The Driving Impact of Digital Innovation Ecosystems on Enterprise Digital Transformation: Based on an Interpretable Machine Learning Model

Abstract

This paper is based on data from Chinese digital creative enterprises from 2015 to 2023. A regression model is constructed to test the driving mechanism of the digital innovation ecosystem on the digital transformation of enterprises. The Shapley Additive exPlanations (SHAP) machine learning method is also employed to reveal the key factors driving enterprise digital transformation in the digital innovation ecosystem. The results show that (1) the digital innovation ecosystem can significantly drive the digital transformation of enterprises. And this driving effect is influenced by the moderating effect of the dynamic capabilities of enterprises. The moderating effect of innovation capability and absorptive capacity on enterprise digital transformation is the most significant. (2) The heterogeneity test finds that among state-owned enterprises and enterprises with a high degree of industry competition, the digital innovation ecosystem significantly drives enterprises’ digital transformation more strongly. (3) The results of SHAP value identification indicate that the digital intelligence foundation is the most important factor in driving enterprise digital transformation. The degree of participant diversity, the enabling capacity of the system environment, and the comprehensive benefits play different roles in driving enterprise digital transformation under different heterogeneous conditions.

1. Introduction

At a time when the wave of digitization is sweeping across the world, digital technology has had a profound impact on global economic development. As the primary actors in economic development, enterprises must urgently achieve digital transformation through systematically applying digital technologies and reconstructing business processes, operational models, and organizational structures. This is essential to maintain competitiveness in intense market competition and realize sustainable development (Zhao et al., 2024) [1]. Digital creative enterprises represent a new form of business entity that utilizes digital technologies as core tools, with their primary operations focused on the production, distribution, and consumption of creative content. In the digital era, these technologically powered and creativity-driven enterprises have become key players in new economic models, transforming supply and demand patterns. The digital transformation of digital creative enterprises is not only of great significance to the goal of building a strong cultural country but also directly affects the competitive position of digital creative industries in the global value chain. However, while empowering the digital transformation of digital creative enterprises, digital technologies have also brought about challenges such as pressure from rapid technological updates, difficulties in technology integration and interoperability, and the volatility of user demands. Therefore, digital creative enterprises urgently need to integrate new digital production factors and traditional innovation factors to accelerate the achievement of digital transformation.

Digital innovation is an evolving concept that has seen significant development in recent years. As this concept evolves, the digital innovation ecosystem, a novel form of organization in the digital era, is emerging as a significant catalyst for enterprise digital transformation (Liu and Li, 2025) [2]. The digital innovation ecosystem refers to a complex system in which multiple participants in a specific market, technology, and policy environment jointly promote innovative development through resource sharing and technological cooperation. The objective of the digital innovation ecosystem is to integrate resources and promote open collaboration through digital means, ultimately achieving value creation with the support of the system environment. The digital innovation ecosystem consists of multiple key participants, including core enterprises that possess core technologies and resources, consumers that drive market demand, researchers (such as universities and research institutions) that innovate products and provide technologies, enterprise suppliers that provide supporting resources and services, and government agencies that provide institutional support and policy frameworks. The system environment is a support system composed of digital infrastructure (5G/computing power/data platforms), social policies (subsidies/regulations), and market demand. Value creation refers to the comprehensive benefits achieved by the digital innovation ecosystem, manifested in the output results of the technical, economic, and social value dimensions. It mainly includes short-term, quantifiable economic benefits and long-term ecological resilience (risk resistance/talent reserves). Take Pixar Animation Studio (Pixar) as an example. Pixar is a company specializing in computer animation that is affiliated with Disney. It is a typical representative of digital creative enterprises. Pixar’s digital innovation ecosystem deeply integrates digital technology, art, and business innovation. In terms of participants, the core companies are Pixar Studios (responsible for content creation and IP development) and Disney (derivative product development and financial support). Users include audiences and consumers of derivative products. Researchers include independent creators and universities/training institutions (such as CalArts, which collaborates with Pixar to cultivate animation talent). Suppliers are companies that provide software and hardware and collaborate with Pixar (Emeryville, CA, USA), such as NVIDIA (Santa Clara, CA, USA), which helps optimize rendering technology. In terms of system environment, the Presto animation system provide Pixar with powerful digital technology support. The U.S. Intellectual Property Rights Protection Policy provides institutional guarantees for Pixar to obtain IP revenue. Consumer demand for high-quality 3D animation has expanded Pixar’s user market. In terms of comprehensive benefits, Pixar has not only achieved economic gains through its digital innovation ecosystem but has also driven technological progress and cultural dissemination within the industry. For example, Pixar’s Frozen has grossed over $14 billion at the global box office and generated over $1 billion in annual revenue from merchandise, driving the development of industry standards for 3D animation. It is evident that the digital innovation ecosystem, characterized by multiple participants and a systemic development environment, enhances the comprehensive benefits of enterprises through technological empowerment and collaborative cooperation (Huang and Mao, 2024; Xin et al., 2023) [3,4], thereby providing rich support conditions for the digital transformation of enterprises. Consequently, it is imperative to investigate the utilization of the constituent elements of the digital innovation ecosystem to facilitate enterprise digital transformation. This has important practical significance for further improving the innovation efficiency of digital creative enterprises and building sustainable competitive advantages.

In summary, the present paper will focus on the following questions: The present study seeks to investigate the impact of the digital innovation ecosystem on the digital transformation of digital creative enterprises. Do these influences have heterogeneous characteristics and mechanism effects? How can we identify critical factors in digital innovation ecosystems that influence corporate digital transformation and objectively quantify their relative importance? The contribution of this paper is mainly first, from the perspective of the digital innovation ecosystem, it discloses the moderating mechanism of dynamic capabilities in the digital innovation ecosystem to propel the digital transformation of enterprises. This approach will contribute to the enrichment of the research framework for digital innovation ecosystems within the context of the digital economy. Second, considering the constituent elements of the digital innovation ecosystem, evaluation indicators are constructed based on dimensions such as participant diversity, the degree of empowerment of the system environment, and comprehensive benefits, providing an objective basis for exploring its driving mechanisms for enterprise digital transformation. Third, using the SHAP (Shapley Additive Explanations) value algorithm of machine learning, we identified the key factors driving enterprise digital transformation in the digital innovation ecosystem, breaking through the limitations of traditional linear models. The findings provide actionable guidance for enterprises to optimize digital innovation ecosystems and implement differentiated support measures.

The remainder of this paper is structured as follows: Section 2 presents the literature review and research hypotheses. Section 3 presents the research design. Section 4 presents the empirical tests of the proposed model. Section 5 presents the identification and analysis of influencing factors. Finally, we summarize the conclusions and managerial insights in Section 6.

