A Study of the Influence Mechanism of Enterprise Innovation Ecosystems on Digital Innovation Capabilities

Abstract

Amid rapid digital transformation in the global economy, digital innovation capability has emerged as a strategic asset for enterprises to cultivate competitive advantages. However, challenges such as the complexity of digital resource integration, limitations of traditional innovation paradigms, and ecosystem fragmentation hinder the development of this capability. Combining innovation ecosystem theory with dynamic capability perspectives, this study analyzed data from 312 Chinese enterprises to examine how enterprise innovation ecosystems shape digital innovation capability. Key findings include the following: (1) the composition, operation, and relationships of enterprise innovation ecosystems exert significant positive effects on digital innovation capability, with the composition dimension demonstrating the most substantial impact.; (2) digital innovation capability comprises three hierarchical components—Digital technological capability (Foundational Capability), Digital technology-enabled Innovation Capability (Evolutionary Capability), and Value reconfiguration capability (Transformational Capability)—all dynamically supported by ecosystem interactions; (3) dynamic capabilities serve as critical mediators between ecosystems and digital innovation capability. Theoretically, this research integrates ecosystem and capability theories to reveal a transformation pathway of “ecosystem embeddedness → dynamic capabilities → digital innovation”. Practically, it offers guidance for resource allocation, ecosystem optimization, and governance, advancing sustainable digital innovation ecosystems.

1. Introduction

Amid rapid digital transformation in the global economy, digital innovation has become a pivotal force reshaping industries and driving economic growth. Leveraging digital innovation has become a decisive pathway for enterprises to navigate the forefront of the digital economy era [1,2,3,4].

Despite development being the trend of development, developing digital innovation capabilities faces persistent systemic challenges. First, the multidimensional attributes of digital resources introduce configuration complexities. In contrast, the technical barriers to digital–physical resource convergence render integration processes intricate and heterogeneous [5]. Second, traditional linear innovation models prove inadequate in adapting to dynamic competitive landscapes, where isolated ecosystem niches frequently impede the market valorization of individual firms’ technological breakthroughs [6]. These constraints collectively hinder progress in enterprise digital innovation.

Innovation ecosystem theory provides a framework for addressing these challenges. From a systemic perspective, enterprises no longer operate as isolated entities within singular industries but function as interconnected nodes spanning multiple industrial ecosystems [7]. Within such ecosystems, firms progressively enhance their capabilities through innovation: they engage in co-opetition to deliver products, fulfil evolving customer demands, and sustain iterative innovation [8]. It is precisely through these dynamic interactions that digital innovation capabilities emerge and mature. Thus, investigating the developmental trajectories of digital innovation capability through the lens of enterprise innovation ecosystems holds substantial significance for guiding strategic transformation, securing novel competitive advantages, and providing actionable operational frameworks.

Research on digital innovation capability has progressed through three phases. Initially, technological determinism dominated, focusing on digital tools’ functional applications [9]. Subsequently, studies shifted to organizational capabilities, emphasizing technology integration and resource orchestration [10]. Recent ecosystem theory highlights the network effects of multi-agent collaborative innovation [11]. Despite these advancements, three critical limitations persist in current scholarship. First, excessive emphasis on technological dimensions has overshadowed the pivotal role of organizational dynamic adaptation. The generative nature of digital innovation necessitates symbiotic co-evolution mechanisms with ecosystem partners [12]. Yet, extant studies inadequately integrate dynamic capability theory. Second, innovation ecosystem research predominantly focuses on structural descriptions, lacking mechanistic analyses of capability transformation pathways and failing to establish substantive dialogue with the ecosystem theory [13].

In summary, this study used empirical research methods to explore how the enterprise innovation ecosystem affects the enterprise’s digital innovation capability. Specifically: firstly, the enterprise innovation ecosystem was deconstructed into three dimensions—composition, operation and relationships—and the digital innovation capability was deconstructed into three dimensions—digital technology capability (Foundational Capability), digital technology-enabled innovation capability (Evolutionary Capability), and value reconfiguration capability (Transformational Capability)—to methodically explore how enterprise innovation ecosystems shape organizational digital innovation capability; secondly, we introduce the dynamic capability as a key intermediary variable to open the “black box” embedded in the enterprise ecosystem to enhance digital innovation capability.

The theoretical contributions of this study are mainly as follows: First, bridging innovation ecosystem theory with the dynamic capabilities perspective, it establishes an analytical framework for digital innovation capabilities that incorporates organizational attributes, a dimension neglected in the existing literature’s predominant focus on technological attributes. Second, it identifies micro-level mechanisms through which enterprise innovation ecosystems (via composition, operation, and relationships) shape digital innovation capabilities, addressing gaps in understanding multi-agent interaction dynamics. Third, it reveals dynamic capabilities as a critical amplification pathway for such ecosystems, offering insights for advancing organizational innovation research, particularly in resolving the “capability rigidity paradox” during digital transformation.

2. Literature Review

2.1. Innovation Ecosystem

When Hannan pioneered the application of ecological perspectives to enterprise activities, laying the groundwork for subsequent innovation ecosystem research [8,12]. This study adopts a micro-level lens, focusing specifically on enterprise innovation ecosystems. Contemporary scholarship has converged on a refined tri-dimensional deconstruction: system architecture, operational mechanisms, and relational networks [8,14,15]. Building upon this analytical progression, this study operationalizes innovation ecosystems through three core dimensions: composition, operation, and relationships.

From a theoretical development perspective, scholars have constructed conceptual frameworks using digital enablement, data factorization, theoretical manifestations, and system governance [16,17,18]. From the capability and evaluation perspective, scholars have explored competencies through digital performance, digital platforms, and value creation [19,20]. However, the following limitations persist: lagging contextualization and China-specific localization studies, existing theories are predominantly based on experiences from developed economies, and their generalizability to latecomer economies like China remains to be validated; underdeveloped micro-level interaction mechanisms, research on the micro-level mechanisms through which inter-enterprise ecosystems enhance digital competitive capabilities remains deficient and requires further exploration.

Existing research demonstrates that enterprise innovation ecosystems provide structural scaffolding and dynamic evolutionary impetus for digital innovation capability development through multi-agent collaborative networks [21,22]. From the resource-based view (RBV) perspective, innovation ecosystems overcome enterprise innovation bottlenecks by aggregating heterogeneous digital resources [23,24]. From the dynamic capability perspective, non-digital-native manufacturing firms enhance resource acquisition and value creation by constructing or joining innovation ecosystems [25]. The co-evolutionary lens further suggests that ecosystem expansion drives capability upgrading. Scholars posit that manufacturers achieve a digital capability leapfrogging through ecological boundary reconfiguration [26]. Collectively, these studies substantiate that enterprise innovation ecosystems systematically enable digital innovation capability development through structural recombination, process reinvention, and relational governance. These conceptual foundations—the resource-based view, dynamic capabilities, and co-evolutionary theory—collectively inform this study’s investigation of ecosystem-driven digital innovation capability mechanisms, as shown in

Table 1. Different theoretical perspective.

Theoretical Perspective	Main Conclusions	Supporting Literature
Resource-based view	Ecosystems overcome innovation bottlenecks through heterogeneous resource aggregation	Wang, 2016 [23]
Dynamic capability	Ecosystems reconfigure architectures to amplify digital innovation.	Zhang, 2023 [25]
Co-evolutionary	Digital capability leapfrogging via boundary reconfiguration	Liu, 2021 [26]

.

