A Study on the Determinants of the Service-Oriented Transformation of Manufacturing Enterprises

Abstract

As global manufacturing pursues green and low-carbon transitions, servitization offers a critical pathway for firms to advance sustainability goals. Yet the “servitization paradox” persists, dampening firms’ willingness and capacity to adopt service-oriented strategies, while existing research offers limited explanations for why some transformations succeed and others fail. Drawing on embeddedness theory, this study applies fuzzy-set qualitative comparative analysis (fsQCA) to a sample of 110 A-share listed servitization demonstration manufacturers in China (2015–2022). It systematically investigates how configurations of cognitive, relational, and environmental embeddedness jointly drive servitization outcomes. The findings reveal the following. (1) The service-oriented transformation of manufacturing enterprises is the result of the synergistic interaction of multiple antecedent conditions, with relational embeddedness playing a core driving role in SSP and strategic flexibility playing a core driving role in SSC. (2) There are multiple equivalent pathways for the service-oriented transformation of manufacturing enterprises; the three configuration pathways for high SSP are innovation-driven gap filling–flexible supply chain collaboration, competition-led–product–customer coupling, and institutional backstop–relationship embedded compensation. The three pathways for high SSC are: low supply dependency–autonomous flexibility, institutional guidance–flexible compensation, and innovative customer lock-in–loose market empowerment. This study enhances the understanding of service transformation in manufacturing enterprises and provides theoretical guidance and decision-making references for enterprises to select service transformation pathways tailored to their specific circumstances.

1. Introduction

Under the dual pressures of intensifying global climate change and tightening resource constraints, sustainable development has become a central issue facing human society in the 21st century. As a major contributor to resource consumption and carbon emissions, the transformation and upgrading of the manufacturing sector is of decisive importance for achieving sustainable development goals [1]. However, against the backdrop of profound restructuring of global value chains and tightening resource and environmental constraints, intensified competition due to product homogenization leads to resource waste; high-energy-consumption and high-emission production methods exacerbate environmental pressures; and reliance on foreign core technologies constrains industrial resilience. Consequently, the traditional extensive growth model of manufacturing enterprises faces severe challenges [2].

To address these challenges, servitization has emerged as a key pathway to drive the transformation and upgrading of the manufacturing sector, thereby achieving sustainable development. Servitization refers to the shift by manufacturing enterprises from simply providing products to offering “product–service bundles” [3]. By embedding service elements throughout the entire product lifecycle, enterprises can effectively reduce resource consumption, extend product lifespans, and enhance value-added along the industrial chain, thereby achieving the synergistic optimization of economic, environmental, and social benefits. This transformation model aligns closely with the Sustainable Development Goals and provides a crucial pathway to resolving the sustainability dilemma facing the manufacturing sector [4]. However, the existence of the “servitization paradox” makes this transformation process fraught with uncertainty [5], whether for small- and medium-sized enterprises or for manufacturing giants such as Intel and Boeing. On the one hand, some enterprises forcefully pursue servitization despite lacking core technological capabilities and organizational foundations, neglecting the differentiated needs arising from their own resource endowments and strategic positioning, which leads to a structural mismatch between strategic choices and actual capabilities. On the other hand, some enterprises overlook the constraints imposed by the external institutional environment and resource dependencies. They fail to carefully assess the profound impact of stakeholder networks, market competition, and institutional pressures on the service-oriented transformation, thereby triggering a systemic disconnect between strategic decisions and the institutional context. This makes it difficult for the service-oriented transformation to secure external legitimacy, ultimately increasing the risks of resource dissipation and strategic failure. Therefore, identifying the key factors influencing the successful implementation of service-oriented strategies and analyzing how these factors interact holds significant theoretical value and practical significance for promoting efficient resource utilization and a circular economy, enhancing the performance of enterprises’ service-oriented transformation, driving the synergistic optimization of economic, environmental, and social benefits in the manufacturing sector, and supporting the global manufacturing industry’s transition toward low-carbon and sustainable models.

Currently, scholars have conducted extensive discussions on the drivers of service-oriented performance in manufacturing enterprises. Internally, the service-oriented transformation of manufacturing enterprises primarily involves the enterprise’s own resources and capabilities, as well as effective management of the service-oriented process. For example, organizational structure and processes [6], specific capabilities and service-oriented organizational cultural transformation [7], employee service skills and cross-functional collaboration mechanisms [8], service innovation processes and customer relationship management systems [9] are considered to play a key role in advancing the service-oriented process and achieving satisfactory service-oriented performance. Externally, stakeholders such as suppliers [10], customers [11], and partners [12], as well as the broader market environment [13] and sociocultural systems [14] in which the enterprise operates, have also been found to exert significant influence on the enterprise’s service transformation.

While these existing studies provide valuable insights and references, they still exhibit significant limitations across three dimensions. First, there is the issue of conceptual homogenization. Most of the literature treats servitization as a single-dimensional, holistic construct, overlooking the fundamental differences among various types of servitization in terms of value propositions, risk characteristics, and capability requirements. These differences may suggest that the conditions necessary for advancing different types of servitization also vary. Consequently, treating servitization as a monolithic construct—particularly in empirical studies—is one reason why scholars have failed to reach consistent conclusions. It also prevents enterprises from developing targeted capabilities and implementation safeguards based on their specific environments, thereby hindering the smooth progression of transformation. Second, a one-sided research perspective. Existing studies often adopt a single perspective, focusing on the impact of a specific type of factor on the servitization of manufacturing enterprises. In reality, manufacturing enterprises are embedded in both internal and external environments, and servitization is often the result of the synergistic interaction of multiple factors, including internal perceptions and capabilities, external relationships, and market and institutional environments. Consequently, previous studies have tended to be one-sided, lacking systematic explanations for how enterprises can overcome servitization challenges and achieve a servitization transformation. They also fail to account for why different enterprises, despite sharing certain preconditions, exhibit vastly different levels of servitization. Third, methodological limitations. The existing literature primarily employs marginal analytical methods, such as regression analysis, to examine the relationship between driving factors and servitization, focusing on identifying the net effect of individual factors while overlooking the multiple interdependencies among different factors and the complex causal mechanisms arising from the concurrent interaction of multiple conditions [15]; consequently failing to deeply explore and test whether different preconditions possess distinct, equivalent coupling mechanisms capable of driving enterprise servitization. This also makes it difficult to address questions such as whether multiple parallel and non-contradictory pathways to enterprise servitization exist, and whether different enterprises can select appropriate pathways based on their own conditions.

To overcome these limitations, this study adopts a configurational perspective and, grounded in embeddedness theory, constructs a three-dimensional analytical framework of “cognitive embeddedness–relational embeddedness–environmental embeddedness” to systematically explore how combinations of multiple embeddednesses synergistically influence the performance of product-oriented and customer-oriented service transformation in manufacturing enterprises. Using data from China’s service-oriented manufacturing demonstration enterprises from 2015 to 2022 and employing the fuzzy set qualitative comparative analysis (fsQCA) method, the study identified multiple equivalent pathways that drive high-level product-oriented and Services Supporting the Customer’s Activity. This research contributes to the existing literature in three aspects: First, it breaks through the homogenized research paradigm of servitization by distinguishing different types of servitization and identifying their respective differentiated driving mechanisms; Second, by integrating internal and external perspectives to construct a comprehensive theoretical framework, the study systematically considers multi-level embedded factors that influence a firm’s level of servitization. This overcomes the shortcoming of existing research, which often focuses solely on the impact of factors at a single level, and provides richer research content for explaining servitization transformation from the perspectives of the micro-, meso-, and macro-environments in which firms are embedded. Third, by applying the fsQCA method to examine the complex configuration effects of multiple antecedent conditions on different types of servitization, this study reveals the multiple equivalent pathways for achieving servitization. This provides a theoretical foundation for enterprises to select the appropriate type of servitization based on their specific circumstances and to successfully advance their servitization efforts.

2. Theoretical Analysis and Model Construction

2.1. Service Transformation of Manufacturing Enterprises and Its Types

The service transformation of manufacturing enterprises refers to the strategic transition process by which manufacturing enterprises shift from providing products alone to offering integrated product–service systems [16]. As theoretical research has deepened, scholars have gradually recognized that the servitization of manufacturing enterprises is not a single, homogeneous phenomenon, but rather one that encompasses multiple implementation pathways and manifestations. To explore the heterogeneous characteristics of servitization strategies and the mechanisms underlying their differentiation in greater detail, this study classifies manufacturing servitization into two types, Services Supporting the Product and Services Supporting the Customer’s Activity, based on differences in service targets and content [17].