2. Literature Review and Theoretical Hypotheses

2.1. Literature Review

In the context of the digital economy, digital technology as a new production factor is comprehensively influencing and participating in innovation activities, promoting the improvement of the traditional innovation ecological theory (Li et al., 2022; Nambisan et al., 2017) [5,6]. Digital transformation has expanded the theory of innovation ecosystems, prompting scholars to reflect on digital innovation ecosystems.

Existing studies mainly focus on the connotation characteristics and functional positioning of digital innovation ecosystems in detail. In regard to connotation characteristics, extant research has demonstrated that the digital innovation ecosystem, in the context of digital technology, exhibits not only the dynamic nature of an innovation ecosystem and resource heterogeneity (Chae, 2019) [7]. It also has important characteristics such as participant diversification, resource sharing, and efficient production (Ma et al., 2022; Purbasari et al., 2023) [8,9]. These new characteristics will enhance the synergy between multiple actors, promote the linkage and reorganization of factors, and also lead to fundamental changes in the basic structure of relationships and operating rules in the innovation ecosystem (Benitez et al., 2020) [10]. With respect to functional positioning, digital innovation ecosystems have the capacity to facilitate the optimal allocation of resources and the empowerment of digital technologies by reshaping the behavioral relationships between participating entities (Chae, 2019) [7]. Huang and Mao (2024) introduced energetics into the framework of digital innovation ecosystems [3]. Wang et al. (2024) found that there are coordination dilemmas such as risk sharing and competing stakeholders in digital innovation ecosystems [11]. Digital innovation ecosystems have the potential to promote enterprise digital transformation by accelerating the integration of manufacturing and digital technologies, thereby enhancing the value of products and services [11].

The pervasive implementation of digital technology has precipitated the digital transformation process of enterprises (Hanelt et al., 2021) [12]. In the context of the ongoing digital transformation, the significance of enterprise digital transformation in various fields has become increasingly evident. Different industries have launched in-depth studies on enterprise digital transformation based on their characteristics and needs. Existing research mainly focuses on the connotation characteristics and realization methods of enterprise digital transformation. First, regarding the connotative characteristics of enterprise digital transformation. Browder et al. (2024) suggest that enterprise digital transformation is the process by which enterprises use digital technology to continuously improve their business operations [13]. The goal of transformation is to realize enhanced customer experience, create new business models, etc. According to Berghaus and Back (2016) [14], enterprise digital transformation includes not only process digital transformation focused on efficiency but also digital capability transformation focused on enhancing product innovation. The digital transformation of enterprises is a complex value-enhancing process, specifically going through the development stages of digital conversion, digital upgrading, and digital transformation (Verhoef et al., 2021) [15]. Digital technology has emerged as the primary catalyst for enterprise digital transformation; disruptive innovation is an important condition, and improving productivity and realizing value creation is the purpose of transformation (Zhu and Li, 2023) [16]. Secondly, about the realization of enterprise-based transformation. Digital technology is an important condition to support the digital transformation of enterprises, and digital components, digital platforms, and digital infrastructure construction are the core components of digital technology [6]. Both the cognitive ability and attitude of enterprise organizations are key factors in promoting enterprise digital transformation. Among them, the digital technology application ability, innovation ability, and organizational relationship of the enterprise participating subjects are the driving elements to achieve enterprise digital transformation (Troise, 2022) [17]. The digital economy is characterized by a high degree of uncertainty and ambiguity in the development environment. Enterprises must pay close attention to the volatile environmental changes in technology and markets to effectively realize digital transformation (Chen and Tian, 2022) [18].

In summary, existing studies have conducted preliminary discussions on digital innovation ecosystems and enterprise digital transformation. The research value of the digital innovation ecosystem and enterprise digital transformation is fully affirmed. However, existing research lacks a systematic analysis of the intrinsic mechanisms of digital innovation ecosystems affecting the digital transformation of enterprises. This has resulted in a paucity of effective theoretical guidance and practical basis for enterprises in the process of utilizing digital innovation ecosystems to achieve digital transformation. In particular, there is a dearth of research on the application of digital innovation ecosystems to specific industry contexts. Therefore, the present study takes digital creative enterprises as its research object, conducts a profound analysis of the development level of their digital innovation ecosystem, and reveals the driving path and achievement logic of their constituent elements for enterprise digital transformation. The objective of this study is to provide a scientific foundation and reference point for the systematic promotion of digital transformation in enterprises.

2.2. Theoretical Hypotheses

The digital innovation ecosystem refers to a complex system in which multiple participants in a specific market, technology, and policy environment jointly promote innovative development through resource sharing and technological cooperation. The objective of the digital innovation ecosystem is to integrate resources and promote open collaboration through digital means, ultimately achieving value creation with the support of the system environment. In accordance with the aforementioned definitions and extant research findings [3], the present study principally analyzes the three constituent elements of digital innovation ecosystems: participants, system environment, and comprehensive benefits (see Figure 1). In the context of the digital economy, diverse participants such as core enterprises, researchers, consumers, and producers carry out collaborative cooperation and jointly promote the achievement of digital transformation of enterprises with the support of policy, market, and innovation environment empowerment and comprehensive benefits.

First, the digital innovation ecosystem consists of multiple participants, including core enterprises, researchers, consumers, suppliers, and governments, which are interconnected and collaborate across boundaries to promote the digital transformation of enterprises (Huang et al., 2023) [19]. For example, digital technology providers are an important source of technology for the digital innovation ecosystem. They continue to develop and provide advanced technologies, such as big data analytics, cloud computing, and blockchain, to provide technical support for the digital transformation of digital creative enterprises. Government departments improve the basic conditions for the digital transformation of enterprises by formulating industrial policies, tax incentives, financial subsidies, and other measures (Cai and Yu, 2022) [20]. Digital creative enterprises that possess both innovative capabilities and significant resource advantages are well-positioned to leverage the cluster effects, thereby encouraging and enabling both upstream and downstream enterprises to participate in and achieve digital transformation.

Second, the digital innovation ecosystem can provide the necessary development environment and basic conditions for the digital transformation of enterprises. The openness and sharing of the technological environment can make it easier for enterprises to access and apply new technologies and accelerate the process of digital transformation (Pronchakov et al., 2022) [21]. Changes in the market environment have stimulated the intrinsic motivation of digital creative enterprises for digital transformation, driving them to continuously realize innovative transformation and comprehensive development. A standardized market order has been demonstrated to be conducive to the healthy development of the entire digital innovation ecosystem. It is imperative to provide a stable institutional guarantee for the digital transformation of enterprises.

