How does performance feedback enable enterprise digital transformation? Explicating the role of institutional pressures Open Access

Digitalization is reshaping business logics across industries (Menz et al., 2021; Zhang et al., 2025). Both born digital and traditional firms are increasingly applying digital technologies to all aspects of business operations to compete for the valuable opportunities created by digital transformation (DT). At the same time, DT has also shown a tendency to be institutionalized under the pursuit and worship of the whole society (Hinings et al., 2018). Nevertheless, many firms remain either unwilling or unable to carry out DT; some are even reducing their investments in digital initiatives. As such, scholars have explored antecedents of DT to better facilitate firms’ effective engagement in DT. Existing studies have explored a range of driving factors, e.g. strategy, executive characteristics, firm capabilities and network position (Björkdahl, 2020; Hu et al., 2023; Jackson and Dunn-Jensen, 2021; Yan and Yang, 2025; Ying and Meng, 2023). However, only few studies focused on the role of performance feedback based on behavioral theory of the firm (BTOF) (e.g. Li et al., 2024; Zhang et al., 2025). Given the uniqueness, complexity and novelty of DT as a distinctive form of risking taking (Adner et al., 2019; Amit and Han, 2017; Bailey et al., 2022; Pihlajamaa et al., 2023), it is surprising that existing research has not deeply explored how performance feedback shapes firms’ willingness to engage in DT.

According to principals of satisfaction and search in BTOF (Cyert and March, 1963), firm’s decision of risk taking is based on the comparison between actual and aspirational performance (Greve, 1998). Early research primarily focused on problemistic search, suggesting that when actual performance falls below aspirational levels, firms search for (typically local) solutions and cease searching once performance recovers (Cyert and March, 1963; Greve, 1998; Posen et al., 2018). Conversely, because “success breeds slack” (Cyert and March, 1963; Titus et al., 2021), superior performance can also trigger slack search, which motivates managers to undertake risk taking. Consequently, scholars have shifted to explore how firms respond to dual scenarios of above-aspiration performance and below-aspiration performance (e.g. Eggers and Kaul, 2018; Ref et al., 2021; Xu et al., 2019). More recent studies have further investigated the directionality of such responses (e.g. Cao et al., 2024).

However, existing conclusions about the relationship between performance feedback and organizational responses remain inconsistent, ranging from linear to nonlinear to mixed effects (Kotiloglu et al., 2021). A key reason lies in the confusion and conflation between slack search and problemistic search (Kotiloglu et al., 2021; Xu et al., 2019). Although some scholars have attempted to distinguish the two by conceptualizing them as “capacity logic” and “motivation logic”, respectively (Xu et al., 2019; Ref et al., 2024), these efforts often present only a partial picture, typically emphasizing either on motivation logic or on below-aspiration performance. Building on this line of inquiry, the present study seeks to integrate both logics across the entire spectrum of performance feedback. Specifically, we argue that when performance is close to the aspiration level, motivation logic dominates and triggers problemistic search. In contrast, when performance deviates substantially from the aspiration level, capacity logic becomes dominant, prompting slack search. The transition and tension between these two logics give rise to a distinctive nonlinear relationship between performance feedback and organizational responses.

Boundary conditions are also important to present a complete picture of the relationship between performance feedback and risk taking. Prior studies have devoted uneven attention to organizational characteristics (Xu et al., 2019; Kotiloglu et al., 2021), such as organizational imprinting (Goyal, 2025), shared belief (Cao et al., 2024) and TMT characteristics (Kolev and McNamara, 2022). Given the institutionalization of digitalization (Hinings et al., 2018), institutional factors external to the firms are likely to play an even more important role in pushing firms toward isomorphism in digitalization (DiMaggio and Powell, 1983). In emerging markets, governments have gone further by politicizing digitalization (Luo et al., 2024; Liu et al., 2024), mobilizing firms to align with national strategic agendas. More broadly, the institutional environment continuously conditions firms’ strategic responses (Peng and Luo, 2000). Moreover, in the absence of strong regulative or coercive requirements, the institutionalization of digitalization primarily makes firms comply to normative and mimetic pressures (DiMaggio and Powell, 1983). However, among the limited studies that investigate performance feedback in the context of DT, some fail to differentiate these pressures (Li et al., 2024), while others overlook institutional forces entirely (Zhang et al., 2025). Importantly, given the agency of individual firms, responses to digitalization-related institutional pressures may be heterogeneous under different scenarios of performance feedback (Greenwood et al., 2011; Oliver, 1991; Schilke, 2018). Therefore, the ways in which performance feedback shapes firms’ willingness to engage in DT, particularly under varying institutional conditions, merit deeper exploration.

To answer this question, this paper integrates BTOF and neo-institutional theory to explore the impact of performance feedback on enterprise DT and the moderating effect of institutional pressures. Based on data from Chinese A-share listed firms in Shanghai and Shenzhen from 2007 to 2020, we find that there is an inverted U-shaped relationship between performance below aspiration (PBA) and DT and a U-shaped relationship between performance above aspiration (PAA) and DT. Furthermore, normative pressure attenuates both the U-shaped and inverted U-shaped relationships, while mimetic pressure amplifies these nonlinear effects.