Services Supporting the Product focuses on enhancing product functionality and improving product competitiveness as its core objectives. It emphasizes integrated product development, packaging the value users require into the product itself, and realizing that value through a one-time transaction, with the relationship between supplier and customer weakening upon the conclusion of the transaction. Services Supporting the Customer’s Activity, on the other hand, focuses on meeting customers’ personalized needs and achieving co-creation of value. It emphasizes the exploration of the “post-product” long-tail service market, using the product as an entry point to uncover users’ latent needs, and providing scenarios and materials to assist users in realizing personalized experiences. Value is co-created by the enterprise and consumers during the consumption process, and the relationship between the two parties deepens further after the product transaction [18]. These two types of service-oriented models differ in their service content and may also exhibit significant variations in the potential benefits and risks they present to enterprises, as well as in the internal and external prerequisites required. Therefore, they warrant separate discussion.

2.2. Embeddedness Theory

Embeddedness theory originated from Polanyi’s historical analysis [19] and was later systematically elaborated by Granovetter. This theory posits that economic actors are neither completely autonomous individuals nor passive puppets bound by social norms; rather, they are embedded within ongoing networks of social relationships. Their economic behavior is profoundly constrained by their position within these networks, providing a suitable overarching analytical framework for this study [20]. In other words, the outcomes of manufacturing firms’ implementation of service-oriented strategies are influenced by both the internal and external environments in which they are embedded. Previous researchers have categorized embeddedness into different levels and types. For example, Zukin and DiMaggio systematically integrated the diverse research traditions on embeddedness within economic sociology, classifying embeddedness into four interrelated types: cognitive embeddedness, cultural embeddedness, structural embeddedness, and political embeddedness [21]. Granovetter elucidates the influence of social networks on individual behavior from the perspectives of relational and structural embeddedness [20]. Hagedoorn, meanwhile, distinguishes three levels of embeddedness—environmental, interorganizational, and bilateral—providing an integrative theoretical perspective for understanding the mechanisms of embeddedness across different analytical levels [22]. In this study, taking into account both relevance and systematicity, we classify the internal and external environments in which enterprises are embedded into three levels according to the micro-, meso-, and macro-perspectives: cognitive embeddedness, relational embeddedness, and environmental embeddedness. Among these, cognitive embeddedness reflects the enterprise’s internal strategic cognition and resource allocation logic; relational embeddedness embodies the characteristics of the enterprise’s network connections with key stakeholders; and environmental embeddedness represents the external institutional and market contexts in which the enterprise operates. We argue that these three levels of embeddedness collectively shape the differentiated pathways of service transformation.

Triple embedding does not operate in a homogeneous manner during servitization transformation; rather, it exhibits an asymmetric driving logic driven by fundamental differences in the nature of service types. Specifically, the key distinctions between Services Supporting the Product and Services Supporting the Customer’s Activity—in terms of value creation boundaries, relationship depth requirements, and strategic response logic—determine the differentiated mechanisms through which cognitive embedding, relational embedding, and environmental embedding influence each transformation path.

At the cognitive embedding level, Services Supporting the Product focuses on enhancing product functionality. Its value creation boundaries are relatively closed, as service elements depend on existing product architectures. Thus, the role of cognitive embedding is primarily to provide an extensible technological foundation and cognitive framework for servitization. In contrast, Services Supporting the Customer’s Activity aims to meet personalized customer needs and achieve value co-creation, requiring enterprises to transcend existing product logic and delve into customer operational processes. Here, the role of cognitive embedding emerges as a core capability that drives organizational paradigm shifts and dynamic resource reorganization.

At the relational embedding level, service modules in Services Supporting the Product are characterized by standardization and repeatability, requiring stable upstream and downstream collaboration networks to ensure delivery continuity. Therefore, deepening relational embedding is its core logic. Services Supporting the Customer’s Activity, by contrast, features highly customized service solutions, requiring enterprises to continuously gather information on diverse customer needs. In this case, the core logic is the flexibility of relational embedding, which requires striking a dynamic balance between deep customer engagement and network openness.

At the environmental embedding level, Services Supporting the Product is characterized by scalability and is easily imitated by competitors. The role of environmental embedding here is that market competitive pressures compel enterprises to establish full-lifecycle service systems, while institutional pressures ensure compliance and legitimacy. Services Supporting the Customer’s Activity, however, is characterized by customization, involving high upfront investments and long return cycles. The role of environmental embedding manifests in moderate market competition that protects service profit margins, while strong institutional support reduces innovation risks and incentivizes the exploration of high-value-added services.

In summary, through differentiated matching and dynamic interaction, the triple embedding mechanism jointly shapes the heterogeneous generation logic of Services Supporting the Product and Services Supporting the Customer’s Activity. This theoretical framework provides testable theoretical expectations for subsequent configuration analysis.

2.2.1. Cognitive Embedment Factors and Servitization

Cognitive embedment focuses on decision-makers’ mental models, cognitive frameworks, and heuristic decision-making mechanisms, emphasizing that economic actors are not entirely rational calculators but rather rely on existing cognitive structures to perceive, interpret, and respond to the external environment. At the cognitive embedment level, key factors influencing a firm’s service transformation include product innovation and strategic flexibility. It not only reflects the company’s investment in technology and R&D resources but also demonstrates its cognitive framework for breaking through existing paradigms and identifying new technology-market opportunities, as well as its ability to explore new technologies and markets. It lays the cognitive foundation for service-oriented transformation and determines a firm’s capacity to identify the strategic value of service-oriented approaches [23]. Strategic flexibility refers to a company’s ability to allocate resources across different time periods. At its core, it is the organization’s cognitive adaptability to perceive environmental changes and reconfigure resources; adaptive resource allocation depends on managers’ cognitive assessment of environmental uncertainty and heuristic decision-making. Therefore, companies with high strategic flexibility typically possess a more open and resilient organizational cognitive framework, enabling them to rapidly adjust the pace and path of their service-oriented strategy implementation in the face of environmental uncertainty [24].

Specifically, product innovation influences service-oriented transformation through two mechanisms [25]. First, the technical knowledge accumulated through product innovation determines the boundaries of technical feasibility for a company’s service-oriented transformation. The technical resources and knowledge reserves generated by product innovation provide the technical vehicle and value-creation foundation for service-oriented transformation, enabling the company to shift from purely selling products to providing integrated product and service solutions. Second, the managerial cognitive frameworks shaped by product innovation influence the selection and implementation of service-oriented strategies [26]. Managers with innovation-oriented mindsets are more inclined to view products as the starting point for service value creation rather than the endpoint of a transaction, thereby driving the enterprise’s shift toward a service-oriented paradigm. Product innovation provides the technological foundation and functional vehicle for services, while technological breakthroughs create entirely new service possibilities.

Strategic flexibility, in turn, supports service-oriented transformation by enabling the reallocation of resources. Strategic flexibility refers to an organization’s ability to perceive environmental changes and rapidly reallocate resources and adjust strategic direction to capitalize on emerging opportunities [27]. Manufacturing firms with high strategic flexibility can respond swiftly to environmental threats and opportunities while maintaining their core identity and value. Particularly in the context of service-oriented transformation, strategic flexibility supports the shift from traditional manufacturing logic to service logic, ensuring that the organization can effectively integrate the delivery processes of new products and services. Strategic flexibility enables enterprises to shift resources from traditional manufacturing to service development and delivery quickly and at low cost, thereby reducing the resource rigidity constraints of service-oriented transformation [28].

In summary, product innovation and strategic flexibility, as core dimensions of cognitive embedded factors, jointly drive the service-oriented transformation of manufacturing enterprises by shaping managers’ cognitive frameworks, providing a technological resource foundation, and enabling dynamic resource allocation. These two dimensions represent, respectively, the state of resource endowment and the cognitive adaptability required for a firm’s service transformation, constituting the key cognitive embeddedness factors for understanding the prerequisites of service transformation in manufacturing firms.

2.2.2. Relational Embeddedness Factors and Servitization

Relational embeddedness focuses on how economic actors are embedded within ongoing social networks, emphasizing the profound constraints that the quality, depth, and history of specific relationships impose on economic behavior [29], and concentrating on the characteristics of direct links between actors—particularly dimensions such as trust mechanisms, norms of reciprocity, and investments in relationship-specificity. Supplier concentration and customer concentration, which respectively characterize the degree of power asymmetry in upstream and downstream relationships, serve as key indicators of relational embeddedness and influence a firm’s service transformation [30]. However, due to the potential “embeddedness paradox” [31]—where moderate concentration facilitates the establishment of deeply embedded cooperative relationships and promotes the acquisition of key knowledge and technologies required for servitization, while excessive concentration weakens a firm’s bargaining power and resource acquisition flexibility in service innovation, leading to a loss of ability to respond to diverse market demands [32]—there remains uncertainty regarding the extent and direction of their impact on servitization.