Finally, the digital innovation ecosystem, by leveraging the synergistic effect of multiple subjects and environmental safeguards, brings multifaceted and comprehensive benefits to the digital transformation of digital creative enterprises and further promotes the sustainable transformation and development of enterprises. Improvements in system effectiveness increase market share and profitability and provide sustained momentum for digital transformation (Li et al., 2023) [22]. In an ecosystem, enterprises can utilize shared infrastructure, market channels, and other resources to reduce unit costs. The realization of economies of scale enhances the development strength of digital creative enterprises. Enterprises have more resources to invest in innovation in the process of digital transformation, which in turn accelerates the realization of transformation. Based on the above analysis, the following research hypotheses are proposed:

Hypothesis 1 (H1).

The digital innovation ecosystems can effectively drive the digital transformation of enterprises.

Dynamic capability refers to the ability of an enterprise to adapt to rapidly changing environments and create new competitive advantages by integrating, building, and reconfiguring internal and external resources. Its core is the dynamic process of enterprises continuously perceiving market opportunities, seizing opportunities, and rapidly transforming to achieve innovation (Teece, 1997) [23]. Innovation capability is defined as the ability of an enterprise to create new products, services, and technologies by integrating internal and external resources. Absorption capacity is defined as the ability of an enterprise to identify, acquire, and digest external knowledge and transform it into its applications. Coordinative capacity is defined as a company’s ability to optimize the combination of existing resources and improve operational efficiency. Dynamic capabilities provide a new perspective for gaining a deeper understanding of the driving mechanisms of digital innovation ecosystems for enterprise digital transformation. Firstly, it is imperative to acknowledge that innovation capability represents a pivotal attribute for enterprises, as it pertains to their capacity for technological innovation and the subsequent implementation of these innovations (Wang and Ahmed, 2007; Crossan and Apaydin, 2010) [24,25]. Enterprises with strong innovation capabilities can proactively explore the application scenarios of new technologies in the digital innovation ecosystem. By applying them to the production and optimization of creative content, they can further accelerate the realization of digital transformation. Second, Strategic coordination capability serves as a critical dynamic capability that enables enterprises to achieve competitive advantage and sustain growth (Wang et al., 2015) [26]. The digital innovation ecosystem is in a constantly changing environment. The ability to coordinate is also an important manifestation of an enterprise’s dynamic capabilities. It enables enterprises to flexibly adjust their strategic decisions under different environmental conditions. Ultimately, it ensures that the enterprise always stays on the right track in the digital transformation process. Third, enterprises with strong learning and absorptive capacities are better able to understand and transform knowledge from other actors in the ecosystem (Schilke, 2014) [27]. Enterprises transform assimilated knowledge and resources into innovation capabilities, thereby enhancing resource conversion efficiency. This process provides critical technological support for accelerating digital transformation.

Hypothesis 2 (H2).

Dynamic capabilities play a moderating role in the process whereby the digital innovation ecosystem drives enterprise digital transformation.

3. Research Design

3.1. Econometric Modeling

Based on the theoretical analysis and the panel data characteristics of the research sample. We choose to construct a fixed effect model to test the driving role of the digital innovation ecosystem on the digital transformation of enterprises. The formula is as follows:

$$D{T}_{it}=\alpha +\beta DI{E}_{it}+\theta X_{it}+{\mu }_i+{\delta }_i+{\epsilon }_{it}$$

(1)

where ${DT}_{it}$ is the explained variable, representing the degree of digital transformation of enterprise i in year t. ${DIE}_{it}$ is the explanatory variable that represents the development level of the digital innovation ecosystem of enterprise i in year t. $X_{it}$ denotes the set of enterprise control variables; ${\mu }_i$ and ${\delta }_i$ represent the enterprise fixed effect and year fixed effect, respectively. ${\epsilon }_{it}$ is the model residual term.

The essence of the moderating effect is the examination of the conditions under which the influence of X on Y is altered. The present study is chiefly concerned with the conditions under which the digital innovation ecosystem exerts its influence on the digital transformation of enterprises. Referring to Gupta et al. (2024) [28], this study uses dynamic capabilities as a moderating variable to examine their moderating role in the impact of the digital innovation ecosystem on the digital transformation process of enterprises. The formula is as follows:

$$D{T}_{it}=\alpha +{\beta }_{1}DI{E}_{it}+{\beta }_{2}D{C}_{it}+{\beta }_{3}DI{E}_{it}\times D{C}_{it}+\theta X_{it}+{\mu }_i+{\delta }_i+{\epsilon }_{it}$$

(2)

where ${DC}_{it}$ is the dynamic capability of the enterprise. ${DIE}_{it}\times {DC}_{it}$ is the interaction term between the digital innovation ecosystem and the dynamic capability. The meanings of the remaining variables are the same as in Equation (1).

3.2. Variable Measurement

3.2.1. Explained Variable

Enterprise Digital Transformation (DT): The digital transformation process of enterprises is characterized by dynamism, complexity, and iteration. Existing studies have mainly used qualitative research (Vial, 2021) [29] and textual analysis methods (Liu et al., 2023) [30] to measure the degree of digital transformation of enterprises. Among them, text analysis of corporate annual reports using web crawler technology is currently a more mature method. The use of the Jieba lexicon can take the natural language environment of Chinese text into more consideration. Therefore, this paper chooses to use the text mining method of machine learning to measure the degree of digital transformation of digital creative enterprises. The process is as follows (Figure 2):

(1)

Annual report text PDF parsing. Using Python 3.5 software, the annual report data of 556 listed Chinese digital creative enterprises from 2015 to 2023 were web-crawled and transformed into text format. Finally, we obtained a total of 5004 annual report samples.

(2)

Establish a thesaurus of keywords. Based on reviewing a large number of policy documents, academic reports, and literature related to the development of digital creative industries. Establish a thesaurus of characteristic words for measuring the digital transformation of digital creative enterprises. Referring to related research (Yildirim, 2025) [31], the thesaurus is constructed in terms of virtual technology, virtual application, and virtual management. Then, according to the content of the thesaurus, the invalid text information in front of the keywords is eliminated and processed by utilizing the text extraction function.

(3)

Text quantization processing. Based on the research of Liu et al. (2018) [32]. According to the constructed feature thesaurus, the selected effective text is subject to participle processing and word frequency statistics. Finally, the term frequency-inverse document frequency (TF-IDF) of the feature words in the text of each enterprise’s annual report is calculated by using the Jieba split-word package. The word frequency statistics are used to represent the digital transformation index of the sample enterprises. The following equation was constructed [32]:

$$T{F}_{ij}={\displaystyle \frac{n_{ij}}{{\displaystyle {\sum }_k{n}_{kj}}}}$$

(3)

$$ID{F}_{ij}=\lg {\displaystyle \frac{\left|D\right|}{1+\left|\left\{j:t_i\in d_i\right\}\right|}}$$

(4)

$$TFIDF={\displaystyle \frac{n_{ij}}{{\displaystyle {\sum }_k{n}_{kj}}}}\cdot \lg {\displaystyle \frac{\left|D\right|}{1+\left|\left\{j:t_i\in d_i\right\}\right|}}$$

(5)

where $n_{ij}$ denotes the frequency of occurrence of word i in text j. ${\sum }_k{n}_{kj}$ denotes the total number of words in the text. $\left|D\right|$ denotes the total number of texts in the corpus lexicon. If a keyword is not in the corpus, i.e., the number of texts containing i is zero, it is replaced by $1+\left|\left\{j:t_i\in d_i\right\}\right|$.