2.2. Dynamic Capability

Dynamic capability theory examines the mechanisms through which firms continuously reconfigure resources to address emergent market dynamics with agility [27]. In 1997, Teece introduced the concept of dynamic capabilities, characterizing them as a firm’s capacity to integrate, structure, and rearrange internal and external resources to respond to rapidly evolving environments [28]. With the evolution of dynamic capability theory, the definition and dimensions of dynamic capabilities have also continuously evolved. Existing scholars have primarily defined the conceptual connotation of dynamic capabilities from three perspectives: element-based viewpoint [28,29,30], process-based [31,32,33], and hierarchy-based [30,34,35]. Drawing on Teece et al.’s perspective, this paper measures dynamic ability [36].

Scholars widely acknowledge that dynamic capabilities influence enterprises’ sustained competitive advantages [37,38,39,40]. Building dynamic capabilities facilitates organizational adaptation to rapidly changing environments, enhancing operational efficiency and flexibility [29]. Additionally, dynamic capabilities elevate innovation performance [41,42,43]. While numerous scholars have addressed the question of “how to leverage dynamic capabilities to enhance corporate competitiveness,” research remains scarce regarding empirical exploration at the micro-level of innovation ecosystems concerning enterprises’ digital capabilities.

In the context of rapid environmental transformations, many theories collectively emphasize the openness brought by digital technologies and the complexity of network relationships. The convergence of digital technologies has rendered participating entities and resource flows more intricate [11], with actors forming innovation ecosystems characterized by complex network relationships [44]. Such highly dynamic ecosystems necessitate the development of new dynamic capabilities by enterprises. Existing research indicates that digital-related dynamic capabilities and the ecosystems in which firms operate mutually reinforce each other [45]. In other words, within digital contexts, enterprise innovation ecosystems can enhance dynamic capabilities. For instance, Midea leveraged data from its ecosystem to reveal latent market opportunities, ultimately driving the development of an all-in-one microwave–steam–oven [46]. Dynamic capabilities represent an advanced organizational capacity central to navigating innovation processes and managerial challenges in contemporary digital-economic ecosystems [47,48].

2.3. Digital Innovation Capability

Annarelli first introduced digital innovation capability based on digital technologies, defining it as the innovation capability specifically oriented toward digital technologies themselves [9]. As the concept has evolved, its connotation has been extended. Many researchers emphasize the influence of digital technology attributes on digital innovation capability [10,49,50]. Given that digital technologies possess the “aggregate capacity for spontaneous change driven by vast, diverse, and uncoordinated user communities” [51] and serve as enabling technologies that catalyze new technology-driven products, processes, and services [45], academia generally recognizes digital innovation capability as fundamentally rooted in digital technologies. Furthermore, the characteristics of data homogeneity and reprogrammability in digital technologies allow firms to profit from iterative processes that induce “innovation waves” at relatively low costs [52]. The affordances of digital technologies also enable enterprises to capture value-appropriable conditions from digital innovations [53]. Effectively leveraging digital technologies to enhance innovation capability is critical for enterprises to secure competitive advantages in the digital economy [1]. Amazon Web Services exemplifies this mechanism: Its elastic cloud infrastructure facilitates dynamic resource orchestration, supporting rapid AI deployment across 1M+ global enterprises. This capability directly translates to a competitive market advantage.

While academic discussions have deeply explored the technological attributes of digital innovation capability, they have somewhat overlooked its organizational dimensions [54]. Therefore, the traditional measurement of digital innovation ability only through digital technology is one-sided and single. Moreover, digital innovation empowers innovators to transcend physical spatial constraints, enabling their integration into unified innovation network platforms [55]. With the continuous iteration and upgrading of digital technologies, their high editability and diversified innovation models introduce unprecedented uncertainty into organizational innovation processes [56]. Simultaneously, advancements in digital technologies have expanded the scale and scope of innovation activities as more participants engage, forming broader and more complex innovation networks [54]. This trend profoundly reshapes the organizational structures and boundaries of digital innovation capability [57]. Consequently, compared with traditional innovation, digital innovation has shifted from single-organization dominance to multi-organizational collaboration, with innovation processes and outcomes exhibiting intertwined complexity [11].

As the cornerstone of digital innovation capability, digital technological capability reflects an enterprise’s mastery and application of digital technologies themselves, constituting the foundational stage of all digital innovation endeavors [58]. The robustness of technological infrastructure provides solid support for subsequent innovative practices, ensuring that firms can anchor themselves in the digital transformation wave and lay the groundwork for value creation [59]. Thus, technological accumulation and integration represent merely the initial phase—capability foundation—in digital innovation capability development. Given that digital technologies catalyze more radical innovations [60], their pervasive integration into enterprise operations transforms traditional industrial development models, accompanied by the emergence of upgraded convergent industries and nascent sectors [61]. This evolution underscores the increasingly prominent enabling role of digital technologies, wherein digital-enabled innovation capability forms the core of capability enhancement. The advancement of this enabling capability signifies the progression into the second developmental phase—capability elevation. Ultimately, the maturation of digital innovation capability inevitably triggers transformative shifts in organizational attributes. As digital innovation processes advance, cross-domain collaboration among diverse innovators becomes prevalent, resource accessibility is democratized, and allocation efficiency is markedly improved [62]. Driven by these dynamics, enterprises undergo structural reinvention, ecosystem reconfiguration, and business model leapfrogging, marking the final stage—capability renewal—in digital innovation capability evolution.

In conclusion, this study contends that when deconstructing the dimensions of digital innovation capability, it should not be reduced to a mere reflection of technological advancement but rather a synthesis of organizational adaptability and transformative capacity. This study conceptualizes digital innovation capability as comprising three integral components: digital technological capability (Foundational Capability), digital technology-enabled innovation capability (Evolutionary Capability), and value reconfiguration capability (Transformational Capability). In addressing measurement challenges, such as temporal misalignment of indicators due to rapid digital iterations and potential partiality in traditional scale items, this study leverages methodologically robust frameworks that accommodate dynamic conceptualizations. Detailed methodological support is elaborated in Chapter 4.

The integrated framework diagram of this paper is shown in Figure 1.

3. Research Hypotheses

3.1. The Impact of Enterprise Innovation Ecosystems on Digital Innovation Capability

According to the resource-based view, the sustainable competitive advantage of an enterprise comes from its valuable, scarce, difficult to imitate, and irreplaceable heterogeneous resources and capabilities [43,63]. However, in the context of digital innovation highly dependent on cross-field technology integration and dynamic scene adaptation, the resource endowment of a single enterprise is often difficult to meet the needs of systematic innovation [64]. Enterprise innovation ecosystems greatly expand the strategic resource boundary that a single enterprise can obtain and integrate through multi-subject collaboration. This gives the members of the ecosystem an opportunity to grow. Enterprise innovation ecosystems leverage their core attributes—symbiosis, resource sharing, openness, and collaboration [65,66]—to influence internal relational governance, knowledge governance, and complementary innovation within the ecosystem [67,68,69]. These drivers minimize resource allocation burdens and innovation-related vulnerabilities for digital transformation initiatives, simultaneously boosting performance efficacy. An increasing number of enterprises are enhancing their digital innovation capabilities by constructing or embedding themselves within innovation ecosystems [50,70,71]. Specifically, the composition, operation, and relationships of enterprise innovation ecosystems collectively drive the development of digital innovation capability, as illustrated in Figure 2.

