The moderating effect of management behavior for Lean and process improvement

Lean is a popular concept for improving operational performance in production environments (Cua et al. 2001; Shah and Ward 2003) and service environments (LaGanga 2011; Meredith et al. 2011; Swank 2003) by means of process improvements (Claycomb et al. 1999; Fullerton et al. 2003; Shah and Ward 2003; Shah and Ward 2007). Lean is therefore broadly classified under the umbrella of process improvement and world-class operations (Shah et al. 2008). Though it can be described at different levels of abstraction (e.g., it can be defined as a philosophy, as a set of principles and as bundles of practices) it is usually defined as a collection of practices that work together synergistically to create a high quality, streamlined system that produces finished products with little or no waste at the rate of customer demand (Shah and Ward 2003). Practices commonly associated with Lean include the ability to create flow and just-in-time production, including set-up time reduction and pull control (Cagliano et al. 2006; Cua et al. 2001; Sakakibara et al. 1997), quality control (Flynn et al. 1995; Narasimhan et al. 2006; Samson and Terziovski 1999) and human resource development including leadership and employee involvement (Ichniowski et al. 1997). Indeed, next to technical Lean practices (e.g., value stream mapping and improvement, benchmarking, and pull) various behavioral practices are distinguished, amongst which management commitment, employee empowerment and an open culture. Samson and Terziovski (1999) found that these types of practices impact performance differently. Mann (2009 p.15) also states that ‘implementing tools represents at most 20 percent of the effort in Lean transformations; the other 80 percent of the effort is expended on changing leaders’ practices and behaviors, and ultimately their mindset’. In addition, Marodin and Saurin (2015) listed 14 barriers that hamper successful Lean implementation, amongst which lack of communication throughout the organization, lack of knowledge, insecure operators to carry out improvement projects, demotivated staff, lack of management support and lack of meaning.

Leadership is the intentional process to influence other people in order to guide, structure and facilitate activities and relationships in a group and/or an organization (Yukl 1989). It is the process by which an individual guides others in their pursuits, often by organizing, directing, coordinating, supporting and motivating their efforts. Lean leadership is a set of leadership competencies, practices and behaviors to successfully implement and exploit a Lean management system (Poksinska et al. 2013). Lean management is associated with leadership that facilitates and stimulates the continuous initiation and execution of improvement initiatives and coordination of change projects (Choo et al. 2007; Wu and Chen 2006). Typical Lean leadership behaviors and actions are paradoxical in nature (Choi and Eboch 1998; Lewis et al. 2014); it incorporates technical aspects like setting ambitious goals (Linderman et al. 2006), actively steering on performance improvement (Bodek 2008), managing on facts (Choi and Eboch 1998), utilizing objective data (Waldman et al. 1998), and adhering to the standard operating procedure for sake of efficiency and effectiveness on the one hand and social, follower-related aspects like improvement stimulation, promotion of employee responsibility, empowerment and collaboration to facilitate creativity and stimulate innovation on the other hand (Spear and Bowen 1999). Worley and Doolen (2006) stated that management must particularly create organizational interest in Lean by means of envisioning the Lean organization (Cua et al. 2001), and must clearly communicate both the objective of Lean and the required change to everyone in the organization (Laohavichien et al. 2011). Indeed, Lean leadership implies the creation of a True North vision, i.e. the organization’s philosophical objective and long term goals (Liker and Convis 2011). By communicating objectives, shared values and principles of the Lean organization, managers influence employees’ attitudes towards Lean and encourage continuous improvement (Poksinska et al. 2013). Managers accordingly control, measure and reward performance aspects that reflect values and contribute to the fulfilment of these goals (Poksinska et al. 2013, p.893). Indeed, Lean leadership implies both steering on practices (e.g. the use of standard work) and performance (e.g. the improvement of processes and the reduction of waste) and focusing all improvement efforts towards a common goal and consequently getting immediate results (Boyer and Sovilla 2003). Typical Lean leadership behavior of higher management is having clear performance expectations, communicating a clear direction, having a vision about Lean and being personally involved (Taylor et al. 2013). Indeed, management must develop a vision and a roadmap for Lean, set targets and also follow-up on these targets (Netland 2016). We therefore conclude that i) envisioning and communicating the meaning of Lean, ii) setting goals and active steering on performance improvement metrics (Bourne et al. 2005; Flynn and Saladin 2006) and iii) encouraging continuous improvement (Choi 1995; Kaynak 2003) are important management behaviors for Lean.

We used data collected from business school participants in the period 2012/2013 to test the hypothesized research model. Participants were predominantly middle managers. We employed a web-based survey approach, which participants filled out before they attended an advanced Operational Excellence / Lean related course. We explicitly remarked that we would use the results anonymously during the course. 80% of the participants filled out the questionnaire resulting in 205 questionnaires, of which 198 were useful for research. About 10% of the respondents had a higher management position, 66% of the respondents were middle-managers, the remaining respondents had no management position but had job titles equivalent to internal advisors, logistics engineer and operations controller. The respondents averaged 8.5 years’ work experience with their current organization; see Table 1. Non-response bias was evaluated by testing responses of 21 non-informants for significant differences during the courses, where they were asked to respond verbally to five substantive items related to key constructs of the whole survey. There were no significant differences (p < .05) in responses to any item, leading to the conclusion that non-response bias was not a problem.