3.2.2. Explanatory Variables

Digital Innovation Ecosystem (DIE): Regarding the quantitative measurement of the digital innovation ecosystems, existing studies mainly construct evaluation indicators based on the components of the system. Based on the theoretical connotation and research hypothesis of the digital innovation ecosystem. It is known that the participants in the digital innovation ecosystem are its core driving force, and their diversity is the basis for evaluation. The system environment is the supporting condition for the digital innovation ecosystem, and digital infrastructure (5G/computing power/data platforms), institutional policies (subsidies/standards/regulations), and market demand together constitute the necessary conditions for the operation of the digital innovation ecosystem. Comprehensive benefits are the core evaluation indicators for assessing the actual effectiveness of the digital innovation ecosystem, reflecting the system’s value-creation capabilities. Therefore, based on research on the constituent elements of innovation ecosystems, this study constructs evaluation indicators from three aspects: diversity of system participants, supportiveness of the system environment, and comprehensive benefits, to comprehensively measure the development level of digital innovation ecosystems. And the entropy method (Li et al., 2025) [33] was used to calculate the indicators. The indicators are shown in

Table 1. Indicators for measuring the development level of the digital innovation ecosystems.

Category	Indicator	Measurement
Diversity of the participants	Technical personnel	$TP=\left({\displaystyle \frac{\mathit{Number} \mathit{of} \mathit{Technical} \mathit{Staff}}{\mathit{Total} \mathit{Employees}}}\right)\times 100\%$
	Internet infrastructure	$\mathit{Count}={\displaystyle \sum \left(\mathit{Company}\text{-}\mathit{Owned} \mathit{Domains} \mathit{in} \mathit{Target} \mathit{Region}\right)}$
	User scale	$CR={\displaystyle {\sum }_{i=1}^N\frac{\mathit{Revenue} \mathit{from} \mathit{Customer} i}{\mathit{Total} \mathit{Revenue}}}\times 100\%$
	Supplier size	$SCR={\displaystyle {\sum }_{i=1}^N\frac{\mathit{Purchases} \mathit{from} \mathit{Supplier} i}{\mathit{Total} \mathit{Purchases}}}\times 100\%$
	Digital platforms	Statistical data published by the big data trading service platform in the enterprise’s location
Supportiveness of the environment	Innovation inputs	Amount of R&D investment = Expense-based investment + capitalized investment
	Digital intelligence foundation	Artificial Intelligence (AI) investment level = AI assets/total asset
	Government support	Government grants to enterprises, including monetary and non-monetary assets.
	Market foundation	$\mathit{Price}\text{-}\mathit{to}\text{-}\mathit{Sales} \mathit{Ratio}=\left({\displaystyle \frac{\mathit{Market} \mathit{Capitalization}}{\mathit{Revenue}}}\right)\times 100\%$
Comprehensive benefits	Risk responsiveness	$\mathit{Debt}\text{-}\mathit{to}\text{-}\mathit{Asset} \mathit{Ratio}=\left({\displaystyle \frac{\mathit{Total} \mathit{Liabilities}}{\mathit{Total} \mathit{Assets}}}\right)\times 100\%$
	Operational capabilities	Operating Revenue = Core Business Revenue + Other Business Revenue
	Growth capacity	Sustainable Growth Rate = Return on Equity × (1 − Dividend Payout Ratio)
	New quality productivity of enterprises	$\mathit{Total} \mathit{factor} \mathit{productivity} \left(\mathit{TFP}\right) \mathit{growth} \mathit{rate}=\Delta InY-\alpha \Delta InK-\beta InL$

.

3.2.3. Moderating Variables

Dynamic Capability (DC): Based on the above theoretical analysis, this paper constructs the evaluation index system of dynamic capability of enterprises from the three dimensions of innovative capability (Ic), absorptive capability (Ac), and coordinating capability (Cc). In particular, innovation capability is measured by the natural logarithm of an enterprise’s total annual patent applications, specifically including the counts of invention patents, utility model patents, and design patents. Absorptive capacity indicates the intensity of an enterprise’s R&D investment and is measured by the ratio of an enterprise’s R&D investment to its operating revenue. Coordination capacity is expressed in terms of an enterprise’s total asset turnover ratio, which is measured by the ratio of an enterprise’s operating income to the closing balance of total assets.

3.2.4. Control Variables

In accordance with extant literature (Han and Wei, 2025) [34], the following control variables were selected: (1) Age of the enterprises (Age): The measurement is determined by the value of the number of years the business has been in operation. (2) Return on Assets (Roe): Roe = (Net profit/Average total assets) × 100%. (3) Enterprise size (Size): Measured by the logarithm of the enterprise’s total assets. (4) Administrative expenses (Mag): Measured by the total amount of line item amounts of the enterprise’s administrative expenses for the year. (5) Capital Liquidity (Mob): Mob = Operating revenue − Cash costs − Income tax. (6) R&D Basic (Basic): Basic = (Total number of R&D personnel/Employees) × 100%.

3.3. Data Selection and Sample Sources

This paper follows the classification criteria for the digital creative enterprise in the National Economic Industry Classification published by the China National Bureau of Statistics [35]. By comparing and matching the information on the main business and business scope of listed enterprises. The listed enterprises of digital creativity in China from 2015 to 2023 are selected from the China Stock Market Accounting Research Database (CSMAR) as research samples. Then, the initial sample data are cleaned and processed as follows: (1) Samples with operating anomalies are eliminated. (2) Eliminate samples with serious missing key financial indicators. Finally, the data of 556 listed digital creative enterprises are selected as research samples. The data for the sample of enterprises are from CSMAR, and the data for regional enterprises are from the China Statistical Yearbook. To mitigate the influence of outliers on model results, all continuous variables were winsorized at the 1% level.

4. Empirical Results

4.1. Results of Descriptive Statistics

Before performing the correlation regression analysis, descriptive statistics were performed on the data of the variables of interest in the study sample. The aim is to get an initial grasp of the basic correlations and statistical characteristics between the analyzed variables.

Table 2. Descriptive statistical results.