3.1.1. How the Composition of Enterprise Innovation Ecosystems Influences Digital Innovation Capability

The composition of enterprise innovation ecosystems lays the foundation for digital innovation capability through multi-agent collaboration. Led by core enterprises, these ecosystems aggregate diverse actors—including suppliers, research institutions, complementary product manufacturers, and customers—to form an open network characterized by technological complementarity and resource sharing [14,24]. For instance, core enterprises establish platform-based resource pools that integrate global technology developers and hardware suppliers to accelerate digital technology integration, as exemplified by Haier’s industrial internet platform. Research institutions and universities provide cutting-edge technological knowledge in blockchain and IoT, while intermediaries facilitate technology transfer, creating diversified knowledge sources [72]. Concurrently, customers and ecosystem partners engage in product iteration through demand feedback, collectively constructing a “technology R&D–R&D-application implementation” closed-loop.

Thus, this study proposes the following hypotheses and sub-hypotheses:

H1. 

The composition of enterprise innovation ecosystems positively affects digital innovation capability.

H1a. 

The composition of enterprise innovation ecosystems positively affects digital technological capability.

H1b. 

The composition of enterprise innovation ecosystems positively affects digital technology-enabled innovation capability.

H1c. 

The composition of enterprise innovation ecosystems positively affects value reconfiguration capability.

3.1.2. How the Operation of Enterprise Innovation Ecosystems Influences Digital Innovation Capability

The operational mechanisms of enterprise innovation ecosystems drive digital innovation efficiency through dynamic integration and rule design. Core enterprises facilitate resource openness and sharing to reduce technological trial-and-error costs, while employing admission mechanisms and profit distribution rules to select high-quality partners and incentivize sustained engagement [73]. These systems dynamically adapt to technological iterations, enabling rapid market responsiveness through modular collaboration that optimizes digital technology R&D processes. Such mechanism designs transform fragmented resources into systemic innovation momentum, enhancing end-to-end efficiency from digital technology development to commercialization.

Thus, this study proposes the following hypotheses and sub-hypotheses:

H2. 

The operation of enterprise innovation ecosystems positively affects digital innovation capability.

H2a. 

The operation of enterprise innovation ecosystems positively affects digital technological capability.

H2b. 

The operation of enterprise innovation ecosystems positively affects digital technology-enabled innovation capability.

H2c. 

The operation of enterprise innovation ecosystems positively affects value reconfiguration capability.

3.1.3. How the Relationships of Enterprise Innovation Ecosystems Influence Digital Innovation Capability

The relational governance of enterprise innovation ecosystems accelerates the application and implementation of digital technologies through synergistic symbiosis. Core enterprises establish long-term trust relationships based on strategic compatibility, driving technological standardization and ecosystem expansion through intellectual property sharing models [74], as exemplified by Google’s collaborative application development with Android manufacturers. Concurrently, customers engage deeply in demand feedback to co-create value networks. This relational network strengthens multi-dimensional synergies across technology, market, and organizational domains, ensuring rapid scenario adaptation and scalable penetration of digital innovation outcomes.

Therefore, this study proposes the following hypotheses and corresponding sub-hypotheses:

H3. 

The relationships of enterprise innovation ecosystems positively affect digital innovation capability.

H3a. 

The relationships of enterprise innovation ecosystems positively affect digital technological capability.

H3b. 

The relationships of enterprise innovation ecosystems positively affect digital technology-enabled innovation capability.

H3c. 

The relationships of enterprise innovation ecosystems positively affect value reconfiguration capability.

3.2. Mediating Role of Dynamic Capabilities

3.2.1. The Effect of the Composition of Enterprise Innovation Ecosystem on Dynamic Capabilities

From the perspective of compositional dimensions in enterprise innovation ecosystems, multi-agent participation fosters resource complementarity, substantially enhancing technological diversity. Innovation actors systematically shape dynamic capabilities through resource integration and knowledge recombination mechanisms. Specifically, Multi-agent participation (e.g., suppliers, clients, research institutions, governments) establishes heterogeneous knowledge reservoirs. By assimilating differentiated technologies, market intelligence, and managerial expertise from external entities, organizations heighten their environmental sensing acuity [75]—the foundational precursor to dynamic capabilities. Resource complementarity reduces internal resource redundancy. Through modular resource bundling [76], firms gain flexibility in reconfiguring resource allocation strategies, such as transitioning from conventional production to digital service paradigms. Technological diversity expands organizational “technological repertoire”, enabling rapid reconfiguration of technical competencies via combinatorial innovation when confronting technological disruptions. For instance, developing digital twin systems reinforces adaptive advantages within dynamic capabilities [38].

Consequently, this study posits the following hypotheses and sub-hypotheses (as illustrated in Figure 3):

H4. 

The composition of enterprise innovation ecosystems positively affects dynamic capabilities.

H4a. 

The composition of enterprise innovation ecosystems positively affects the enterprise’s sensing capability.

H4b. 

The composition of enterprise innovation ecosystems positively affects the enterprise’s seizing capability.

H4c. 

The composition of enterprise innovation ecosystems positively affects the enterprise’s reconfiguring capability.

3.2.2. The Effect of the Operation of Enterprise Innovation Ecosystems on Dynamic Capabilities

From the perspective of the operational dimension of enterprise innovation ecosystems, collaborative mechanisms, knowledge sharing efficiency, and conflict coordination mechanisms drive the continuous upgrading of dynamic capabilities through organizational learning mechanisms and iterative feedback mechanisms. Specifically, the effective operation of enterprise innovation ecosystems improves production efficiency, creates robust systems, and continuously enhances systemic creativity [14]. The collaborative mechanisms established in enterprise innovation ecosystems effectively reduce transaction costs in cross-organizational cooperation [77], enabling enterprises to rapidly integrate external innovation capabilities and thereby enhance resource integration efficiency [78], which forms the foundation for dynamic capability iteration. Knowledge sharing efficiency directly strengthens enterprises’ absorptive capacity [79], supporting the development of dynamic capabilities. Conflict coordination mechanisms shorten the feedback cycle from “problem identification” to “solution implementation”, thereby accelerating the update frequency of dynamic capabilities [31].

Therefore, this study proposes the following hypotheses and sub-hypotheses (as shown in Figure 3):

H5. 

The operation of enterprise innovation ecosystems positively affects dynamic capabilities.

H5a. 

The operation of enterprise innovation ecosystems positively affects enterprises’ sensing capabilities.

H5b. 

The operation of enterprise innovation ecosystems positively affects enterprises’ seizing capabilities.

H5c. 

The operation of enterprise innovation ecosystems positively affects enterprises’ reconfiguring capabilities.