We adapted the familiar operationalization of Shah and Ward (2007) as a measure of Lean practices, including set-up reduction (Cronbach alpha = .85), statistical process control (Cronbach alpha = .74), pull (Cronbach alpha = .87), flow (Cronbach alpha = .68), involved employees (Cronbach alpha = .70), customer involvement (Cronbach alpha = .64), and supplier feedback (Cronbach alpha = .64). We evaluated the unidimensionality, reliability and convergent validity of the scale for Lean using confirmatory factor analysis with the software package AMOS 22. For satisfactory convergent validity, the estimated parameters between the latent variables and their indicators should be at least 0.50 (Hair et al. 1998). Some items have therefore been removed from the final scales and omitted from Table 6 of online Appendix A. The final first order measurement model fits the data sufficiently:χ²= 184.750, df = 131, p = .001, CFI = 0.957, IFI = 0.958, TLI/NNFI = 0.944, RMSEA = 0.046.

With respect to the constructs related to management behavior for Lean, we operationalized communicating the meaning of Lean (Cronbach alpha = .89), encouraging continuous improvement (Cronbach alpha = .77) and steering on performance improvement (Cronbach alpha = .72) using items from Cua et al. (2001), Douglas and Judge (2001) and Flynn et al. (1999). The first order measurement model with these three constructs fits the data sufficiently: χ²= 82.980, df = 41, p = .000, CFI = .955, IFI = .956, TLI/NNFI = .940, NFI = .916, RMSEA = .072; see Table 7 of online Appendix A. Finally, we developed a scale for process improvement (Cronbach alpha = .67) with items from Samson and Terziovski (1999) and Kohlbacher and Gruenwald (2011) including ‘We have systematically analyzed processes and eliminated all non-value adding (dysfunctional) activities’ as an example.

To test for discriminant validity, we performed chi-square difference tests on the result of CFA’s for an unconstrained model versus a constrained model for each pair of constructs of the scales for Lean and management behavior, where the covariance parameter of the pair of constructs was constrained at 1 in the constrained model and set to be free in the unconstrained model. For satisfactory discriminant validity, the χ²-difference values should be greater than 3.84 (Bagozzi 1981; Bagozzi and Phillips 1982; Kim et al. 2012; Ng et al. 2015). The χ²-difference values ranged from 37.21 to 119.65 indicating satisfactory discriminant validity; see Table 8 of online Appendix B. In addition, the correlation matrix for all constructs with Cronbach alpha values on the diagonal, is presented in Table 2; a full correlation matrix including all sectors is presented in Table 9 in online Appendix B. Cronbach’s alpha exceeds .60 for all constructs, which indicates satisfactory reliability (Cronbach 1951). In addition, all Cronbach alpha values are greater than the correlations, which also indicates satisfactory discriminant validity (see e.g., Kaynak 2003).

We used size, tenure in the organization, position and type of industry as control variables. Smaller organizations, for instance, typically have fewer resources for process improvement practices and Lean (Cao and Zhang 2011). Size of the organization was measured by the number of employees (logarithmized). We controlled for hierarchical level effects by including two dummy variables for position as perceived management behavior for Lean may differ across (non)management levels. Dummies are also included to control for the type of industry; because of missing values for this variable the valid number of respondents for regression analysis is 178.

Procedural methods were applied to minimize the potential for common method bias since both the independent and dependent measures were obtained from the same source (Podsakoff and Organ 1986). Our sample consisted predominantly of mid- to senior-level managers with significant levels of relevant knowledge, which tends to mitigate single source bias (Mitchell 1985). Common method bias was also reduced by separating the dependent and independent variable items over the length of the survey instrument and by assuring participants that their individual responses would be kept anonymous (Podsakoff et al. 2003). We also applied a common latent factor analysis with respect to the full measurement model in AMOS 22 to assess whether common method bias exists, but since this resulted in a small common variance and all changes in factor loadings of the items are smaller than 0.2, we conclude that the tests of reliability, validity, overall model fit and common method bias provide adequate support for the appropriateness of the model constructs.

We developed and tested a research model in order to examine leadership behaviors of management to operate on performance improvement and encourage Lean management; Table 5 gives an overview of the direct and indirect testing effects. Consistent with the predictions, we found that encouraging continuous improvement is positively related to both Lean (Hypothesis 2) and the level of process improvement (Hypothesis 3), and Lean is positively related to the level of process improvement (Hypothesis 1). Though communicating the meaning of Lean is positively related to Lean, we did not find a relationship between the interaction term communicating the meaning of Lean × encouraging continuous improvement and Lean. We therefore reject Hypothesis 4.

