عوامل حیاتی موفقیت برای خروجی های منابع انسانی در حوادث کایزن: یک مطالعه تجربی

Kaizen events are an increasingly common organizational improvement mechanism aimed at work area transformation and employee development. While many anecdotal design prescriptions exist, there is little empirical evidence of which input and process factors are most strongly related to Kaizen event outcomes in practice. This paper uses results from a field study of 51 events in six manufacturing organizations to identify the set of input and process factors that most strongly relate to the development of employee attitudinal outcomes and problem-solving capabilities in Kaizen events. These results are used to develop guidelines for organizations and identify directions for future work.

The study of improvement programs has long been a focus of the operations management (OM) and industrial engineering community (e.g., Chan et al., 2005; Dar-El, 1997; Guimaraes, 1997; Gunasekaran et al., 1994; Hales and Chakravorty, 2006; Herron and Braiden, 2006; Launonen and Kess, 2002; McIntosh et al., 2001; Van Landeghem, 2000; Vits and Gelders, 2002). Recently, lean manufacturing (Womack et al., 1990) has become a—if not the—dominant improvement paradigm, leading to a variety of studies examining this topic (e.g., Matusi, 2007; Panizzolo, 1998; Simons and Taylor, 2007; Warnecke and Huser, 1995). Within lean manufacturing, one increasingly utilized mechanism is the Kaizen event, a focused and structured continuous improvement project, using a dedicated cross-functional team to address a targeted work area, to achieve specific goals in an accelerated timeframe (usually 1 week or shorter) ( Farris et al., 2008b). In addition to potential, direct improvements in the target work area, Kaizen events are purported to serve as a “just-in-time” training mechanism for participating employees ( Drickhamer, 2004a), helping these employees develop new problem-solving capabilities and increased motivation to participate in future improvement activities. However, despite their popularity and potential benefits, Kaizen events have not been widely studied to date ( Bateman, 2005; Melnyk et al., 1998). Many guidelines for Kaizen event design exist, primarily in the practitioner literature; however, these guidelines do not appear to have been tested through empirical research. Until the determinants of Kaizen event outcomes are well understood, organizations will not be able to systematically manage Kaizen events to consistently achieve positive results. This paper presents findings from a field study of 51 Kaizen events in six manufacturing organizations, where multiple regression was used to test the relationships between Kaizen event input and process factors and employee attitudinal and problem-solving capability outcomes. Findings are used to develop design guidelines for organizations using Kaizen events and to lay a foundation for future research. Section 2 reviews the literature related to this topic, Section 3 presents the research methodology, Section 4 presents results, and Section 5 discusses study findings, limitations, and directions for future research.

5.1. Discussion of findings Following the conclusion of the study, findings were reported to the participating organizations to evaluate the validity of study conclusions and to allow the organizations to benefit from the results. In general, participating organizations found the study results convincing and used study feedback to evaluate the effectiveness of their current practices and to identify potential changes. As mentioned, the study provided partial support for the three broad research hypotheses. Specific findings of the study are as follows: • Internal processes and goal clarity were the strongest predictors of both kaizen capabilities and attitude. • Management support and team functional heterogeneity were significant predictors of attitude, but not kaizen capabilities. • Team autonomy, affective commitment to change, goal difficulty, work area routineness, team kaizen experience and team leader experience were significant predictors of kaizen capabilities but not attitude. • Some variables proposed to affect Kaizen event outcomes, i.e., action orientation, tool quality, tool appropriateness, and event planning process, showed no significant relationship to either outcome in this study. Internal processes was by far the strongest predictor of both attitude (β=0.694) and kaizen capabilities (β=0.465), suggesting that maintaining positive group dynamics may be the most important factor for generating employee motivation to participate in future improvement activities and for developing employee problem-solving capabilities. This result is aligned with previous empirical team research, which has consistently reported a positive relationship between internal processes and human resource outcomes (e.g., Bailey, 2000; Jehn, 1995; Pinto et al., 1993; Vinokur-Kaplan, 1995). Acting indirectly through internal processes, goal clarity is also one of the strongest predictors of attitude and kaizen capabilities, accounting for approximately 60% of the direct effect of internal processes. This finding is aligned with several previous studies in team effectiveness reporting a positive relationship between goal clarity and effectiveness (e.g., Doolen et al, 2003a; Koch, 1979; Pritchard et al., 1988; Van Aken and Kleiner, 1997). This finding also supports the majority of published Kaizen event accounts, which have suggested using clear and specific goals (e.g., Adams et al., 1997; Bradley and Willett, 2004; Rusiniak, 1996), and contradicts the few that have suggested using more loosely defined goals ( Kumar and Harms, 2004; Wittenberg, 1994). The finding that management support had a significant positive relationship to attitude, but not to kaizen capabilities, partially supports the consistent emphasis on management support in published Kaizen event accounts (e.g., Adams et al., 1997; Bateman, 2005; Bicheno, 2001; Bradley and Willett, 2004; Kumar and Harms, 2004; Martin, 2004; McNichols et al., 1999; Patil, 2003; Sheridan, 1997; Tanner