Variable	N	Mean	SD	Min	Max
DT	5004	−0.012	0.938	−0.695	3.98
DIE	5004	−0.024	0.829	−1.389	3.236
Age	4857	0.004	0.89	−1.651	2.874
Roe	4870	0.001	0.034	−0.264	0.057
Size	4896	−0.079	0.364	−0.297	3.162
Mag	4898	−0.074	0.313	−0.293	2.696
Mob	4635	0.002	0.693	−2.188	2.54
Basic	4710	−0.074	0.575	−0.434	4.562

presents the results of descriptive statistics for each variable. The report primarily comprises the statistical outcomes of the mean, standard deviation, maximum, and minimum values of the study variables.

Multicollinearity means that there may be a high degree of correlation between the explanatory variables that affect the accuracy of the model estimation, and severe covariance problems may lead to distorted results. Based on this, we performed multiple covariance tests on all sample data.

Table 3. VIF test results.

Variable	VIF	1/VIF
DIE	1.96	0.51
Ic	1.51	0.66
Ac	1.23	0.81
Cc	1.11	0.90
Age	1.02	0.98
Roe	1.00	0.99
Size	5.57	0.18
Mag	5.52	0.18
Mob	1.02	0.98
Basic	2.13	0.47
Mean	2.21

shows the results of the variance inflation factor (VIF) test. The results indicate that the average VIF for all variables is 2.21, and there is no multicollinearity problem between the variables.

4.2. Baseline Regression Results

Table 4. Regression test results.

Variables	DT
	(1)	(2)
DIE	0.137 ***	0.068 ***
	(5.721)	(2.996)
_cons	−0.008 ***	0.014 ***
	(−14.757)	(16.364)
Controls	No	Yes
Enterprise/Year FE	Yes	Yes
N	4537	5004
R2	0.017	0.053

Note: t statistics in parentheses, * p < 0.1, ** p < 0.05, *** p < 0.01.

shows the results of the benchmark regression on the impact of the digital innovation ecosystem on the digital transformation of enterprises. Model (1) shows the results of the test without the inclusion of control variables, and model (2) shows the results of the test with the inclusion of control variables. The results show that all the estimated coefficients of the digital innovation ecosystem are positive. And they are significant at the 1% statistical level (β = 0.137, p < 0.01; β = 0.068, p < 0.01). It shows that the higher the development level of the digital innovation ecosystem, the higher the degree of digital transformation of digital creative enterprises. The theoretical presupposition that the digital innovation ecosystem has a driving effect on the digital transformation of enterprises is verified. Research hypothesis H1 is supported.

4.3. Endogeneity Test and Robustness Tests

To solve the problem of possible bias in the regression results. In this section, we will perform endogeneity treatment and robustness tests on the model. It aims to test the reliability of the regression results.

(1)

Lagged treatment of explanatory variables. The digital transformation of enterprises is a long-term development process, and there may be a lagged effect in the impact that the digital innovation ecosystem has on enterprises. To eliminate potential endogeneity effects and test for lag levels. Based on the existing literature (Luan et al., 2025) [36], the explanatory variable DIE lagged one period (L.DIE) is added to the regression model. Column (1) of

Table 5. Endogeneity and robustness test results.

Variables	DT
	(1)	(2)	(3)	(4)
DIE		0.113 ***	0.063 ***	0.070 **
		(4.021)	(2.713)	(2.579)
L.DIE	0.074 ***
	(3.326)
Age	0.086 ***	0.733 ***	0.161 ***	0.212 ***
	(2.926)	(10.058)	(5.345)	(8.116)
Roe	−0.004 ***	0.014 ***	−0.005 ***	−0.003
	(−4.261)	(2.626)	(−4.312)	(−1.178)
Size	−0.028	−0.080	−0.054	−0.034
	(−0.612)	(−1.098)	(−1.026)	(−0.700)
Mag	0.023	0.047	0.046	0.048 *
	(0.813)	(1.287)	(1.063)	(1.781)
Mob	−0.020 **	−0.015	−0.021 **	−0.016 *
	(−2.240)	(−0.921)	(−2.060)	(−1.783)
Basic	0.100 ***	0.129 **	0.125 **	0.114 **
	(2.860)	(2.307)	(2.394)	(2.556)
_cons	0.026 ***	−0.009 ***	0.024 ***	−0.060
	(22.076)	(−10.719)	(19.549)	(−1.231)
Year FE	Yes	Yes	Yes	Yes
Enterprise FE	Yes	Yes	Yes	Yes
Ind FE	No	No	No	Yes
N	4448	2224	3782	4996
R2	0.022	0.159	0.048	0.093

Note: t statistics in parentheses, * p < 0.1, ** p < 0.05, *** p < 0.01.

reports the results of the test. It can be seen that the estimated coefficient on the lagged digital innovation ecosystem is positive and significant at the 1% statistical level (β = 0.074, p < 0.01). The results show that after controlling for endogeneity, the digital innovation ecosystem still significantly contributes to enterprise digital transformation. Supports that the main findings of the study remain robust and reliable after adopting instrumental variables.

(2)

Excluding exceptional year data. Considering the COVID-19 period, the populations, the environment, and the combined benefits of the digital innovation ecosystem will tend to develop conservatively. During this period, the digital transformation of enterprises will be affected to some extent. Therefore, to ensure the robustness of the test results. We choose to exclude the 2019–2023 data to conduct the empirical test again. Column (2) of Table 5 reports the test results. The results show that the driving effect of the digital innovation ecosystems on enterprises’ digital transformation still exists after excluding the special period data (β = 0.113, p < 0.01), which proves the robustness of the regression results.

(3)

Excluding data from the sample of enterprises in municipalities. Considering that enterprises located in municipalities directly under the central government have the advantage of a digital innovation foundation and resource endowment. The digital transformation process of enterprises in municipalities directly under the central government is more strongly influenced by regional support. Therefore, we exclude the sample of enterprises located in municipalities and then run the regression test again. Column (3) of Table 5 reports the test results. The results show that after excluding the data from the sample of municipality enterprises, the driving effect of the digital innovation ecosystems on enterprises’ digital transformation is positive and significant at the 1% statistical level (β = 0.063, p < 0.01). It indicates that the driving role of the digital innovation ecosystem for enterprises’ digital transformation still exists after excluding the data from the sample of enterprises in municipalities. The robustness of the benchmark regression results is verified.

(4)

Controlling for multidimensional fixed effects. Consider that industry variables that change over time may impact enterprise digital transformation. Therefore, we choose to add control for industry-fixed effects to test the robustness of the results. Column (4) of Table 5 reports the test results. The study shows that after controlling for enterprise-time-industry fixed effects, the estimated coefficient of the digital innovation ecosystem remains positive and statistically significant at the 5% level (β = 0.070, p < 0.05). It shows that controlling for industry fixed effects, the digital innovation ecosystem still significantly facilitates enterprises’ digital transformation. Therefore, it proves that the regression results and conclusions of this study are robust after considering macro factors.

