Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers...

39
Post-print. This is the Authors’ Original Manuscript of an article published in the International Journal of Operations and Production Management on January 14, 2021. For the final version, see https://doi.org/10.1108/IJOPM-02-2020-0092 Learning lean: Rhythm of production and the pace of lean implementation Torbjørn Netland 1 , Jason Schloetzter 2 , and Kasra Ferdows 2 1 Department of Management, Technology, and Economics, ETH Zurich, Zurich, Switzerland 2 McDonough School of Business, Georgetown University, Washington D.C., USA ABSTRACT Purpose: Why some assembly factories implement a lean program faster than others is an enduring puzzle. We examine the effect of a fundamental characteristic of every assembly factory—its rhythm of production. Approach: We designed a multi-method study and collected data from a leading global equipment manufacturer that launched a lean program across its factory network. We use quantitative data gathered from internal company documents to test our hypothesis that production rhythm affects the pace of lean implementation. We then analyze qualitative data from interviews and factory visits to derive theoretical explanations for how production rhythm affects lean implementation. Findings: Consistent with our hypothesis, we present evidence that factories with faster production rhythms implement lean faster than those with slower rhythms. This evidence is consistent with learning theories as well as the literature on organizational routines and forms of knowledge. We propose a theory of the relation between rhythm and learning in lean implementation. Research implications: The hitherto unexplored relation between production rhythm and lean implementation raises intriguing questions for scholars and ushers new insights into how organizations learn to implement lean. Practical implications: Organizations need to calibrate their expectations for lean implementation pace when their factories have widely different production rhythms and find ways to mitigate any adverse effects slower rhythms may have. Organizations can alleviate the unfavorable context of slower rhythms by inculcating practices in the factory that emulate the learning environment present in faster-paced factories. Originality: We contribute novel quantitative and qualitative evidence that production rhythm affects lean implementation through learning-based mechanisms. Keywords: Lean, Operational excellence, Organizational learning, Production rhythm, Takt time, Multi-method design

Transcript of Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers...

Page 1: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

Post-print. This is the Authors’ Original Manuscript of an article published in the International Journal

of Operations and Production Management on January 14, 2021. For the final version, see

https://doi.org/10.1108/IJOPM-02-2020-0092

Learning lean: Rhythm of production and the pace of lean implementation

Torbjørn Netland1, Jason Schloetzter2, and Kasra Ferdows2

1Department of Management, Technology, and Economics, ETH Zurich, Zurich, Switzerland 2 McDonough School of Business, Georgetown University, Washington D.C., USA

ABSTRACT

Purpose: Why some assembly factories implement a lean program faster than others is an enduring puzzle. We examine the effect of a fundamental characteristic of every assembly factory—its rhythm of production. Approach: We designed a multi-method study and collected data from a leading global equipment manufacturer that launched a lean program across its factory network. We use quantitative data gathered from internal company documents to test our hypothesis that production rhythm affects the pace of lean implementation. We then analyze qualitative data from interviews and factory visits to derive theoretical explanations for how production rhythm affects lean implementation. Findings: Consistent with our hypothesis, we present evidence that factories with faster production rhythms implement lean faster than those with slower rhythms. This evidence is consistent with learning theories as well as the literature on organizational routines and forms of knowledge. We propose a theory of the relation between rhythm and learning in lean implementation. Research implications: The hitherto unexplored relation between production rhythm and lean implementation raises intriguing questions for scholars and ushers new insights into how organizations learn to implement lean. Practical implications: Organizations need to calibrate their expectations for lean implementation pace when their factories have widely different production rhythms and find ways to mitigate any adverse effects slower rhythms may have. Organizations can alleviate the unfavorable context of slower rhythms by inculcating practices in the factory that emulate the learning environment present in faster-paced factories. Originality: We contribute novel quantitative and qualitative evidence that production rhythm affects lean implementation through learning-based mechanisms. Keywords: Lean, Operational excellence, Organizational learning, Production rhythm, Takt time, Multi-method design

Page 2: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

1

1. INTRODUCTION

Lean production creates a greater competitive advantage if it is implemented faster

(Derfus et al. 2008; Su et al. 2015). Yet, implementing lean remains elusive and slow for most

manufacturers (Bateman 2005; Pay 2008; Sim and Rogers 2008; Sadun et al. 2017; Womack

2017). A recent World Management Survey of 12,000 firms concludes as follows:

“Achieving operational excellence [e.g., lean manufacturing] is still a massive challenge for many organizations. Even well informed and well-structured companies often struggle with it. This is true across countries and industries—and in spite of the fact that many of the managerial processes we studied are well known.” (Sadun et al. 2017, p. 123)

This enduring puzzle motivates our research. The literature suggests a host of factors

that might help or hinder lean implementation, including management commitment (Rodgers

et al. 1993; Emiliani and Stec 2005), plant size (Shah and Ward 2003), plant age (Shah and

Ward 2003; Furlan et al. 2011), the role of national culture (Wiengarten et al. 2015), product

complexity and variety (MacDuffie et al. 1996; Browning and Heath 2009), the role of unions

(Fast 2014; Fullerton et al. 2014), prior experience with process improvement programs

(Fullerton et al. 2014; Netland and Ferdows 2016), the type of management control practices

in the plant (Netland et al. 2015), and environmental dynamics, measured by market growth

(Eroglu and Hofer 2011), rate of change in technology (Chavez et al. 2015), or demand

variability (Bortolotti et al. 2013). We contribute to this rich literature by proposing that a

hitherto unexplored variable, the rhythm of production in the factory, also impacts lean

implementation.

