Introduction

The rapid development of digital technology has accelerated the significant upgrade of the manufacturing industry (Liu et al. 2024; Zhu et al. 2023). As a critical component of digital tech, the industrial internet is vital in promoting enterprise performance and competitive advantage (Qi et al. 2023). Digital transformation is a comprehensive process that encompasses various aspects of an organization (Wang & Esperança, 2023), with the industrial internet acting as a key enabler and accelerator. For small and medium-sized enterprises (SMEs), however, challenges such as immature technology, insufficient resources, and ineffective management of digital transformation stages in the corporate strategy exist (Wu & Li, 2024). Therefore, it is essential to investigate the impact of industrial internet adoption on enterprise performance and competitive advantage through various stages of digital transformation (Frank et al. 2024).

Although it has been suggested that digital transformation is not a technical issue but a dynamic and continuous process of organization and management (Prajapati et al. 2024), existing empirical studies have primarily focused on IT-driven transformation and enterprise performance generated by IT (Belhadi et al. 2024; Kumar et al. 2023; Lin, 2022), while few have verified the complete process of digital transformation. Furthermore, the industrial internet is defined as an open global network based on connecting devices, people, and data analytics to facilitate the convergence of the cyber and physical worlds and is a key support for the digital transformation of enterprises (Yin & Zhao, 2024). As an important subject in the application of the industrial internet (Dou et al. 2022), whether the manufacturing industry has the application measures and application modes of the industrial internet becomes the key to the application of the industrial internet (Troise et al. 2022).

Our study aims to offer a new perspective by leveraging the Technology-Organization-Environment (TOE) framework and the Fuzzy Set Qualitative Comparative Analysis (fsQCA) methodology. The TOE framework, extensively applied in technology innovation adoption research (Tornatzky et al. 1990; Van Phuoc, 2022), offers a comprehensive perspective for examining the technological, organizational, and environmental factors influencing various digital transformation stages, not only the adoption stage but also integration, application, and driven stages. This extension is based on our findings that TOE influences the entire digital transformation process, especially in the context of the industrial internet. This understanding enriches existing literature on TOE’s influence on digital transformation (Eze et al. 2020; Shahzad et al. 2023; Wassie & Lakatos, 2024), extending the applicability of the TOE framework to the research environment of digital transformation via the industrial internet (AlShamsi et al. 2022; Deng et al. 2021). Furthermore, the fsQCA methodology, conversely, suits our research by revealing how these factors combine in complex ways to influence enterprises’ digital transformation (Fiss, 2011; Ragin, 2008; Wang & Esperança, 2023). Therefore, this study focuses on:

RQ1: Whether the technological, organizational, and environmental factors affect industrial internet adoption among SMEs in the manufacturing sector?

RQ2: In what ways does industrial internet adoption enhance the performance and competitive advantage of SMEs?

The objectives of this study are: (1) Develop a conceptual framework to reflect the states of digital transformation combining industrial internet and manufacturing; (2) Identify the key factors for SMEs to start the digital transformation with the help of the industrial internet; and (3) Reveal the impact pathways of the industrial internet on the performance and competitiveness of SMEs. After the introduction, the paper reviews relevant literature, proposes hypotheses, describes design and methods, presents results, and draws conclusions.

Literature review and hypotheses

Theoretical framework

Technology-Organization-Environmental framework

The TOE framework, initially proposed by Tornatzky et al. (1990), offers a comprehensive perspective on how technological, organizational, and environmental factors influence an organization’s decision to adopt technological innovations.

Technological factors encompass the IT capabilities of the enterprise and its members, the collection, storage, management, and application of data, and the establishment and application of IT systems (Wassie & Lakatos, 2024). Organizational factors involve management’s awareness and support for digitalization, enterprise culture, acceptance of technical teams, informatization leadership, and organizational digital technology application experience (Elsharnouby et al. 2024). Environmental factors incorporate government support policies and technical standards related to digitization, such as incentive policies encouraging manufacturing enterprises to adopt the industrial internet and data security usage standards (Van Phuoc, 2022).

The general framework of TOE has been widely applied in organizational IT adoption studies (Gökalp et al. 2022). Al Hadwer et al. (2021) found that the technology dimension is a crucial link affecting enterprise cloud adoption using a systematic review of the TOE framework. Eldalabeeh et al. (2021) proposed a validated cloud accounting model studying the influencing factors of cloud computing adoption in the financial sector using TOE. Van Phuoc (2022) proposed strategies for guiding enterprises to prepare for AI application by extracting key factors affecting the adoption of AI technology from three aspects: technology, organization, and environment. Awuah et al. (2022) confirmed that technical, organizational, and environmental readiness drive the adoption of computer-aided technologies by exploring the transformation of enterprise audit processes using the TOE framework. AlShamsi et al. (2022) pointed out that TOE is one of the most common models for studying blockchain adoption and is also mostly used in organizational-level adoption studies of digital technologies.