3.2.3. The Effect of the Relationships of Enterprise Innovation Ecosystems on Dynamic Capabilities

From the relational dimension perspective of enterprise innovation ecosystems, trust levels, network embeddedness depth, and long-term collaboration commitments strengthen the stability and sustainability of dynamic capabilities through relationship-specific investment mechanisms and social capital accumulation mechanisms. Specifically, high trust levels reduce opportunistic behavior risks [80], encouraging enterprises to invest in relationship-specific assets. These assets exhibit high stickiness characteristics and can solidify dynamic capabilities through resource commitment effects [81]. Network embeddedness depth enhances enterprises’ control over critical resources [82]. For instance, firms deeply embedded in cloud computing ecosystems gain priority access to computing power resources, securing structural advantages in digital capability reconfiguration. Long-term collaboration commitments establish “shared future expectations” [83], driving enterprises and partners to co-develop incremental innovation capabilities. This shifts dynamic capabilities from “passive adaptation” to proactively leading transformations [84].

Therefore, this study proposes the following hypotheses and sub-hypotheses (as shown in Figure 3):

H6. 

The relationships in enterprise innovation ecosystems positively affect dynamic capabilities.

H6a. 

The relationships in enterprise innovation ecosystems positively affect enterprises’ sensing capabilities.

H6b. 

The relationships in enterprise innovation ecosystems positively affect enterprises’ seizing capabilities.

H6c. 

The relationships in enterprise innovation ecosystems positively affect enterprises’ reconfiguring capabilities.

Based on long-term academic research, in technology-driven environments, dynamic capabilities drive innovation, efficiency and sustain competitive relevance [27,28], providing theoretical support for enterprises to strengthen their digital innovation capabilities. Digital innovation is a comprehensive management activity based on digital technologies and deeply integrated with enterprise operations, culture, workflows, and other aspects, placing greater emphasis on the innovation process [71]. Furthermore, the main body of digital innovation has shifted from single-organization dominance to collaborative efforts among multiple organizations, while the innovation processes and outcomes exhibit intertwined complexity [11]. It is evident that with the rapid development of digitalization, enterprises must adapt to the openness brought by digital technologies and the complexity of network relationships to improve digital innovation capabilities. Supported by dynamic capabilities, enterprises can continuously monitor external environments and search for and mine information about current mainstream market user demands [85] to capture new market needs and opportunities, thereby laying the foundation for developing digital innovation capabilities. Resource-based theory contends that inimitable, scarce, value-laden, and non-fungible resources constitute the fundamental source of competitive superiority for organizations [86]. Therefore, while interacting with other stakeholders, enterprises need to acquire the technologies and resources essential for digital innovation capabilities. The dynamic capabilities of enterprises facilitate the rapid sharing and coordination of resources [87,88] and enhance the efficiency and performance of digital innovation. Additionally, in dynamic environments, enterprises will struggle to develop without corresponding capability building. Thus, besides obtaining necessary resources internally and externally, enterprises must maximize resource utilization and integration through dynamic capabilities [89]. This promotes the realization of technological and product innovations by enterprises [41], ultimately fostering the formation of digital innovation capabilities.

Therefore, this study proposes the following hypotheses:

H7. 

Dynamic capabilities positively affect enterprises’ digital innovation capabilities.

H8. 

Dynamic capabilities mediate the relationship between the composition of enterprise innovation ecosystems and digital innovation capabilities.

H9. 

Dynamic capabilities mediate the relationship between the operation of enterprise innovation ecosystems and digital innovation capabilities.

H10. 

Dynamic capabilities mediate the relationship between the relationships in enterprise innovation ecosystems and digital innovation capabilities.

Based on the above hypotheses, Figure 4 presents the conceptual model of this study.

4. Research Design

4.1. Sample and Data

This study is based on a sample of Chinese companies. To ensure the relevance to the theme of the questionnaire and the accuracy of the data, this study randomly selected some enterprises to conduct a pre-survey. Given the research objectives and practical constraints, a non-probability sampling approach was employed, specifically combining convenience sampling with targeted invitations. The sample firms are distributed across industries in major cities in eastern, central, and western China to ensure that the sample is sufficiently representative. These firms include wholly foreign-owned enterprises (WFOEs), Sino-foreign joint ventures (JVs), state-owned enterprises (SOEs), and privately owned enterprises (POEs). The industries in which the firms are located include pharmaceutical and medicine industries, machinery manufacturing industries, food manufacturing industries, information technology industries, and financial services industries. Crucially, to enhance sample diversity and target respondents with relevant knowledge, targeted invitations were extended. Researchers and collaborators actively identified and contacted potential respondents (primarily mid-to-senior managers or functional specialists knowledgeable about innovation, digital transformation, and organizational capabilities) in companies fitting the desired profile across different regions, ownership types, industries, and sizes. Respondents were also encouraged to share the survey link with suitable contacts within their professional networks (snowball sampling).

A total of 380 questionnaires were distributed through the QuestionStar platform. This data collection method yielded 358 preliminary responses, resulting in a response rate of 94.2%. After implementing validity screening protocols that excluded 46 non-conforming instruments, a final sample of 312 validated questionnaires was retained for statistical analysis, achieving an effective response rate of 82.1%.

4.2. Measures of Variables

To ensure psychometric robustness, established maturity scales were contextually adapted for this investigation, with full instrumentation detailed in Appendix A (

Table A1. The scale of the main variables.

Variables		Item
the composition of enterprise innovation ecosystems		Q1	The company engages in frequent communication with innovation partners.
		Q2	The company dedicates substantial resources to collaborative initiatives with innovation partners.
		Q3	The company pursues comprehensive multi-project collaborations with innovation partners.
		Q4	The company’s partnerships with innovation stakeholders are mutually beneficial.
the operation of enterprise innovation ecosystems		Q5	Key technical personnel are readily integrated into joint R&D teams.
		Q6	Funding for collaborative technology development is efficiently secured.
		Q7	Partners maintain consistent intent throughout collaborative R&D processes.
the relationships of enterprise innovation ecosystems		Q8	The company provides platforms and related services to enhance interactions with collaborative enterprises.
		Q9	The company facilitates information sharing among partner firms via dedicated platforms and services.
		Q10	The company enables cross-partner knowledge contributions through platform-based infrastructure and support.
Dynamic capabilities	Sensing	Q1	The company possesses the capability to promptly respond to changes in the market environment.
		Q2	The company’s operations can swiftly react to competitors’ strategic actions.
		Q3	The company excels at leveraging business capabilities to capitalize on external opportunities.
	Seizing	Q4	The company has abundant opportunities and demonstrates proficiency in exploiting them.
		Q5	The company is adept at seizing unexpected opportunities when external environments shift.
		Q6	The company maintains vigilant monitoring of market trends, consumer behaviors, and competitor activities, enabling timely implementation of countermeasures.
	Reconfiguring	Q7	The company employs novel organizational norms to modernize work patterns or management frameworks.
		Q8	The company frequently adjusts its organizational structure in alignment with emerging organizational norms.
		Q9	The company consistently updates its business processes to reflect new organizational values.
Digital innovation capability	digital technological capability	Q1	The company holds critical competencies in digital technologies.
		Q2	The company is capable of identifying emerging digital opportunities.
		Q3	The company is equipped to address challenges posed by digital transformation.
		Q4	The company can utilize digital technologies to develop innovative products, services, or processes.
		Q5	The company possesses mastery over cutting-edge digital technologies.
	Digital Technology-Enabled Innovation Capability	Q1	Relative to competitors, the company leverages digital technologies to pioneer novel products/services.
		Q2	Compared to competitors, the company actively introduces new digital technology-driven products into existing markets.
		Q3	Relative to competitors, the company’s newly launched digital products have achieved commercial success.
		Q4	The company outperforms competitors in rapidly exploiting latent digital opportunities within markets.
		Q5	The company effectively utilizes digital technology-enabled distribution channels compared to competitors.
	value reconfiguration capability	Q1	The company delivers novel combinations of products, services, and information.
		Q2	The company proactively integrates new and diverse ecosystem participants.
		Q3	The company is committed to expanding profit models and redefining business boundaries.
		Q4	The company strives to establish innovative transaction mechanisms or operational workflows.
		Q5	The company continuously optimizes resource allocation in response to market information dynamics.
		Q6	The company constructs value co-creation networks through data sharing and platform complementarities.
		Q7	Overall, the company’s business model innovation exhibits novelty and dynamic adaptability.