We uncovered the potential relation between production rhythm and the pace of lean

implementation during a four-year collaborative research project with a leading global

equipment manufacturer (GEM; a pseudonym) in the commercial vehicles industry. GEM had

launched a corporate-wide lean program across its global production network, but despite

Page 3: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

2

sustained efforts by corporate headquarters, the pace of lean implementation varied

considerably among the factories. We combine quantitative and qualitative data gathered from

GEM’s formal factory-level lean assessments and our own factory visits to examine the

potential role of rhythm first-hand.

2. BACKGROUND

We noticed the possible influence of production rhythm on lean implementation after

visiting about two dozen factories in GEM’s global factory network. While the factories shared

similar characteristics, their production rates varied significantly—some assembled a unit

every few minutes while others assembled a similar unit every few hours. Even when we

accounted for local management commitment and experience, process technologies, factory

size, and factory age, there appeared to be a relation between a factory’s production rhythm

and its pace of lean implementation. This pattern was particularly clear when we visited two

GEM factories that were located next to each other. Both factories assembled heavy vehicles

of similar complexity, using similar technology, with similarly-skilled shop floor employees.

But they had distinctly different production rhythms: one factory assembled two units per shift

and the other 15 per shift, and lean implementation was noticeably faster in the latter factory.

This was an intriguing observation that, upon reflection, made intuitive sense from both

a theoretical and practical standpoint. As we discuss in more detail later in our study, learning

curve theory, supplemented by the theories of organizational routines and codification of tacit

knowledge, suggests that managers and operators in factories with faster production rhythms

are likely to have more opportunities to observe, learn, and implement lean practices than those

in factories with slower rhythms. From a practical standpoint, production rhythm is a deep,

structural parameter that influences the learning opportunities on the factory floor—it can be

considered the “heartbeat” of an assembly line. Rhythm influences the layout of a factory, the

choice of equipment, the design of shop-floor jobs, even the culture in the factory (as we argue

Page 4: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

3

later). Therefore, rooted both in theory and in practical observation, it is reasonable to expect

a relation between the rhythm in a factory and how fast it inculcates the extensive practices

prescribed by a lean program. Yet, despite the vast literature on lean manufacturing, the

importance of production rhythm has remained unnoticed and unexplored.

Three clarifications are needed. First, we focus on typical discrete manufacturing

processes in which humans and machines assemble a product. We do not address highly

automated assembly processes, such as fast-moving consumer goods industries or process

industries in which “assembly” is often continuous. In the classic product-process model of

Hayes and Wheelwright (1984), we only study the assembly line. Our findings may provide

insights into other process designs, but we are not arguing that our results will persist in these

settings.

Second, our research setting involves factories with distinctly different production

rhythms, ranging from few minutes to several hours between the production of consecutive

products. Our study does not investigate the impact of small differences in production rhythms

across factories or day-to-day fluctuations of rhythm in the same factory. Instead, we examine

how lean implementation is affected in factories with distinctively different rhythms.

Finally, we use “rhythm” instead of “takt time” to describe the pace of assembly to

avoid potential confusion. In fact, rhythm is the original meaning of “takt time.” Takt time

stems from the German word “Taktzeit,” which refers to the grouping of musical notes within

the same beat that underlies the rhythm of music. German producers of military aircraft in the

1930s were the first to introduce this notion in manufacturing (Overy 1994; Holweg 2007).

The Junkers aircraft factory in Dessau, Germany, organized production into segments of equal

production times. The aircraft moved to the next station at fixed intervals, creating a steady

rhythm of production. Thus, the original notion of takt time was essentially a way to organize

work inside a mass-production factory and was not directly related to outside demand.

Page 5: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

4

It was the Toyota Production System that linked takt time with the market demand of a

factory. As Baudin (2002, p. 43) states, “assuming we complete the product one unit at a time

at a constant rate during the net available work time, the takt time is the amount of time that

must elapse between two successive unit completions in order to meet demand [italics added].”

In a lean factory, the takt time should ideally equal the pace of demand (Ohno 1988; Hopp and

Spearman 2004). We agree, but this is not what we study in this research, and we choose the

term “rhythm” instead of “takt time” because we are investigating the effect of the general pace

of production inside the factory—regardless of how this pace is set. We also opted for not using

the term “cycle time” because it is sometimes used to convey throughput time and not the

rhythm of production (e.g., Hopp and Spearman 2011).