Supplier concentration reflects the power dependency structure of buyers toward upstream resource providers. From the perspective of social exchange theory, moderate supplier concentration leads manufacturing firms to rely on suppliers for critical components and technical support, while suppliers depend on the manufacturing firms for order volume and market channels. This balance of power encourages both parties to make relationship-specific investments based on the principle of reciprocity, thereby driving the deepening of service transformation [33]. However, excessive supplier concentration leads to a severe power imbalance, manifested specifically in suppliers potentially exploiting their monopolistic position to capture excess profits generated by servitization, or reducing service quality in joint service delivery while demanding a higher share of revenue [34]. The instability of this exchange relationship forces manufacturing firms to allocate substantial resources to relationship governance and monitoring mechanisms rather than to service innovation itself, thereby inhibiting servitization performance.

Customer concentration and the level of service-oriented transformation in manufacturing firms also exhibit a complex nonlinear relationship. From the perspective of social exchange theory, the positive effects of moderate concentration stem from the norms of reciprocity and bilateral commitment mechanisms within the exchange relationship. When manufacturing firms make relationship-specific investments for customers, these investments induce reciprocal behavior from customers—that is, major customers are willing to pay a premium for in-depth services and provide long-term, stable order commitments, forming a balanced structure of mutual dependence. However, once customer concentration exceeds a critical threshold, the power structure between suppliers and customers undergoes a fundamental reversal. A small number of major customers exploit the firm’s investments in relationship-specificity to engage in opportunistic behavior, such as price-squeezing, extending payment terms, or demanding additional value-added services. This marks a qualitative shift in the exchange relationship from a reciprocal to an exploitative model. At this stage, the excessive strengthening of customer bargaining power not only squeezes service profits, but continued investment by the firm also increases the risk of being locked into assets, while withdrawal entails massive sunk cost losses. Ultimately, this suppresses the firm’s motivation to transition toward high-value-added services, trapping the service transformation in a vicious cycle of deteriorating performance [11].

In summary, supplier concentration and customer concentration, as core dimensions of relational embeddedness, jointly influence the service transformation of manufacturing firms by shaping the power-dependence structures between firms and their upstream and downstream partners, affecting decisions regarding relationship-specific investments, and determining the quality of exchange relationships. They constitute key environmental embeddedness factors for understanding the preconditions of service transformation in manufacturing firms.

2.2.3. Environmental Embeddedness Factors and Servitization

Environmental embeddedness emphasizes that corporate behavior is profoundly influenced by the characteristics of the macro-environment, focusing on how external institutional contexts and market structures shape firms’ strategic choices and organizational forms. Environmental embeddedness is distinguished into two dimensions: the national institutional environment and market competition. Significant differences among countries in terms of marketization levels, the sophistication of legal systems, cultural values, and other factors, as well as the intensity of market competition, collectively constitute the external contextual constraints on firms’ strategic decision-making.

As a core feature of market institutions, market competition provides the economic impetus and efficiency incentives for service-oriented transformation. Faced with intense market competition and pressure to commoditize, manufacturing firms achieve differentiation by incorporating service elements, which serves as a key strategy for responding to competition. From an institutional theory perspective, market competition is not only an efficiency-driven economic process but also an institutional force that shapes corporate strategic choices. The institutional environment constrains organizational behavior through three types of pressure: regulatory, normative, and cognitive [35]. Regulatory pressure manifests as stricter industry oversight and heightened standardization requirements; manufacturing firms demonstrate compliance with institutional rules by offering services such as compliance maintenance, thereby securing regulatory legitimacy. Normative pressure stems from the endorsement of the “modern manufacturing service provider” identity by industry associations, leading firms, and professional networks. Firms emulate competitors’ service-oriented practices to align with industry norms and secure normative legitimacy; cognitive pressure manifests in competition fostering a consensus that the service-oriented transformation of manufacturing firms is a natural and inevitable trend. Firms that cling to a purely manufacturing model risk being perceived as backward or out of step with industry trends, facing a crisis of cognitive legitimacy [36]. In highly competitive markets, firms face dual constraints from efficiency pressures and legitimacy pressures. On the one hand, intense competition requires firms to achieve differentiation and profit growth through service-oriented transformation; on the other hand, the institutional environment shapes the specific pathways and models of service-oriented transformation through regulatory policies, industry norms, and cognitive frameworks [37].

The institutional environment influences manufacturing firms’ servitization decisions and performance through both formal institutional support (such as policies, regulations, and fiscal incentives) and informal institutional pressures (such as industry norms and perceptions). However, a servitization paradox exists between institutional pressures and servitization performance: excessive regulation by the institutional environment may trap firms in a compliance trap, thereby undermining the flexibility required for service innovation [38]. When coercive pressures are too strong, firms may offer only symbolic responses to government calls while continuing to operate strictly within a manufacturing-centric model, resulting in superficial servitization. Excessive imitative pressures can similarly lead to strategic homogenization; when all firms pursue servitization based on the same template, service supply becomes oversaturated and differentiation disappears, thereby undermining the economic returns of servitization. Effective institutional strategies should leverage regulatory pressure to break path dependencies, utilize normative pressure to build service capabilities, and employ imitative pressure to disseminate best practices, while maintaining appropriate institutional flexibility to accommodate the iterative nature of service innovation. This approach helps avoid ritualistic compliance and strategic rigidity caused by institutional pressures.

In summary, market competition and the institutional environment, as core dimensions of environmental embeddedness, jointly drive the service transformation of manufacturing enterprises by shaping the economic efficiency pressures and institutional legitimacy constraints they face. Market competition provides the economic impetus and differentiation incentives for service transformation, while the institutional environment shapes the specific pathways and models of service transformation through three mechanisms: regulation, norms, and cognition. The interplay between these two constitutes the key environmental embeddedness factors for understanding the prerequisites of service transformation in manufacturing enterprises.

2.3. Model Construction

Based on the above analysis, this paper constructs a theoretical framework for the influencing factors of service transformation in manufacturing enterprises using embeddedness theory and the concept of configuration. It explores the complex interactive causal relationships among the eight variables, namely the six antecedent conditions (product innovation, strategic flexibility, supplier concentration, customer concentration, market competition, and the institutional environment) falling under three dimensions of embeddedness, and the two outcome variables (services supporting the product and services supporting the customer’s activity). The research model is shown in Figure 1.

3. Research Design

3.1. Research Methodology

This study aims to investigate how multiple antecedent conditions interact to drive the service-oriented transformation of manufacturing enterprises. Given the complexity of the research questions, the asymmetry of causal relationships, and the diversity of transformation pathways, traditional econometric methods face significant limitations. Therefore, this study adopts Fuzzy Set Qualitative Comparative Analysis (fsQCA) as its core research method.

3.2. Data Sources

As the world’s largest manufacturing nation, China has ranked first globally in manufacturing value-added for 16 consecutive years and is a major contributor to global carbon emissions and resource consumption. Against this backdrop, the green transformation of Chinese manufacturing enterprises holds significant importance for advancing the global sustainable development process. Since the introduction of China’s “dual carbon” goals, the Chinese government has positioned service-oriented manufacturing as a key strategy to address the high energy consumption and emissions challenges in the manufacturing sector and to achieve the synergistic optimization of economic, environmental, and social benefits. This provides a unique institutional context for examining the sustainability effects of the service-oriented transformation. At the same time, China is in a critical phase of economic transformation, characterized by a dynamically evolving institutional environment and increasingly fierce market competition. The service-oriented paradox faced by enterprises is particularly pronounced, providing an ideal research setting for this study to reveal how multiple factors synergistically drive sustainable service-oriented transformation.

As rigorously selected industry benchmarks, these model enterprises embody the typical models and advanced practices of service transformation in China’s manufacturing sector. By studying these high-performing enterprises, we can clearly identify the key factors driving sustainable service transformation. The transformation pathways validated by these benchmark enterprises serve as direct references and offer practical value for general manufacturing firms, providing a theoretical foundation and decision-making guidance for the green service transformation of a broader range of enterprises. Therefore, to ensure the representativeness of the research findings and their theoretical guidance and practical utility, this study selects listed companies from the five batches of service-oriented manufacturing demonstration enterprises announced by the Ministry of Industry and Information Technology of the People’s Republic of China as the research subjects. After excluding samples in *ST and ST status as well as those with severe data gaps, this study uses data from 2015 to 2022 for a total of 110 A-share listed service-oriented manufacturing demonstration enterprises as the sample, with data primarily sourced from the CSMAR database.

Since the evaluation criteria for service-oriented manufacturing demonstration enterprises are based on the company’s revenue and other data from the three years prior to the application for selection, this study uses data from the three years preceding the enterprises’ designation as service-oriented manufacturing demonstration enterprises as the sample data. Given the potential bidirectional interaction between the service-oriented transformation of manufacturing enterprises and its antecedent conditions—specifically, the possibility of endogeneity where independent and dependent variables may cause each other in general panel data models—this study employs a time-lagged design to break this feedback loop. Since strategic adjustments by manufacturing firms (such as R&D investments, supply chain restructuring, and deepening customer relationships) typically take an annual business cycle to manifest as changes in service revenue structure, and given that the sample period for this study is 2015–2022, an excessively short lag period (e.g., the same period) cannot avoid reverse causality. However, using a lag of t-2 or longer would shorten the effective sample period available for analysis (e.g., a t-2 design would cover only 2017–2022), thereby reducing the robustness of the path model. Therefore, this study sets the observed values of each antecedent variable are set to the level from the previous period (t-1) of the dependent variable, thereby establishing the directionality of the causal chain [39].

