients remain within a low to moderate range. Additionally, the variance inflation factors (VIFs) across the regression models ranged from 3.18 to 3.24, well below the threshold of 10, further confirming the absence of multicollinearity issues.
Descriptive statistics and correlations.
Notes: n = 8225; *p < 0.05.
Table 2 presents the estimation results for the impact of DT on firm performance, along with the mediating effects of innovation capability and cost stickiness. To test H1 and H2, which propose that these factors act as mechanisms linking DT to firm performance, we conducted formal mediation tests. Model 1 serves as the baseline, including only control variables. Model 2 adds DT to assess its overall impact on firm performance (ROA). Models 3 and 4 examine the effect of DT on innovation capability (IC) and its mediating role, while Models 5 and 6 examine DT’s influence on cost stickiness (CS) and its mediation effect. Model 7, the full model, incorporates DT, both mediators, and all controls.
Digital transformation, value creation, and firm performance.
Notes: Non-standardized coefficients are reported; numbers in parentheses are standard errors; one-tailed tests for hypothesized variables and two-tailed tests for control variables; † p < 0.1, *p < 0.05, ⁎⁎p < 0.01, ⁎⁎⁎p < 0.001.
As shown in Table 2, Model 2 confirms a significant positive relationship between DT and ROA (β=0.0037, p < 0.01), providing the basis for testing mediation effects. Model 3 shows that DT positively influences IC (β=0.0699, p < 0.001), while Model 4 shows that IC is positively related to ROA (β=0.0048, p < 0.001). The coefficient of DT in Model 4 (β=0.0033) decreases relative to Model 2 (β=0.0037), suggesting that DT improves ROA through IC, supporting H1. Similarly, Model 5 indicates that DT significantly reduces CS (β=–0.0101, p < 0.01), and Model 6 shows a negative relationship between CS and ROA (β =–0.1048, p < 0.001). The coefficient of DT in Model 6 (β=0.0026) is lower than that in Model 2, suggesting that DT mitigates CS, which subsequently reduces ROA, supporting H2. Finally, Model 7, which includes both mediators, reveals a further weakened positive relationship between DT and ROA (β=0.0023, p < 0.05), reinforcing the mediating roles of IC and CS.
Moderating effects of customer concentration and relative capital intensityModerating effects of customer concentrationTable 3 reports the moderating effects of customer concentration (CC) on the relationship between DT and innovation capability (IC), cost stickiness (CS), and firm performance. Models 8 and 9 examine if CC moderates the impact of DT on IC and CS, respectively. The negative and significant interaction of CC and DT in Model 8 (β =–0.0352, p < 0.01) suggests that CC weakens DT’s positive effect on IC, supporting H3a. By contrast, Model 9 shows an insignificant interaction (β=–0.0040, p > 0.05), indicating that CC does not have the moderating effects of DT on CS; therefore, H3b is not supported. A possible explanation is that the DT-driven cost stickiness reduction may be sufficiently robust to mitigate the risk of high customer dependence. Under DT, firms may actively develop supplementary strategies, such as diversifying product portfolios, expanding into new market segments, or establishing multi-channel sales networks, to mitigate rigidity caused by customer concentration. In addition, DT-enabled data integration and predictive analytics help firms to better anticipate key customers’ demands and adjust resource allocation proactively. These mechanisms may reduce the constraints of customer concentration and neutralize the moderating effect we had originally proposed.
Digital transformation, customer concentration, and firm performance.
Notes: Non-standardized coefficients are reported; numbers in parentheses are standard errors; one-tailed tests for hypothesized variables and two-tailed tests for control variables; *p < 0.05, ⁎⁎p < 0.01, ⁎⁎⁎p < 0.001.
Models 10–13 further test the moderating effects of CC in the DT–ROA link. In Model 10, the interaction term DT × CC is negative and significant (β=–0.0046, p < 0.001), suggesting that CC weakens the positive effect of DT on ROA. The coefficient of DT (β=0.0036) is also slightly lower than in Model 2 (β=0.0037). When IC and CS are introduced as mediators in Models 11 and 12, the interaction effects remain significant (Model 11: β=−0.0038, p < 0.01; Model 12: β=–0.0044, p < 0.001), and DT’s coefficient further decreases (Model 11: β=0.0032, p < 0.01; Model 12: β=0.0025, p < 0.01). This suggests that CC disrupts DT’s value creation mechanisms, especially through IC. Finally, Model 13, which includes both mediators and the interaction item, shows a negative moderation effect (β=–0.0042, p < 0.001), with a lower coefficient than in Model 10, providing further support for H3.