Our study makes three contributions. First, unlike prior studies on BTOF that mainly address organizational characteristics as moderators, our study takes the institutionalization of DT into the analysis (Hinings et al., 2018). In doing so, we provide a more comprehensive understanding of the relationship between performance feedback and enterprise DT by highlighting the moderating roles of institutional pressures, specifically normative and mimetic pressures. This integration also extends and enriches BTOF by situating performance feedback within broader institutional contexts.

Second, we contribute to neo-institutional theory by differentiating normative and mimetic pressure arising from institutionalization of DT. The reason why the “iron cage” was revisited lies in the shift in the underlying causes of bureaucratization and rationalization (DiMaggio and Powell, 1983). Similarly, digitalization is undergoing a transformation in its drivers, from efficiency-oriented to legitimacy-oriented logics (Hinings et al., 2018; Luo et al., 2024). Our findings show that the decision to implement DT is shaped not only by performance comparisons but also by institutional pressures. In particular, firms adopt heterogeneous strategies when responding to DT-related institutional pressures, depending on the specific performance scenario they face. Thus, the isomorphism of DT does not diffuse in a smooth or uniform manner. By explicating the heterogeneous effects of institutional pressures in concrete contexts, this study extends the boundary conditions of neo-institutional theory.

Third, we extend research on performance feedback by clarifying the role of problemistic search and slack search. Existing studies on performance feedback and risk taking ignore the essential difference between problemistic search and slack search (Kotiloglu et al., 2021; Li et al., 2024). Building on this differentiation, our study applies both logics across the entire performance spectrum and specifies the conditions under which each becomes dominant. In doing so, we provide a more nuanced understanding of how performance feedback shapes organizational responses, thereby reconciling prior inconsistencies in the literature.

Practically, our findings provide important implications for both managers and policymakers. For managers, it is essential to recognize that comparisons between actual and aspirational performance significantly shape DT, particularly when performance is slightly above or below the aspiration level. Managers should also pay close attention to digitalization-related policies and the digital practices of peer firms, as both exert strong institutional pressures. For policymakers, beyond formulating targeted policies to promote DT, it is equally important to remain vigilant about the potential negative consequences of media coverage and peer competition, which may impose excessive pressure on certain firms. Governments should therefore seek to balance incentives and pressures, preventing unintended adverse effects. In short, both motivation and capacity logics play different but complementary roles in driving digitalization and effective management of these dynamics is critical for successful transformation.

Cyert and March (1963) laid the foundation for BTOF by examining how organizational structures and routines shape goals, aspirations and decisions. Central to BTOF is performance feedback and search behavior (Argote and Greve, 2007). When performance fall short of the aspiration level, firms may initiate a problemistic search for improvements (Gavetti et al., 2012) or adjust aspirations downward (Posen et al., 2018). In contrast, when performance exceeds the aspiration level, firms may implement slack search (Greve, 2003; Pitelis, 2007), including the continuous redeployment of underused resource into new productive opportunities (Nason and Wiklund, 2018). In general, the greater the performance gap below or above aspirational level, firms would more engage in risk taking, such as bribery or R&D (Xu et al., 2019). Although the theorization of performance feedback appears straightforward, prior research has produced inconsistent or even contradictory findings, reporting linear, nonlinear or mixed effects (Kotiloglu et al., 2021). Beyond contextual differences, the core issue lies in the frequent confusion or conflation of problemistic search and slack search (Kotiloglu et al., 2021; Xu et al., 2019). Because problemistic search ceases once the performance gap is closed, some scholars have assumed that firms with above-aspiration performance are less likely to engage in risk taking. However, slack search does not disappear; rather, it is fueled by the logic of “success breeds slack” (Cyert and March, 1963; Titus et al., 2021). Slack resources, in turn, provide firms with the capacity to pursue risk taking. While some scholars have acknowledged the role of slack search, they often treat it as merely an auxiliary moderator of problemistic search (Xu et al., 2019). More recent research has sought to differentiate the logics underlying these activities: problemistic search is driven by motivation logic, whereas slack search is triggered by capacity logic (Eggers and Kaul, 2018; Xu et al., 2019; Ref et al., 2024). Yet, to achieve more consistent predictions, it is necessary to integrate these perspectives, as prior studies typically present only half of the story, focusing either on motivation logic or on incomplete performance scenarios.

In this study, we hope to take both logics into consideration in the whole performance feedback spectrum. Specifically, when actual performance approaches aspirational level, motivation logic tends to dominate, leading organizations to engage in problemistic search. In contrast, when actual performance deviates substantially from aspirations, capacity logic becomes more salient, giving rise to slack search. The dynamic transition and inherent tension between these logics reveal a potentially non-linear pattern linking performance feedback with subsequent strategic responses.