3.3. Variable Measurement

To ensure the scientific rigor and operational feasibility of the empirical study, this research systematically elaborates and standardizes the definitions and operational measurement methods for each core variable. This is based on a thorough review of measurement methods from authoritative domestic and international literature, combined with the data disclosure characteristics of listed Chinese manufacturing companies.

With regard to the outcome variables, this study distinguishes servitization into two dimensions based on differences in service content: Services Supporting the Product (SSP) and Services Supporting the Customer’s Activity (SSC). Following Mathieu’s (2001) dichotomy, this study classifies service revenues directly associated with product functionality, such as spare parts maintenance and repair, installation, product monitoring, and training, as product-oriented service revenues. Service revenues oriented toward customer business processes, such as consulting, operational support, solutions, and design, are classified as customer-oriented service revenues. Service revenues were manually compiled and categorized according to the degree of association between service content and products, with specific data sourced from the “Composition of Main Business Revenue” items in listed companies’ annual reports. Upon completion of classification, the ratios of product-oriented service revenue and customer-oriented service revenue to total main business revenue were calculated to measure the levels of product servitization and customer servitization, respectively [17].

Regarding the independent variables, product innovation (PI) refers to the process by which a firm analyzes the market environment and supply–demand relationships, rationally integrates, utilizes, and allocates internal and external factors and resources, and develops or innovates corporate technology to create entirely new products or improve the performance of existing products [40], measured by the proportion of R&D investment to total operating revenue [41]. Strategic Flexibility (SF) is an enterprise’s ability to formulate alternative action plans by leveraging internal flexible elements (such as resource flexibility and coordination flexibility) in complex and dynamic environments to enhance adaptability, effectively achieve corporate objectives, and improve competitiveness, thereby forming an adaptive system to manage market changes [42], measured by a standardized composite index derived from five-dimensional indicators, Specifically, this includes the inventory-to-revenue ratio, net fixed assets ratio, sales expenses-to-revenue ratio, administrative expenses-to-revenue ratio, and financial leverage ratio(results of the five-dimensional factor analysis: KMO = 0.812, p < 0.001, Cronbach’s α = 0.834). The variance of each indicator over the [T−1, T+1] time window is calculated, standardized against the industry’s annual average, and then summed as absolute values [43]. Supplier concentration (SC) reflects the closeness of the relationship and the degree of dependence between the enterprise and its suppliers [44], measured by the proportion of procurement from the top five suppliers relative to total annual procurement [45]. Customer concentration (CC) reflects the closeness of the cooperative relationship between customers and the enterprise [46], measured by the proportion of sales to the top five customers relative to annual revenue [47]. Market competition (MC) reflects the intensity of market competition and is measured by 1 minus the Herfindahl–Hirschman Index (HHI) [48,49]; the institutional environment (IE) is an external institutional framework composed of policies and regulations, market mechanisms, and social norms, which exerts both constraining and incentivizing effects on corporate strategic choices [50] and is measured using a marketization composite index [51].

3.4. Data Calibration

Given the large standard deviations and distribution of outliers in the data for each variable in this study, if traditional fixed theoretical thresholds or subjectively set anchors are used, the membership assignments might deviate from the actual distribution due to sample-specific characteristics. That is, some cases with practical significance might be misclassified as “full membership” or “full non-membership,” thereby losing the sample’s variability information and even distorting the results of the configurational analysis.

To address these issues, this study draws on existing literature and adopts a quantile-based calibration method based on the sample distribution. Specifically, the 95th quantile of each variable is used as the threshold for full membership, and the 5th quantile as the threshold for no membership [52]. This approach minimizes the loss of the sample’s primary variation range while eliminating the interference of extreme outliers, ensuring that the calibrated fuzzy sets accurately reflect the actual differences among the sample enterprises [53]. Using the sample mean as the crossover point (0.5 membership degree) represents the average level of the sample and effectively distinguishes between groups of cases performing above or below average. By calibrating the variables to fuzzy set membership scores within the 0–1 range using the anchor thresholds of the 95th percentile, mean, and 5th percentile, with no extreme 0 or 1 values, satisfying the fsQCA requirement for fuzzy sets to be “neither zero nor one,” and the fuzzy set distributions of each variable exhibit good discriminative power, laying a robust data foundation for subsequent truth table analysis and configurational path exploration.

Table 1. Variable Calibration Anchors.

Variable		SSP Model			SSC Model
		0.95	Mean	0.05	0.95	Mean	0.05
Independent variable	PI	11.028000	4.317337	0.701000	11.537500	5.244065	0.515000
	SF	1.284836	0.849047	0.529337	3.293996	1.146640	0.518674
	SC	62.137000	26.022289	7.022000	62.487500	24.867243	5.682500
	CC	54.044000	23.790301	4.934500	63.320000	28.101589	5.825000
	MC	0.968328	0.887283	0.693022	0.969977	0.889772	0.699917
	IE	10.660000	8.899458	5.380000	10.645000	9.135850	5.680000
Dependent variable	SSP	67.742520	17.932033	0.000000	-	-	-
	SSC	-	-	-	84.097450	26.561697	0.000000

presents the variable calibration anchors.

4. Results

4.1. Single-Condition Necessity Analysis

Based on the fsQCA research criteria, a necessity test was conducted on the antecedent variables. The final results are shown in

Table 2. Necessity Analysis.

Variable	High SSP		High SSC
	Consistency	Coverage	Consistency	Coverage
PI	0.665	0.752	0.677	0.720
~PI	0.549	0.845	0.535	0.796
SF	0.673	0.757	0.855	0.700
~SF	0.547	0.847	0.316	0.809
SC	0.697	0.772	0.669	0.696
~SC	0.507	0.804	0.525	0.807
CC	0.666	0.759	0.659	0.712
~CC	0.525	0.800	0.524	0.763
MC	0.429	0.844	0.394	0.741
~MC	0.723	0.705	0.785	0.726
IE	0.535	0.825	0.502	0.789
~IE	0.670	0.756	0.692	0.709
PI	0.665	0.752	0.677	0.720

. The consistency levels were all below 0.9, indicating that there were no necessary conditions in this study [54].

4.2. Analysis of the Sufficiency of Condition Configurations

4.2.1. Analysis of Condition Configurations for High SSP

After determining that there are no necessary conditions for product-oriented service transformation as an outcome, the study model was run with a case frequency threshold set to 5, an original consistency threshold set to 0.85, and a PRI consistency threshold set to 0.70. During the analysis, we assumed that all antecedent conditions (including their presence and absence) could contribute to the outcome. We included all counterfactual combinations consistent with theoretical logic, ultimately generating three solution types: complex, reduced, and intermediate solutions. The complex solution includes all conditions appearing in the truth table, representing the most conservative combination of conditions. However, these conditions are overly stringent, potentially overlooking other equally valid combinations, and fail to distinguish between core and peripheral conditions; The intermediate solution retains both core conditions and theoretically relevant marginal conditions [55], combining theoretical soundness with explanatory simplicity, and is therefore adopted as the final report result; The reduced solution simplifies the complex solution to minimal sufficient conditions, retaining only the core conditions shared across paths. Therefore, by comparing the nesting relationship between the reduced solution and the intermediate solution, one can identify the core conditions that appear in both the reduced and intermediate solutions, as well as the peripheral conditions that appear only in the intermediate solution. Among these, core conditions indicate a strong causal relationship with the outcome, while peripheral conditions play a supporting role. Finally, a table of configurational analysis results was plotted based on the simplified and intermediate solutions.

Table 3. Configurational Analysis Results for SSP.

Condition	P1	P2	P3
PI	⨂	·	·
SF	·		⨂
SC	·		·
CC		·	·
MC		●
IE			●
Consistency	0.913	0.893	0.911
Raw Coverage	0.406	0.324	0.258
Unique Coverage	0.195	0.121	0.045
Overall Consistency	0.876
Overall Coverage	0.608

● indicates that the condition exists and is a core condition; · indicates that the condition exists but is not a core condition; ⨂ indicates that the condition does not exist and is a core condition; ⊗ indicates that the condition does not exist and is not a core condition; a blank indicates that the condition may or may not exist and has no impact on the final result.

presents the results of the configuration analysis for conditions that generate high SSP.

By comparing the reduced solution with the intermediate solution, this study identified three configuration paths leading to a high level of product-oriented service transformation. The results are shown in Table 3.

Based on the above classification results, this study adheres to the principle of representative sampling (set membership > 0.8, performance outcomes > 0.8, rich data, and diversity in industries and ownership structures) to select representative cases from each pathway. The following sections provide a detailed discussion and case analysis of the three classification pathways.