Moderating effect of relative capital intensityTable 4 presents the moderating effects of relative capital intensity (RCI) on the link between DT and innovation capability (IC), cost stickiness (CS), and firm performance (ROA). Models 14 and 15 examine if RCI moderates the impact of DT on IC and CS, respectively. In Model 14, the interaction term DT × RCI is positive and significant (β=0.0471, p < 0.01), showing that RCI strengthens the positive effect of DT on IC. In Model 15, the interaction is negative and significant (β =–0.0062, p < 0.05), indicating that RCI also strengthens the negative effect of DT on CS. These results support H4a and H4b.
Digital transformation, relative capital intensity, and firm performance.
Notes: Non-standardized coefficients are reported; numbers in parentheses are standard errors; one-tailed tests for hypothesized variables and two-tailed tests for control variables; *p < 0.05, ⁎⁎p < 0.01, ⁎⁎⁎p < 0.001.
Models 16–19 further explore RCI’s moderating role in the DT–ROA link. Model 16 shows a positive and significant interaction of DT and RCI (β=0.0040, p < 0.001), suggesting that RCI strengthens DT’s positive effect on ROA. When IC and CS are added as mediators in Models 17 and 18, the interaction terms remain significant but slightly reduced (Model 17: β=0.0030, p < 0.01; Model 18: β=0.0038, p < 0.001). At the same time, the coefficients of DT decrease across these models (Model 16: β=0.0032, p < 0.01; Model 17: β= 0.0032, p < 0.05; Model 18: β=0.0024, p < 0.05), confirming that RCI enhances DT’s value creation mechanisms, particularly through IC and CS. Finally, Model 19, which includes both mediators and the interaction, still shows a positive moderation effect (β=0.0036, p < 0.01), with a slightly smaller magnitude than in Model 16, further supporting H4.
We summarize the hypothesized relationships in Fig. 2, which illustrates the effects of DT on firm performance through innovation capability and cost stickiness, along with the moderating roles of customer concentration and relative capital intensity.
Robustness tests
We conducted additional analyses to assess the robustness of our results and ensure the validity of our findings across diverse specifications. First, in the main regression, we gauged DT by counting the total word frequency related to DT. Although this method has been widely used in recent research focused on DT, it may be influenced by the selection of keywords, and the length of the text in annual reports. To address this concern and ensure robustness, we constructed two alternative measures. First, we followed Yang et al. (2024) and normalized the frequency of DT-related keywords by the length of the management discussion and analysis (MD&A) section, which is widely regarded as reflecting firms’ strategic intentions and implementation priorities. Second, we developed a patent-based proxy by extracting DT-related keywords from firms’ patent application texts, thereby capturing technological efforts and resource commitments in digital initiatives (Li et al., 2024). Compared with disclosure-based measures, this indicator is less subject to managerial rhetoric. The results of our robustness tests are consistent with our main findings, supporting the validity of the impact of DT’s value creation mechanisms on firm performance (H1 and H2). Moreover, to address the potential correlation between IC and CS, we controlled for their lagged values when testing DT’s impact on each mediator, including each mediator as a control when examining its effect on firm performance. The results remain consistent.
To further validate the moderating effects of CC and RCI, we conducted additional robustness checks using alternative measures and subgroup regressions. Specifically, for CC, we replaced the share of sales of the top five customers with the proportion of sales of the largest customers. Firms were then divided into high- and low-CC groups based on whether this proportion exceeded the industry-year average. The results show that the effects of DT on value creation efficiency and firm performance are significant only in the low-CC group. For RCI, we first calculated annual industry-level capital intensity and classified industries as high or low capital-intensive depending on whether their average exceeded the overall cross-industry mean. Subgroup regressions revealed that DT’s impact on value creation mechanisms and firm performance is only significant /more significant in industries with higher capital intensity. These findings further support H3, H4, and H5. All results in relation to robustness checks will be available upon request.