Innovation diffusion theory

Innovation diffusion theory (IDT) is primarily used to explain how, why, and how quickly different people and organizations adopt new ideas and technologies (Rogers et al. 2014). Many studies based on IDT have explained the diffusion of new ideas/technologies and related behavioral changes in different research contexts, including technology acceptance or rejection, and these different research scenarios influence the adoption of new technologies through different characteristics (Hartley et al. 2022). IDT identifies five stages in the innovation process: understanding, persuasion, decision, implementation, and confirmation. Tetteh et al. (2022) conducted a study on the adoption of digital management systems by local government agencies based on IDT and found that regulation, environment, and organization are positive factors that promote adoption, while technology and social culture are negative factors that hinder it. Salah et al. (2021) integrated IDT and the TOE framework to identify digital management system adoption determinants from three technical, organizational, and environmental perspectives.

The digital transformation paths chain

The digital transformation paths chain was developed through a literature review and a review of related theories (as shown in Fig. 1). The four boxes represent the four stages of the enterprise digital transformation paths chain: adoption, application, driven, and performance (Yin & Zhao, 2024; Wu et al. 2024). The right arrow indicates the link between the four stages, while the downward arrow indicates the rejection/negative reply in the current stage. In the four-stage theoretical framework of digital transformation, this study defines the following four stages according to IDT and the TOE framework (Liu et al. 2024).

Fig. 1
figure 1

Theoretical framework of the digital transformation paths chain.

Full size image

Stage 1: Adoption. In this stage, enterprises form adoption and the origin stage of digital transformation. According to TOE, three aspects that influence adoption are shaped by technology, organization and environment (Prajapati et al. 2024). At this stage, enterprises must face two questions, “Why do you need to carry out digital transformation?” and “Is digital transformation worth it?” After that, the enterprise decides to adopt or avoid a digital transformation strategy, terminating or postponing the transformation process.

Stage 2: Application. Adoption is the action phase after the adoption phase (Schöggl et al. 2024). At this stage, enterprises begin integrating digital technology into the enterprise production and operation processes, starting with the application stage. In the application stage, enterprises still need to overcome the resistance from technology, organization, and environment, requiring significant manpower, material, and financial resources. This stage will also lead enterprises to think about related questions such as “What returns can the enterprise get after digital application?”, “Can it achieve the expected goal after digital application?” If the answer is no, it terminates the process.

Stage 3: Driven. After finishing phased digital application and integration, digital transformation comes to a critical period of harvesting results. For example, at this stage, manufacturing enterprises will carry out whole-process changes in design and development, material procurement, manufacturing, quality control, marketing promotion, and after-sales management. Enterprises also begin to drive their operations and decision-making through digitalization gradually (Elsharnouby et al. 2024). Suppose the enterprise at this stage has a negative answer to the questions of “Whether to continue digital application and integration,” “Whether adherence to digital transformation can be rewarded,” and “Is the input and output proportional.” In that case, it will deny the digital driven and terminate digital transformation.

Stage 4: Performance. In this stage, enterprises use digital technology to build a competitive advantage and extract returns through digital transformation. With the support of digital technology, manufacturing enterprises may achieve profitable returns regarding production process upgrades, product life cycle improvements, flexible and intensive production, efficient supply chain collaboration (Sivan et al. 2022), and collaborative manufacturing implementation, thereby enhancing their competitive advantage and enterprise performance (Wu et al. 2024).

Hypotheses

Industrial internet technology, organization, environment and adoption

Industrial internet adoption (IIA) is a strategic move towards digital transformation, enabling enterprises to engage effectively with the industrial internet (Awuah et al. 2022; Wu & Li, 2024). Understanding that individual innovation does not directly impact enterprise technology adoption behavior is important. Instead, this is influenced by technology readiness, organizational readiness, and environmental factors (Awuah et al. 2022). Among the factors affecting the adoption of emerging technologies, organizational policies, organizational culture, lack of knowledge and management support, and lack of collaboration and coordination can create barriers to the use of digital technologies (AlShamsi et al. 2022). Furthermore, the adoption to consolidate the decision further of digital technologies is hindered by several technological aspects, such as technological immaturity, reluctance to change (Sabatini et al. 2022), interoperability issues, and scalability issues (Kim & Kim, 2021). Combining the three aspects involved in the TOE framework, technological compatibility, comparative advantage, technological complexity, technological capabilities, management capabilities, organizational readiness, government participation, market uncertainty, and supplier partnerships will all significantly affect the use of artificial intelligence.

Industrial internet technology (IIT) evaluates technical dimensions such as technical resources, capabilities, and capital enterprises required to adopt the industrial internet (Xiufan et al. 2024). The technological environment is important for improving organizational effectiveness, efficiency (Sivan et al. 2022), productivity, and product/service quality (Chang, 2020). IT infrastructure provides assurance for enterprise digital application development, information sharing across products/services, business processing, and cross-organizational collaboration (Eldalabeeh et al. 2021; Prajapati et al. 2024; Van Phuoc, 2022). IT technical capabilities play a key role in successful digital transformations, and failed digital transformations are caused by complex technologies and inadequate organizational readiness (Awuah et al. 2022). The interrelationships between these factors lead us to propose:

H1: Industrial internet technology has a positive effect on industrial internet adoption.