We found something similar for active steering on performance improvement by management. Although we found a direct relationship between active steering on performance improvement and process improvement, we did not find an indirect relationship via encouraging continuous improvement. Hence, we reject Hypothesis 6. However, we did find that active steering on performance improvement has a reinforcing effect on the relationship between Lean and process improvement (Hypothesis 5). For respondents with a low level of steering on performance improvement, Lean does not lead to process improvement, while it does for respondents with average and high levels of steering on performance improvement.

The practical implication of this study is that management must actively facilitate the implementation of Lean by communicating the meaning of Lean and to actively promote and stimulate the use of Lean tools and practices to continuously improve processes and activities. Communicating the meaning of Lean means that management continuously promote the reason and purpose of Lean, explain the True North of the Lean-organization (see for instance Rother 2009; Worley and Doolen 2006) and stress the importance to build and strengthen Lean capabilities. This confirms the claim of Lakshman (2006) that the degree to which leaders communicate the importance of continuous improvement of work processes and outcomes will be positively related to their effectiveness and to the unit’s performance. In addition, management must encourage continuous improvement by means of promotion of employee responsibility, employee empowerment, provision of training, promotion of teamwork and team-based problem solving. Ensuring that people in various parts of the organization continuously strive to improve work processes to increase business performance is indeed an important management task in introducing and implementing Lean.

The results of this study also show that with the absence of active steering on performance improvement by management, Lean does not result in process improvement; and that the more management operates on performance improvement, the more Lean results in process improvement. This finding fits the hard technical perspective of Lean leadership that includes setting ambitious goals and steering on performance improvement (Linderman et al. 2006); managing by facts by utilizing objective data (Choi and Eboch 1998), and providing timely feedback on performances (Waldman et al. 1998). However, the results appear to be contrary to what is commonly thought of appropriate management behavior for continuous improvement and the building of a culture of continuous improvement. Excessive focus on performance improvement and performance feedback might hamper continuous improvement (Anand et al. 2009), which concurs with the claim of Browning and Heath (2009) that too much top management involvement and support of lean implementation can contribute to pushing lean implementation into the region of negative returns. Deming (1986) also contend that short-term performance goals had to be replaced with clear leadership vision concerning how the organization can achieve quality by continuous process improvement. The role of top management in culture of continuous improvement initiatives is to provide a vision that guides the formulation of goals at the middle and lower managerial levels (Anand et al. 2009). Hence, management must ensure employees’ commitment through their involvement and participation in goal-setting. Communicating the organization’s vision and the meaning of Lean aims to align multiple sub-goals. Subsequently, encouraging continuous improvement, the delegation of authority and the facilitation of participatory decision making leads to higher employee commitment, motivation and satisfaction, and consequently higher improvement performance. However, this does not mean that management should not actively steer on performance improvement; the question whether active steering on performance improvements and providing feedback is counterproductive on employee morale and well-being and consequently on building a culture of continuous improvement depends on, amongst others, whether employees perceive it as a job demand (something that requires sustained effort that is therefore associated with certain costs) or a job resource (something that is functional in achieving personal work goals); see for instance Bakker and Demerouti (2007). An important factor that might influence employees’ performance management and feedback as a job demand or a job resource, is the attribution employees make about management’s purpose in implementing performance management and feedback (Giesbers et al. 2015), that can be either commitment-focused or control-focused; see for instance Arthur (1994). Commitment-focused attributions connote positive consequences for employees, while control-focused attributions connote negative consequences (Giesbers et al. 2015). However, more research is needed to find out how to manage this crossover point, how to balance ‘control’ and ‘exploration’ structuring behaviors (Douglas & Judge, 2001) and how to combine decentralized goal-determination with result-oriented management.

A cross-sectional research design generally limits the extent to which cause-effect relationships can be inferred. Additional research to collect and analyze longitudinal data is also recommended for this research, for instance through multiple case studies using multiple informants and multiple sources. That will also help overcome possible problems with common method bias and increase overall criterion validity as we used of single respondents’ perceptual data.

Future research into Lean leadership and appropriate management behavior is recommended. This research shows that management commitment and corresponding visible management actions positively impacts the success of Lean, which corresponds with the findings of McLachlin (1997) and Mann (2009). However, as there is little empirical work that tests the type of leadership (e.g., transformational, servant, instrumental and empowering leadership) at different hierarchical levels and for different stages of Lean implementation, we would welcome future research on these Lean leadership issues.

Another interesting topic for future research is to study Lean and related management behavior (i.e. steering on performance management and feedback) on employees’ well-being, as performance measures such as employee satisfaction and improvement in employee morale may have an enhancing influence on the relation between Lean and organizational performance. Also the effects of the use of technical Lean practices on job design and the consequences on employee motivation is an interesting topic for future research. Future research should also test the differences and interconnectedness of technical Lean practices and continuous improvement as an organizational infrastructure (in coherence with a culture of continuous improvement) as the results of this paper suggests that these are different, though related constructs (i.e. Lean practices predominantly refer to a more technical construct, while continuous improvement is something that occurs more organically). Indeed, this study shows that the introduction of Lean practices does not automatically lead to process improvement, but that management must actively manage that: but does that also apply for a culture of continuous improvement and infrastructural practices of continuous improvement?