). Survey items employed a 7-point Likert scale anchored by extreme poles: 1 (Strongly Disagree) → 7 (Strongly Agree), with midpoint 4 (Generally).

4.2.1. Enterprise Innovation Ecosystem

Based on existing research, this study divided the enterprise innovation ecosystem into composition, operation, and relationships. For the composition of enterprise innovation ecosystem, this study refers to the study of Zhang Hui-Qin et al. [90] based on the study of Pan Song-Ting et al. [91], which adopts four question items of linkage strength scale for measurement; for the operation of enterprise innovation ecosystem, Den Hartigh proposes that the health of the core enterprise partners is one of the important indicators for evaluating the degree of the operation of the enterprise innovation ecosystem [92], so this study refers to the studies by Jehn [93] and Chen Zhenhong [94] to use three question items for measurement; for the relationship of enterprise innovation ecosystem, this study refers to the studies by Liu Haixin and Liu Renjing [95] to use three question items for measurement.

4.2.2. Dynamic Capabilities

This paper drew on existing research [96,97,98]. Dynamic ability was classified into sensing, seizing, and reconfiguring. Then, this study utilized nine questions from Xi Yujuan’s study to measure dynamic abilities [99].

4.2.3. Digital Innovation Capabilities

Through synthesizing existing research, this study categorized digital innovation capabilities into three dimensions: digital technological capability, digital technology-enabled innovation capability, and value reconfiguration capability. Digital Technological Capability: Measured using five items adapted from Zhou and Wu’s research [100]. Digital Technology-Enabled Innovation Capability: Operationalized through digital innovation performance, assessed via five items drawn from Pesch R [101] and Tang H [102]. Value Reconfiguration Capability: Evaluated through novel business model innovation capabilities (defined as designing new business models and user experiences to transform existing transaction structures, exchange mechanisms, and economic activities, thereby enabling value capture and value creation mechanisms). This dimension employs six items based on Han Wei and Gao Yu’s study [103].

This study controlled for firm age (years since establishment), nature (wholly foreign-owned, Sino-foreign joint ventures, state-owned, private, other), industry (12 categories including pharmaceuticals, machinery manufacturing, food manufacturing), size (employee count), revenue (previous fiscal year operating income), and location (headquarters region) when assessing innovation ecosystem impacts on digital innovation capabilities.

4.3. Models and Data Analysis Procedure

This research developed an OLS regression framework using empirical survey data to statistically validate the hypothesized relationships. Specifically, this study verifies the positive impact of the three dimensions of enterprise innovation ecosystem on digital innovation capability (H1, H2, and H3) by constructing Equations (1)–(3). Digital Innovation Capability (DIC), Composition (C), Operation (O) and Relationships (R) are the terms used in the equations. The coefficients a1, b1, and c1 represent the extent to which the composition, operation and relationships of the enterprise innovation ecosystem affect digital innovation capability. If a1, b1, and c1 > 0 and statistically significant (p < 0.05), then H1, H2, and H3 hold. In addition, Equations (4)–(6) were constructed to verify the influence of enterprise innovation ecosystems on the three dimensions of digital innovation capability (H1a–H1c, H2a–H2c, and H3a–H3c), where DIC (1–3) represents the three dimensions of digital innovation capability.

DIC = a₀ + a₁ × C + a₂ × controls + ε

(1)

DIC = b₀ + b₁ × O + b₂ × controls + ε

(2)

DIC = c₀ + c₁ × R + c₂ × controls + ε

(3)

DIC_(1–3) = a₀’_(1–3) + a₁’_(1–3) × C + a₂’_(1–3) × controls + ε

(4)

DIC_(1–3) = b₀’_(1–3) + b₁’_(1–3) × O + b₂’_(1–3) × controls + ε

(5)

DIC_(1–3) = c₀’_(1–3) + c₁’_(1–3) × R + c₂’_(1–3) × controls + ε

(6)

To examine the mediating role of dynamic capabilities (DCs) between each enterprise innovation ecosystem dimension (composition, operation, relationships) and digital innovation capability, Equations (7)–(10) were constructed. Results indicated mediation by DC when d1 was significantly positive and e1 < a1 (supporting H6 for composition); d1 was significantly positive and f1 < b1 (supporting H7 for operation); and d1 was significantly positive and g1 < c1 (supporting H8 for relationships).

DIC = d₀ + d₁ × DC + d₂ × controls + ε

(7)

DIC = e₀ + e₁ × C + e₂ × DC + e₃ × controls + ε

(8)

DIC = f₀ + f₁ × O + f₂ × DC + f₃ × controls + ε

(9)

DIC = g₀ + g₁ × R + g₂ × DC + g₃ × controls + ε

(10)

This study developed Equations (11)–(13) to further investigate the effects of the composition, operation, and relationships of enterprise innovation ecosystems on the three dimensions of dynamic capabilities. DC(1–3) represent sensing, seizing, and reconfiguring, respectively.

DC_(1–3) = h_0(1–3) + h_1(1–3) × C + h_2(1–3) × controls + ε

(11)

DC_(1–3) = i_0(1–3) + i_1(1–3) × O + i_2(1–3) × controls + ε

(12)

DC_(1–3) = j_0(1–3) + j_1(1–3) × R + j_2(1–3) × controls + ε

(13)

5. Results

5.1. Descriptive Statistics

This study employed SPSS 22.0 software to conduct descriptive statistical analysis on the survey data (

Table 2. Descriptive statistics of key variables.

Variables	n	Min	Max	Mean	S.D.	Mdn
Firm age	312	1.000	4.000	2.454	1.040	2.000
Firm nature	312	1.000	5.000	2.773	1.442	3.000
Industry background	312	1.000	12.000	6.537	3.724	7.000
Firm size	312	1.000	10.000	5.150	2.977	5.000
Revenue	312	1.000	6.000	3.476	1.888	4.000
Geographic Region	312	1.000	8.000	3.923	2.238	4.000
Composition of enterprise innovation ecosystems	312	1.250	7.000	4.708	1.728	5.000
Operation of enterprise innovation ecosystems	312	1.000	7.000	4.294	1.860	4.330
Relationships of enterprise innovation ecosystems	312	1.000	7.000	4.415	1.837	4.330
Dynamic capabilities	312	1.777	7.000	4.447	1.800	4.220
digital innovation capability	312	1.487	7.000	4.612	1.668	4.297

). All measured constructs demonstrated central tendency values within statistically plausible intervals. For instance, the mean value of enterprise age was 2.454, indicating that the majority of surveyed enterprises had been operational for over six years. These enterprises demonstrated greater maturity and likely possessed more refined organizational structures. The mean value of enterprise ownership types was 2.773, suggesting a predominance of state-owned enterprises and Sino-foreign joint ventures. The standard deviations for all variables exceeded 0.8, signifying that the acquired data hold research validity without stagnation, thereby enabling analysis of variations among model variables.

