Introduction
In the global landscape of the manufacturing industry, governments and enterprises recognize the significance of accelerating the servitization process to foster high-quality industry development (Salonen, Saglam & Hacklin, Reference Salonen, Saglam and Hacklin2017). By expanding the range of value-added services built upon their products, they aim to cater to the diverse and customized needs of customers while also propelling the industry toward the higher echelons of the industrial chain (Vandermerwe & Rada, Reference Vandermerwe and Rada1988). However, substantial variation exists in the degree of servitization across different countries. The 2016 Global Input−Output Database reveals that developed economies generally exhibit higher levels of servitization in their manufacturing sectors, with an average of 9.9%. Notably, there are significant disparities among countries, with South Korea (0.7%) and Cyprus (1.4%) having the lowest levels, while Luxembourg (42.5%), the Netherlands (20.1%), and Sweden (16%) boast the highest levels (Friesenbichler & Kügler, Reference Friesenbichler and Kügler2022). China’s manufacturing industry still faces a low level of servitization, particularly in critical areas such as R&D and design services, system integration services, overall solutions, and personalized customization services. This deficiency in high-end service industry competitiveness has emerged as a prominent factor hindering the development of the real economy, thereby impacting overall production efficiency and the innovation and growth capabilities of the industrial system as a whole. In response, the national government has issued a series of policy opinions in recent years to promote the deep integration of advanced manufacturing and modern service industries. One of the key directions for the future development of China’s manufacturing industry is the adoption of a service-oriented manufacturing approach. Given the intensifying global industrial competition, conducting comprehensive research on the servitization path of manufacturing enterprises holds substantial practical significance for constructing a modern industrial system.
Currently, the academic community has extensively studied the servitization of the manufacturing industry, focusing on its impact on various factors such as industrial upgrading (Huang, Ma, Xietian & Huang, Reference Huang, Ma, Xietian and Huang2024), enterprise value chain enhancement (Dai, Zhao, Yang & Chen, Reference Dai, Zhao, Yang and Chen2024), technological progress (Liu & Hua, Reference Liu and Hua2024), and carbon emission intensity (Tang, Zhu, Luo & Duan, Reference Tang, Zhu, Luo and Duan2022). However, relatively limited attention has been given to the factors that promote or hinder the transformation of manufacturing enterprises toward servitization (Khanra, Dhir, Parida & Kohtamäki, Reference Khanra, Dhir, Parida and Kohtamäki2021). Senior executives, as the key decision-makers in corporate strategic planning, possess significant influence over the adjustment and implementation of various strategies. This influence has been affirmed by extensive theoretical research over the past few decades (Finkelstein, Hambrick & Cannella, Reference Finkelstein, Hambrick and Cannella2009; Hambrick & Mason, Reference Hambrick and Mason1984). Consequently, it is essential to investigate whether and how executive factors impact servitization, which requires further attention from the academic community (Ahamed, Inohara & Kamoshida, Reference Ahamed, Inohara and Kamoshida2013). However, a literature search reveals a scarcity of relevant studies. Existing studies have provided insights into how executive factors influence enterprise servitization from different perspectives. For instance, based on a case study, Nie and Kosaka (Reference Nie and Kosaka2014) found that executives expedited the servitization process by fostering shared vision creation with employees and employing purposeful leadership styles. Kim and Toya (Reference Kim and Toya2019) examined the impact of CEO leadership styles on enterprise servitization within the institutional context of Japanese manufacturing firms. They discovered that a charismatic leadership style promotes servitization by enhancing employees’ identification with organizational goals. While these studies shed light on the influence of executive factors on enterprise servitization, they do have limitations. First, they primarily rely on case analyses of specific enterprises and industries, lacking empirical analysis based on long-term panel data, which hinders the provision of more general explanations regarding how executive characteristics affect manufacturing enterprise servitization (Khanra et al., Reference Khanra, Dhir, Parida and Kohtamäki2021). Second, there is a dearth of research on the role of executive factors in corporate servitization within the Chinese context, limiting the ability to offer insights for improving China’s existing policy framework and harnessing the unique role of senior management teams to expedite the servitization process.
This paper examines the influence of executive cognition on enterprise servitization by analyzing data from listed manufacturing enterprises in Shanghai and Shenzhen spanning the years 2010–2020. The study aims to uncover the mechanism behind this influence as well as explore potential heterogeneity in the context. According to the upper echelons theory (Finkelstein et al., Reference Finkelstein, Hambrick and Cannella2009), executives’ cognition plays a pivotal role in organizational performance, particularly through their strategic and lasting influence on the organization. In this paper, TMT Service Cognition refers to the senior management team’s stable understanding and perception of service activities, which is shaped by their knowledge and experience. This cognition is of significant importance as it promotes a service-oriented approach and serves as a leading factor in resource allocation and organizational structure within enterprises.
The paper contributes in several ways. First, it addresses a gap in the existing research by examining whether and how executive factors impact the servitization of manufacturing enterprises within the Chinese context. This helps to fill the shortcomings of previous studies that have primarily focused on developed countries. Second, by utilizing long-term microdata from listed companies, the paper conducts an empirical analysis to investigate the impact of senior management factors on enterprise servitization and its underlying mechanisms. This approach overcomes the limitations of previous research that heavily relied on case studies. Lastly, a key theoretical contribution of this study lies in establishing executive service cognition as the core explanatory variable for enterprise servitization. Departing from prevailing paradigms that prioritize demographic characteristics (e.g., age, education, tenure, functional experience) or leadership styles (e.g., transformational/transactional behaviors) as primary drivers – as documented extensively in extant literature – this research advances the analytical lens to the cognitive stratum of strategic decision-makers. By exploring these aspects, this research aims to provide a deeper understanding of how executive factors, particularly executive service cognition, influence the servitization process of manufacturing enterprises.
The paper is structured as follows. The second section provides a theoretical analysis and presents the research hypotheses. The third section focuses on the research design, outlining the methodology used to test the proposed model and describing the variables under consideration. The fourth section presents the results of the benchmark empirical analysis and performs robustness tests. In the fifth section, the paper delves into the mechanism analysis results, shedding light on the underlying mechanisms through which executive cognition influences enterprise servitization. Additionally, it explores potential heterogeneity in these relationships from various dimensions. Finally, the last section encompasses the conclusion and policy implications.