Innovation-Driven Gap Filling–Flexible Supply Chain Collaboration Path

In configuration P1 (~PI*SF*SC), ~PI is the core condition, while SF and SC are peripheral conditions that are present but not essential. Configuration P1. indicates that when a company faces the dilemma of insufficient product innovation, it can compensate for this shortcoming through the synergistic effect of internal strategic flexibility and external supplier integration, thereby achieving a high level of Services Supporting the Product. The company relies on flexible resource allocation capabilities to respond rapidly to market changes, while simultaneously establishing stable vertical collaborative relationships through high supplier concentration to obtain resource support, forming a compensatory mechanism of internal and external synergy. This path can be viewed as an “innovation-driven gap filling–flexible supply chain collaboration” servitization transformation pathway.

The transformation practices of Xi’an Shaanxi Blower Power Co., Ltd. (Xi’an, Shaanxi, China) (Stock Code: 601369) from 2015 to 2016 corroborate Configuration P1. As a leading domestic manufacturer of turbomachinery, Shaanxi Blower Power adopted a strategy of “deepening technological expertise + optimizing existing assets” rather than a breakthrough innovation strategy, with R&D investment falling below the industry average of 8–10%. At the same time, the company demonstrated high strategic flexibility. It transformed its sales model by establishing a solutions-based approach led by engineering and services, while driving the upgrade of its business structure from an industrial economy to a service economy and building an intelligent cloud service platform to achieve full product lifecycle management. Regarding the supply chain, the company spearheaded the establishment of the Shaanxi Blower Complete Equipment Technology Collaboration Network, forming deeply integrated relationships with core suppliers to provide technical support for value-added services such as remote monitoring and fault diagnosis. Under these configuration conditions, Shaanxi Blower Power was selected as a National Model Enterprise for Service-Oriented Manufacturing in 2017, successfully transitioning from a “single-product supplier” to a “full-product-lifecycle service provider”.

Competition-Led–Product–Customer Coupling Path

In configuration P2 (MC*PI*CC), MC is the core condition, while PI and CC are boundary conditions. Configuration P2 indicates that in a highly competitive market environment, enterprises can achieve a high level of Services Supporting the Product through the close coupling of product innovation advantages with key customers. Market competition serves as the external pressure driving the enterprise’s servitization transformation, while product innovation and customer concentration constitute the core capability combination for responding to competition. This path can be viewed as a “competition-led–product–customer coupling” servitization transformation pathway.

The transformation practices of Dehua Tubabao Decorative New Materials Co., Ltd. (Deqing, Zhejiang, China) (Stock Code: 002043) from 2019 to 2021 validate Configuration P2. As a leader in the decorative panel industry, Tubabao increased its R&D investment year-on-year from 2019 to 2020, continuously launching functional panels and accelerating the development of whole-home customization products to establish a differentiated competitive advantage. The company has long maintained strategic partnerships with key clients such as Vanke and Evergrande; this high client concentration ensures stable orders and provides precise insights into end-user demand. At the same time, competition in China’s decorative panel industry has intensified, with frequent price wars and accelerating channel transformations, leaving companies facing dual pressures of homogenized competition and channel fragmentation. Under these conditions, Tubao was selected as a National Model Enterprise for Service-Oriented Manufacturing in 2022, successfully transitioning from a “single-product panel supplier” to a “full-value-chain service provider for panel products”.

Institutional Backstop–Relationship Embedded Compensation Path

In configuration P3 (PI*~SF*SC* CC*IE), ~SF and IE are core conditions, while SC, CC, and PI are boundary conditions. Configuration P3 indicates that, supported by a sound institutional environment, even enterprises lacking strategic flexibility can achieve a high level of Services Supporting the Product by combining product innovation with dual upstream and downstream integration of suppliers and customers. The provision of formal institutions reduces environmental uncertainty and substitutes for the buffering function of internal flexibility, enabling enterprises to ensure strategic performance through structured supply chain governance rather than dynamic adjustments. This path can be viewed as the “institutional backstop–relationship embedded compensation” servitization transformation pathway.

The transformation practices of Sichuan Changhong Electric Co., Ltd. (Mianyang, Sichuan, China) (Stock Code: 600839) from 2017 to 2020 corroborate Configuration P3. As a long-established state-owned home appliance enterprise, Changhong increased its R&D investment from 1.2 billion yuan to 2.1 billion yuan between 2017 and 2019, continuously advancing the development of smart products such as smart TVs and IoT-enabled home appliances. In terms of the supply chain, the company established long-term strategic partnerships with panel suppliers such as BOE and CSOT, while maintaining deep ties with retail chains like Suning and Gome, resulting in a dual-high concentration characteristic. Regarding the institutional environment, national strategies such as “Made in China 2025” and “Industrial Internet”, along with local policy support, provided safeguards for the company’s transformation. As a mature state-owned enterprise, Changhong’s strategic adjustments proceed at a steady pace. The company responds to market changes by integrating existing channel resources rather than through internal organizational flexibility, exhibiting characteristics of low strategic flexibility. Under these conditions, Changhong was designated a National Model Enterprise for Service-Oriented Manufacturing in 2021, successfully transitioning from a “single hardware manufacturer” to a “full-lifecycle service provider for smart products”.

4.2.2. Analysis of the Conditional Configuration for High SSC

Table 4. Configurational Analysis Results for SSC.

Condition	C1	C2	C3
PI		⨂	●
SF	·	·	·
SC	⨂
CC			●
MC			⨂
IE		●
Consistency	0.812	0.904	0.842
Raw Coverage	0.459	0.336	0.432
Unique Coverage	0.125	0.061	0.129
Overall Consistency	0.811
Overall Coverage	0.686

● indicates that the condition exists and is a core condition; · indicates that the condition exists but is not a core condition; ⊗ indicates that the condition does not exist and is a core condition; ⊗ indicates that the condition does not exist and is not a core condition; a blank indicates that the condition may or may not exist and has no impact on the final result.

presents the results of the conditional configuration analysis for SSC.

Based on the configuration results described above, this study also follows the principle of representative sampling to conduct a detailed discussion and case analysis of the three configuration paths.

Low Supply Dependency–Autonomous Flexibility Path

In configuration C1 (SF*~SC), ~SC is the core condition and SF is the boundary condition. Configuration C1 indicates that when a firm possesses high strategic flexibility and avoids over-reliance on a small number of suppliers, it can rapidly respond to market changes through flexible resource allocation capabilities. At the same time, by maintaining a decentralized supplier network, it preserves the diversity of resource acquisition and bargaining power, thereby achieving a high level of customer-oriented service orientation. The enterprise relies on strategic flexibility to dynamically allocate internal resources and rapidly reorganize organizational practices. At the same time, by maintaining low supplier concentration, it avoids the risk of resource lock-in caused by relationship-specific investments, forming a service-oriented transformation path characterized by strategic flexibility dominance and an open relationship network. This path can be viewed as the “low supply dependency–autonomous flexibility” servitization transformation pathway.

The transformation practices of Shenzhen Great Wall Development Technology Co., Ltd. (Shenzhen, Guangdong, China) (Stock Code: 000021) from 2019 to 2021 validate Configuration C1. As a leading global electronics manufacturing services provider, Shenzhen Great Wall Technology responded to the restructuring of the global electronics supply chain by optimizing its business structure to shift from “order-based production” to “customized services”, thereby advancing the transformation toward a “manufacturing + services” model. Concurrently, the procurement share of the company’s top five suppliers decreased from 64.90% to 27.87%. By diversifying procurement channels for key components and adopting localized sourcing, the company built a flexible supply chain network, thereby breaking free from lock-in in specific supply chain relationships. Under these configuration conditions, Shenzhen Technology’s customer-oriented service level surged from 0 in 2019 to 81.73% in 2021. In just three years, the company achieved a transformation from a “single-service contract manufacturer” to a “comprehensive service provider of manufacturing services and solutions”, validating the effectiveness of the strategy-driven flexibility–autonomous path.

Institutional Guidance–Flexible Compensation Path

Configuration C2: ~PI*SF*IE~, where PI and IE are core conditions and SF is an edge condition. Configuration C2 indicates that when enterprises face the challenge of insufficient product innovation, they can compensate for innovation shortcomings through the synergistic interaction of internal strategic flexibility and external institutional support, thereby achieving a high level of customer-oriented service transformation. By leveraging flexible resource allocation capabilities to actively utilize policy incentives and institutional dividends, while relying on robust institutional safeguards (such as intellectual property protection and service standard guidance) to mitigate service innovation risks, enterprises can establish a service-oriented transformation path characterized by internal-external synergistic responsiveness. This path can be regarded as an “institutional guidance–flexible compensation” servitization transformation pathway.