Industrial internet organization (IIO) embodies the internal evaluation of organizational dimensions such as attitudes, readiness, and experience for adopting the industrial internet (Frank et al. 2024). Organizational characteristics refer to organizational factors that influence innovation adoption decisions and are enablers for the use and success of digital management systems (Tetteh et al. 2022). Successful innovation implementation was found to be significantly related to organizational and environmental contexts (Ahmad et al. 2022). Top management support is key in cloud service adoption, especially in guiding resource allocation, service integration, and process reengineering (Al Hadwer et al. 2021). Top management needs to create a buffer period between new computer-aided systems and current business practices, and a lack of top management commitment becomes a major obstacle to the successful implementation of the system (Awuah et al. 2022). Recognizing these interrelationships, the study hypothesizes:

H2: Industrial internet organization have a positive effect on industrial internet adoption.

The industrial internet environment (IIE) evaluates environmental dimensions such as government support policies, technical standards, and partnerships related to enterprises’ adoption of the industrial internet (Kim & Kim, 2021; Malik et al. 2021). Environmental dimensions also determine the adoption of digital technologies by businesses, among which are competitive pressures, partners, and government incentives (Gökalp et al. 2022). Firms must understand customer needs and requirements to win customer satisfaction, and digital management systems can track customers in real-time and provide timely feedback and personalized recommendations based on customer characteristics (Salah et al. 2021). Competition, customer, and industry pressures that lead business leaders to adopt appropriate competitive strategies to gain maximum market share also force firms to adopt digital technologies (Kareem et al. 2024). Considering these interrelationships, the study hypothesizes:

H3: Industrial internet environment has a positive effect on industrial internet adoption.

Industrial internet adoption, integration, application and driven

Industrial internet integration (III) represents the successful assimilation of industrial internet technologies into an enterprise’s existing ICT infrastructure, encompassing both internal and external integration processes (Chen & Chen, 2021; Lin, 2022). This integration is not merely a technical endeavor but a strategic imperative for organizations seeking to fully leverage the transformative potential of the industrial internet (Vial, 2019). The adoption of the industrial internet, signifying a strategic decision to embrace digital transformation, serves as a critical precursor to effective integration (Oliveira & Martins, 2011; Baker, 2012). Industrial internet adoption serves as a catalyst, enabling enterprises to reimagine and reconfigure their resource allocation and utilization patterns (Wu & Li, 2024). The adoption of industrial internet technologies is not merely a technical upgrade but a strategic imperative that lays the groundwork for comprehensive digital transformation (Frank et al. 2024). It provides the necessary technological foundation and organizational readiness for subsequent integration efforts. The adoption stage is particularly crucial for manufacturing enterprises, as it sets the stage for the seamless incorporation of industrial internet capabilities into their core operations (Yin & Zhao, 2024). This decision to adopt often precedes the allocation of resources and the development of capabilities necessary for successful integration (Wang et al. 2016). The integration of industrial internet technologies necessitates a shift in organizational processes, workflows, and potentially even business models, all of which stem from the initial adoption decision (Verhoef et al. 2021). SMEs that strategically adopt digital technologies are more likely to successfully integrate these technologies into their core business processes, leading to enhanced operational efficiency (Elsharnouby et al. 2024). Similarly, the initial adoption of digital technologies serves as a critical enabler for their subsequent integration into circular economy practices within manufacturing firms (Schöggl et al. 2024). Recognizing this inherent dependency and the foundational role of adoption in enabling integration, we posit:

H4: Industrial internet adoption has a positive effect on industrial internet integration.

Industrial internet application (IIAP) encompasses the business operations by which manufacturing enterprises apply the industrial internet in R&D, production, sales, after-sales, and others (Belhadi et al. 2024). The infrastructure and technical resources in the digital environment are the antecedents of enterprise digitalization activities (Salah et al. 2021). The application of the industrial internet is also inseparable from IT hardware, software, and networks. Therefore, when enterprises implement the industrial internet, they should include it in their top-level strategy, adopt it, and prepare for the basic conditions required for deploying it (Hartley et al. 2022). Adopting advanced digital technologies, building data analysis capabilities, and effective governance mechanisms can enhance enterprises’ digital innovation and application capabilities (Chatterjee et al. 2024). Adopting the industrial internet can facilitate access to financial resources (Hasani et al. 2017). Only when a consensus can be formed on adopting the industrial internet strategy from senior managers to grass-roots users within the enterprise can the application of the industrial internet be successfully promoted in enterprises (Van Phuoc, 2022). However, these assumptions have not been fully tested in SMEs’ adoption of the industrial internet (Jones et al. 2021; Wang & Esperança, 2023). Therefore, we consider it necessary to verify the applicability of these hypotheses in the specific context of this study. Considering these interconnected variables, the study hypothesizes:

H5: Industrial internet adoption has a positive effect on industrial internet application.