Theoretical analysis and research hypotheses
Executive service cognition and enterprise servitization
Executives’ cognitive frameworks – how they perceive the external environment, interpret strategic goals, and make decisions – play a central role in shaping corporate strategy and its execution (Porac & Thomas, Reference Porac, Thomas, Pettigrew, Thomas and Whittington2002). Recent scholarship has moved beyond merely cataloging executive traits to exploring how cognitive diversity and capacity actively shape organizational outcomes. However, the literature reveals both consensus and contention regarding the strategic value of executive cognition.
One area of ongoing debate centers on cognitive diversity within top management teams. While some studies argue that diversity fosters innovation and adaptability, the relationship is not universally positive. For example, Younis (Reference Younis2019) identifies an inverted U-shaped relationship between cognitive diversity and firm performance, moderated by organizational culture, suggesting that beyond a certain point, diversity may generate diminishing returns. Similarly, Qu, Zhao, Cao and Cao (Reference Qu, Zhao, Cao and Cao2021) note that while cognitive diversity enhances knowledge sharing and innovation potential, its impact is constrained by hierarchical or differential internal climates. Other studies, such as Kanchanabha and Badir (Reference Kanchanabha and Badir2021), affirm the benefits of cognitive variety on flexibility and innovation, yet fail to account for how contextual variables may disrupt this positive link. This inconsistency highlights a need for further examination into the boundary conditions under which cognitive diversity contributes to strategic outcomes.
Another thread in the literature focuses on the role of cognitive capacity – executives’ ability to interpret external signals and make strategic judgments under uncertainty. Helfat and Peteraf (Reference Helfat and Peteraf2015) emphasize that such capabilities are central to dynamic decision-making and long-term performance. More recently, Yang, Peng and Ge (Reference Yang, Peng and Ge2022) argue that heightened executive awareness of technological trends boosts innovation sustainability, contingent on external regulatory environments. Similarly, Wand and Xie (Reference Wand and Xie2021) show that managers’ perceptions of digital opportunities versus risks lead to divergent outcomes in technology adoption. These findings collectively suggest that cognition is not static but shaped by contextual stimuli, such as regulation and environmental volatility.
Despite this growing body of work, few studies have critically examined how executive cognition specifically influences servitization – a strategic transformation process in manufacturing. Given that servitization demands new business logics and risk perceptions, executive cognitive framing becomes even more consequential. This study seeks to extend the literature by investigating how executives’ service cognition – how they conceptualize and value service transformation – affects strategy implementation and servitization performance. Accordingly, we propose Hypothesis 1:
Hypothesis 1: Executive service awareness enhances the level of servitization in manufacturing enterprises.
Executive cognition, human capital, and enterprise servitization
The cognitive perspective offers a powerful lens for understanding how executive characteristics shape strategic behaviors (Miller, Chiu, Wesley II, Vera & Avery, Reference Miller, Chiu, Wesley II, Vera and Avery2022). Yet a critical gap remains: how do executives’ perceptions translate into enterprise-level strategic action, particularly in complex transformations such as servitization? While prior studies have acknowledged the role of executive cognition, they have often overlooked the mechanisms through which this cognition exerts its influence. This paper addresses that gap by proposing that human capital accumulation serves as a key mediating mechanism between executive service cognition and servitization outcomes.
Although Collings and Mellahi (Reference Collings and Mellahi2009) argue that timely and strategic acquisition of human resources can drive change, this insight has rarely been applied to the domain of servitization in manufacturing. The shift from product-oriented to service-oriented strategies is not merely operational – it requires a fundamental reconfiguration of human resource capabilities. Yet much of the literature assumes a passive role for HR, focusing on structure rather than capability development.
More critically, while scholars such as Levitt (Reference Levitt1983) and Jovanovic, Raja, Visnjic and Wiengarten (Reference Jovanovic, Raja, Visnjic and Wiengarten2019) acknowledge the organizational demands of servitization – such as customer experience, data analytics, and after-sales service – they often treat human capital as a secondary enabler rather than a central strategic asset. This downplays the complex skill sets required to support long-term service contracts and hybrid value propositions.
A few studies have challenged this narrow view. For example, Baines, Lightfoot, Smart and Fletcher (Reference Baines, Lightfoot, Smart and Fletcher2013), in their analysis of leading service-transformed firms like Xerox, identified a need for adaptable, relationship-driven, and innovation-oriented employee skillsets. Similarly, Opazo-Basáez, Vendrell-Herrero and Bustinza (Reference Opazo-Basáez, Vendrell-Herrero, Bustinza and Liu2019) contend that the strategic role of human capital in enabling servitization remains underexplored in both theory and practice. Their empirical evidence from multinational manufacturers suggests a clear link between high-quality human capital and both advanced servitization and stronger financial performance – implying that talent capabilities may be as crucial as technological readiness.
This study builds on that critique by examining how executives’ service cognition may shape their approach to human capital investment and, in turn, how such investments facilitate servitization. In doing so, we extend the literature by integrating cognitive and human capital perspectives into a unified framework for understanding strategic transformation. Accordingly, we propose Hypothesis 2:
Hypothesis 2: Executive service awareness can facilitate the servitization of manufacturing enterprises by enhancing investment in innovation and accumulating human capital.
Centralization of power and servitization of enterprises
Prior research has emphasized the moderating role of executive power concentration in shaping strategic decisions (Finkelstein, Reference Finkelstein1992), yet findings remain fragmented and at times contradictory. While some studies highlight how concentrated power can accelerate strategy implementation, others point to diminishing returns or even counterproductive outcomes, particularly in the context of long-term investments.
For example, green innovation strategies – despite their long-term sustainability benefits – often entail substantial upfront costs. These strategies typically lack short-term payoffs and may deter risk-averse executives. Aibar-Guzmán and Frías-Aceituno (Reference Aibar-Guzmán and Frías-Aceituno2021) found that CEO power significantly influences the adoption and performance of green initiatives, yet their analysis raises questions about whether such power encourages innovation or simply reflects a CEO’s discretion to allocate resources toward noncore goals. Similarly, Li, He, Hu and Zhou (Reference Li, He, Hu and Zhou2023) observed that executives with environmental expertise exert greater influence over firms’ environmental responsibility, but the broader mechanisms behind this relationship remain underexplored.