The transformation practices of Neusoft Group Co., Ltd. (Shenyang, Liaoning, China) (Stock Code: 600718) from 2015 to 2016 corroborate Configuration C2. As a leading IT solutions and services provider in China, Neusoft maintained an R&D expenditure ratio of 9.31–9.93% in 2014–2015, which was lower than the 10–15% average for software companies on the STAR Market. By adopting a strategy of “optimizing existing product lines + deepening service models”, the company avoided excessive investment in basic software R&D. At the same time, the company demonstrated high strategic flexibility: facing opportunities in healthcare IT and smart city development, it dynamically adjusted its business architecture to rapidly reorganize medical resources and IT service capabilities, forming a flexible delivery model of “software + operations + data”. In terms of the policy environment, around 2016, the state intensively rolled out policies such as “Healthy China 2030” and “Smart Cities”, enabling Neusoft, as an industry leader, to secure a large volume of government contracts. Under these conditions, Neusoft’s customer-oriented service level (SSC) reached 80.16% in 2015, successfully transitioning from a “software product supplier” to a “platform operation service provider”.

Innovative Customer Lock-In–Loose Market Empowerment Path

In configuration C3 (PI*SF*CC*~MC), PI, CC, and ~MC are core conditions, while SF is a boundary condition. Configuration C3 indicates that when a company possesses strong product innovation capabilities and has established deep, binding relationships with key major clients, it can rapidly leverage strategic flexibility to convert technological advantages into customized service solutions. Simultaneously, in a moderately competitive market environment, it can avoid price wars that erode service profits, thereby achieving a high level of customer-oriented service transformation. By leveraging dual barriers of technological leadership and exclusive client relationships, enterprises deepen customer ties through flexible service delivery, forming a service-oriented transformation path characterized by deep integration between products and clients. This path can be viewed as an “Innovative Customer Lock-in–Loose Market Empowerment” servitization transformation pathway.

Midea Group Co., Ltd. (Foshan, Guangdong, China) (Stock Code: 000333) has validated the C3 framework through its transformation practices from 2019 to 2022. As a globally leading technology group, Midea has invested nearly 50 billion yuan in R&D over the past five years, focusing on systematic, scenario-based, and platform-oriented innovation. The company has built core platforms such as the “Midea Smart Home” and the Industrial Internet of Things (IIoT) platform, thereby establishing platform-level product innovation capabilities. In terms of strategic flexibility, Midea leverages a multi-brand portfolio (Midea, Little Swan, COLMO, KUKA, etc.) to achieve decentralized responsiveness under centralized control, combining economies of scale with market agility. Regarding customer relationships, Midea embeds product innovation and flexible service capabilities into key account operations through a full-value-chain digital system, achieving deep synergy and value co-creation. In terms of market competition, leveraging its industry leadership and full-industry-chain layout, Midea has carved out new, relatively monopolistic competitive territories in areas such as whole-home smart ecosystems and integrated smart building solutions, thereby creating strategic buffer zones where market competition is mitigated. Under these conditions, Midea has successfully driven a steady improvement in its customer-oriented service capabilities, achieving a transformation from a “mass-production manufacturer” to a global technology service group characterized by “technology leadership, direct-to-consumer engagement, and digital-intelligence-driven operations”.

4.3. Comparative Analysis of Two High-Level Service-Orientation Configurations

Overall, SSP and SSC, as two key models for the service-oriented transformation of manufacturing enterprises, differ in terms of core driving factors and applicable scenarios, reflecting the diverse pathways and complex mechanisms of service-oriented transformation.

Regarding core driving factors, among the three configuration types of SSP, factors at the relationship-embedded level serve as the primary drivers of final performance. Configuration P1 features supplier concentration as the core relational embedding factor, P2 features customer concentration, and P3 is characterized by dual relational embedding involving both supplier and customer concentration. This indicates that, since Services Supporting the Product typically unfolds within existing product domains, it relies heavily on close relationships with key stakeholders. Enterprises obtain resource support and market information through embedded relationships with suppliers or customers, thereby extending services around existing product platforms. This model is suitable for contexts with a high degree of relational embeddedness. In other words, in contexts characterized by a high degree of relational embeddedness, when firms face significant external pressures or constraints, they can leverage their relational embeddedness to secure resource support and achieve a gradual transition toward service-oriented operations.

From the perspective of the three configuration types of SSC, strategic flexibility serves as the primary driver. Although the core constraints vary, all three configurations regard strategic flexibility as a key requirement. This indicates that, given the greater degree of transformation required by customer-oriented service demands and their heavy reliance on the enterprise’s dynamic adaptability, companies need to utilize strategic flexibility to allocate resources flexibly and respond directly to customers’ personalized needs, rather than being constrained by specific relational embedding structures. When a company possesses strong strategic autonomy, ample strategic leeway, or robust resource support, it can leverage strategic flexibility to innovate services directly in response to customer needs, thereby achieving breakthrough or customized service-oriented transformation.

Services Supporting the Product and Services Supporting the Customer’s Activity are two important models for the servitization transformation of manufacturing enterprises; however, their performance depends on different embedded conditions. Enterprises should select an appropriate servitization transformation model based on their own capabilities, network embedding characteristics, and external contextual conditions to successfully transition from traditional manufacturing to a service-oriented model.

4.4. Robustness Analysis

Referring to the QCA stability verification framework [56], this study conducted robustness tests using parameter configurations such as the consistency threshold and the case frequency threshold. When the consistency threshold was lowered from 0.85 to 0.80, or the case frequency threshold was raised to 6, the configuration paths derived from both parameter configurations were consistent with the original model. The results of the robustness analysis are presented in

Table A1. Results of the SSP robustness analysis.

Condition	Consistency → 0.80			Case Frequency → 6
	P1	P2	P3	P1	P2	P3
PI	⨂	·	·	⨂	·	·
SF	·		⨂	·		⨂
SC	·		·	·		·
CC		·	·		·	·
MC		●			●
IE			●			●
Consistency	0.913	0.893	0.911	0.913	0.893	0.911
Raw Coverage	0.406	0.324	0.258	0.406	0.324	0.258
Unique Coverage	0.195	0.121	0.045	0.195	0.121	0.045
Overall Consistency	0.876			0.876
Overall Coverage	0.608			0.608

● indicates that the condition exists and is a core condition; · indicates that the condition exists but is not a core condition; ⊗ indicates that the condition does not exist and is a core condition; ⊗ indicates that the condition does not exist and is not a core condition; a blank indicates that the condition may or may not exist and has no impact on the final result.

and

Table A2. Results of the SSC robustness analysis.

Condition	Consistency → 0.80			Case Frequency → 6
	C1	C2	C3	C1	C2	C3
PI		⨂	●		⨂	●
SF	·	·	·	·	·	·
SC	⨂			⨂
CC			●			●
MC			⨂			⨂
IE		●			●
Consistency	0.812	0.904	0.842	0.812	0.904	0.842
Raw Coverage	0.459	0.336	0.432	0.459	0.336	0.432
Unique Coverage	0.125	0.061	0.129	0.125	0.061	0.129
Overall Consistency	0.811			0.811
Overall Coverage	0.686			0.686

● indicates that the condition exists and is a core condition; · indicates that the condition exists but is not a core condition; ⊗ indicates that the condition does not exist and is a core condition; ⊗ indicates that the condition does not exist and is not a core condition; a blank indicates that the condition may or may not exist and has no impact on the final result.

in Appendix A. In summary, the results of the multiple robustness tests are consistent, and the core configuration is insensitive to changes in the consistency threshold, supporting the reliability of the core conclusions of this study.

4.5. Deviant Cases Analysis

It is important to note that the core feature of the fsQCA method lies in revealing “multiple concurrent causal pathways” and “equivalent paths,” meaning that there is not a single unique combination of conditions leading to the same outcome. The three configuration paths described above are theoretically representative sufficiency paths identified in this study; however, this does not imply that all cases with high outcome variables must strictly conform to one of these three paths. In fact, the sample contains deviant cases that do not conform to the aforementioned three configuration paths but still achieve high levels of the outcome variable. This precisely demonstrates the unique advantages of the fsQCA method and the essence of the complex causal perspective.

For example, Qingdao Double Star (Stock Code: 000599) has a condition configuration of high PI (product innovation present), high CC (high customer concentration), and low MC (low market competition intensity). This configuration does not belong to any of the three configuration paths capable of achieving high SSP. Nevertheless, Qingdao Double Star’s SSP performance is also above 0.8, successfully achieving high product-oriented servitization. A similar case is Dian Guang Technology (Stock Code: 002730), whose condition configuration is high SF (high strategic flexibility), high SC (high customer concentration), and low IE (poor institutional environment). This configuration does not fall under any of the three configuration paths capable of achieving high SSC. Nevertheless, Dian Guang Technology’s SSC performance score is also above 0.8, successfully achieving high customer-oriented service transformation.