5.2. Correlation Analysis

This study analyzed correlations between different variables based on Pearson correlation coefficients.

Table 3. The correlations between the variables.

Variables	1	2	3	4	5	6	7	8	9	10	11
Digital innovation capability(1)	1
Firm age(2)	0.103	1
Firm nature(3)	0.028	−0.019	1
Industry background(4)	−0.005	−0.048	0.101	1
Firm size(5)	−0.021	0.052	0.062	0.137 *	1
Revenue(6)	0.104	0.097	0.033	0.116 *	0.11	1
Geographic Region(7)	−0.038	−0.022	0.092	−0.031	0.045	0.109	1
Composition of enterprise innovation ecosystems(8)	0.844 **	0.116 *	0.043	−0.004	0.036	0.179 **	−0.003	1
Operation of enterprise innovation ecosystems(9)	0.872 **	0.095	0.001	0.015	−0.003	0.113 *	−0.047	0.857 **	1
Relationships of enterprise innovation ecosystems(10)	0.880 **	0.064	0	−0.017	−0.009	0.129 *	−0.017	0.864 **	0.906 **	1
Dynamic capabilities(11)	0.909 **	0.078	−0.018	0.043	−0.034	0.093	−0.077	0.826 **	0.888 **	0.880 **	1

* p < 0.05 ** p < 0.01.

presents Pearson correlation results. At the 1% significance level, significant positive correlations were found: between the enterprise innovation ecosystem dimensions (composition: 0.844, operation: 0.872, relationships: 0.880) and digital innovation capability; between these dimensions (composition: 0.826, operation: 0.888, relationships: 0.880) and dynamic capabilities; and between dynamic capabilities and digital innovation capability (0.909). These significant correlations (all p < 0.01) among the core variables provide preliminary support for subsequent hypothesis testing.

5.3. Reliability and Validity

This study assessed the key variables’ reliability through internal consistency.

Table 4. The results of the reliability and validity analysis.

Variables		Item	Factor Loading	Cronbach’s α	AVE	CR
The composition of enterprise innovation ecosystems		Q1	0.942	0.958	0.817	0.958
		Q2	0.929
		Q3	0.901
		Q4	0.916
The operation of enterprise innovation ecosystems		Q5	0.914	0.945	0.852	0.945
		Q6	0.938
		Q7	0.918
The relationships of enterprise innovation ecosystems		Q8	0.929	0.944	0.852	0.945
		Q9	0.906
		Q10	0.933
Dynamic capabilities	Sensing	Q1	0.934	0.953	0.871	0.953
		Q2	0.939
		Q3	0.927
	Seizing	Q4	0.902	0.938	0.839	0.953
		Q5	0.918
		Q6	0.927
	Reconfiguring	Q7	0.913	0.946	0.858	0.948
		Q8	0.942
		Q9	0.922
Digital innovation capability	Digital technological capability	Q1	0.947	0.968	0.858	0.968
		Q2	0.924
		Q3	0.920
		Q4	0.922
		Q5	0.918
	Digital Technology-Enabled Innovation Capability	Q1	0.912	0.959	0.826	0.960
		Q2	0.942
		Q3	0.873
		Q4	0.902
		Q5	0.913
	Value reconfiguration capability	Q1	0.878	0.968	0.817	0.969
		Q2	0.932
		Q3	0.896
		Q4	0.909
		Q5	0.878
		Q6	0.919
		Q7	0.912

confirms that all main measurement scales exceeded Cronbach’s alpha thresholds of 0.7, meeting reliability standards. For validity, convergent and discriminant validity tests were conducted. All variable measurement items demonstrated factor loadings > 0.7, indicating strong alignment with their latent constructs. With average variance extracted (AVE) values > 0.5, the constructs’ explanatory power significantly surpassed random error. Furthermore, the square roots of AVE exceeded inter-variable correlation coefficients, confirming discriminant validity. Collectively, the scales exhibit satisfactory reliability and validity for subsequent empirical analysis.

5.4. Hypothesis Testing

Ordinary Least Squares (OLS) regression analysis was employed to test the hypotheses. In order to avoid the distortion of regression results caused by the inflation of standard error, the instability of coefficient estimation and the complicated interpretation of the influence of each variable on the dependent variable. Prior to conducting regression analysis, we examined potential multicollinearity between independent and control variables. The results showed Variance Inflation Factors (VIFs) below 5, indicating no multicollinearity issues among variables.

Table 5. Regression results of main effects.

	Digital Innovation Capability			Three Dimensions of Digital Innovation Capability
				Digital Technological Capability			Digital Technology-Enabled Innovation Capability			Value Reconfiguration Capability
	Model 1	Model 2	Model 3	Model 4			Model 5			Model 6
	Β	β	β	β	β	β	β	β	β	β	β	β
Firm age	0.017	0.034	0.081	0.009	0.026	0.074	0.015	0.032	0.082	0.027	0.044	0.086
Firm nature	−0.003	0.034	0.036	−0.007	0.032	0.034	0.008	0.047	0.049	−0.012	0.023	0.025
Industry background	0.004	−0.008	0.005	0.005	−0.008	0.006	0.006	−0.007	0.007	0.001	−0.01	0.002
Firm size	−0.026	−0.011	−0.009	−0.034	−0.018	−0.016	−0.024	−0.008	−0.007	−0.020	−0.006	−0.005
Revenue	−0.037	0.006	−0.011	−0.038	0.007	−0.013	−0.042	0.003	−0.013	−0.031	0.01	−0.006
Geographic Region	−0.021	0	−0.017	−0.014	0.008	−0.009	−0.03	−0.008	−0.026	−0.019	0.001	−0.015
Composition of enterprise innovation ecosystems	0.823 **			0.857 **			0.846 **			0.766 **
Operation of enterprise innovation ecosystems		0.779 **			0.818 **			0.803 **			0.717 **
Relationships of enterprise innovation ecosystems			0.798 **			0.847 **			0.805 **			0.741 **
R2	0.718	0.762	0.779	0.669	0.72	0.754	0.696	0.741	0.729	0.655	0.681	0.708
Adjusted R2	0.711	0.756	0.774	0.661	0.714	0.748	0.689	0.735	0.722	0.647	0.681	0.701
F-value	110.802	139.173	153.456	87.958	112.053	133.537	99.679	124.743	117.01	82.798	92.984	105.62

** p < 0.01.

presents the main effect results. Models 1–3, testing the impact of the three enterprise innovation ecosystem dimensions (composition, operation, relationships) on digital innovation capability, were all significant at the 1% level. The coefficients were composition 0.823 (Model 1, supporting H1), operation 0.779 (Model 2, supporting H2), and relationships 0.798 (Model 3, supporting H3). Coefficient analysis further showed that while all three dimensions significantly enhanced digital innovation capability, the composition dimension (0.823) had the strongest effect and operation (0.779) the weakest.