The same ambiguity applies to corporate social responsibility (CSR). While CSR investments are increasingly viewed as strategic, their financial returns are typically delayed. Jiraporn and Chintrakarn (Reference Jiraporn and Chintrakarn2013) demonstrated an inverted U-shaped relationship between CEO power and CSR, suggesting that excessive concentration may lead to diminishing strategic alignment or myopic overreach. These findings reflect a broader debate: Is executive power a strategic enabler or a source of unchecked autonomy that distorts long-term priorities?
Further complicating this picture is evidence from the innovation domain. Wei, Liu and Yue (Reference Wei, Liu and Yue2015) suggest that power equality among senior executives fosters collaboration but may simultaneously suppress bold innovation. This tension is especially salient in the servitization of manufacturing, where heavy upfront investments in services often fail to yield quick financial returns – a phenomenon known as the ‘servitization paradox’ (Dmitrijeva, Schroeder, Bigdeli & Baines, Reference Dmitrijeva, Schroeder, Bigdeli and Baines2022; Gebauer, Fleisch & Friedli, Reference Gebauer, Fleisch and Friedli2005). In such scenarios, concentrated executive power might either sustain long-term strategic commitment or prematurely curtail servitization efforts in pursuit of short-term gains.
Despite these insights, existing literature lacks consensus on how executive power affects the implementation of complex, uncertain strategies like servitization. This study addresses that gap by examining how power structures within executive teams influence servitization outcomes, particularly under conditions of delayed payoff. In doing so, it contributes to the ongoing debate about the double-edged nature of executive power in strategic transformation. Accordingly, we propose Hypothesis 3:
Hypothesis 3: In enterprises with a higher concentration of power, senior management service cognition is more effective in promoting enterprise servitization.
Regional digital infrastructure and enterprise servitization
Beyond firm-level determinants, growing evidence suggests that the urban context in which a manufacturing enterprise operates can significantly shape its trajectory toward servitization. However, while prior studies have documented positive associations between city-level features and servitization, the mechanisms remain under-theorized, and the relative importance of different urban enablers is still debated.
Hu and Gu (Reference Hu and Gu2020) show that improved administrative efficiency at the city level, when aligned with manufacturing servitization, can enhance resource allocation and reduce urban inefficiencies. While this supports the notion that institutional quality facilitates transformation, their findings raise the question of whether efficiency gains stem primarily from policy improvements or from the synergistic effects of digital and service integration.
A parallel line of research underscores the critical role of digital infrastructure in enabling servitization (Paschou, Rapaccini, Adrodegari & Saccani, Reference Paschou, Rapaccini, Adrodegari and Saccani2020; Zhan, Zhang, Zhang, Zhao & Wang, Reference Zhan, Zhang, Zhang, Zhao and Wang2023). Yet, much of this literature treats digitalization as a background enabler rather than a structural constraint or facilitator that varies by locality. This is problematic because servitization demands both technological capabilities and ecosystem-level connectivity – factors that are not evenly distributed across cities. For instance, digital technologies not only enable firms to capture and respond to customer preferences in real time – as seen in personalized industries like music (Parry, Bustinza & Vendrell-Herrero, Reference Parry, Bustinza and Vendrell-Herrero2012) – but also support the architectural shifts required for service-based business models. These include reorganizing internal operations, coordinating with ecosystem partners, and delivering scalable digital services (Burton, Story, Zolkiewski & Nisha, Reference Burton, Story, Zolkiewski and Nisha2024; Coreynen, Matthyssens, Vanderstraeten & van Witteloostuijn, Reference Coreynen, Matthyssens, Vanderstraeten and van Witteloostuijn2020; Marić, Pejić Bach & Gupta, Reference Marić, Pejić Bach and Gupta2024).
However, a critical tension exists: while the benefits of digital servitization accrue at the firm level, the infrastructure needed to enable it often depends on city-wide investments. This raises a key strategic dilemma for manufacturers: whether to internalize digital capabilities or rely on the broader urban digital infrastructure. Although prior studies have emphasized the cost efficiencies of city-level platforms, few have explored how this shared infrastructure shapes firms’ willingness or ability to invest in servitization strategies.
Addressing this gap, our study contributes a more nuanced understanding of how local digital ecosystems condition the relationship between firm-level digital readiness and servitization outcomes. We examine how city-level digital infrastructure facilitates – not just enables – firms’ transition from product providers to service orchestrators. In doing so, we offer new insights into the interplay between firm strategy and urban digital environments. Based on this synthesis, we propose Hypothesis 4:
Hypothesis 4: A higher level of digital infrastructure in a region enhances the positive impact of senior executives’ service cognition on enterprise servitization.
Methodology
Econometric models
Based on the theoretical analysis presented in this paper, we have formulated a regression model to empirically examine the influence of executives’ service cognition on enterprise servitization. The model is as follows:
\begin{equation}Servic{e_{i,t}} = \alpha + \beta TMTS{C_{i,t}} + \gamma Control{s_{i,t}} + {\varepsilon _{i,t}}\end{equation}In this model, the explanatory variable Service represents the extent of servitization within the enterprise. It is measured as the proportion of service business income relative to the total main business income of the enterprise. The core explanatory variable TMTSC reflects the service cognition of enterprise executives, and the coefficient β quantifies the impact of TMTSC on enterprise servitization. Additional enterprise-level control variables are included as Controls. To account for the potential effects of macro and industry factors on enterprise servitization, we have incorporated fixed effects for province, industry, and year into the regression model.
Data description
This paper focuses on a sample of China’s manufacturing listed companies from 2007 to 2020. The data for this research are obtained from the following sources: (1) The core variables reflecting executives’ service cognition are derived from annual reports of the listed companies during the corresponding years. These variables are obtained through text analysis, following the methodology employed by Yang, Wang, Zhou and Jiang (Reference Yang, Wang, Zhou and Jiang2019). (2) The core variables representing the level of servitization within the listed enterprises are manually calculated by carefully examining the detailed main business income information retrieved from the Wind database. (3) Additional data pertaining to various enterprise characteristics are also sourced from the Wind database. In accordance with Chen, Han and Han (Reference Chen, Han and Han2022), the level of urban digital infrastructure is quantified by the number of mobile phone users per 100 people. The specific data points are obtained from the China Urban Statistical Yearbook for the corresponding year.