The industrial internet driven (IID) enterprise development process can systematically collect, organize, analyze, and generate decision-making suggestions for manufacturing enterprises, thereby improving the overall efficiency and quality of the enterprise (Chen et al. 2022; Haftor et al. 2021). Industrial internet integration can improve processes and practices and drive enterprises to adapt to a highly uncertain and rapidly changing business environment (Kim & Kim, 2021). Industrial internet integration plays a central role in the acquisition, release, and integration of enterprise resources and is a core factor in promoting and improving the driving effect of the industrial internet (Eldalabeeh et al. 2021). The internal integration of the industrial internet includes forming a cross-departmental collaboration system within the enterprise for resources, production, and processes to fulfill customer needs. In contrast, external integration is the close collaboration between the enterprise and its partners to support enterprise development (Chen & Chen, 2021). For the industrial internet to generate value, enterprises must adopt the industrial internet to use it in the processes that drive the enterprise’s growth (Reis et al. 2020). SMEs can take advantage of digital transformation opportunities to enhance their competitiveness (Wang & Esperança, 2023). Drawing from the literature above review, this study posits:

H6: Industrial internet integration has a positive effect on industrial internet driven.

H7: Industrial internet application have a positive effect on industrial internet driven.

Industrial internet driven, enterprise competitive advantage and performance

Enterprise performance (EP) is an evaluation index to measure an enterprise’s operational efficiency and results (Ali, 2023; Kalaitzi & Tsolakis, 2022). Adopting new platforms and systems based on digital technologies can bring many benefits to improved business performance (Salah et al. 2021). The industrial internet can collect a large amount of data in the production process of manufacturing enterprises in real-time, saving many labor costs (Salah et al. 2021). In addition, enterprises can generate reports required for business operations based on the industrial internet and make decision-making suggestions for enterprises based on automated analysis (AlShamsi et al. 2022). Manufacturing enterprises implementing digital transformation with the help of the industrial internet can accumulate digital assets, which can help enterprises analyze how to improve financial and innovation performance (Deng et al. 2021). The study hypothesizes:

H8: Industrial internet driven has a positive effect on enterprise performance.

Enterprise competitive advantage (ECD) is a comprehensive reflection of an enterprise’s market and non-market competitive advantage in social development (Chatterjee et al. 2024; Rahman et al. 2022). The industrial internet, based on digital technology and a large amount of data, is an important driving force for enterprises to quickly update products, which also helps enterprises to improve product development efficiency and potential commercial value (Kim & Kim, 2021; Peng et al. 2021). Companies adopting digital strategies will also tend to digitally drive business operations to improve performance, competitiveness, and value (Chen et al. 2022). Industrial internet-driven enterprises can improve performance and maximize competitive advantage (Chatterjee et al. 2024). Drawing from the interrelationships observed in the literature review, we posit:

H9: Industrial internet driven has a positive effect on enterprise competitive advantage.

When an enterprise conducts digital transformation to achieve higher enterprise performance, it can also bring greater competitive advantage to the enterprise (Kumar et al. 2023). Digital transformation opportunities can create opportunities for enterprises to re-improve their processes and products, improving their overall performance and value (Chang, 2020). The industrial internet is an important driving force for transforming the manufacturing industry. Through the application of the industrial internet, enterprises will move from machine manufacturing to digital manufacturing to realize the transition of enterprise value (David et al. 2022). After integrating the industrial internet, enterprise management becomes more efficient and evidence-based, and improving enterprise management capabilities also increases competitive advantage (Chen & Chen, 2021). In light of the interrelationships observed among the variables, the study hypothesizes:

H10: Enterprise performance has a positive effect on enterprise competitive advantage.

A hypothetical model of the four phases of enterprise digital transformation (Fig. 2).

Fig. 2
figure 2

Research model of the digital transformation paths chain based on the industrial internet.

Full size image

Methods

Instrument

Data were collected through a questionnaire divided into two parts. The first contains the measurement scale of the four-stage research model of enterprise digital transformation, and the questionnaire includes a marker variable (perceived habit) to test the common variance of the research model. The content of the first part contains a total of 9 latent variables, a total of 39 items; for the scale items, the Likert 5-level scale was measured (values 1–5 indicate strongly disagree to strongly agree, respectively). The second part is the basic statistical information of the interviewed enterprises.

To ensure the content validity, all measurement items were adopted or contextualized from published literature (Table 1). Industrial internet technology was adapted from Pinheiro et al. (2022) and Na et al. (2022); Industrial internet organization was adapted from Chatterjee et al. (2024), Zhang et al. (2021), and Malik et al. (2021); Industrial internet environment was adapted from Malik et al. (2021), Kalaitzi and Tsolakis (2022); Industrial internet adoption was adapted from Zhang et al. (2021), and Fonseka et al. (2022); Industrial internet integration was adapted from Chen and Chen (2021), and Lin (2022); Industrial internet application was adapted from Bag et al. (2021), and Belhadi et al. (2024); Industrial internet driven adapted from Haftor et al. (2021), and Chen et al. (2022); Enterprise competitive advantage adapted from Chatterjee et al. (2024), and Rahman et al. (2022); Enterprise performance was adapted from references in Kumar et al. (2023), and Ali (2023); perceived habit (marker variable) was adapted from Zha et al. (2013).

Table 1 Instrument.
Full size table

Data collection and sample

After the initial completion of the questionnaire based on pertinent literature, a meticulous pre-test was implemented to ensure the questionnaire’s clarity and applicability (Eldalabeeh et al. 2021; Van Phuoc, 2022). Before the pre-testing, an expert panel of five scholars and entrepreneurs was invited to review the questionnaire. They provided valuable suggestions on aspects like the clarity of specific terminologies such as “Industrial internet driven”, potential ambiguities in the phrasing of certain variables, and the arrangement of variables. Accordingly, we made the necessary revisions:

  1. 1.