4.4. Heterogeneity Analysis

In this section, we focus on testing whether there is heterogeneity in the driving role of the digital innovation ecosystems on enterprises’ digital transformation. The aim is to test the consistency of the findings.

(1)

Ownership Heterogeneity Analysis. China’s state-owned and non-state-owned enterprises (private or foreign) have significant differences in terms of government support, industry status, and development foundation. These differences may affect the driving role of the digital innovation ecosystems on enterprises’ digital transformation. Therefore, based on the existing literature (Zhang et al., 2022) [37], the research sample of this paper is categorized into state-owned enterprises (SOEs) and non-state-owned enterprises (NSOEs) according to the type of ownership. Columns (1) and (2) of

Table 6. Heterogeneity test results.

Variables	Ownership Heterogeneity		Industry Competition Heterogeneity
	(1) SOEs	(2) NSOEs	(3) HC	(4) LC
DIE	0.132 **	0.105 ***	0.158 ***	0.101 ***
	(2.330)	(4.417)	(5.072)	(2.926)
Age	0.160 **	0.138 ***	0.142 ***	0.179 ***
	(2.530)	(5.113)	(4.305)	(4.251)
Roe	0.001	−0.004 ***	−0.003 ***	−0.002
	(0.161)	(−6.099)	(−2.896)	(−1.267)
Size	−0.069	−0.036	−0.033	−0.084 *
	(−0.698)	(−0.676)	(−0.510)	(−1.867)
Mag	−0.016	0.044	−0.042	0.038 **
	(−0.460)	(0.782)	(−0.543)	(2.410)
Mob	−0.017	−0.026 **	−0.022 *	−0.020
	(−1.158)	(−2.148)	(−1.735)	(−1.413)
Basic	0.266 **	0.103 *	0.099 **	0.239 ***
	(2.121)	(1.833)	(2.201)	(3.154)
_cons	0.025 ***	0.025 ***	0.026 ***	0.027 ***
	(6.854)	(22.112)	(19.094)	(17.280)
Year FE	Yes	Yes	Yes	Yes
Enterprise FE	Yes	Yes	Yes	Yes
N	1365	3172	2256	2273
R2	0.064	0.041	0.083	0.037

Note: t statistics in parentheses, * p < 0.1, ** p < 0.05, *** p < 0.01.

report the test results. The results show that the digital innovation ecosystem of both SOEs and NSOEs has a significant positive contribution to the digital transformation of enterprises. Among them, the positive promotion effect of SOEs is more significant (β = 0.132, p < 0.01). The reason may be that SOEs can get priority access to the government’s digital infrastructure projects, funding, and talent support. These supports can optimize the innovation population, innovation environment, and innovation benefits of the digital innovation ecosystem, improve the system of SOEs, and thus accelerate the drive of the digital transformation of enterprises.

(2)

Industry Competition Heterogeneity. The degree of industry competition is a key factor influencing the efficiency of innovation and change and the innovation environment of enterprises. Within highly competitive industry markets, the frequency of enterprises’ competition is accelerated (Dagnino et al., 2021) [38], which accordingly affects enterprises’ resource allocation and level of digital innovation. Therefore, concerning the existing literature (Kalebe and Gwatidzo, 2025) [39], this paper divides the study sample into high competition (HC) and low competition (LC) groups based on the industry Lerner index. Columns (3) and (4) of Table 6 report the test results. The results show that the digital innovation ecosystems of highly industry-competitive enterprises are more significant in driving enterprises’ digital transformation (β = 0.158, p < 0.01). The reason may be that in a highly competitive industry market, enterprises will actively search for innovation resources due to the survival pressure, and the accumulation speed of enterprises building digital innovation ecosystems is faster. Eventually, enterprises can have a higher digital innovation ecosystem development level in a highly competitive market, which in turn accelerates the drive for digital transformation.

4.5. Moderating Effects Test

In the previous section, we provided evidence to support that the digital innovation ecosystem significantly contributes to the digital transformation of enterprises, and the next step will be to reveal the moderating mechanisms in the impact process. Specifically, this part uses the interaction term between the digital innovation ecosystem and dynamic capabilities as a moderating variable. It verifies the moderating roles of innovation ability, absorptive ability, and coordination ability of enterprises in the process of driving the digital transformation of enterprises by the digital innovation ecosystem.

Referring to the study of Ferreira et al. (2020) [40], the enterprise dynamic capability evaluation index system is constructed from the three dimensions of innovative capability (Ic), absorptive capability (Ac), and coordination capability (Cc). Based on the measurement results of the indicators, use Formula (2) to test the moderating effect of dynamic capabilities. The test results are shown in

Table 7. Regulatory mechanism test results.

Variables	DT
	(1)	(2)	(3)
DIE	0.069 ***	0.061 ***	0.060 ***
	(2.970)	(2.674)	(2.757)
Ic	0.159 *
	(1.799)
Ac		0.116 ***
		(2.641)
Cc
			−0.057 *
DIE × Cc	0.291 **		(−1.828)
	(2.121)
DIE × Ic		0.080 **
		(2.553)
DIE × Ac			0.013
			(0.572)
Age	0.153 ***	0.144 ***	0.134 ***
	(5.760)	(5.461)	(5.339)
Roe	−0.005 ***	−0.004 ***	−0.003 **
	(−4.822)	(−4.104)	(−2.161)
Size	−0.043	−0.043	−0.025
	(−0.903)	(−0.895)	(−0.523)
Mag	0.039	0.036	0.040
	(1.471)	(1.295)	(1.513)
Mob	−0.023 **	−0.024 **	−0.021 **
	(−2.491)	(−2.566)	(−2.392)
Basic	0.126 ***	0.120 ***	0.113 **
	(2.795)	(2.692)	(2.563)
_cons	0.014 ***	0.039 ***	0.012 ***
	(16.423)	(4.321)	(9.351)
Year FE	Yes	Yes	Yes
Enterprise FE	Yes	Yes	Yes
N	4537	4442	4537
R2	0.053	0.054	0.060

Note: t statistics in parentheses, * p < 0.1, ** p < 0.05, *** p < 0.01.