This phenomenon can be understood from the following three perspectives:

First, the principle of “limited diversity” in fsQCA. Based on truth table analysis, fsQCA conducts sufficiency tests only on observed case combinations, excluding unobserved combinations (logical residues) from the analysis. The configurations of the aforementioned anomalous cases may represent rare combinations within the sample or logical residues not covered by intermediate solutions, and thus were not presented in the final three pathways. However, this does not negate the practical validity of this configuration; it merely indicates that it did not form a stable, sufficient path within the coverage of the sample in this study.

Second, fsQCA’s “causal asymmetry.” The generation logic of high-outcome variables is not a simple mirror image of that of low-outcome variables. Similarly, the paths leading to high-outcome variables may involve other equivalent combinations not captured by the current parameter settings. Under more lenient parameter settings or with a larger sample size, a fourth or even more configuration paths for the high outcome variable might emerge.

Third, the principle of “equifinality” in configuration theory. The achievement of organizational performance allows for the coexistence of multiple combinations of conditions; the three paths identified in this study are “discovered” representative paths, not “all” possible paths. These deviant cases precisely validate the complexity and path diversity of service transformation; enterprises can achieve equivalent performance through different combinations of conditions (such as relying on innovative customer lock-in in a low-competition environment), which is fully consistent with this study’s core finding of “multiple equivalent paths.”

In summary, the existence of deviant cases does not negate the conclusions of this study but rather further validates the methodological strengths of the fsQCA approach. The three high-SSP configuration paths and three high-SSC configuration paths identified in this study are sufficient paths that possess statistical robustness and theoretical representativeness under the current sample size and parameter settings; they systematically reveal the logic of differentiated combinations of the three-level factors—cognitive embedding, relational embedding, and environmental embedding—across different types of service transformation. However, the path to successful service transformation is far more diverse and multifaceted than the pathways we have currently identified; these deviant cases precisely constitute important leads for future research.

5. Conclusions and Outlook

5.1. Research Conclusions

This study adopts a configuration perspective and uses embeddedness theory as a framework. Based on resource-based theory, social exchange theory, and institutional theory, a research model was constructed. The study sample consisted of data from 2015 to 2022 for a total of 110 listed manufacturing enterprises—the five batches of service-oriented manufacturing demonstration enterprises announced by the Ministry of Industry and Information Technology of the People’s Republic of China. After employing a time-lagged design to address endogeneity, the fsQCA method was used to conduct a configuration analysis, ultimately identifying multiple equivalent pathways leading to high-performance service-oriented transformation.

Three configurations were identified that lead to high-level product-oriented service-oriented transformation. The Innovation-Driven Gap Filling–Flexible Supply Chain Collaboration Path (~PI*SF*SC) indicates that enterprises lacking prominent product innovation capabilities can achieve successful product-oriented service-oriented transformation through high strategic flexibility and deep supplier embedding. The Competition-Led–Product–Customer Coupling Path (MC*PI*CC) reveals that intense market competition, combined with strong product innovation capabilities and concentrated customer relationships, can catalyze product-oriented service transformation. The Institutional Backstop–Relationship Embedded Compensation Path (PI*~SF*SC*CC*IE) illustrates that even with low levels of strategic flexibility, a favorable institutional environment can enable product-oriented service transformation through relationship governance mechanisms.

There are also three configurations that generate high levels of customer-oriented service transformation. The Low Supply Dependency–Autonomous Flexibility Path (SF*~SC) shows that organizations with high organizational agility can independently develop customer-oriented service capabilities without over-reliance on specific supplier networks. The Institutional Guidance–Flexible Compensation Path (~PI*SF*IE) demonstrates that when strategic flexibility remains high, a favorable institutional environment can compensate for limitations in product innovation and foster customer-oriented service innovation. The Innovative Customer Lock-in–Loose Market Empowerment Path (PI*SF*CC*~MC) reveals that deep customer embedding, combined with product innovation and strategic flexibility, can achieve proactive service co-creation with customers under moderate competitive pressure.

The six configurations described above not only provide practical pathways for manufacturing enterprises to achieve high-performance service transformation but also offer a new theoretical perspective for resolving the service transformation paradox. This study finds that service-oriented performance is not a linear sum of individual conditions, but rather the outcome of the synergistic interaction of multiple conditions. The reason why certain individual advantageous conditions lead to the service-oriented paradox is that these conditions can only be effective within specific configurations. For example, while a firm may possess the advantageous condition of high strategic flexibility, strategic flexibility alone does not automatically promote the success of service-oriented transformation; only when high strategic flexibility synergizes with low supplier concentration (Configuration C1: SF*~SC) or with a favorable institutional environment (Configuration C2: ~PI*SF*IE) can customer-oriented servitization be achieved, respectively. This finding advances servitization research from a linear “resource input–performance output” mindset to a refined decision-making framework of “condition configuration–type adaptation–path selection”.

5.2. Theoretical Contributions

This study advances theoretical research on the service transformation of manufacturing enterprises in four dimensions.

First, it breaks through the homogenized research paradigm of servitization, revealing the differentiated driving mechanisms of Services Supporting the Product and Services Supporting the Customer’s Activity. Existing research often treats servitization as a single, homogeneous phenomenon, focusing on the linear relationship between servitization and performance [38] or exploring general antecedents of servitization transformation [1,5], while neglecting the typological differences in servitization itself and their differentiated driving logic. Although early scholars have identified diverse forms of servitization [17], subsequent research has lacked an in-depth analysis of the driving mechanisms behind different types of servitization. Through a mixed-methods design combining fsQCA and multi-case studies, this study systematically reveals the fundamental differences between the two servitization models in terms of strategic logic and resource allocation: Services Supporting the Product relies on relationship embedding to obtain resource support, while Services Supporting the Customer’s Activity relies on strategic flexibility to achieve dynamic adaptation. This finding explains why identical antecedent conditions produce markedly different effects across different types of servitization, providing theoretical support for understanding the heterogeneity of servitization and resolving the “servitization paradox”.

Second, this study reveals the configuration-based resolution mechanism of the service-ization paradox and promotes a shift in perspective from static capability complementation to dynamic capability matching. The service-ization paradox has long plagued both academia and industry; existing research largely attributes it to inherent contradictions within service-ization strategies or general deficiencies in organizational capabilities, lacking a nuanced explanation of the conditions under which the paradox arises. This study systematically demonstrates that when firms overlook the synergy and equivalence of condition combinations, it leads to the service-oriented paradox—a phenomenon characterized by high condition configuration but low service-oriented performance. This finding shifts the resolution mechanism of the service-oriented paradox from a static perspective of capability completion to a dynamic perspective of condition combination and type adaptation, providing a systematic theoretical basis for explaining the service-oriented paradox.

Third, by integrating a multi-level embeddedness theoretical framework, this study expands the contextualized explanation of servitization transformation and reveal cross-level substitution effects. Existing research often examines the drivers of servitization from a single theoretical perspective, lacking a systematic integration of micro-, meso-, and macro-level factors. Although many scholars have made foundational contributions to embeddedness theory [20,21], research applying this framework to servitization contexts remains scarce. This study incorporates micro-level cognitive embeddedness factors, meso-level relational embeddedness factors, and macro-level environmental embeddedness factors into a unified analytical framework. This integrated framework not only overcomes the limitations of traditional fragmented research but also bridges multi-level analytical perspectives, providing a systematic and contextualized theoretical lens for understanding the strategic choices and implementation of manufacturing firms in complex transformation contexts. More importantly, this study reveals synergy and substitution effects among the three levels of embeddedness through configurational analysis, which represents complex mechanisms that prior regression analyses could not identify. Regarding the substitution effect of relational and institutional embeddedness for cognitive embeddedness, the P3 path (IE*~SF*PI*SC*CC) demonstrates that under favorable institutional conditions, firms can still achieve high SSP through relational embeddedness even when lacking strategic flexibility (~SF). This indicates that relational embeddedness and macro-level institutional embeddedness can compensate for deficiencies in micro-level cognitive embeddedness under certain conditions, thereby challenging the linear thinking of internal capability determinism. Regarding the substitution effect of cognitive embeddedness for relational embeddedness, the C1 path (SF*~SC) demonstrates that when firms possess high strategic flexibility, low supplier concentration (~SC) not only fails to constitute an obstacle but instead promotes SSC by maintaining network openness. This indicates that cognitive embeddedness (strategic flexibility) can substitute for relational embeddedness (supplier dependence) under certain conditions, providing firms with a transformation path that does not rely on specific relational networks. Finally, regarding the double-edged effect of relational embeddedness, the P1 path relies on high supplier concentration (SC), whereas the C1 path relies on low supplier concentration (~SC), indicating that the directional effect of relational embeddedness depends on servitization type and accompanying conditions. This phenomenon of the same condition exerting opposite effects in different configurations can only be revealed through a configurational perspective.