and Roncarti, 1994; Taylor and Ramsey, 1993). In addition, it agrees with previous team studies which have reported a positive relationship between management support and team effectiveness (e.g., Campion et al. 1993; Doolen et al., 2003a; Hyatt and Ruddy, 1997; Pinto and Slevin, 1987). The finding that team functional heterogeneity had a significant negative relationship to attitude, but not to kaizen capabilities, is aligned with previous team research suggesting that cross-functional teams may experience lower levels of enjoyment in working together ( Ancona and Caldwell, 1992; Amason and Schweiger, 1994; Baron, 1990). However, the mediation analysis showed no significant relationship between team functional heterogeneity and internal processes, suggesting that more diverse teams were still able to develop effective communication processes ( Azzi, 1993; Earley and Mosakowski, 2000). This appears to agree with team research indicating that cross-functional teams may experience high quality decision-making processes (e.g., Jackson, 1992; Jehn et al., 1999; Lovelace et al., 2001; McGrath, 1984). The finding that team autonomy had a significant positive relationship to kaizen capabilities, but not to attitude, partially supports the assertions in the published Kaizen event accounts that autonomy promotes positive outcomes (e.g., Adams et al., 1997; Bicheno, 2001; Bradley and Willett, 2004; Foreman and Vargas, 1999; Kumar and Harms, 2004). Some empirical team studies have similarly reported a positive relationship between autonomy and outcomes ( Cohen and Ledford, 1994; Wall et al., 1986). The finding that affective commitment to change had a significant positive relationship to kaizen capabilities, but not to attitude, partially supports the emphasis in the published Kaizen event accounts on creating team buy-in for the improvements targeted by the event ( Bradley and Willett, 2004; McNichols et al., 1999; Melnyk et al., 1998; Taylor and Ramsey, 1993; Wheatley, 1998). In addition, this finding is aligned with the organizational change literature, which suggests a relationship between participant commitment and change initiative effectiveness ( Chan et al., 2005; Keating et al., 1999). The finding that goal difficulty had a significant positive relationship to kaizen capabilities, but not to attitude, suggests that, from the standpoint of employee learning, the more challenging the goals, the more positive the skill development results, which would support the Kaizen event accounts that advocate using highly challenging goals (e.g. Bicheno, 2001; Bradley and Willett, 2004; Cuscela, 1998; Gregory, 2003; Kumar and Harms, 2004; LeBlanc, 1999; Minton, 1998; Rusiniak, 1996; Tanner and Roncarti, 1994; Treece, 1993). However, findings related to technical system outcomes revealed negative relationships between goal difficulty and technical success outcomes ( Farris et al., 2008a), suggesting a trade-off between technical and skill development outcomes. Meanwhile, the finding that work area routineness had a significant positive relationship to kaizen capabilities, but not to attitude, supports the suggestions in published Kaizen event accounts that less complex work areas provide favorable “learning laboratories” for skill development ( LeBlanc, 1999; Bradley and Willett, 2004), and aligns with findings in the organizational learning literature ( Vits and Gelders, 2002). This effect may be due to the fact that more predictable work areas allow more complete implementation of lean practices. The finding that team kaizen experience had a significant negative relationship to kaizen capabilities, but not to attitude, is consistent with the well-known learning curve effect ( Wright, 1936), which predicts decreasing incremental gains in learning from each successive attempt at a given application. The finding that team leader experience had a significant negative relationship to kaizen capabilities, but not to attitude, is unexpected and warrants further investigation. A manager in one of the participating organizations suggested that, based on his experience, teams with more seasoned leaders tended to skip or hurry through steps in the Kaizen event process, under the paradigm that they “already know what they are doing.” In this study, although data were collected on the activities of the team through the team activities log, there was no formal measure of the extent to which teams adhered to the formal Kaizen event process. Recent literature has indicated a link between process adherence and team and organizational outcomes ( Douglas and Judge, 2001; Ghosh and Sobek, 2007; Handfield et al., 1999; Linderman et al., 2006), although some published Kaizen event guidelines propose a less rigid adherence to the organization's Kaizen event format (e.g., Bradley and Willett, 2004), including even omitting training in some cases (e.g., Bicheno, 2001; Gregory, 2003; McNichols et al., 1999). Teams with high team leader experience might also be more likely to jump to solutions similar to those used in previous events that the leader has conducted, thus limiting team creativity and team member participation in designing solutions. Or, more experienced leaders might tend to adopt a more directive leadership style, allowing individual team members less role in shaping the solution. Such directive leadership has been shown to be negatively related to team effectiveness in at least one study ( Durham et al., 1997). Finally, it is possible that in teams with less experienced leaders the Kaizen event facilitator, who is typically highly experienced in problem-solving tools and applications, may tend to become more involved in coaching the team through the solution process, providing on-the-floor training, etc., thus contributing to greater learning. Future research on the relationship between team leader experience, team process adherence, team creativity, leadership style, facilitator involvement, and kaizen capabilities is clearly needed. Finally, some variables hypothesized to affect outcomes, i.e., action orientation, tool quality, tool appropriateness, and event planning