Model 4 confirms that composition, operation, and relationships significantly enhance digital technological capabilities (β = 0.857, 0.818, 0.847; all p < 0.01), supporting H1a, H2a, and H3a. Model 5 shows that all three dimensions positively affect digital technology-enabled innovation capabilities (β = 0.846, 0.803, 0.805; all p < 0.01), validating H1b, H2b, and H3b. Model 6 demonstrates significant positive effects on value reconfiguration capabilities (β = 0.766, 0.717, 0.741; all p < 0.01), confirming H1c, H2c, and H3c.

Analysis reveals that composition exhibits the highest coefficients for digital technological and digital technology-enabled innovation capabilities, indicating its dominant influence. While composition also most strongly affects value reconfiguration capabilities, coefficients decreased significantly overall. Mediating effects are supported (

Table 6. Regression results of the mediating effect.

	Dynamic Capabilities			Digital Innovation Capability
	Model 7	Model 8	Model 9	Model10	Model11	Model12
	β	β	β	β	β	β
Firm age	−0.020	−0.007	0.046	0.030	0.038	0.055
Firm nature	−0.062	−0.022	−0.020	0.035	0.048	0.047
Industry background	0.030	0.017	0.032	−0.015	−0.019	−0.013
Firm size	−0.037	−0.020	−0.019	−0.003	0.001	0.001
Revenue	−0.046	−0.003	−0.019	−0.008	0.008	0.000
Geographic region	−0.049	−0.025	−0.044	0.010	0.016	0.008
Composition of enterprise innovation ecosystems	0.875			0.274 **
Operation of enterprise innovation ecosystems		0.858			0.262 **
Relationships of enterprise innovation ecosystems			0.863			0.314 **
Dynamic capabilities				0.627 **	0.604 **	0.560 **
R2	0.698	0.792	0.784	0.856	0.850	0.858
Adjusted R2	0.691	0.787	0.779	0.852	0.846	0.854
F-value	100.881	166.052	157.941	225.534	214.830	229.429

** p < 0.01.

): Models 7–9 confirm that composition, operation, and relationships each significantly enhanced dynamic capabilities (coefficients: 0.875, 0.858, 0.863; all p < 0.01), supporting H4, H5, and H6. Incorporating dynamic capabilities as a mediator (Model 10) reduced composition’s direct effect on digital innovation capability from 0.823 (p < 0.01) to 0.274 (p < 0.01), confirming its mediating role (H8). Similarly, operation’s effect decreased from 0.779 to 0.262 (Model 11 vs. Model 2), supporting H9. Relationships’ effect decreased from 0.798 to 0.314 (Model 12 vs. Model 3), supporting H10. These results validate the mediating effect of dynamic capabilities across all dimensions.

We used the Bootstrap method in SPSS 22.0 to test the mediation effects (

Table 7. The results of the Bootstrap test.

Item	Mediating Effect	Boot SE	BootLLCI	BootULCI
Composition of enterprise innovation ecosystems=> Dynamic capabilities=> Digital innovation capability	0.549	0.038	0.493	0.645
Operation of enterprise innovation ecosystems=> Dynamic capabilities=> Digital innovation capability	0.518	0.045	0.483	0.663
Relationships of enterprise innovation ecosystems=> Dynamic capabilities=> Digital innovation capability	0.484	0.049	0.435	0.629

). The path mediation effect of Composition → Dynamic Capabilities → Digital Innovation Capability was 0.549 (95% CI: 0.493~0.645). The effect of Operation → Dynamic Capabilities → Digital Innovation Capability was 0.518 (95% CI: 0.483~0.663). The effect of Relationships → Dynamic Capabilities → Digital Innovation Capability was 0.484 (95% CI: 0.435~0.629). None of these confidence intervals contains zero, indicating a significant mediating effect of dynamic capabilities in all three paths. These results further support hypotheses H8, H9, and H10.

This study categorized dynamic capabilities into sensing, seizing, and reconfiguring capabilities (

Table 8. Regression results for different dimensions of dynamic capabilities.

	Dimensions of Dynamic Capabilities
	Sensing			Seizing			Reconfiguring
	Model 13	Model 14	Model 15	Model 16	Model 17	Model 18	Model 19	Model 20	Model 21
	β	β	β	β	β	β	β	β	β
Firm age	−0.032	−0.019	0.035	0.042	0.055	0.107	−0.071	−0.057	−0.002
Firm nature	−0.043	−0.003	−0.001	−0.055	−0.016	−0.013	−0.088	−0.048	−0.045
Industry background	0.014	0.001	0.016	0.047	0.034	0.048	0.030	0.017	0.031 *
Firm size	−0.039	−0.023	−0.021	−0.036	−0.020	−0.018	−0.035	−0.019	−0.017
Revenue	−0.036	0.007	−0.010	−0.058	−0.016	−0.034	−0.043	0.001	−0.015
Geographic Region	−0.053	−0.030	−0.048	−0.062	−0.039	−0.057	−0.032	−0.008	−0.027
Composition of enterprise innovation ecosystems	0.876 **			0.864 **			0.886 **
Operation of enterprise innovation ecosystems		0.862 **			0.846 **			0.866 **
Relationships of enterprise innovation ecosystems			0.867 **			0.861 **			0.862 **
R2	0.642	0.733	0.725	0.658	0.745	0.753	0.660	0.745	0.721
Adjusted R2	0.634	0.727	0.719	0.65	0.739	0.747	0.652	0.739	0.715
F-value	78.196	119.727	115.009	83.94	127.201	132.891	84.510	127.058	112.683

** p < 0.01.

) and verified the impact of enterprise innovation ecosystems on these dimensions. Models 13, 16, and 19 show that Composition significantly enhances sensing, seizing, and reconfiguring capabilities (coefficients: 0.876, 0.864, 0.886; p < 0.01), supporting H4a, H4b, and H4c. Models 14, 17, and 20 indicate that Operation also has significant positive effects (coefficients: 0.862, 0.846, 0.866; p < 0.01), confirming H5a, H5b, and H5c. Similarly, Models 15, 18, and 21 demonstrate that Relationships significantly strengthen all three capabilities (coefficients: 0.867, 0.861, 0.862; p < 0.01), validating H6a, H6b, and H6c.

6. Conclusions

The composition, operation, and relationships of enterprise innovation ecosystems significantly promote digital innovation capabilities. Among these dimensions, composition exerts the strongest influence on digital innovation capabilities. This aligns with the business ecosystem theory proposed by Moore (1993) [8], whereby the integration of multi-agent complementary resources constitutes the fundamental basis of innovation capability. Operation demonstrates the weakest effect, which may be attributed to higher process rigidity and deficient dynamic rule-adjustment capabilities among Chinese enterprises (e.g., lagging profit distribution mechanisms), whereas their Western counterparts prioritize governance agility. This will be the direction of subsequent research.

All three dimensions of digital innovation capabilities—digital technological capabilities, digital technology-enabled innovation capabilities, and value reconfiguration capabilities—are positively influenced by the composition, operation, and relationships of enterprise innovation ecosystems. Specifically, composition exerts the strongest impact across all three dimensions of digital innovation capabilities. Operation shows relative weakness in driving value reconfiguration capabilities. Relationships exhibit influence levels comparable to composition in enhancing digital technological capabilities and digital technology-enabled innovation capabilities.

Dynamic capabilities play positive partial mediating roles in the relationships between the composition of enterprise innovation ecosystems and digital innovation capabilities, the operation of enterprise innovation ecosystems and digital innovation capabilities, and the relationships of enterprise innovation ecosystems and digital innovation capabilities. This verifies the dynamic capability theoretical framework of Teece (1997) [28] and extends the “ higher-order capability “ view of Helfat (2011) [30].