Finally, we introduce a configurational approach to reveal the multiple equivalent pathways of service transformation and expands the methodological boundaries of service-oriented research. Existing research primarily relies on linear methods such as regression analysis, which implicitly assumes the net effect of individual factors while overlooking the multiple interdependencies and complex causal mechanisms among factors, making it difficult to capture the characteristics of multi-condition concurrency and equivalence [15]. Organizational performance is not a linear sum of a single optimal factor, but rather the product of an equivalent combination of multiple factors under specific circumstances [57]. Based on a configurational theory perspective and employing the fsQCA method, this study reveals the multiple equivalent configurations leading to high product-oriented service performance and high customer-oriented service performance, and finds that service performance is not a linear sum of specific conditions, but rather the product of an equivalent combination of multiple factors under specific spatiotemporal circumstances. This finding advances methodological innovation from three dimensions. In terms of equifinality, fsQCA breaks the “necessary condition thinking.” Among the antecedent variables included in this study, servitization transformation does not require any single necessary element, indicating that firms can achieve equivalent performance through different condition combinations, thereby providing diversified strategic choice spaces for firms with varying resource endowments. Regarding causal asymmetry, fsQCA reveals that “failure logic” differs from “success logic.” The causal asymmetry of fsQCA implies that condition combinations leading to high SSP/SSC are not simple mirrors of those leading to low SSP/SSC, which further validates the uniqueness of configurational paths. Concerning boundary conditions, this refinement advances “contextual theorizing.” The six paths identified in this study each correspond to different contextual conditions, pushing servitization research from binary debates of “whether to transform” toward a refined decision-making framework of “which path to choose under what circumstances.” This finding reveals the causal complexity and path diversity of service transformation, propelling service-oriented research from correlation analysis toward the identification of causal mechanisms. It provides a theoretical breakthrough at the methodological level for complex transformation processes while enriching the methodological toolkit of service-oriented research.

5.3. Practical Implications

The research conclusions provide actionable theoretical guidance for managers of manufacturing enterprises in formulating service-oriented decisions.

First, managers must adopt a configuration-based mindset and select a service-oriented path that aligns with the company’s specific profile. Managers should abandon a “one-size-fits-all” approach to service transformation; instead, they should comprehensively assess the company’s resource endowments, network position, and environmental context to identify the degree of alignment between their current conditions and the ideal configuration, and then choose a service-oriented path that suits their specific profile. Specifically, for enterprises with tight supply chain relationships, Services Supporting the Product should be prioritized. By leveraging deep embedded relationships with key suppliers or customers, they can incrementally expand service operations around existing product platforms. For enterprises with strong strategic autonomy, Services Supporting the Customer’s Activity can be explored. Through strategic flexibility, they can respond to customers’ personalized needs and achieve breakthrough service innovation.

Second, enterprises should balance the depth of relational embedding with network openness to avoid the lock-in effects caused by over-reliance. Enterprises should seek an optimal balance through formal contractual safeguards and relationship governance mechanisms, maintaining close embedding with key partners to secure resource support while avoiding the lock-in effects caused by over-reliance. Specifically, for product-oriented servitizing firms, mutually beneficial norms can be established with key suppliers through long-term cooperation agreements, leveraging deep relationship embedding to progressively expand service operations around existing product platforms; For customer-oriented servitizing firms, they should maintain close collaboration with core customers while preserving moderate network openness, diversify risks through a diversified customer portfolio, and achieve proactive service co-creation with customers. Particularly in scenarios of heightened external uncertainty, enterprises should maintain the ability to dynamically adjust network boundaries, reserving room for strategic flexibility.

Third, the government should provide differentiated institutional support, and enterprises should proactively convert policy resources into internal service capabilities. The institutional supply system should be improved to facilitate the transformation of policy resources into enterprise capabilities. Government departments should establish a multi-tiered policy support system, designing differentiated policy toolkits tailored to different service-oriented pathways. For product-oriented servitizing firms, this includes providing supply chain collaborative innovation platforms and industrial alliance support; for customer-oriented servitizing firms, it involves offering digital transformation subsidies and funding for flexible manufacturing capability development. At the same time, the service-oriented standard system, intellectual property protection mechanisms, and tax incentive policies should be improved to reduce the institutional transaction costs of enterprise transformation. Enterprises, in turn, should proactively incorporate these institutional factors into their strategic considerations, actively align with regional industrial internet platform development policies and special funds for service-oriented manufacturing, and establish the ability to perceive and respond to policies. By transforming external institutional support into internal service innovation capabilities, they can achieve a shift from “policy-driven” to “capability internalization.”

5.4. Research Limitations and Future Prospects

First, this study did not segment the sample by industry or corporate characteristics. Although the research subjects are representative of the service transformation of manufacturing enterprises, the study did not differentiate based on the specific manufacturing sector, ownership structure, or life cycle stage of the enterprises. Different types of enterprises may exhibit distinct service-oriented transformation drivers. Future research could incorporate multi-level factors—including individual, organizational, and industrial dimensions—into the analysis. Additionally, cross-national comparative studies could be conducted to explore the similarities and differences in the configurations of service-oriented transformation drivers among manufacturing enterprises in developed and emerging markets, as well as across different cultural contexts, thereby validating the cross-cultural applicability of the research findings.

Second, this study is subject to limitations related to sample selection bias and survivor bias. The sample consists of 110 national-level service-oriented manufacturing demonstration enterprises recognized by the Ministry of Industry and Information Technology of the People’s Republic of China. While this selection ensures the representativeness of the research subjects and the availability of data, it inevitably introduces sample selection bias and survivor bias. The selection criteria for demonstration enterprises inherently include performance thresholds, meaning the sample systematically excludes a large number of manufacturing enterprises that have not yet achieved breakthroughs in service-oriented transformation, have failed in their transformation efforts, or are still in the exploratory stage. Therefore, the six high-performance configuration pathways identified in this study may have differing service-oriented driving mechanisms for various enterprise types, particularly small- and medium-sized enterprises (SMEs) and non-demonstration enterprises. Future research could incorporate non-demonstration enterprises into the analysis, conduct comparative studies of successful and failed enterprises, explore the similarities and differences in the service-oriented antecedent configurations of manufacturing enterprises at different performance levels, and validate the cross-group applicability of the research conclusions.

Third, the selection of antecedent variables has limitations. Numerous factors influence the service transformation of manufacturing firms; this transformation is the result of the synergistic interaction of multiple factors rather than the independent effect of a single factor. The exclusion of many factors from the analytical framework of this study may lead to an incomplete explanation of the mechanisms driving service transformation. Although this study has identified six equivalent pathways to resolve the “servitization paradox” and demonstrated that the paradox largely stems not from servitization itself but from the neglect of the equivalence of condition combinations, limitations in the dimensions of the antecedent variables mean that certain condition combinations may still exist that were not included in the analysis. In specific contexts, these could lead to the servitization paradox arising even under identical condition configurations. For example, factors such as the degree of digital transformation, internationalization experience, entrepreneurial cognitive styles, and industry technological dynamism may moderate or alter the causal efficacy of existing configuration pathways, yet these factors have not been systematically examined in this study. Future research could expand the theoretical boundaries and dimensions of antecedent variables from a multidisciplinary perspective, incorporate the aforementioned potential factors into the analytical framework, and further test and extend the configuration paths identified in this study. This would provide a more comprehensive understanding of the boundary conditions and resolution mechanisms of the service-ization paradox, thereby offering enterprises more precise support for transformation decision-making.

Finally, this study has certain methodological limitations. This study mainly relies on literature research, fsQCA empirical analysis, and case analysis. Although this combination of methods helps reveal complex causal relationships among multiple concurrent factors, many research methods from multiple theoretical perspectives, analytical levels, and interdisciplinary fields have not been incorporated into this study’s framework. In terms of causal inference, although this study adopts a time lag design (pairing antecedent variables at t-1 with outcome variables at t) to alleviate reverse causality problems, fsQCA is essentially still a cross-sectional analytical technique that identifies coexistence relationships between condition combinations and outcomes at specific time points, rather than dynamic evolutionary processes. Specifically, fsQCA cannot model the cumulative effects of antecedent conditions over time, path dependence, and the sequential order of transformation stages, nor can it exclude alternative explanations where outcome variables inversely shape antecedent conditions. Therefore, although the time lag design logically ensures that antecedent variables are observed prior to outcome variables, this temporal sequence evidence remains weaker than strict dynamic process modeling. Given that this study has panel data structure from 2015 to 2022, future research could attempt to apply temporal QCA (tQCA) or multi-period fsQCA methods to explicitly incorporate the time dimension into configurational analysis, examining the stability and evolutionary patterns of paths. Future research could also conduct deep mutual verification and complementary integration of quantitative and qualitative research to perform multi-dimensional cross-validation and robustness tests on findings, thereby establishing more solid causal temporal evidence. Regarding measurement refinement, future research could attempt to add more detailed antecedent variable measurement indicators, for example, subdividing the measurement indicators of institutional environment into manufacturing servitization-specific policies to achieve more precise external measurement standards.