To make the prediction of performance feedback more complete, subsequent research explore the boundary conditions (Kotiloglu et al., 2021). More emphasis is on organizational contingencies, including firm age (Audia and Greve, 2006), size (Greve, 2008), family ownership (Kotlar et al., 2013) and board diversity (Kolev and McNamara, 2022). However, firms do not exist in a vacuum, nor does performance feedback. Environment “impacts managers’ cognitions of performance and risk and hence their motivations for risk taking” (Xu et al., 2019, p. 1231). Among the limited environmental contingencies, institutional environment remains underexplored even with growing evidence of their influence (Dong et al., 2022). For example, Cheng et al. (2022) and Xu et al. (2019) found the impact of institutional development on the relationship between performance feedback and responses. Recently, Gao et al. (2023) investigate how institutional forces suppress problemistic search. However, these studies primarily draw on North’s institutional economics perspective, rather than neo-institutional theory, which emphasizes institutional isomorphism (DiMaggio and Powell, 1983; North, 1990). Even though Li et al. (2024) mention the role of institutional pressure, little is known about how coercive, mimetic and normative pressures differentially affect the performance feedback–response linkage. This lack of dialogue between BTOF and neo-institutional theory remains an important gap in the literature.

Enterprise DT refers to the process through which firms leverage information technologies such as mobile internet, artificial intelligence, cloud computing, blockchain and the Internet of Things to explore business opportunities, enhance user experience, streamline operational processes and create new business models (Warner and Wäger, 2019), which triggers profound organizational changes (Vial, 2019). Although DT helps firms to improve efficiency and affords potential for firms to achieve hyperspecialization and hyperscalability (Giustiziero et al., 2023), it also presents many challenges and uncertainties for firms (Bodrožić and Adler, 2022). Not only does DT require the devotion of substantial resources, but it also necessitates the reconfiguration of organizational structure and business models (Warner and Wäger, 2019). Moreover, to benefit more from digital innovation, firms have to engage stakeholders in the synchronization (Yan and Yang, 2025). However, this is particularly difficult and beyond the control of the firms. Therefore, compared to existing risk behaviors, DT appears more unique and complex.

Ironically, DT has also been regarded as elixir and inevitability, which drags firms into a dilemma: failing to undertake DT amounts to waiting for organizational obsolescence; undertaking it, by contrast, is akin to courting operational peril. Firms are confronting significant uncertainties originated from DT (Luo et al., 2024). Given the increasing governmental initiatives and worship of the whole society (Luo et al., 2024; Hu et al., 2023), DT has shown a tendency to be institutionalized, presenting institutional pressures for firms to reduce uncertainty (Hinings et al., 2018). Thus, performance feedback would influence firms’ willingness to engage in digital innovation and institutional pressures would regulate how firms to make sense performance feedback.

In this study, based on the integration of BTOF and neo-institutional theory, we aim to uncover how performance feedback impact risk taking (specifically enterprise DT) under the influence of institutional pressures.

Whereas existing studies have tended to focus on either motivation logic or isolated segments of the performance spectrum, we examine the transition and tension between motivation logic and capacity logic across the entire spectrum of performance feedback. Specifically, in the scenario of above-aspiration performance, improvements in performance are expected to initially reduce the degree of DT, but further increases should eventually promote it. Conversely, in the scenario of below-aspiration performance, declines in performance are expected to initially encourage DT, but deeper declines should ultimately hinder it.

When actual performance is slightly above aspiration level, firms tend to follow motivation logic, triggering problemistic search. Satisfied with current outcomes, managers often fall into organizational myopia (Posen et al., 2018; Xu et al., 2019) and express high-level risk aversion. Shareholders may also exert pressures on top management teams to maintain conservative strategies and avoid overinvestment (Schimmer and Brauer, 2012). Moreover, DT entails high levels of uncertainty, difficulties in risk evaluation and ambiguous goals. It also requires substantial investment, resource reallocation and even shifts in organizational identification or belief. These challenges are difficult to meet when slack resources remain limited despite modest surplus performance.

In contrast, when performance is far above aspiration, capacity logic dominates and slack search is activated. Abundant slack resources create room for experimentation (Bentley and Kehoe, 2020) and foster a long-term, opportunity-oriented perspective (Xu et al., 2019). This encourages firms to invest in DT, shifting from production to servitization models (Amit and Han, 2017) and mobilizing ecosystem resources via digital technologies (Altman et al., 2021), thereby creating and appropriating greater value. While motivation logic remains present, it is reinforced by capacity logic, ensuring that performance continues to keep pace with rising expectations (Xu et al., 2019). Thus, we propose that:

When performance is slightly below aspiration, firms follow motivation logic and trigger problemistic search. Realizing operational issues, managers express higher risk tolerance under shareholder pressure, pursuing quick-fix strategies to restore performance (Posen et al., 2018; Xu et al., 2019). Although DT requires long-term investment, its potential to achieve “leapfrogging” makes firms confident and willing to overlook risks. Recoverable slack and potential slack can be redeployed to support such risk taking (Carnes et al., 2019), while government policies that favor DT provide additional resources (Wang et al., 2023).