Variable description
Dependent variables
The measurement of the degree of servitization in the manufacturing industry lacks consistency. Some studies have used the consumption coefficient method with the World Input–Output Database to calculate enterprise servitization (Dai et al., Reference Dai, Zhao, Yang and Chen2024; Liu & Hua, Reference Liu and Hua2024; Liu & Wang, Reference Liu and Wang2016; Roos, Reference Roos, Agarwal, Selen, Roos and Green2015). In contrast, other studies (Xiao, Reference Xiao2023) have relied on questionnaires and field interviews to assess servitization levels. The former method is generally employed for macro-level research, analyzing the impact of servitization on export competitiveness, carbon emission intensity, and other factors. However, the latter method provides enterprise-level survey data, albeit with a relatively limited sample size.
Following the practices of Zhao (Reference Zhao2021) and Zhang, Zhan, Yin and Huang (Reference Zhang, Zhan, Yin and Huang2021), this paper measures enterprise servitization using two variables: service business income (Serrev) and its proportion to total main business income (Serrevr). Data on the top 5 components and income of listed companies’ main businesses are obtained from the Wind database. The process involves several steps: First, service categories related to the enterprise’s main business are identified according to the national economic industry classification standard (GB/T 4754-2011), including technical support, consulting, training, research and development, and more. Next, the business income corresponding to these service-related categories within the manufacturing enterprise’s main business composition is considered as service income. The higher the proportion of this service income to total main business income, the greater the degree of enterprise servitization. To account for cases where listed companies do not have service-related businesses, a binomial dummy variable (Serrevdum) is constructed. It takes a value of 1 when the enterprise includes service business as part of its main operations and 0 otherwise. This variable is used for robustness testing.
Independent variables
The annual report of a company plays a crucial role in reflecting the understanding of senior executives. This study focuses on analyzing the ‘Management Discussion and Analysis’ section of annual reports from China’s listed manufacturing enterprises between 2000 and 2020. Our objective is to construct the Senior Executives’ Service Awareness (TMTSC) index by utilizing natural language processing techniques to extract service-related descriptions. To achieve this, we create a dictionary of executive service-related terms.
Since there is no specialized dictionary for manufacturing service-oriented terms and a lack of uniform classification and conceptual systems for derivative services in manufacturing enterprises, we conducted searches on the websites of the Chinese government, the National Development and Reform Commission, the Ministry of Industry and Information Technology, and other relevant ministries and commissions. Additionally, we referred to significant domestic literature from recent years (Kohtamäki, Parida, Oghazi, Gebauer & Baines, Reference Kohtamäki, Parida, Oghazi, Gebauer and Baines2019; Li, Ma & Huang, Reference Li, Ma and Huang2015; Lin, Luo, Ieromonachou, Rong & Huang, Reference Lin, Luo, Ieromonachou, Rong and Huang2019; Luo, Rao & Chen, Reference Luo, Rao and Chen2021). Following the approach of Baines and Lightfoot (Reference Baines and Lightfoot2013), we categorized services into basic services (fdsv) and advanced services (adsv), resulting in the creation of a manufacturing servitization term dictionary (Table 1).
Table 1. Construction of executive servitization cognition index
Next, we conducted text analysis on the relevant sections of the annual reports. Using the ‘Jieba’ Chinese thesaurus in Python, we expanded the identified service-oriented keywords and employed machine learning methods to analyze the ‘Management Discussion and Analysis’ part of the annual reports of listed companies. This allowed us to calculate the frequency of the identified keywords in the respective annual reports of each enterprise. Subsequently, we derived the Senior Executives’ Service Awareness index. Considering the varying lengths of annual reports and the ‘Management Discussion and Analysis’ sections across different enterprises, we used the frequency proportion of service-oriented keywords within each enterprise as a measure of senior executives’ service awareness. To account for the right-skewed distribution of the data, we applied a logarithmic transformation, resulting in three executive service cognition indicators (Lnsv, Lnfdsv, and Lnadsv).
Control variables
This study draws upon existing research on senior management teams and the servitization of the manufacturing industry (Bhaskarabhatla, Cabral, Hegde & Peeters, Reference Bhaskarabhatla, Cabral, Hegde and Peeters2021; Zhao, Reference Zhao2021). The primary focus is on three categories: characteristics of the senior management team, characteristics of listed companies, and characteristics of the industry. The first category pertains to governance structure variables. It includes the virtual variable (Duality) representing whether the chairman also serves as the CEO. A value of 1 indicates the chairman is the CEO, while 0 indicates otherwise. Equity concentration is measured as the ratio of shares held by management to the total share capital (Tmtshhratio), and the shareholding ratio of the largest shareholder (Topshhratio). Board size (Lnbdsize) is expressed as the logarithm of the number of directors plus 1.
The second category encompasses business variables. The market value of the enterprise is measured using the TobinQ ratio (Tobinq). The debt-to-asset ratio (Lev) represents the ratio of total liabilities at the end of the year to the total assets at the end of the year. Profitability (ROA) is calculated as the ratio of net profit to the average balance of total assets. The third category focuses on innovation investment variables. The scale of enterprise human capital (Lnmstabv) is expressed as the logarithm of the total number of employees with a master’s degree or above plus 1. The intensity of R&D investment (RDint) is computed as the ratio of R&D investment to total operating income. Lastly, the fourth category comprises other firm, industry, and regional control variables. Enterprise size (Lnsize) is expressed as the logarithm of total assets plus 1. Enterprise age (Age) is determined by subtracting the year of establishment from the corresponding year. Government-related variables (Lngov) are represented by the current government subsidy amount plus 1. The degree of competition in the industry (Lerner) is indicated by the Lerner index of the respective industry. Urban digital infrastructure (Lnintnet) is expressed as the logarithm of the number of urban Internet users plus 1. To maintain consistency and robustness, the top and bottom 1% of observations were trimmed for the relevant continuous variables.