Sensing, seizing, and reconfiguring capabilities are all significantly enhanced by the composition, operation, and relationships of enterprise innovation ecosystems. Notably, composition, operation, and relationships exhibit the most pronounced effects on reconfiguring capabilities. The core influence of reconfiguring capabilities originates from their fundamental role in overcoming resistance to deep organizational change [104]. According to Lewin’s three-stage model of organizational change [105], digital transformation first necessitates unfreezing entrenched legacy systems—including hierarchical structures, obsolete technological paradigms, and inertial cognition [106]. Reconfiguring capabilities serve as the critical lever for executing this unfreezing process. Thereby validating the organizational change axiom: “No Disruption, No Transformation” [107].

6.1. Theoretical Contributions

The theoretical contributions are manifested in three key aspects:

First, it expands the theoretical connotation of digital innovation capabilities. Existing studies predominantly focus on the instrumental attributes of digital technologies while neglecting their organizational dynamism and ecosystem synergies. From a progressive “Foundational Capability → Evolutionary Capability → Transformational Capability” perspective, this study deconstructs digital innovation capabilities into digital technological capabilities, digital technology-enabled innovation capabilities, and value reconfiguration capabilities. This framework reveals that digital innovation capabilities are not merely outcomes of technological accumulation but also manifestations of organizational adaptability and ecosystem coordination.

Second, it unveils the dynamic transformation mechanism between enterprise innovation ecosystems and digital innovation capabilities. Prior literature on innovation ecosystems predominantly remains at the structural description level, lacking in-depth analysis of capability transformation pathways. By introducing dynamic capabilities as mediators, this study constructs a theoretical model of “ecosystem embedding → dynamic capability cultivation → digital capability leapfrogging”, elucidating how multi-agent collaboration, operational mechanism optimization, and relational governance drive digital innovation through the mediating roles of sensing, seizing, and reconfiguring capabilities. This discovery deepens the theoretical dialogue between innovation ecosystems and dynamic capabilities, offering a mechanistic explanation to address the “ecological niche isolation” dilemma.

Third, it advances dynamic and systemic approaches in digital innovation research. Countering the static tendencies in existing studies, this research employs empirical methods to capture the dynamic evolutionary processes of capabilities within ecosystems, validating the differential impacts of ecosystem dimensions on digital innovation capabilities. Findings indicate that the composition dimension provides the strongest foundational support, while the operation dimension may become a bottleneck due to process rigidities. These conclusions offer methodological implications for studying the dynamic evolution of digital innovation capabilities, emphasizing the criticality of systemic element synergy and dynamic adaptation.

6.2. Managerial Implications

Based on empirical findings, this study proposes targeted recommendations for enterprises to enhance digital innovation capabilities through innovation ecosystems:

(1) Prioritize Ecosystem Resource Composition to Strengthen Digital Technological Foundations. Data analysis reveals that the composition dimension exerts the strongest influence on digital innovation capabilities (β = 0.823, ** p < 0.01), particularly on digital technological capabilities (β = 0.857, ** p < 0.01).

Enterprises should actively build or join open innovation platforms and focus on attracting diverse participants (suppliers, universities, research institutes, complementors, lead users) with complementary digital resources and expertise. Examples include establishing industry-specific digital R&D consortia or participating in national/regional innovation platforms like those promoted under China’s “Integration of Innovation Chains, Industrial Chains, Capital Chains, and Talent Chains” strategy.

(2) Optimize Ecosystem Operation Mechanisms to Overcome Process Rigidities. While the operation dimension shows weaker direct effects (β = 0.779, ** p < 0.01), it significantly improves innovation efficiency via dynamic capabilities (β = 0.858, ** p < 0.01). Recommended strategies are the following: Modular division of labor: Adopt Huawei’s 5G ecosystem model by decomposing complex technologies into standardized interfaces to reduce collaboration costs. Implement profit-sharing mechanisms (e.g., Apple’s App Store revenue model) to motivate ecosystem partners. Data-driven iteration: Leverage user analytics for rapid product upgrades, as demonstrated by Midea’s micro-steam-oven development based on youth preference data [41].

(3) Deepen Relational Networks to Enhance Value Reconfiguration Capabilities. The Relationships dimension significantly impacts value reconfiguration (β = 0.741, ** p < 0.01). Enterprises should perform the following:

Conduct strategic compatibility assessments when selecting core ecosystem partners. Prioritize partners with aligned long-term digital visions and complementary assets, fostering trust and commitment akin to the Google–Android standardization collaboration.

Formalize partnerships through strategic investments. Consider cross-shareholding or joint venture arrangements with key technology providers or platform enablers to deepen embeddedness and secure preferential access to critical resources like computing power.

(4) Systematically Cultivate Dynamic Capabilities to Address Technological Uncertainties. Significant mediation effects of dynamic capabilities (effect size: 0.484–0.549, ** p < 0.01) highlight reconfiguring capabilities’ prominence (β = 0.862–0.886). Enterprises should establish dedicated ecosystem intelligence units; task teams with continuously scanning the external environment (market trends, tech disruptions, competitor moves) using insights pooled from diverse ecosystem partners; implement agile organizational structures; and foster internal adaptability through mechanisms like Haier’s “micro-enterprise” model or dedicated cross-functional digital innovation teams empowered for rapid decision-making and resource recombination.

6.3. Limitations

While this study enriches the literature on enterprise innovation ecosystems and digital innovation capabilities, the following limitations warrant further investigation:

(1)

Industry and Regional Constraints of Samples

The generalizability of findings is constrained by the geographic and industrial concentration of the sample. While this design ensures intra-national diversity, it may limit extrapolation to non-Chinese contexts where institutional frameworks (e.g., policy support, market mechanisms) differ significantly. For instance, state-led initiatives like China’s “Broadband China” infrastructure policy uniquely shape ecosystem dynamics, whereas Western ecosystems rely more on market-driven coordination.

To replicate this study in non-Chinese environments, researchers should recalibrate institutional variables, replace China-specific controls (e.g., SOE classification) with locally relevant equivalents (e.g., public/private ownership distinctions in OECD economies), reparameterize ecosystem dimensions, and emphasize service-oriented actors in advanced economies and integrate longitudinal designs to trace capability evolution.

(2)

Cross-Sectional Data Limitations

The cross-sectional design (single-timepoint survey) cannot track long-term ecosystem co-evolution. For instance, Haier’s COSMOPlat platform required years to transition from “resource integration” to “ecosystem expansion”—dynamics unobservable in snapshot data. Future studies should employ multi-wave panel designs tracking 3–5-year capability maturation cycles, or leverage computational methods (e.g., agent-based modeling) to simulate “composition → operation → relationship” synergy pathways.

(3)

Endogeneity and Self-Selection Bias

While this study confirms the positive effects of enterprise innovation ecosystems on digital innovation capabilities, potential endogeneity issues require further attention. This could lead to the overestimation of ecosystem impacts due to reverse causality—where superior capabilities drive ecosystem engagement rather than solely resulting from it. Future research should adopt quasi-experimental designs (e.g., Heckman two-stage models or instrumental variable approaches) to disentangle causation by explicitly modeling the selection mechanism of ecosystem participation.