. The results in columns (1) and (2) show that the coefficients of the interaction terms of the digital innovation ecosystem with innovation capacity and absorptive capacity are all positive and significant at the 5% statistical level. It indicates that the innovation capacity and absorptive capacity of enterprises can effectively regulate the driving effect of the digital innovation ecosystem. However, the results in column (3) show a positive but statistically insignificant coefficient for the interaction term of digital innovation ecosystem development level and coordination capacity. The reason may be the developmental characteristics of digital innovation ecosystems and the stage-dependent digital transformation of enterprises. On the one hand, digital innovation ecosystems emphasize openness and knowledge flows, which are a direct match to innovative and absorptive capabilities, while coordination capabilities work better in strongly centralized systems. On the other hand, innovative and absorptive capabilities play an important role in the full cycle of a firm’s digital transformation, while coordination capabilities may be more critical for the maturity of the transformation. In conclusion, the research hypothesis H2 is not fully supported.

5. Impact Factor Identification and Analysis

5.1. Retesting Regression Models Based on Machine Learning Methods

In light of the aforementioned analytical findings, this part will use machine learning methods such as the Random Forest algorithm (Random Forest), Support Vector Machine (SVM), and Extreme Gradient Boosting (XGBoost) to reexamine the model regression results. Based on the optimal modeling algorithm, combined with the Shapley additive explanations (SHAP) methodology, it reveals the key factors of the digital innovation ecosystem’s energetic elements driving the digital transformation of enterprises. SHAP is a widely used method for interpreting the predictions of machine learning models (Younisse et al., 2022) [41]. Among them, the SHAP value is the core element of the SHAP interpretation method. It can measure the degree of contribution of each feature value to the result and visualize and present the contribution value of each feature. The advantage of the SHAP algorithm is that it can evaluate the importance of each feature more accurately, thus providing a scientific explanation for the analysis of the results (Yang et al., 2024) [42]. The objective of this chapter is to identify the factors that effectively drive the digital transformation of enterprises. Therefore, SHAP is chosen to explain the impact of various elements of the digital innovation ecosystem on the digital transformation of enterprises. Figure 3 shows the basic process of the SHAP model. The specific calculation steps are as follows [41]:

(1)

Calculate the SHAP value of feature variable i. Referring to the Lundberg et al. (2020) [43] study, the tree-based interpretable algorithm is used to calculate the SHAP value of each sample feature. The formula is as follows:

$${\phi }_i={\displaystyle \sum _{R=\Re }\frac{1}{M!}\left[f_x\left(P_i^R\cup i\right)-f_x\left(P_i^R\right)\right]}$$

(6)

where ${\phi }_i$ denotes the degree of contribution of the ith sample feature to the label value. R is the set of ordered permutations containing the attributes of the selected sample features. $P_i^R$ is the subset of features listed before i samples in an ordered permutation of the set R. M is the sum of sample features of the model. $f_x$ denotes the conditional expectation function.

(2)

Calculate model output SHAP values for all sample features. This step mainly utilizes the linear combination g(x) consisting of features x to interpret the machine learning results. If g(x) > 0, it indicates that the ith sample feature contributes positively to the labeled-value result. If g(x) < 0, it indicates that the ith sample feature value has a negative impact. The formula is as follows:

$$g(x)=y_{base}+{\displaystyle \sum _{i=1}^M{\phi }_i{x}_i}$$

(7)

where $y_{base}$ represents the conditional expectation before not incorporating the input values. The rest of the variables have the same meaning as in Equation (6).

Substitute the explanatory variables (DIE) and the explained variable (DT) into the above formula. The training and test sets of the models were randomly selected from the full sample in the ratio of 70% and 30%. The optimal parameter tuning objective of each algorithm was determined using the randomized grid search method. Regression analysis is performed on the training and test sets of each model by tuning the optimization settings. Then, the RMSE, MAE, and R² outputs are compared to conclude that XGBoost is the best-performing algorithmic model. Therefore, in the analysis of SHAP values, the interpretation and visual presentation of the importance of feature combinations will be performed with the XGBoost algorithm. It aims to represent more visually the importance of each element of the digital innovation ecosystem in driving the digital transformation of enterprises.

5.2. Based on the SHAP Value Feature Importance Analysis

5.2.1. SHAP Value Interpretation Results

Based on the high-precision regression model established by XGBoost, combined with the evaluation indexes in Table 1. The SHAP model was used to identify the degree of importance of each element of the digital innovation ecosystem in influencing the digital transformation of enterprises. The results are shown in Figure 4. The results show that digital intelligence foundation, internet infrastructure, digital innovation capability, growth capability, and government support are the five most important factors driving enterprise digital transformation. First, the digital intelligence foundation is the most important feature value that drives enterprises’ digital transformation. And it has the strongest positive driving effect on enterprise digital transformation. It shows that AI technology has become a core driver influencing the digital transformation of enterprises through technology penetration. In general, AI technology deeply drives the digital transformation of enterprises at production, management, and service ends through core technologies such as deep learning, machine vision, and natural language processing. The digital intelligence foundation and internet infrastructure can open up the data-knowledge-decision closed loop so that the operational efficiency of enterprises can be significantly improved. Ultimately, it drives enterprises to move from the application of digital tools to the reconstruction of intelligent ecosystems.

Second, in terms of the constituent dimensions of the digital innovation ecosystem level of competence, the system environment, and the comprehensive benefits have a stronger role in driving the digital transformation of enterprises. In terms of the system environment, the digital innovation ecosystem creates a perfect technological environment, institutional environment, and market environment. It can provide enterprises with the necessary technological infrastructure, institutional legitimacy, market demand signals, and cultural cognitive framework to promote the process of enterprise digital transformation. The comprehensive benefits are driven by risk sharing, growth potential, and value creation. Digital technologies can help enterprises build resilience mechanisms (Boh et al., 2023) [44]. The collaborative system of multiple subjects provides a complete risk-sharing mechanism for enterprise development so that enterprises can also quickly build digital solutions (Nambisan et al., 2017) [6] and then realize digital transformation.

5.2.2. SHAP Value of Competitive Heterogeneity in the Industry

Recognizing that ecosystem-driven digital transformation effects differ by industry competitiveness, we assess the relative weights of components in the development of digital innovation ecosystems under different levels of market competition. The results will inform precise, evidence-based interventions. The results are shown in Figure 5. Figure 5a shows the results of identifying the driving factors of high-industry competition enterprises. It can be seen that digital intelligence foundation, digital innovation capability, internet infrastructure, government support, and technical personnel are the top five important factors driving the digital transformation of high-industry competition enterprises. Figure 5b shows the results of identifying the drivers of low-industry competition enterprises. The results show that digital intelligence foundation, internet infrastructure, growth capability, user scale, and digital platforms are the top five important factors driving digital transformation for low-industry competition enterprises. The comparison reveals that the system population and system environment of high-industry competition enterprises are important factors driving the digital transformation of enterprises. For low industry competition enterprises, system population and comprehensive efficiency are important factors driving enterprise digital transformation. The reason may be that in a competitive market environment, enterprises will face stronger threats to survival and the need to optimize efficiency. There is a strong need for enterprises to maintain a competitive advantage by enhancing the operational efficiency of digital innovation and accelerating product iteration (Shahadat et al., 2023) [45]. As a result, the competitive pressure drive and innovation catch-up effect have forced enterprises to realize digital transformation by strengthening main body synergy, improving the system environment, and enhancing comprehensive benefits. However, in low-competition industries, enterprises lack the proactive motivation to innovate driven by market pressure. Their digital transformation tends to rely more on resource pooling at the population level (Hannan and Freeman, 1977) [46]. For instance, enterprises need to leverage industry associations for technology sharing, supply chain collaboration platforms, etc., to enhance digital technology innovation through knowledge spillovers within the population.