    Definitions for key terms like “Industrial internet driven” were added in the Variables section for better comprehension.

  2. 2.

    Adjustments were made in phrasing six variables to express their connotations accurately.

  3. 3.

    The collection of enterprise characteristic data was placed at the end of the scale, serving as control variables.

Subsequently, an online questionnaire was disseminated among 30 managers from various manufacturing industries for a preliminary survey. The feedback indicated that the respondents clearly understood the scale variables, and the questionnaire was smoothly filled out, demonstrating good validity (Cronbach’s alpha >0.7). Afterward, a formal large-scale survey was carried out.

The purpose sampling technique was adopted in this study to select enterprises from China, to improve the theory of the research results and the flow of the data collection process (Chaudhuri et al. 2021). Since the government of China issued the call to promote enterprise digitalization in 2021, it has successfully carried out many publicity and public welfare training activities, and even awarded 15% of investment funds to enterprises that carry out enterprise digitalization transformation. A series of measures by the local government has made many achievements in enhancing the intentions and effectiveness of enterprises in digital transformation (He et al. 2022).

Questionnaire Star produced and distributed the questionnaires online among the entrepreneurs association and the professional managers association. A total of 401 questionnaires were distributed, with a response rate of 78.3%. The survey lasted a month, and 314 manufacturing enterprises returned with valid responses (Table 2). The questionnaire also collected data about the age and size of the enterprises.

Table 2 Demographic profile of sample firms.
Full size table

Common method variance

Acknowledging the potential for common method variance (CMV), we implemented strategies to mitigate its impact based on previous literature, spanning pre-survey, during-survey, and post-survey phases (Wang & Esperança, 2023). Post hoc tests indicate that CMV does not significantly influence our results, lending validity to the causal relationships between the antecedents and outcomes. Concretely, we made efforts to reduce the influence of CMV in three aspects:

  1. 1.

    Before the survey, we organized expert panel discussions and conducted pre-surveys to ensure the clarity and relevance of the questions.

  2. 2.

    During the survey, we emphasized that all data would be kept strictly confidential and used solely for research purposes. The survey did not involve any personally identifiable information, and participants were informed that their participation was voluntary, with the freedom to withdraw at any time.

  3. 3.

    After the survey, we conducted post hoc tests, including the single factor test and the marker variable technique (Wang & Esperança, 2023). The “Perceived habit,” unrelated to the research content, was selected as the marker variable, and we tested the relationship between this marker variable and other variables in the model. The results revealed no significant relationship between the marker variable and other latent variables in the model, indicating that there was no CMV in the results of this study (Hair et al. 2022).

Data analysis technology

To enhance the methodological rigor of our study, we employed a mixed-method approach combining partial least squares structural equation modeling (PLS-SEM) and fsQCA. This methodological triangulation allows us to capture both linear relationships and complex configurational patterns in our data (Zhang, 2024). We selected these analytical techniques based on their suitability for our research context and their ability to provide complementary insights (Wang & Esperança, 2023). PLS-SEM is particularly appropriate for our exploratory study, which involves developing a new theoretical model with multiple mediating variables. It is well-suited for theory development and can handle complex models with minimal demands on measurement scales and sample size (Chatterjee et al. 2024). fsQCA, on the other hand, allows us to identify complex causal patterns that may not be apparent through traditional linear analyses, providing a nuanced understanding of the factors influencing digital transformation in SMEs (Fiss, 2011; Ragin, 2008; Schöggl et al. 2024).

PLS is particularly well-suited for our research, which is exploratory and involves developing a new theoretical model (Zhang, 2024). PLS, a structural equation modeling technique, is frequently employed in exploratory studies due to its minimal demands on measurement scales, sample size, and residual distributions (Wang and Zhang, 2024a). It is especially useful for theory development as it can handle complex models with multiple mediating variables and has better predictive power than other techniques (Wang & Zhang, 2024b). PLS-SEM has also been widely used in similar studies (Kalaitzi & Tsolakis, 2022; Schöggl et al. 2024).

To circumvent the limitations of relying on a single method, we also utilized fsQCA in our research. fsQCA is a method that identifies causal patterns in the data that lead to an outcome (Fiss, 2011; Ragin, 2008). The fsQCA approach is particularly useful when uncovering complex causal relationships that may not be readily apparent with other analytical techniques. Combining PLS and fsQCA uncovered more nuanced and interesting insights, providing a richer understanding of the phenomena under study.

The combination of PLS and fsQCA in our data analysis approach allowed for a robust and comprehensive exploration of our research questions (Wang & Esperança, 2023).

Analyses and results

Structural equation modelling analysis

Measurement model

The PLS-SEM measurement model was tested for discriminant, content, and convergent validity (Hair et al. 2022). All measurement items in this study are derived from existing literature or adapted from previous research results and have been verified by pre-research. Therefore, the measurement model of this study can be considered to have good content validity. The reliability of the measurement model can be judged by testing the composite reliability (CR) and Cronbach’s alpha. Table 3 shows that both values are greater than 0.8, which is much higher than the critical value of 0.7, so the measurement model in this study has good reliability. The convergent validity of the measurement model was judged by verifying the average variance extracted (AVE) value. Table 3 shows that the average extraction variance of each latent variable is greater than 0.594, and none is less than 0.5, indicating that the measurement model in this study has good convergent validity. The discriminant validity can be obtained by comparing the square root of the AVE of each latent variable and the correlation coefficient between the other latent variables (Table 4). Therefore, the discriminant validity of the measurement model in this study is ideal (Zhang, 2024).