When performance falls far below aspiration, capacity logic dominates and slack search emerges. Fearing further resource erosion, firms become rigid and conservative (Wennberg et al., 2016), prioritizing strict resource allocation to ensure short-term survival (Ref and Shapira, 2017). Workplace pessimism further constrains internal and external resource mobilization. Although problemistic search persists (Posen et al., 2018), firms often focus on immediate remedies rather than long-term initiatives (Xu et al., 2019). Yet, the success of DT requires a long-term orientation. Severe resource scarcity also prevents firms from effectively signaling commitment to DT to leverage government support. Thus, we propose that:

According to neo-institutional theory, organizations operate within institutional environment (Meyer and Rowan, 1977), where their behaviors are constrained by prevailing structures, practices and rules (DiMaggio and Powell, 1991). These constraints originate from institutional pressures that generate a legitimacy demand, pushing organizations to conform to reduce uncertainty and enhance visibility (Suddaby et al., 2017). Institutional pressures typically take three forms: coercive, normative and mimetic (DiMaggio and Powell, 1983). Coercive pressures define what organizations must do, with noncompliance leading to sanctions. Normative pressure reflects what organizations should do, requiring adherence to established norms and professional standards (Boxenbaum and Jonssonm, 2008). Mimetic pressure, which encourages organizations to emulate their peers, arises under conditions of goal ambiguity and environmental uncertainty. In such contexts, organizations imitate the practices of others, whether optimal or not, as a means of reducing perceived uncertainty (DiMaggio and Powell, 1983; Scott, 1995). Yet, despite facing institutional pressures, some organizations resist enrollment in isomorphic processes. As Oliver (1991) emphasized, organizational agency cannot be overlooked. Accordingly, a more recent trend in neo-institutional research has been to examine how motivations, cognitions, backgrounds and behaviors of decision-makers shape organizational responses to institutional pressures (Schilke, 2018). In line with the motivation logic and capacity logic proposed in BTOF, responses to institutional pressures depend not only on willingness but also on ability (Durand et al., 2019). Before taking actions, managers need to assess the uncertainty in local situation and evaluate the costs and benefits of conforming to institutional expectations (Schilke, 2018; Durand et al., 2019).

The revisit of “iron cage” reflects shifting drivers of bureaucratization and rationalization (DiMaggio and Powell, 1983). In a similar vein, digitalization is experiencing a comparable transformation, with its underlying logic evolving from efficiency orientation to legitimacy orientation (Hinings et al., 2018; Luo et al., 2024). Fueled by both academic research and practical discourse, DT is increasingly institutionalized, thereby generating institutional pressures (Hinings et al., 2018). Unlike coercive pressures, DT is not mandatory; rather, it gives rise to normative pressure through social consensus and to mimetic pressure in response to uncertainty surrounding the transformation process. Yet, existing studies rarely examine how these institutional pressures promote DT. Although Li et al. (2024) explored how institutional pressures impact enterprise DT, they did not differentiate between normative and mimetic pressures. More recently, Liu et al. (2024), drawing on a case study of Chinese SOEs, probed the distinct roles of institutional pressures in shaping DT. Given the agency of firms, it is possible for firms to respond to institutional pressures with deliberate strategies depending on their specific performance feedback scenarios (Greenwood et al., 2011; Oliver, 1991; Schilke, 2018). Thus, incorporating normative and mimetic pressures into the analysis of performance feedback and DT can yield deeper insights into the mechanisms underlying this phenomenon (Menz et al., 2021).

Normative pressure stems from “what is widely considered a proper course of action or even a moral duty” (Boxenbaum and Jonssonm, 2008, p. 80). It involves the social pressures to conform to a particular group (e.g. industry associations, media and stakeholders) or society’s norms and standards (e.g. CSR, ESG) (Durand et al., 2019; Shankar et al., 2025). To acquire legitimacy and social status, organizations give in to normative pressure (DiMaggio and Powell, 1983). Normative pressure leads organizations to adopt practices such as using intelligent product–service systems (Kropp and Totzek, 2020) or investing in environmentally relevant innovation activities (Berrone et al., 2013).