5.2.3. SHAP Value of Enterprise Ownership Heterogeneity

Considering the premise of different enterprise natures, there is a differential impact of the digital innovation ecosystem on the driving role of enterprise digital transformation. We will identify the important factors that drive the digital transformation of SOEs and NSOEs based on the heterogeneity of enterprise nature. The results are shown in Figure 6. Figure 6a shows the results of driver identification for SOEs. It can be seen that digital intelligence foundation, internet infrastructure, digital innovation capability, government support, and technical personnel are the top five important factors driving the digital transformation of SOEs. Figure 6b shows the results of identifying the drivers of NSOEs. The results show that digital intelligence foundation, digital innovation capability, user scale, internet infrastructure, and digital platforms are the top five important factors driving the digital transformation of NSOEs. The reason is that the digital innovation ecosystems of enterprises of different natures have different structural characteristics and value logics. The digital transformation of SOEs relies more on government-led ecological synergy (Liu et al., 2024) [47] and focuses more on strengthening digital core technologies and infrastructure scale to match national strategies. Therefore, the system environment and system population are crucial to driving the digital transformation of SOEs. In the case of NSOEs, the digital transformation of NSOEs is more dependent on market-driven ecological competition. Highly uncertain market systems and environments make corporate digital investments more focused on short-term returns (Brynjolfsson et al., 2021) [48]. Therefore, the system environment and overall benefits are even more important to drive the digital transformation of NSOEs. Digital technology and productivity became a priority option for NSOEs. This enables enterprises to respond quickly to changes in market demand and also achieve incremental marginal gains through data assetization to drive their digital transformation.

6. Discussion and Conclusions

This study constructed a regression model to analyze the driving mechanisms of digital innovation ecosystems on enterprise digital transformation. Using machine learning explanatory methods, it comprehensively revealed the important factors that influence enterprise digital transformation in digital innovation ecosystems. The main findings are as follows: (1) The digital innovation ecosystem can significantly drive the digital transformation of enterprises. And this driving effect is influenced by the moderating effect of enterprises’ dynamic capabilities. Among them, innovation capability and absorptive capacity have the most significant moderating effect on enterprise digital transformation. (2) Heterogeneity testing reveals that the digital innovation ecosystem has a differentiated impact on promoting enterprise digital transformation. The effects vary by enterprise ownership and industry competition intensity. SOEs show stronger driving effects than NSOEs. Enterprises in highly competitive industries exhibit greater driving effects than those in less competitive industries. (3) The results of the driver identification show that the digital intelligence foundation is the most important factor in driving the digital transformation of enterprises. System population and system environment are most important for driving digital transformation in SOEs and high-industry competition enterprises. System population and integrated benefits are important factors for driving digital transformation in low-industry competition enterprises. System environment and integrated benefits are important factors for driving digital transformation in NSOEs. In light of the aforementioned conclusions, the following management implications are hereby proposed:

Optimize the population relationship and development environment of the digital innovation ecosystem and improve the digital innovation ecosystem. First, enterprises, research institutions, users, and other diversified subjects are actively encouraged to strengthen cooperation within the ecosystem, build a mechanism for collaborative innovation among industry, academia, and research institutes, and enhance the synergistic effect of the ecosystem as a whole. Second, the government should implement measures to fortify the construction of digital smart infrastructure, promote collaboration between the government and enterprises, and bolster the fundamental conditions for the digital transformation of enterprises. Finally, it should formulate industry digitization standards, cultivate digital innovation incubators, reduce enterprise innovation costs, and provide comprehensive system benefits.

Enhance enterprise dynamic capabilities and strengthen the internal driving force of digital transformation. First, enhance the innovation capacity of enterprises and stimulate the vitality of the digital innovation ecosystem. Enterprises should increase investment in digital technology research and development, creative product development, and other aspects. Such as building an open innovation platform to promote the rapid transformation of scientific research results into enterprise productivity. Second, enhance the absorption capacity and optimize the efficiency of enterprises’ utilization of ecological resources. Enterprises should establish an internal knowledge management system to promote the circulation and dissemination of knowledge within the enterprise so that the enterprise can quickly transform the absorbed external resources into its digital transformation capabilities. Finally, coordination capacity should be strengthened to promote synergistic development within the ecosystem. Enterprises should strengthen the adjustment and optimization of the organizational structure of internal and external ministries, break down the barriers between industries and departments, and stimulate the system-driven role of coordination capacity.

Strengthen key drivers based on enterprise heterogeneity. First, increase investment in new digital infrastructure technologies such as AI, 5G, cloud computing, and the industrial internet. Lower the technological threshold for enterprise digital transformation and give full play to the enabling role of digital technology. Second, SOEs should actively guide related enterprises to gather and develop, form industrial clusters, and drive enterprises upstream and downstream of the industrial chain to realize digital transformation together. For NSOEs, the government can provide more supportive policies to help NSOEs better utilize the resources in the digital innovation ecosystem. Finally, for enterprises in highly competitive industries, it is imperative to strengthen their digital intelligence infrastructure, enhance collaborative relationships with other digital creative entities, promote the sharing and flow of digital innovation resources, and leverage the leading role of frontrunner enterprises. For enterprises operating in less competitive industries, governments should proactively provide financial, talent, and technological support to stimulate their digital innovation vitality, thereby fostering a conducive market environment for digital transformation.

Limitations and Future Research Directions

Although effective progress has been made in this paper, there are still some limitations. In the future, it can be supplemented and improved in the following aspects. First, this study is mainly based on data from enterprises in specific industries of digital creativity, which may not fully reflect the differences in the digital transformation of enterprises in different countries and economic environments. In the future, the sample could be expanded to include more multinational enterprises or emerging economy data. Second, this study examines the unidirectional impact of the digital innovation ecosystem on enterprise digital transformation from an ecosystem perspective yet lacks discussion on their bidirectional causal relationship. Future research could investigate how enterprise digital transformation conversely influences the digital innovation ecosystem, thereby revealing their mutually reinforcing dynamics.