Descriptive statistics of variables
Table 2 displays the descriptive statistics for the main variables. The data indicates that the average word frequency for annual report service (Lnsv), basic service (Lnfdsv), and senior service (Lnadsv) – reflecting the cognitive perspective of senior executives – are 2.444, 1.435, and 2.120, respectively. This suggests that within the sampled enterprises, senior executives pay more attention to advanced service businesses like design, customization, and supply chain, compared to basic services such as logistics, maintenance, and assembly. The mean values for Lnserrev and the proportion of service revenue in total operating income (Serrevr) are 6.974 and 5.358, respectively, with standard deviations of 9.166 and 12.769. These figures indicate that, overall, the level of service orientation among the sampled enterprises is relatively low, and there are significant variations among different companies. Additionally, the firm-level control variables, such as the integration of two positions, board size, and profitability, align closely with prior studies. Looking at Fig. 1, it is evident that from 2007 to 2020, the proportion of service business revenue and the number of service-oriented enterprises in China experienced a substantial upward trend. The proportion of service business revenue increased from 3.26% in 2007 to 7.48% in 2017, with a notable rise of 3.46 percentage points between 2013 and 2017. Similarly, the proportion of service-oriented enterprises grew from a mere 9.31% in 2007 to 50.01% in 2020, representing a significant increase of 40.7 percentage points. The surge in listed service firms (2020: 50% vs 2007: 9%) partly reflects China’s registration-based IPO reform accelerating tech listings. To address structural shifts, we conducted size-stratified regressions (see the ‘Heterogeneity test of enterprise scale’ section), confirming that the impact of executives’ service cognition on servitization is more pronounced in large enterprises.
Figure 1. Proportion of service business revenue and servitized enterprises from 2007 to 2020 in listed companies.
Table 2. Description of main variables and descriptive statistics
Empirical analysis and results
Benchmark regression results
Table 3 presents the overall test results regarding the impact of executive perception on the servitization of manufacturing firms. In column (1), the regression results include only the core explanatory variable (Serrevr). It is evident that the frequency of service words in the annual report, represented by the coefficient of Lnsv (3.0985), has a significant positive impact on the servitization of manufacturing enterprises (Lnserrev). Moving to column (2), the explanatory variable is the proportion of enterprise service business revenue. Here, we find that senior executives’ service awareness, with a coefficient of 2.5059, is statistically significant at the 1% level. In column (3), the explanatory variable is replaced with a binary variable (Serdum), indicating enterprise servitization. In this case, the executive service perception coefficient is 0.1620, which also displays a significant positive effect. Columns (4)–(6) present the regression results after including the main control variables. It is worth noting that the coefficients for executives’ service cognition remain statistically significant at the 1% level, with values of 3.0266, 2.4131, and 0.1592, respectively. Regarding the control variables, among the sampled listed manufacturing enterprises, higher levels of servitization are associated with a greater number of companies receiving government subsidies and larger enterprise assets. Conversely, a higher employee count is linked to lower levels of servitization. The impact of the enterprise asset-liability ratio and board size is not statistically significant.
Table 3. Executive cognition and servitization of manufacturing enterprises
*** Note: The robust standard errors are denoted inside the parentheses, with **, and * indicating the significance levels of 1%, 5%, and 10%, respectively. The subsequent tables follow the same convention.
Addressing endogeneity
Constructing the instrumental variable Lewbel_IV
This study constructs instrumental variables (IVs) following Lewbel (Reference Lewbel1997) and Li, Li and Chen (Reference Li, Li and Chen2020) methodologies to address endogeneity concerns. By leveraging data heteroscedasticity to generate internally valid IVs, it overcomes the constraints of acquiring traditional exogenous instruments, thereby enhancing the credibility of causal identification. It utilizes the cubic difference between the mean difference of the enterprise servitization index and the industry secondary coded servitization index as the IV (Lewbel_IV). Table 4, Panel A (1), displays the impact of these IVs on the core explanatory variable (Lnsv) in the first stage. The coefficient is 0.0062, which is significantly positive at the 1% level. Moreover, the first-stage F-statistic exceeds 10, indicating that there are no concerns regarding weak IVs. Moving to column (2), the IV (Lewbel_IV) does not exhibit a significant effect on the dependent variable, thereby satisfying the second requirement of instrumental variable testing for exogeneity. Columns (3)–(5) present the results of the second-stage IV analysis, revealing coefficients of 5.7341, 5.9822, and 0.2996, respectively, for the executive service cognitive variable (Lnsv). These coefficients remain significant at the 1% level, demonstrating consistency with the benchmark regression results.
Table 4. Results of instrument variables and Heckman two-stage model
*** The robust standard errors are denoted inside the parentheses, with **, and * indicating the significance levels of 1%, 5%, and 10%, respectively.
Heckman two-stage model test
This study adopts the Heckman two-stage modeling approach (Di, Luo & Jiang, Reference Di, Luo and Jiang2020) to mitigate sample selection bias, correcting for endogeneity arising from nonrandom sampling and thereby enhancing the consistency of parameter estimates. Specifically, we investigate the factors influencing the as-a-service transformation (Serdum) in the first stage, which is influenced by various enterprise characteristics, government support, and the macro-environment. A set of control variables is included in the binary Probit regression. In the second stage, we include the inverse Mills ratio (IMR) estimated from the first stage for regression analysis. Referring to the results presented in columns (6)–(8) of Panel B in Table 4, we find that the coefficients for the three enterprise servitization indicators (Lnserrev, serrevr, and serdum) are 3.0198, 2.4092, and 0.1590, respectively. All of these coefficients are significant at the 1% level, consistent with the benchmark regression results, and provide further support for our findings.
Robustness tests
Testing for reverse causality test
To address the potential issue of reverse causality, wherein the improvement in enterprise servitization level may be driven by the enterprise’s own high servitization level (e.g., a higher servitization degree in the previous period leading to a higher degree in the current period), this study employs the dynamic panel model Generalized Method of Moments (GMM) for testing. The regression model includes lagged values (1 and 2 periods behind) of service business revenue amount (Lnserrev) and service business revenue proportion (serrevr) as control variables. Table 5 presents the results in columns (1)–(4) of Panel A, indicating that the coefficients of the core explanatory variables are 3.0210, 3.0400, 2.4160, and 2.4201, respectively. These coefficients are significant at the 1% level, confirming the validity of the study’s conclusions.
Table 5. Results of dynamic panel model, alternative independent variables, and omission of variables
*** The robust standard errors are denoted inside the parentheses, with **, and * indicating the significance levels of 1%, 5%, and 10%, respectively.