Witnessing the power of digital technologies and the success of born-digital firms, society increasingly regards DT as a panacea for all problems and emphasizes that all organizations should adopt it, regardless of its actual effectiveness. Amplified discussions and encouragement from governments, media, industry associations and academic institutions intensify this normative pressure, prompting organizations to internalize DT as an institutional myth (Meyer and Rowan, 1977). Nevertheless, exceptions remain, as some organizations possess both the motivation and capabilities to resist normative pressure (Durand et al., 2019). As discussed above, when experiencing slightly above- or below-aspiration performance, the motivation logic would be dominant; when experiencing significant above- or below-aspiration performance, the capacity logic would be dominant. Yet, the rise of normative pressure reshapes this landscape. Specifically, normative pressure strengthens the motivation to implement DT when performance is slightly above aspirations, as they highlight the potential to enhance social status and secure external resources (DiMaggio and Powell, 1983). For firms performing only modestly above aspirations, often positioned at the margins of the institutional field, the perceived benefits of digitalization outweigh its costs (Greenwood et al., 2011). By contrast, normative pressure suppresses the motivation to implement DT when performance is slightly below aspirations. In this context, firms focus primarily on restoring performance and insist on retaining discretion in choosing risk taking strategies. They may perceive calls for digitalization from various institutional actors as intrusive, thereby triggering resistance (Brehm and Brehm, 1981). Even when firms initially intended to pursue digitalization, they may instead turn their attention to alternative actions as a means of asserting autonomy.

When performance is significantly above the aspiration level, normative pressure amplifies the capacity gap. Unlike average performers, star firms attract greater attention from stakeholders, who tend to impose more demanding and stringent expectations on them (King and McDonnell, 2013). In the context of digitalization, firms with outstanding performance may be required to promote and empower the whole ecosystem or supply chain to undergo DT, which exceed the resources and capability of the focal firms. Confronted with these “growing pains,” high-performing firms may slow the pace of digitalization or even pursue it covertly to maintain sustainable growth. Conversely, when performance is significantly below the aspiration level, normative pressure can supplement limited capacity or mitigate capacity-related concerns. On the one hand, firms may adopt a decoupling strategy, symbolically complying with DT requirements to maintain legitimacy (Durand et al., 2019); on the other hand, they may take advantage of related funding programs or subsidies to promote substantive DT. In such cases, normative pressure may facilitate a potential turnaround. Thus, we propose that:

Mimetic pressure arises from goal ambiguity and environmental uncertainty (DiMaggio and Powell, 1983). Under such conditions, organizations tend to imitate prevailing practices or emulate competitors perceived as successful (Boxenbaum and Jonssonm, 2008; Haveman, 1993). Although mimetic isomorphism is intended to reduce perceived uncertainty, imitation does not always generate optimal outcomes for imitators (DiMaggio and Powell, 1983). Like responses to normative pressure, organizations may also shield themselves from mimetic pressure if they can resolve uncertainty through alternative mechanisms. Even when solutions are lacking, firms may strategically redirect their attention to more certain issues, thereby mitigating the influence of mimetic pressure (Krause et al., 2019).

As a novel form of risk taking, DT is characterized by high uncertainty (Bodrožić and Adler, 2022). It not only involves the application of digital technologies but also triggers organizational change. Without careful evaluation, some firms may even fall into a “digital trap”. Recent shocks, such as the uncertainty triggered by COVID-19 or broader economic sluggishness, have further accelerated DT. In addition, competitive pressures heighten uncertainty and encourage imitation (Banerjee, 1992). As more firms implement DT and some even achieve notable profits, mimetic pressure becomes increasingly intensified. However, the perception of uncertainty is contingent upon local contexts, such as firms’ performance scenarios, leading firms to respond to mimetic pressure in heterogeneous ways. Specifically, when performance is slightly above aspiration, mimetic pressure would worsen the motivation to engage in digitalization. Risk-averse firms prefer to maintain the status quo, viewing DT as highly uncertain compared to the relative certainty of established organizational routines. Mimetic pressure can also reinforce the mindset of stabilizing competition dynamics, further discouraging investment in digitalization. Conversely, when performance is slightly below aspiration, the motivation to enable digitalization would be promoted. Performance shortfall creates a condition of uncertainty. Even though DT comes with risks, peers’ practices could eliminate such concern. Moreover, mimetic pressure even makes firms perceive the possibility of deteriorating performance without imitating peers (Barreto and Baden-Fuller, 2006), thereby motivating them to engage in digitalization.

When performance is significantly above or below aspirations, mimetic pressure can either reinforce or distort firms’ capacity for DT. Owing to the network effects of digital technologies, the digitalization efforts of peers strengthen the overall digital infrastructure, yet the ability to leverage such affordances depends on firms’ own resources and knowledge (Autio et al., 2018). Compared with average performers, star firms are better positioned to capitalize on these infrastructures. Recognizing heightened uncertainty in the competitive environment, they may accelerate digitalization through imitation. By contrast, when firms are trapped in severely adverse situations, mimetic pressure may exacerbate managerial anxiety (Xu et al., 2019). In these cases, firms not only strive to reverse poor performance but also feel compelled to quickly catch up with competitors. Although peers’ digital practices provide a reference point, the long-term orientation and inherent uncertainty of DT may relegate it to a secondary option. Instead, firms may devote greater attention to illegal or unethical actions aimed at boosting performance by undermining rivals (Xu et al., 2019). Thus, we propose that (Figure 1):

We select the data of Chinese A-share listed firms in Shanghai and Shenzhen from 2007 to 2020 for testing the hypotheses. Based on prior research (Cheng et al., 2022), the data are processed as follows:

• firms in the financial industry are excluded;

• firms delisted are excluded;

• firms labeled by ST, ST*, SST are excluded; and

• to ensure that the sample does not have missing data for at least five consecutive years, firms listed in 2015 and after are excluded.