Table 3 Summary of the measurement model.
Full size table
Table 4 Discriminant analysis.
Full size table

Factor loading between each latent variable and its measurement item is greater than the cross-factor loading between the latent variable and other latent variables (Table 5). It can be concluded that the measurement model of this study has ideal convergent and discriminant validity (Hair et al. 2022). The analysis results with the help of SmartPLS 3 show that the heterotrait-monotrait ratio (HTMT) of the measurement model in this study is less than 0.663, further indicating that the discriminant validity of the measurement model is relatively good.

Table 5 Factor loading and Cross-loading.
Full size table

Structural model

The research results of the four-stage research model of enterprise digital transformation are shown in Fig. 3.

Fig. 3
figure 3

Results of the digital transformation paths chain based on industrial internet.

Full size image

The coefficients of determination (R2) of industrial internet driven, enterprise performance and enterprise competitive advantage are 42.5%, 37.1%, and 46.4%, respectively, indicating that the model in this study has a moderate explanatory power (Hair et al. 2022). Standardized root mean square residual (SRMR) can evaluate the fit of the research model. The result of SRMR is 0.046, less than the critical value of 0.08, indicating that the model fits well (Hair et al. 2022). In this study, the model Q2 was calculated using the blindfolding algorithm. The Q2 of enterprise competitive advantage in the current four-stage model of enterprise digitization is 0.293 (greater than 0.15), showing a moderate predictive correlation.

Hypotheses H1 to H10 in the basic model were supported (Fig. 3). Control variables, including enterprise age, enterprise size, and the enterprise’s development stage, had no significant effect on enterprise competitive advantage.

Cross-stage analysis

This study tested the rationality of the four stages of digital transformation through cross-stage analysis. The results of the data analysis show that IIA has no effect on IID (p > 0.5), IIA has no effect on ECA (p > 0.5), and IIAP has no effect on ECA (p > 0.5). The above results indicate that Stage 1 cannot directly influence Stages 3 and 4, which further justifies the division of the four stages in the digital transformation of enterprises.

fsQCA

Calibration

In the first stage of using fsQCA, this study uses the calibration function in the fsQCA 3.0 software to calibrate the sample data to a continuous set from 0 to 1, and then convert the data into degrees of full membership, cross-over, and full non-membership (Wang & Esperança, 2023). The calibration results are shown in Table 6.

Table 6 Calibration.
Full size table

Necessity analysis

The second phase of fsQCA is necessity condition analysis (NCA), which analyzes the combination of causal conditions through measures of consistency and coverage (Dahms et al. 2023), to explore eight prerequisites on the existence of derived dependent variables. The results of the necessity analysis in Table 7 show that no single condition is necessary to produce the “Enterprise competitive advantage” result.

Table 7 Results of necessity analysis.
Full size table

Sufficient condition analysis

In our fsQCA analysis, we employed a raw consistency threshold of 0.80 and a PRI (Proportional Reduction in Inconsistency) consistency threshold of 0.50, with a minimum case frequency of 2. These thresholds were chosen based on recommendations in the fsQCA literature to ensure robust and meaningful results (Fiss, 2011; Ragin, 2008). Table 8 shows the 12 groups of configurations of the sufficient condition analysis results, the solution consistency is 0.877 (greater than 0.8), and the solution coverage rate is 0.519 (above 0.5), indicating that the results have acceptable reliability and interpretation power (Wang & Esperança, 2023). The results of groups 4, 5, 6, 11, and 12 indicate that the combination of the above conditions is a sufficient condition to bring about the competitive advantage of the firm. Groups 1, 4, 5, 10, 11, and 12 cover the four stages of digital transformation, although they are less than eight conditions. Most of the groups also indicate that the drive for enterprise competitive advantage cannot be separated from the elements in the TOE framework.

Table 8 Sufficient Condition Analysis.
Full size table

To ensure the robustness of our findings, we conducted several additional analyses. For the fsQCA results, we conduct a robustness test on the configuration of the competitive advantage of enterprises, increasing the PRI consistency threshold from 0.50 to 0.60 and increasing the case frequency from 2 to 3, resulting in similar configuration results. Therefore, the findings are robust (Shehzad et al. 2024).

Discussion and academic contribution

Digital transformation paths chain

This research proposes a “Digital transformation paths chain” framework, and a research model of enterprise industrial internet adoption and impacts based on the TOE framework and IDT through cross-analysis of relevant literature.