DISCUSSIONTheoretical contributionsThis study explored how DT influences firm performance from the value creation perspective within the VCA model. Empirical results from a dataset of Chinese listed manufacturing firms revealed that DT enhances innovation capability and decreases cost stickiness, two critical determinants of a firm’s value creation efficiency that respectively enhance and constrain firm performance. We further explored how two fundamental contingencies—customer concentration and relative capital intensity—moderate these mechanisms. Empirical results showed that customer concentration weakens the positive effect of DT on innovation capability but does not significantly influence its negative effect on cost stickiness, thereby weakening DT’s overall value creation efficiency and limiting its positive effect on firm performance. We also found that relative capital intensity strengthens both the positive and negative effects of DT on innovation capability and cost stickiness, thereby strengthening DT’s overall value creation efficiency and further improving firm performance. This study makes three key contributions.
First, although extensive research has emphasized the strategic importance of DT for firm performance from the RBV and dynamic capability perspectives, it has mainly explained why DT matters without exploring the specific mechanisms. Building on the value equation in the VCA framework (Lepak et al., 2007), this study verifies theoretically and empirically how DT affects key value creation efficiency determinants through which DT drives superior firm performance.
This study also extends prior research on DT’s value creation effects on firm performance. Existing studies have typically examined efficiency outcomes in isolation (e.g., Gaspar et al., 2024; Li et al., 2022a), overlooking the interdependence among multiple value creation activities that jointly determine firm performance. In contrast, we believe DT’s overall value creation effect is contingent on the underlying mechanisms that shape the value equation. By extracting the key determinants of value creation efficiency from the VCA framework and testing their mediating roles, this study opens the value creation “black box” wherein DT, by enhancing efficiency in value creation processes, is the key to firm performance.
Second, our study advances value creation research as it identifies innovation capability and cost stickiness as the two key determinants of value creation efficiency in the VCA framework. Prior studies emphasize customer-side outcomes, such as personalized services, engagement, and co-creation (Matarazzo et al., 2021), and even within the RBV literature, there is a focus, albeit limited, on innovation protection and competitive barriers (Barney et al., 2021) rather than internal efficiency mechanisms. Leveraging the VCA framework, we identify innovation capability and cost stickiness as two key determinants that jointly drive value creation efficiency. Specifically, greater innovation capability drives customers’ WTP, and lower cost stickiness leads to cost efficiency. Positioning these drivers in the context of DT, this study provides new insights into how value-based strategies shape value creation.
These findings align with prior evidence from both developing and developed economies that DT fosters firms’ innovation capacity by enhancing information processing, facilitating resource integration, and accelerating knowledge recombination (e.g., Chen et al., 2024; Nambisan et al., 2017). Extending the literature beyond this, we identify the role of DT in reducing cost stickiness. Existing studies have primarily examined cost stickiness from the perspective of managerial expectations, resource adjustment costs, or agency problems (e.g., Banker et al., 2014), rarely considering the role of DT. By introducing cost stickiness into DT analysis, this study provides a novel perspective that complements the innovation-focused literature and provides a more systematic understanding of how DT translates into improved value creation efficiency and firm performance.
Third, we enrich the contingent view of strategy implementation in transitional economies, where both customer and resource bases play important roles. Existing research on DT’s boundary conditions emphasizes governance, competition, and institutional factors (Guo et al., 2023; Li, 2022) but has largely overlooked the structural foundations underpinning digital value creation. Fundamentally, value creation relies on identifying target customers and deploying resources to serve them (Lepak et al., 2007). These foundations differ markedly between advanced and developing economies. In developing economies, customer bases are often highly concentrated, while resource bases are labor intensive. By incorporating customer concentration and capital intensity as moderators within a unified value creation framework, this study develops a more nuanced and comprehensive understanding of how DT influences firm performance.
Our findings reveal that customer concentration negatively moderates DT’s value creation effects, aligning with earlier evidence that overreliance on a few customers constrains innovation responsiveness and adaptability (e.g., Zhong et al., 2021). Unlike prior works emphasizing cost efficiency or supply chain stability, this study shows that concentrated structures restrict the scalability and flexibility of digital initiatives. By contrast, relative capital intensity positively moderates DT’s value creation effects, consistent with the evidence that capital-intensive structures support technological upgrading and process innovation (e.g., Chen et al., 2024; Woodard et al., 2013). These findings highlight that structural contingencies shape digital contexts differently from traditional settings. By integrating structural contingencies, we provide a holistic perspective of how DT translates to value creation and improved firm performance.