process, showed no significant relationship to either outcome in this study. This finding does not prove that these variables are unimportant to the development of attitude and kaizen capabilities, although it does suggest that they might be less important than other variables. It should also be noted that these variables may be related to technical system outcomes or to social system outcomes not studied in this research. Although admittedly preliminary, the findings of this research can be further used to develop initial guidelines for organizations (to be tested in future research), including the following: • To obtain maximum impact on both employee attitude and problem-solving ability development, these results suggest that organizations should seek to maintain a high level of positive internal team dynamics, through the use of structured mechanisms (e.g., team ground rules, ice breakers, training, charters) and facilitator coaching. A key structural mechanism is the development of explicit goal statements in advance of the event. • To further develop employee motivation to participate in events, these results suggest that organizations must maintain strong and visible management support during events. Meanwhile, organizations should recognize that cross-functional heterogeneity may have a negative impact on team member affect toward events, and potentially compensate for this effect through other variables, e.g., internal processes, management support. • To further increase team problem-solving capability development, these results suggest that teams should be allowed a high degree of autonomy and that the organization should use mechanisms designed to increase participant buy-in for the event (e.g., clearly describing the reasons for the event, demonstrating how the event will positively impact both individual and organizational interests). Organizations should also recognize that the use of difficult goals could increase team learning, although other findings suggests that goal difficulty may compromise technical system results (Farris et al., 2008a). • Finally, these results suggest that organizations should recognize that more complex work area may contribute to a lower degree of team learning, and potentially compensate for this effect through other mechanisms, e.g., autonomy, goal clarity, etc. Organizations should also recognize that greater team and team leader experience do not necessarily result in greater team learning, and should consider compensating for this effect through other mechanisms or through using less experienced personnel. 5.2. Limitations and future research To the authors’ knowledge, this study was the first to empirically investigate the quantitative relationships between input and process variables and initial human resource outcomes in Kaizen events, using multiple events from multiple organizations. Additional research is needed to confirm and extend these findings. Other study limitations include: • Although data were collected from 51 Kaizen events and over 300 participants, the sample represented only six organizations. Furthermore, the selection criteria deliberately resulted in the selection of manufacturing organizations that were relatively experienced in the strategic use of Kaizen events. Thus, it is possible that findings might not generalize to markedly different organizational contexts. However, the six organizations spanned a variety of industries and approximately 16% of the events were in non-manufacturing work areas. • Events were selected using a random sampling process designed to capture the mixture of events occurring naturally within the participating organizations; the researchers did not attempt to control the percentages of specific event types sampled. Most of the events in this research (71%) were general processes improvement events, often using the “standard work” methodology. A smaller number of events were of other types: total productive maintenance (TPM) (14%), single minute exchange of die (SMED) (8%), 5S (6%) and value stream mapping (VSM) (2%). (Note, these percentages total 101% due to rounding.) In addition, most events included implementation of the solution during the event (75%), targeted manufacturing processes or work areas (84%), and were highly successful with 100% goal achievement (69%). Thus, it is possible that findings might not generalize to events of markedly different types. • This research only included event-level variables. It is possible that some organizational-level effects also contribute to event-level outcomes. While GEE accounts for correlation in regression residuals between teams within the same organization, it also assumes uniformity of regression slopes across organizations. Hierarchical linear modeling (HLM) has the capability to model differences in effects of both organization and event-level variables across organizations. However, as noted, a larger data set are needed to construct models of any size (Raudenbush and Byrk, 2002). • Finally, this research only investigated the determinants of initial event outcomes, while sustainability of both human resource and technical outcomes is of great interest, and many organizations appear to struggle to sustain initial gains from Kaizen events (e.g., Bateman and David, 2002; Bateman 2005; Laraia et al, 1999). In summary, this research has systematically investigated the empirical relationships between input and process variables and human resource outcomes using a sample of 51 events from six organizations. In closing, the authors reiterate the call from Melnyk et al. (1998) regarding the need for additional research on this topic. If the proliferation of books, websites, conference proceedings and non-refereed articles is any indication, organizational interest in Kaizen events only appears to have increased since the late 1990s, and little scholarly research has been conducted to date. It is the responsibility of the operations management and industrial engineering research community to help organizations better understand this phenomenon and how it contributes to longer-term organizational success.