Existing research on enterprise industrial internet adoption mainly focuses on the adoption of the technology or the impact of the industrial internet on enterprise performance (Prajapati et al. 2024; Wassie & Lakatos, 2024), lacking empirical support for the processes and linkages that facilitate the two separate stages. Our findings on the multi-stage nature of digital transformation align with and extend recent work by Wu and Li (2024), who emphasized the dynamic and continuous nature of digital transformation in Chinese enterprises. While they focused on entrepreneurship and disruptive innovation, our study provides a more granular view of the transformation stages, particularly in the context of SMEs adopting the industrial internet. Our research on the positive impact of industrial internet adoption on enterprise performance and competitive advantage aligns with recent work by Sivan et al. (2022), who highlighted the critical role of Industry 4.0 technologies in driving efficiency and competitiveness in the logistics sector.

This study considers enterprise industrial internet adoption as a process of technological innovation development. Drawing on the excellent results of many TOE frameworks (AlShamsi et al. 2022) and IDT (Hartley et al. 2022), we propose a four-stage model to study the complete process of enterprises from industrial internet adoption, integration, and application to finally driving enterprise development.

The research results confirm that adopting the industrial internet significantly positively affects enterprise performance and competitive advantage (Wu et al. 2024). It is important to note the stages of industrial internet adoption, application, driven, and performance, which need to be completed to impact the competitive advantage of enterprises. The four-stage model of industrial internet adoption is an important theoretical contribution of this study, which helps to increase the research literature in the field of enterprise industrial internet adoption. This study provides a framework and evidence for manufacturing enterprises to implement industrial internet adoption and enriches the research results on industrial internet adoption in manufacturing enterprises. We have further elucidated the applicability and importance of our findings in SMEs adopting the industrial internet (Wang & Esperança, 2023).

Industrial internet can play the effect through four stages of adoption, application, driven, and performance, ultimately affecting enterprise performance and competitive advantage. This finding opens up the enterprise industrial internet adoption black box for stakeholders, detailing what and how each stage is related. This result shows that the implementation of industrial internet adoption by enterprises can improve enterprise performance (Inegbedion et al. 2024) and competitive advantage (Chen et al. 2022). While previous research on industrial internet adoption has focused more on the influencing factors of enterprise performance (Lin, 2022), this study expands on the specific impact and process of industrial internet adoption on enterprise performance and competitive advantage.

Roles of technology, organization and environment

Digital technology, organization, and environment are three essential dimensions influencing enterprise industrial internet adoption (Awuah et al. 2022). This study reaffirms that these three dimensions play a pivotal role as the antecedents of industrial internet adoption. Moreover, our research has discovered that these dimensions also represent the initial stage of the adoption process. Our investigation into the critical role of technological, organizational, and environmental factors throughout the digital transformation process supports and builds upon recent findings by Elsharnouby et al. (2024), who identified critical internal enablers for SMEs’ export performance in Qatar. While their study focused on export performance, our research extends these insights to the specific context of industrial internet adoption and its impact on overall enterprise performance and competitive advantage. We also found that organizational factors play a crucial role in the adoption and implementation of digital technologies, which is consistent with Sabatini et al. (2022), who emphasized the importance of organizational vision and strategy in driving business model innovation through digital technologies in Italian SMEs.

Applying the TOE framework, our study underscores its importance for enterprises adopting industrial internet strategies. We propose a pathway to enhance enterprise performance and competitive advantage through these three dimensions. Importantly, our study leverages the fsQCA methodology to reveal complex combinations of these factors that influence the enterprise’s digital transformation (Fiss, 2011; Ragin, 2008). This approach is particularly suitable for our research, as it allows us to account for multiple causal pathways and provides a more nuanced understanding of the factors driving industrial internet adoption. The findings on the positive impact of industrial internet adoption on enterprise performance and competitive advantage are consistent with recent work by Wang and Esperança (2023), who found that digital transformation can improve market and ESG performance in Chinese SMEs. However, our study goes further by delineating the specific pathways through which this impact occurs, namely through the stages of adoption, application, driven, and performance.

Considering the optimal strategy for enhancing enterprise competitive advantage, our fsQCA results indicate that one or several TOE elements are included in most of the portfolio regardless of the sufficient combination of conditions. This finding further attests to the critical role of digital technology, organization, and environment in the inception and outcome of enterprise industrial internet adoption. These study’s findings are consistent with and extend the work of Elsharnouby et al. (2024), who studied critical internal enablers of SME export performance. While they focused on general organizational factors, our study specifically highlights how these factors interact in the context of industrial internet adoption. Furthermore, our results complement the research of Wassie and Lakatos (2024) on the future of internal audit functions in the era of artificial intelligence by demonstrating how similar technological, organizational, and environmental factors influence the broader digital transformation process in SMEs.

Our research found that the TOE framework affects companies’ adoption of digital transformation and corporate performance through integration and application. This finding enriches the existing literature on the impact of TOE in various stages of digital transformation and extends the applicability of the TOE framework in research on the digital transformation of the industrial internet (AlShamsi et al. 2022). Our results on the importance of the industrial internet for enhancing SME competitiveness align with recent work by Schöggl et al. (2024), who examined the role of digital technologies in implementing circular economy practices in manufacturing firms. While their focus was on sustainability, our study complements their findings by highlighting how similar digital technologies can drive overall performance and competitive advantage. We emphasize that this finding expands the scope of the application of the TOE framework in research on enterprise digital transformation and accumulates new evidence for research in this field (Gökalp et al. 2022).