Replacement of independent variables
In this study, the service keywords from annual reports of listed companies are categorized into two groups: advanced services and basic services. Through text analysis, the frequencies of advanced service words (Lnadsv) and basic service words (Lnfdsv) are derived and used as substitute variables for the core explanatory variables (Lnsv) in the regression analysis. The results in Table 5, specifically columns (5)–(10) of Panel B, indicate significantly positive coefficients at the 1% level for the two alternative explanatory variables: 2.8344, 2.3455, 0.1491, 3.7722, 3.0564, and 0.1975, respectively. These results align with the previous findings. Notably, across the three different dependent variables, the coefficients associated with the frequency of advanced service words are consistently higher than those of basic service words. This suggests that a stronger senior service cognition among executives is associated with a greater proportion of enterprise service business revenue, a higher proportion of service business revenue, and a higher likelihood of servitization transformation.
Testing for omitted variables
Despite including a range of control variables in the benchmark regression, it is possible that important factors influencing enterprise servitization transformation have been overlooked. To address this concern, the regression model incorporates interactive fixed effects of year–city and industry–city, effectively controlling for time-varying macroenvironmental and heterogeneous industry factors at the city level. The results in Table 5, specifically columns (11)–(13) of Panel C, demonstrate significant coefficients for the executive cognitive variables: 2.9469, 2.3485, and 0.1556, respectively, at the 1% level. These findings are consistent with the benchmark regression results, providing further support to the study’s conclusions.
Further analysis
Analysis of mediation mechanisms
This paper argues that executive service cognition influences the level of enterprise servitization through two key mechanisms: human capital investment and R&D investment. On one hand, when executives have a higher awareness of servitization, it encourages companies to increase their investment in human capital related to servitization, thereby enhancing the level of enterprise servitization. On the other hand, the transition to servitization requires manufacturing enterprises to intensify their R&D investment to support the development of the service business. Building upon Eq. (1), we construct the following mediating effect model:
\begin{equation}Mediatio{n_{i,t}} = {\beta _0} + {\beta _1}TMTS{C_{i,t - 1}} + {\beta _2}Control{s_{i,t - 1}} + Yea{r_{t - 1}} + Secto{r_{t - 1}} + {\varepsilon _{i,t}}\end{equation}
\begin{align}Servic{e_{i,t}} &= {\lambda _0} + {\lambda _1}TMTS{C_{i,t - 1}} + {\lambda _2}Mediatio{n_{i,t - 1}} + {\lambda _3}Control{s_{i,t - 1}} + Yea{r_{t - 1}} \nonumber\\
&\quad + Secto{r_{t - 1}} + {\varepsilon _{i,t}}\end{align}In this model, the mediation variable includes human capital investment (Lnmstabv), measured by the number of employees with a master’s degree or above, and R&D investment intensity (RDint), expressed as the proportion of R&D investment to operating income. To examine the mechanism of human capital accumulation, we first analyze the impact of executive service cognition on enterprise human capital investment. As shown in column (1) of Panel A in Table 6, the coefficient of executive service cognition (LNSV) is 2.3154, significantly positive at the 5% level. This indicates that executive service cognition promotes the accumulation of high-quality human capital within enterprises.
Table 6. Mechanism test results
*** The robust standard errors are denoted inside the parentheses, with **, and * indicating the significance levels of 1%, 5%, and 10%, respectively.
Next, we explore the influence of human capital accumulation on the servitization level of enterprises. The results, presented in columns (2)–(4), reveal significantly positive coefficients for human capital investment at the 5% level. These findings confirm that executive service cognition drives investment in human capital, thereby improving the level of enterprise servitization. Furthermore, we investigate the mediating role of R&D investment (RDint). The regression results, shown in Panel B of Table 6, demonstrate that in column (5), the coefficient of executive service cognition is 1.0269, significantly positive at the 1% level. This illustrates that executive service cognition significantly promotes enterprises to increase their R&D investment. Columns (6)–(8) further reveal that an increase in enterprise R&D investment intensity substantially enhances the level of enterprise servitization. These findings support the established mechanism in the theoretical analysis, indicating that executive service cognition drives enterprise R&D innovation, leading to an elevated level of servitization.
Analysis of moderating effects
This paper delves into exploring the moderating effects of executive power concentration and urban Internet penetration on the relationship between executive cognition and the transformation of enterprise servitization. First, to examine the moderating effect of executive power concentration, two variables were introduced: the proportion of shares held by the largest shareholder (Topshhratio) and the proportion of management shares in total equity (Tmtshhratio). Interaction terms between executive service cognition and these variables (Lnsv × Topshhratio, Lnsv × Tmtshhratio) were included in the regression model. The results, shown in Table 7, Panel A, indicate significant positive coefficients of the interaction terms (0.0166, 0.0104, 0.0008, 1.3141, 2.2083, and 0.0442) across three different explanatory variables. These findings suggest that executive power concentration plays a significant positive moderating role in the relationship between executive service cognition and the level of enterprise servitization.
Second, the paper measures the urban Internet penetration rate, represented by the number of Internet users in the city where the enterprise is located. The interaction term between executive service cognition and Internet penetration rate (Lnsv × Lnintnet) was added to the regression model. The results displayed in Table 7, Panel B, reveal significantly positive coefficients for the interaction terms (0.3554, 0.1188, and 0.0191). This indicates a positive moderating effect of Internet penetration rate on the relationship. Furthermore, the paper investigates the moderating effect of industry competition by introducing the interaction term between the industry Lerner index and executive service cognition. The results demonstrate a generally positive moderating effect of industry competition. Specifically, when the binomial value of enterprise servitization is used as the explanatory variable, the moderating effect is significantly positive. This suggests that industry competition can significantly amplify the impact of senior executives’ cognition on whether the enterprise adopts servitization. However, when the service business revenue and its proportion are considered as the dependent variables, the moderating effect of industry competition is not significant.
Table 7. Results of moderating effect test
*** The robust standard errors are denoted inside the parentheses, with **, and * indicating the significance levels of 1%, 5%, and 10%, respectively.