Practical implicationsThis study offers valuable insights for managers to effectively leverage DT in the digital economy. First, these findings indicate that DT enhances firm performance by simultaneously improving innovation capability and reducing cost stickiness. Managers should, therefore, view DT not only as a technological upgrade but also as a driver of organizational renewal. This requires allocating sufficient resources to foster digital-driven innovation, and embedding digital tools in R&D, product design, and service delivery. Firms can also use digital platforms to monitor costs in real time, identify underutilized resources, and increase operational flexibility, thereby reducing the risk of resource rigidity.
Second, the study offers guidance on tailoring DT strategies to fit firms’ structural characteristics. Specifically, high customer concentration weakens the effects of DT’s value creation efficiency because reliance on a narrow customer base limits firms’ incentives and capacities to exploit DT innovation. To address this, managers should diversify their customer portfolio, expand market exploration, and employ digital tools for better customer segmentation and relationship management. For firms that inevitably operate in markets with high customer concentration, managers can leverage DT to deepen collaboration and co-innovation through shared digital platforms, real-time data exchange, or joint product design. They can also use advanced analytics to anticipate evolving demands, building more agile supply and service systems that mitigate risks from customer concentration.
Third, this study highlights that firms with higher fixed-asset bases can effectively leverage technology upgrades to foster innovation cycles and optimize operations, capturing greater performance benefits from DT. In such firms, managers should proactively integrate Internet of Things and predictive analytics into production to improve equipment utilization, reduce downtime, and enhance overall efficiency. They can also embed AI into capital-intensive processes, such as logistics and resource allocation, to strengthen innovation and efficiency. Firms with lower capital intensity can turn to cloud-based platforms, digital partnerships, and data-driven services to offset their limited asset base and gain from DT adoption.
Limitations and further research directionsNotwithstanding the study’s contributions, several limitations, which indicate avenues for future research, are acknowledged. First, although China’s manufacturing sector provides an ideal setting for examining the performance implications of DT from a value creation perspective, reliance on a single-country geographical context inevitably limits the generalizability of findings. Hence, future studies could expand the empirical scope to other countries and industries to validate the robustness of these results.
Second, although this study adopts textual analysis of annual reports to measure digital transformation and conducts robustness checks using keyword frequency ratios and digital patent, keywords selection remains a challenge. It is difficult to ensure that the chosen keywords fully capture the scope of DT or accurately reflect contextual implications. Future research could enhance measurement validity by integrating textual analysis with alternative data sources, such as surveys or case studies, to provide a comprehensive and nuanced assessment of DT.
Third, despite using lagged independent variables and including extensive controls to mitigate endogeneity, potential endogeneity issues may persist, particularly in the mediation and moderated mediation models, where establishing causal relationships is inherently challenging. Future research could employ rigorous causal inference approaches, such as instrumental variables, propensity score matching, or natural experiments, to validate and extend our findings.
Fourth, while this study identifies and verifies the mechanisms of innovation capability and cost stickiness as key determinants of value creation efficiency within the VCA framework, the value appropriation dimension remains underexplored. As firm performance is jointly shaped by value creation and appropriation, future research could adopt a more integrative perspective to explore how DT reshapes value appropriation mechanisms alongside value creation.
ConclusionThis study investigates the impact of DT on firm performance through value creation within the VCA framework. Evidence from China’s manufacturing firms shows that DT improves performance by enhancing innovation capability and reducing cost rigidity, with these effects contingent on structural factors: customer concentration weakens, whereas relative capital intensity strengthens them. These findings clarify the value creation mechanisms linking DT to firm performance, highlight the role of structural contingencies, and provide practical guidance for managers addressing issues linked to DT, such as innovation, cost management, customer diversification, and resource alignment. Future research could broaden the scope across countries and industries, refine DT measurement, and integrate value appropriation. Overall, this study provides a comprehensive framework for understanding how DT translates into superior value creation efficiency and firm performance.
Declaration of generative AI and AI-assisted technologies in the writing processDuring the preparation of this work the author(s) used ChatGPT (GPT-4) to improve manuscript clarity and grammar. After using this tool/service, the author(s) reviewed and edited the content and take(s) full responsibility for the content of the publication.
FundingThis study was supported by the National Natural Science Foundation of China [72402016], the Humanities and Social Sciences Research of Ministry of Education of China [24YJC630025], and the China Postdoctoral Science Foundation [2024M752742].
CRediT authorship contribution statementLulu Cheng: Writing – original draft, Methodology, Funding acquisition, Data curation, Conceptualization. Nan Mei: Visualization, Project administration, Investigation. Guodong Yi: Writing – review & editing.