Managerial implications

Implications for business

Enterprises should formulate strategies for digital transformation, enhance learning within and outside the organization, and improve their digital capabilities to adapt to the rapid development of the current digital wave (Xiufan et al. 2024). Enterprises must be aware that digital transformation does not happen overnight as this is a path chain. Therefore, enterprises can develop long-term, medium-term, and short-term phased strategies in the digitalization process to alleviate the pressure of digital transformation in capital, workforce, and technology (Elsharnouby et al. 2024). From outside the organization, enterprises should focus on customer needs to improve business agility and the success rate of digital transformation.

Enterprises should flexibly adjust their organizational structure from inside the organization to accommodate digital transformation (Sivan et al. 2022). First, in terms of human resources, enterprises should recruit new professionals and share the latest digital knowledge with veteran employees to help them prepare for digital transformation (Liu et al. 2024). Second, in terms of organizational culture, enterprises encounter internal resistance to new technologies and strategies and should foster a culture that encourages digital innovation and helps enterprises remove the burden of thinking about digital transformation (Sabatini et al. 2022). A culture of innovation can help enterprises be more willing to experiment with new technologies and ideas, thus gaining more opportunities to improve corporate performance and competitive advantage (Kareem et al. 2024). Third, in terms of organizational security, enterprises need to develop a data security manual, standardize data permission settings, strengthen data backup and management, and protect their digital assets and trade secrets.

Implications for relevant stakeholders

First, the R&D enterprises of the industrial internet should improve technical solutions, provide technical support and training promptly, and help enterprises solve technical problems (Frank et al. 2024). The service providers of the industrial internet should not only provide hardware and software equipment, but also provide training for technical application talents to provide complete solutions for enterprises relying on digital transformation. Digital solutions service providers must help enterprises complete security training to address data security issues. Digital solution providers should also be able to innovate the way they charge. For example, providers can use digital technologies like cloud services to provide enterprises with on-demand services such as SaaS, thus avoiding large upfront hardware investments (Schöggl et al. 2024).

Second, to solve enterprises’ weak digital technology capabilities, the government can support the agglomeration of industrial internet R&D enterprises to improve the technical service capabilities of the industrial internet. Governments should speed up formulating data security laws and regulations and coordinate multiple departments to issue data security standards based on sufficient research to ensure the security of enterprise digital assets and business secrets (Dou et al. 2022). The competent government departments can entrust industry associations to organize group and bulk purchases to obtain lower purchase costs and, at the same time, introduce corresponding industrial internet implementation incentive plans to give corresponding subsidies (Bordoloi et al. 2022). The government’s bulk procurement measures and subsidies can help enterprises alleviate their digital transformation concerns and financial pressure and help companies improve the success rate of digital transformation.

Limitations and future research

This study achieves a set of significant results. However, it also highlights the need for further research. This study collects cross-sectional data and does not address the differences between enterprises before and after digital transformation; subsequent research can consider adopting a hierarchical tracking method, that is, using a multi-group structural equation model to analyze the impact of the two samples on improving enterprises’ performance and competitive advantage. Research subjects with different characteristics may produce more interesting research results. Future research could delve more deeply into the differences between enterprises at different development stages and in different industries during the digital transformation process. While our sample is representative of Chinese SMEs, the generalizability of our findings to other contexts should be tested in future research. Cross-cultural comparisons could provide valuable insights into how the digital transformation process may vary across different institutional environments.

Conclusion

This study has contributed to our understanding of digital transformation pathways in SMEs, particularly in the context of the industrial internet. Through the development of a “Digital transformation paths chain” framework and research model, we have illuminated the key factors driving SMEs to adopt digital transformation and the impact pathways through which the industrial internet influences performance and competitiveness. Our research has revealed that the adoption of the industrial internet is not a single event but a complex, multi-stage process comprising adoption, application, driven, and performance stages. Each of these stages plays a crucial role in the overall digital transformation journey of SMEs. The study underscores the positive impact of industrial internet adoption on enterprise performance and competitive advantage, highlighting the importance of a holistic approach to digital transformation. The findings emphasize the critical roles of TOE in driving industrial internet adoption. The application of the TOE framework in this context provides a comprehensive perspective on the factors influencing various digital transformation stages. This insight is particularly valuable for SMEs, which often face unique challenges in their digital transformation efforts due to resource constraints and organizational characteristics. The use of both PLS-SEM and fsQCA methodologies has allowed us to uncover nuanced relationships and complex combinations of factors influencing digital transformation. This multi-method approach provides a more comprehensive understanding of the digital transformation process in SMEs. Our study offers valuable insights for businesses, policymakers, and other stakeholders involved in digital transformation initiatives. For businesses, it provides a roadmap for navigating the complex process of digital transformation, highlighting the need for a staged approach and the importance of considering technological, organizational, and environmental factors. For policymakers, our findings underscore the need for supportive policies and initiatives that address the specific challenges faced by SMEs in their digital transformation journeys. This research not only advances our theoretical understanding of digital transformation in SMEs but also provides practical guidance for implementing effective digital strategies. By emphasizing the importance of a comprehensive approach to digital transformation, our study underlines the potential for SMEs to enhance their performance and competitive advantage through the strategic adoption and implementation of industrial internet technologies.