We ultimately obtain 24,550 firm-year observations for 1,914 listed firms. All data are sourced from China Stock Market and Accounting Research Database.

DT. As a kind of complex and unique risk behavior, DT necessitates organizational primary attention and covers unstructured data and novel practice, which makes annual reports suitable to capture the willingness and practice of engage in digital innovation (Zhu and Yu, 2024; Luo et al., 2024). Consistent with the prior research (Chen and Tian, 2022; Hu et al., 2023; Luo et al., 2024; Zhu and Yu, 2024), we use text analysis to measure enterprise DT. Specifically, we calculated the word frequencies of relevant keywords in firms’ annual reports from 2007 onward across four dimensions: artificial intelligence, blockchain, cloud computing, big data and digital technology application. We then aggregated the keyword frequencies for each firm by year and applied a logarithmic transformation to construct the DT indicator. A higher value of this indicator reflects a greater degree of enterprise digitalization.

Performance feedback. Performance feedback refers to the gap between an organization’s actual performance and aspiration (Cyert and March, 1963). Specifically, aspirational performance is sum of historical and social aspiration:

where $HAi,t$is calculated as:$HAi,t$= $βPi,t-1+(1-β)HAi,t-1$⁠; $SAi,t$ is estimated as: $SAi,t=(∑i≠jPj,t)/N-1$⁠, where $Pi$ denotes the ROA of firm i and N −1 denotes the number of other firms j in the same industry other than firm i at time t. To determine the values of α and β, a grid search is used by iterating through possible values at 0.1 intervals. The values enable the best model fit is 0.4 and 0.3, respectively.

We then calculated PAA and PBA of firm i, respectively:

Normative pressure (NP). Since normative pressure is not obligatory, it mainly comes from news media, decentralized groups such as the public and specific organizations such as accreditation bodies (Kodeih and Greenwood, 2013). With reference to Cheng et al. (2022), we measure normative pressure in terms of media attention. Specifically, this study counted news reports related to the digital economy on an annual basis and then applied a logarithmic transformation. The larger the value, the stronger the normative pressure.

Mimetic pressure (MP). Mimetic pressure mainly comes from other firms in the same industry and we measured the mimetic pressure faced by the focal firm by calculating the mean value of the degree of DT of other firms in the same industry (Mehrabi et al., 2021). Larger values indicate that firms face more pressure to imitate.

Control variables. To control for other factors affecting DT, referring to prior research (i.e. Prügl and Spitzley, 2021), we incorporate variables at the firm and corporate governance levels. At the firm level, we control for firm size (SIZE), measured by taking the logarithm of the firm’s total assets, firm age (AGE), measured by taking the logarithm of the difference between the firm’s inception and the target year, financial leverage (LEV), which is calculated by the ratio of firms’ liabilities to total assets, cash flow intensity (CASH), calculated as the ratio of cash and its cash equivalents to total assets and ownership (OWN), which is valued as 1 if the firm’s beneficial owner is a SOE and 0 otherwise. At the corporate governance level, we control for board size (BOARD), measured by the number of board members, duality (DUAL), assigned a value of 1 when the chairman of the board and CEO are the same person and 0 otherwise, audit opinion (AUDIT), which is assigned a value of 0 when the accounting firm issues a standardized unqualified opinion and 1 otherwise and shareholding concentration (SC), calculated by the proportion of shares held by the first largest shareholder.

Tables 1 and 2 depict the results of descriptive statistics and correlation analysis, respectively. We calculated the variance inflation factors (VIF), showing an average VIF of 1.27 and a maximum VIF of 1.59, indicating that multicollinearity is not a severe issue. We analyzed the relationship between performance feedback and enterprise DT using a fixed-effects model that clusters standard errors at the firm level to test our hypotheses. To alleviate the endogeneity problem, we lagged the explanatory and moderator variables by one period following prior research (i.e. Zeng, 2024). Moreover, all continuous variables were winsorized at 1% level in both tails to mitigate the outlier issue.

Table 3 illustrates the results of PAA and enterprise DT. Model (1) incorporates enterprise DT and all control variables. Firm size (SIZE), firm age (AGE) and board size (BOARD) are all significantly positively associated with enterprise DT, whereas financial leverage (LEV), cash flow intensity (CASH) and concentration of the first largest shareholder (SC) are significantly negatively related to firms’ DT, which is generally in line with the previous studies (Chen and Tian, 2022; Prügl and Spitzley, 2021). The results of model (2) and model (3) show that PAA is positively correlated to enterprise DT at 10% level and after adding the quadratic term of PAA, the linear term of PAA is significantly negative (φ = −0.696, p < 0.01) and the quadratic term of PAA is significantly positive (φ = 1.912, p < 0.01). Moreover, we performed u-test command and plotted their relationship. It is suggested that with performance increasing above aspiration, the willingness of enterprises to engage in DT will decrease first and then increase. Thus, Hypothesis 1 is supported (Figure 2).