Heterogeneity analysis
Heterogeneity test of position integration
It is anticipated that integrating the two positions will amplify the positive influence of senior executives’ service cognition on the level of enterprise servitization. In the case of manufacturing enterprises, the transformation toward a service-oriented approach is a strategic endeavor that significantly impacts their medium and long-term development. It involves a multitude of foreseeable and unforeseeable risks. Without a high level of consensus at the top decision-making level of the enterprise, this strategy may not be implemented at all. Moreover, even if it is implemented, it may face regression due to substantial risks, as previously demonstrated in existing studies on executive cognition and enterprise innovation (Yang et al., Reference Yang, Wang, Zhou and Jiang2019), executive team attention and enterprise internationalization strategy (Kolagar, Reim, Parida & Sjödin, Reference Kolagar, Reim, Parida and Sjödin2022; Pisani, Muller & Bogăţan, Reference Pisani, Muller and Bogăţan2018), and executive team structure and enterprise performance (Combs, Ketchen, Perryman & Donahue, Reference Combs, Ketchen, Perryman and Donahue2007). Accordingly, this paper examines the effect of senior executives’ service cognition on the transformation of enterprise servitization under different configurations of position integration. The results, presented in Table 8, Panel A, reveal that the coefficients of executive service cognition in the two-in-one enterprise sample are 3.0180, 3.3779, and 0.1565, respectively. These values are significantly higher than those observed in the two-position separation enterprise sample. These findings indicate that senior executives’ service cognition positively contributes to enhancing the level of enterprise servitization when the two positions are integrated.
Heterogeneity test of industry
Industries differ in terms of their level of technology intensity, which can influence the impact of senior executives’ service cognition on the level of enterprise servitization. On one hand, the development of service-oriented businesses relies heavily on strong technological support. Even for the most fundamental service offerings, enterprises need digital technology to enable service delivery and dynamic information collection, thereby enhancing the quality and scope of their services (Rymaszewska, Helo & Gunasekaran, Reference Rymaszewska, Helo and Gunasekaran2017). On the other hand, industries with higher levels of technology intensity tend to experience greater competition and face stronger market pressures, compelling executives to enhance their service awareness and drive the adoption of servitization strategies (Lee, Yoo & Kim, Reference Lee, Yoo and Kim2016). To explore this further, this paper follows the approach taken by Yang et al. (Reference Yang, Peng and Ge2022) and classifies a total of 31 manufacturing categories from C13 to C43 into high-tech and non-high-tech industries, utilizing the Classification of Strategic Emerging Industries (2012) (Trial) and the Guidelines for the Industry Classification of Listed Companies (Revised in 2012). Table 8, Panel B, presents the results of this analysis. The coefficients of executive service cognition in the high-tech industry sample are 3.3436, 3.0020, and 0.1727, respectively. These values are significantly higher compared to those observed in the non-high-tech industry sample. These findings indicate that in high-tech industries, the service cognition of executives plays a more influential role in driving the transformation of enterprise services.
Table 8. Heterogeneity test results
*** The robust standard errors are denoted inside the parentheses, with **, and * indicating the significance levels of 1%, 5%, and 10%, respectively.
Heterogeneity test of enterprise ownership
The nature of enterprise ownership has a significant impact on business strategies and resource allocation. It is commonly believed that state-owned enterprises exhibit a stronger path dependence orientation and relatively weaker incentives for internal innovation (Naughton, Reference Naughton2018). To examine the differing effects of executive service cognition on the level of enterprise servitization across different ownership structures, this paper conducts a sub-sample regression analysis comprising 1,213 state-owned enterprises and 2,183 private enterprises. The results, presented in Table 8, Panel C, demonstrate noteworthy findings. The coefficients of executive service cognition in the state-owned enterprise sample are 3.5667, 2.9242, and 0.1813, respectively. Surprisingly, these values are significantly higher compared to those observed in the private enterprise sample, contrary to commonly held views.
A plausible explanation for this disparity can be attributed to two factors. First, state-owned enterprises enjoy certain advantages in resource acquisition, including fewer external financing constraints. Consequently, when corporate executives possess a stronger service-oriented mindset, they are better positioned to drive the implementation of service-oriented transformation strategies. Second, executives in state-owned enterprises typically have stronger political connections and are more responsive to government policy guidance, which often emphasizes the promotion of manufacturing servitization, the development of service-oriented manufacturing, and the fulfillment of social responsibilities (Bhaskarabhatla et al., Reference Bhaskarabhatla, Cabral, Hegde and Peeters2021).
Heterogeneity test of enterprise scale
As previously mentioned, the process of servitization transformation requires substantial human capital and capital investment, while also posing unpredictable market risks. Small and medium-sized enterprises, in comparison to their larger counterparts, often face resource limitations and possess weaker risk mitigation capabilities. Consequently, they encounter greater difficulty in making high-risk strategic adjustments and may lack internal incentives to drive such changes effectively. To delve deeper into this matter, this paper classifies the entire sample of enterprises based on their size. Using the median standard, it divides them into 1,362 large enterprises and 828 small and medium-sized enterprises. Subsequently, a group regression analysis is conducted. The findings, presented in Table 8, Panel D, reveal that the impact of senior executives’ service cognition on the level of enterprise servitization is more pronounced in the sample of large enterprises.
Conclusion
Drawing on data from listed manufacturing enterprises in China’s Shanghai and Shenzhen stock markets spanning the period from 2007 to 2020, this study employs text mining technology to construct an enterprise-level executive service cognition index. Additionally, the study manually compiles an index to measure the degree of enterprise servitization. Based on these indices, empirical tests are conducted to examine the impact of executive service cognition on the servitization transformation of manufacturing enterprises and its underlying mechanism. The main conclusions are as follows: (1) The overall regression results unequivocally demonstrate that executive service cognition significantly enhances the level of enterprise servitization, making it a crucial driving force for accelerating the servitization transformation of manufacturing enterprises. (2) The analysis of mediating effects reveals that executive service cognition promotes servitization transformation by fostering the accumulation of human capital and increasing the intensity of R&D investment. (3) The analysis of moderating effects indicates that the concentration of executive power and the Internet penetration rate of cities have a notable influence on the relationship between executive service cognition and enterprise servitization. (4) To gain a deeper understanding of the heterogeneous impact of executive service cognition on the level of enterprise servitization, this study estimates subsamples based on four dimensions: industry technology type, enterprise ownership, and enterprise scale. The findings suggest that executive service cognition has a greater impact on the level of enterprise servitization in high-tech industries, state-owned enterprises, and large enterprises. These conclusions hold significant implications for business managers and policymakers.