After adding normative pressure, the interaction term between NP and the quadratic term of PAA has a significant negative relationship with enterprise DT (φ = −1.419, p = 0.044). H3a is supported. In addition, the interaction term of MP and the quadratic term of PAA is significantly and positively related to enterprise DT (φ = 1.932, p = 0.001). H4a is supported.

Table 4 demonstrates the results of PBA and enterprise DT. There is a significant negative relationship between the linear term of performance blow aspiration and the DT of enterprises and after adding the quadratic term of performance blow aspiration, the linear term of performance blow aspiration is significantly negative (φ = −2.118, p = 0.000) and the quadratic term of performance blow aspiration is significantly negative (φ = −4.550, p = 0.000), indicating that with performance deteriorating below aspiration, the willingness of enterprise to engage in DT will increase first and then decease. Therefore, H2 is supported (Figure 3).

Incorporating the moderating effect of normative pressure, the interaction term of normative pressure and the quadratic term of performance blow aspiration is positively but significantly related to enterprise DT (φ = 2.216, p = 0.013), which supports H3b. Meanwhile, after including mimetic pressure as a moderating variable, the interaction term of mimetic pressure and the quadratic term of performance blow aspiration is significantly and negatively related to enterprise DT (φ = −4.618, p = 0.000). H4b is supported. To display the effects of the interactions visually, we also plotted the moderating effects in Figures 4–7. Specifically, normative pressure from media coverage weakens the U-shaped (the inverted U-shaped) relationship between PAA (PBA) and enterprise DT. However, mimetic pressure from peer’s digitalization strengthens the U-shaped (the inverted U-shaped) relationship between PAA (PBA) and enterprise DT.

The impact of performance feedback on DT may be affected by the potential endogeneity problems:

• Unobserved or difficult-to-quantify factors may simultaneously influence both performance feedback and corporate digitization behavior, leading to omitted variable bias.

• Bidirectional causality may exist between performance feedback and DT, i.e. a firm’s digital maturity may also influence its performance outcomes.

In the preceding main effects model, we controlled for this issue using fixed effects and lagged variables. To ensure robustness, we followed the prior research (Li and Zhao, 2024) to use one-period lagged performance feedback as an instrumental variable in the regression. After incorporating this instrumental variable and testing for non-identifiability and weak instrumental variable, the findings in Table 5 remain in agreement with the results from the main model.

We removed the sample data from 2007–2010 to 2020 to exclude the effects of international financial crisis and the COVID-19 pandemic (Li et al., 2024) and the results of the study are presented in Table 6. The findings are broadly consistent with the previous section, suggesting that the findings are robust.

In the previous calculation on aspiration, β is assigned to 0.3 to obtain the optimal model fit. Consistent with prior study (Xu et al., 2019), we modified the assignment of β in the performance feedback and recalculate PAA and PBA. The regression test results are shown in Table 7. The results are generally consistent with the prior section, indicating that the findings are reliable.

We also used a proxy variable for DT to test the robustness of our findings. Specifically, drawing on prior research (Zhang et al., 2025), we used the ratio of the total number of digital vocabulary frequencies to the number of MD&A words in annual reports as a proxy variable for DT. We reran the regression models, and the results in Table 8 are consistent with our findings.

Our study contributes to the existing literature in several important ways. First, we provide a more comprehensive understanding of the relationship between performance feedback and enterprise DT by revealing the moderating role of institutional pressures. When considering boundary conditions, studies grounded in the BTOF have primarily focused on organizational characteristics as moderators (Desai, 2015; Li et al., 2024), while paying insufficient attention to environmental constraints and triggers, particularly the institutional environment (see Xu et al., 2019). This gap is notable because, in emerging markets such as China, institutions play a critical role in shaping firms’ behavior (Peng and Luo, 2000). Moreover, unlike other forms of risk taking, DT has demonstrated a strong tendency toward institutionalization, driven by widespread societal endorsement (Hinings et al., 2018). Importantly, this institutionalization process resembles the evolution of social norms more than the enactment of coercive legislation. We argue and evidence that normative and mimetic pressures affect the relationship between performance feedback and enterprise DT, respectively. By incorporating institutional pressures into the research framework, our study not only unravels the complexity unique to DT but also advances BTOF by embedding performance feedback within its broader institutional context.

Second, this study extends neo-institutional theory by explicating the distinct effects of institutional pressures in the context of DT. DiMaggio and Powell (1983) emphasized that organizational isomorphism is driven more by legitimacy than efficiency, identifying three types of institutional pressures. Although prior research has examined how firms respond to institutional pressures in various domains, such as CSR (Li and Lu, 2020) and quarterly earnings guidance (Park and Patterson, 2021), limited attention has been paid to digitalization. Yet, given the growing institutionalization and even politicization of DT (Hinings et al., 2018; Li et al., 2024), firms must simultaneously account for efficiency and legitimacy concerns. The societal belief in digitalization creates normative pressure, while peer adoption generates mimetic pressure. Our findings further show that firms respond to these institutional pressures in heterogeneous ways, depending on their performance scenarios. The isomorphism of DT thus does not diffuse smoothly across organizations. Specifically, when faced with significant above-aspiration performance and slightly below-aspiration performance, firms tend to shield from normative pressure which demands greater resource and intervention; when faced with slightly above-aspiration performance and significantly below-aspiration performance, firms are more likely to resist mimetic pressure, which amplifies uncertainty and managerial anxiety. By revealing these nuanced dynamics, this study revitalizes neo-institutional theory in the digital era and demonstrates how institutional pressures interact with organizational performance to shape heterogeneous strategic responses.

Finally, this study helps resolve the inconsistent findings regarding the impact of performance feedback on organizational responses. Prior research on this relationship has yielded non-uniform results (Kotiloglu et al., 2021), largely because it has overlooked the fundamental differences between two distinct search mechanisms and misapplied them across the performance spectrum. Problemistic search is underpinned by motivation logic, whereas slack search is driven by capacity logic (Xu et al., 2019). Although Li et al. (2024) explored the impact of performance feedback on DT, problemistic search is underpinned by motivation logic, whereas slack search is driven by capacity logic. We investigate the impact of performance feedback on enterprise DT by distinguishing between problematic search and slack search. Specifically, when actual performance is near the aspiration, motivation logic dominates and problematic search is triggered. In contrast, when actual performance is far from the aspiration, capacity logic dominates and slack search emerges. The transition and tension between motivation logic and capacity logic would showcase distinctive non-linear relationship between performance feedback and response. By clarifying these mechanisms, our study offers a more coherent explanation for prior inconsistencies in the literature about performance feedback.

Our research has some practical implications. For firm managers, our study underscores that decisions regarding DT are shaped by comparisons between actual and aspirational performance. When the actual performance is slightly higher than the aspiration, managers should proactively deploy digital technologies and prepare the necessary talent and resources to secure a first-mover advantage in the digital era. When performance is substantially above aspirations, managers should focus on the effective integration of digital technologies with business models, ensuring that DT supports sustainable growth rather than short-lived gains. When the actual performance is slightly lower than the aspiration, managers should embed digital technologies into core strategic processes and align them with strategy to achieve meaningful digitalization. In contrast, when performance is far below the aspiration, managers should adjust their mindset and actively mobilize ecosystem resources beyond organizational boundaries, using DT as a lever to reverse performance decline. In addition, managers should remain attentive to the influence of media coverage and peer adoption of digital technologies, recognizing both the opportunities and the pressures these institutional forces create.

For governments, our findings suggest that policies guiding enterprise DT should be designed with reference to firms’ relative performance against aspiration levels. Particular attention should be given to firms whose actual performance is slightly above and far below the aspiration, as these firms often lack either the motivation or the capacity to engage effectively in DT. For firms whose performance is far below aspirations, governments should take substantive actions to support turnaround efforts, such as introducing digitalization experts, providing financial subsidies and offering technological assistance, rather than relying on purely symbolic exhortations. For firms performing well above aspirations, governments can encourage them to document and disseminate their DT experiences, while helping them sustain momentum and avoid complacency. For firms slightly below aspirations, governments should provide targeted support in the form of talent development, financial resources and technological guidance, enabling them to quickly recover performance through DT. The government should also provide some financial and technological support to firms whose actual performance is slightly lower than the aspiration, to help them quickly repair their performance with the help of DT. Noticeably, policy makers should avoid imposing burdens and intervention on firms whose performance is slightly above and below the aspiration level. Meanwhile, they should also take substantive actions to relieve competitive pressure faced by firms whose performance is slightly above and significantly below the aspiration, allowing them to pursue DT more proactively. In short, the role of government policy should be to either trigger motivation or cultivate capacity for DT, depending on where firms are situated along the performance spectrum.

Unavoidably, there are limitations in our study. First, although we have found that the motivation logic and capacity logic would get dominant alternatively, how such transition occurs is still “black box”. Further research is suggested to uncover this based on attention-based view (Berchicci and Tarakci, 2022; Ocasio, 1997) or other microfoundation perspectives (Felin et al., 2015). Second, with respect to methodology, we sought to address endogeneity by using lagged independent variables, fixed-effect models and instrumental variable method. Nonetheless, such efforts may not fully rule out potential biases. Future research is encouraged to adopt machine learning techniques, natural experiments or field experiments to provide more rigorous causal identification. Finally, our study is mainly based on Chinese listed firms, which may constrain the generalizability of our findings. Future research could validate our findings with data from international samples and other countries.