First, it is crucial to enhance the service cognition of senior executives through systematic mechanisms. The transformation of enterprises toward servitization is a complex strategic process influenced by multiple factors, and executives, as decision-makers, play a vital role in driving this transformation. To achieve this, it is necessary to encourage the active participation of executives with a service-oriented mindset in business decision-making. By doing so, the cognitive structure of management can be diversified, minimizing limitations and short-sightedness in decision-making and promoting the formulation of strategic decisions that align with the long-term development interests of the enterprise. Additionally, fostering effective communication between senior executives and the service business domain is essential. Executives should gain a deep understanding of the business model, value creation methods, and the various aspects involved in the journey of service-oriented transformation, such as potential paths, implementation mechanisms, and business architecture.
Second, substantial support for human capital and R&D investment is crucial in facilitating the service-oriented transformation of enterprises. Servitization transformation is not only a strategic endeavor but also a technical challenge that necessitates corresponding human capital capabilities and a strong focus on research and development. It is imperative to strengthen the accumulation of human capital in alignment with industry trends, product characteristics, and market competition. Particular emphasis should be placed on investing in human resources related to overall solutions, supply chain management, personalized customization, R&D and design, and data analysis. Moreover, the service-oriented transformation places higher demands on the technology research and development capabilities of enterprises. Therefore, it is necessary to systematically reinforce R&D investment, establish dedicated R&D platforms, and conduct scientific and technological research to address the technical challenges arising from the integration of manufacturing and services, as well as the integration of products and services.
Third, it is essential to dynamically adjust the allocation of executive power based on the strategic goals of the enterprise. While executive power allocation can be centralized or decentralized in a general sense, a more nuanced approach is required when considering the micro-level dynamics. The distribution of power within the senior management team should be tailored to align with industrial development trends, the unique characteristics of the company, and the competitive landscape. Employing a contingency strategy, power allocation should be adjusted to meet the strategic goals of the company and adapt to both internal and external changes. Servitization as a strategic priority necessitates embedding service-dominant logic into the top management team’s cognitive framework. These adjustments will better facilitate the implementation of the service-oriented transformation strategy, particularly in organizations with significant organizational inertia.
Fourth, the government has a crucial role to play in providing effective policy guidance and infrastructure support. In the Chinese context, industrial policy significantly influences enterprise development. To promote the transformation toward a service-oriented economy, the government should enhance policy support by establishing a clear roadmap for the development of China’s manufacturing service sector. This should involve creating an integrated industrial policy system that consolidates the functions of industry management departments and formulating cohesive support policies to bridge the policy gaps between the service industry and the manufacturing industry. Additionally, the government should actively support research and development in common technologies, such as 3D printing, and invest in improving urban information infrastructure networks. These efforts will create a favorable technical environment for manufacturing enterprises to expand their service-oriented businesses.
This study’s findings reveal both contextual commonalities and institutional divergences relative to international research. Regarding commonalities, the pivotal role of executive cognition aligns with Coreynen et al. (Reference Coreynen, Matthyssens, Vanderstraeten and van Witteloostuijn2020) on European manufacturing and Jovanovic et al. (Reference Jovanovic, Raja, Visnjic and Wiengarten2019) on U.S. firms; the mediating effect of human capital resonates with Baines et al. (Reference Baines, Lightfoot, Smart and Fletcher2013); the moderating function of digital infrastructure corroborates Zhan et al. (Reference Zhan, Zhang, Zhang, Zhao and Wang2023). Conversely, China’s state-led digital infrastructure governance diverges from Western market-oriented paradigms (Hu & Gu, Reference Hu and Gu2020).
While this study provides valuable insights, it is important to acknowledge its limitations. First, the data used in this study focus on Chinese listed companies, examining the impact of executive service cognition on the servitization transformation of enterprises and its mechanism. This study’s conclusions reveal both contextual commonalities and institutional divergences relative to international research. It is essential to recognize that different countries and regions have distinct institutional environments and market backgrounds. Therefore, it is necessary to broaden the investigation to include various contexts, exploring the cognitive model and service transformation constraints for executives. Second, in addition to the mediating effect and moderation mechanisms examined in this paper, it is crucial to explore other mechanisms through which executive service cognition influences the transformation of service-oriented enterprises. Factors such as organizational inertia and customer demand status may play significant roles and should be further investigated. Third, it should be explicitly annotated that the conclusions are strictly generalizable only to listed companies with well-established governance protocols. Furthermore, future measurement of enterprise servitization levels ought to prioritize the accurate capture of embedded service revenue streams (e.g., product-service bundles), and emphasize the catalytic role of firm-level digital infrastructure in enabling servitization transformation. Fourth, while widely adopted, our measures – servitization level (Serrev/Serrevr) and executive cognition (text mining) – have limitations. Servitization proxies may be distorted by strategic revenue reporting or divestitures (e.g., IBM’s hardware sale), while annual reports’ PR orientation may inflate service rhetoric (e.g., ‘Made in China 2025’). Nevertheless, this study reveals an economically significant correlation between executive cognition and servitization in Chinese listed firms, providing foundational evidence for future causal inquiry. Lastly, this paper measures the service cognition of executives using text mining technology. This methodological approach aligns with prevailing practices in extant literature (e.g., Kohtamäki et al., Reference Kohtamäki, Parida, Oghazi, Gebauer and Baines2019). However, critical limitations warrant acknowledgment, e.g., reliance on firm-disclosed revenue classification risks strategic impression management, potentially distorting data validity, and annual reports’ inherent public relations orientation enables symbolic adherence to policy trends, which may decouple from substantive strategic commitments. Notwithstanding these constraints, this research contributes novel empirical evidence revealing an economically significant correlation between executive service cognition and servitization transformation within China’s listed firms. This finding advances understanding of servitization drivers and establishes a stepping stone for future causal inquiry into the cognitive microfoundations of servitization.
Acknowledgements
The study is supported by the National Natural Science Foundation of China (71872166 and 72274174).
Conflicts of interest
The authors declare none.
Open access