3. LITERATURE REVIEW AND HYPOTHESIS DEVELOPMENT

Transforming an organization with the help of a lean program involves inculcating a

culture of continuous improvement and adopting a wide range of specific practices throughout

the factory. Despite a vast literature on lean implementation (e.g., Shah and Ward 2003; Shah

et al. 2008; Bortolotti et al. 2015; Netland et al. 2015; Netland and Ferdows 2016; Galeazzo

and Furlan 2018; Knol et al. 2019), the majority of companies are struggling in their lean

journeys (Pay 2008; Sadun et al. 2017). Prior work has suggested a long list of factors that

support or hinder the implementation of lean, but no study has yet devoted attention to the role

of production rhythm. This may be an unfortunate oversight, as learning curve theory (e.g.,

Yelle 1979; Adler and Clark 1991; Fogliatto and Anzanello 2011) suggests that production

rhythm relates directly to learning.

Studies of production improvement recognize the critical role of learning in quality

programs (see, in particular, Deming 1986; Linderman et al. 2004; Choo et al. 2007; Anand et

al. 2009) and lean implementation (e.g., MacDuffie 1997; Spear and Bowen 1999; Ballé et al.

Page 6: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

5

2016; Netland and Powell 2016; Powell and Coughlan 2020). For example, anecdotal evidence

reveals that Toyota referred to their famous TPS as an acronym for the “Thinking People

System.” Explaining the underlying foundations of the TPS, Spear and Bowen (1999, p. 5)

suggest, “They [Toyota] use a teaching and learning approach that allows their workers to

discover the rules as a consequence of solving problems.” While these studies help understand

the role of learning in lean, none of them explicitly links the notion of learning to the production

rhythm inside a factory.

A key observation from the learning curve theory literature is the presence of a strong

relation between the volume of output, repetition of tasks, and organizational learning (Wright

1936; Yelle 1979; Adler and Clark 1991; Lapré et al. 2000). This theory implies that an

assembly factory with a faster production rhythm should be able to learn new practices faster

than a factory with a slower production rhythm. In other words, when two assembly factories

embark on a lean journey, the factory with the faster production rhythm is likely to learn and

implement lean practices faster because its workers repeat the practices more often.

Studies of organizational routines and the codification of tacit knowledge inform our

understanding of how the production of higher volumes of output—hence more frequent

repetition of tasks—expedites learning in an assembly factory (cf. Anand et al. 2009). Feldman

and Pentland (2003, p. 113) highlight the importance of repetition in the theory of

organizational routines, particularly when routines “create variations that other participants

recognize as legitimate instances of the ostensive aspect of the routine.” In the lean context,

these new ostensive routines are typically the day-to-day manufacturing practices in the factory

that the lean journey aims to change for the better (cf. Shah and Ward 2003; Shah et al. 2008;

Netland and Ferdows 2016). Implementing lean requires employees to learn new routines (e.g.,

Knol et al. 2019) and “internalize” them (Nonaka 1994). As these routines are repeated more

Page 7: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

6

frequently in factories with faster rhythms, they can reveal their positive results to all

employees at a faster pace and, hence, inculcate the new routines faster in the factory.

Studies of the codification of tacit knowledge add an explanation as to why factories

with faster production rhythms can learn faster. Tacit knowledge is subjective, experience-

based, and cannot be explicitly expressed (Polanyi 1967). Transforming tacit knowledge into

codified or explicit knowledge makes the knowledge easier to teach and transfer to others.

Nonaka and von Krogh (2009) posit that the conversion of tactic knowledge into “explicitness”

happens via action and practice, which occurs more frequently in factories with faster

production rhythms. Furthermore, a faster rhythm necessitates the assignment of fewer tasks

to be performed at each workstation, which reduces the complexity of codification and

“standardization” of the production method at that workstation. In contrast, a slower production

rhythm can obscure differences in the methods followed by operators for doing the same task

and keep their know-how in tacit form.

In short, learning curve theory, supplemented by the theories of organizational routines

and codification of tacit knowledge, confirm our observations from the factory floor: managers

and operators in factories with faster production rhythms are likely to have more opportunities

to observe, learn, and implement lean practices than those in factories with slower rhythms.

This occurs for at least three reasons. First, the frequency of repetition associated with faster

rhythms provides more opportunities to learn. Second, faster rhythms entail fewer tasks at each

workstation, thereby reducing the complexity of codifying, standardizing, and learning the

tasks. Third, faster rhythms reveal any deviations in the methods used to perform tassks more

quickly, which fosters both attention and motivation to rectify them. We formulate our

hypothesis as follows:

Hypothesis 1: Assembly factories with faster production rhythms implement lean programs faster than assembly factories with slower production rhythms.

Page 8: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

7

4. RESEARCH SETTING AND METHOD

We use a mixed-method, deductive-inductive research design that complements

hypothesis testing with a structured analysis of qualitative data. Our research setting is a four-

year collaboration with GEM (a pseudonym), a leading global equipment manufacturer in the

commercial vehicles industry (e.g., heavy construction vehicles, trucks, and buses). In 2019,

GEM reported approximately $43 billion in revenue, with 104,000 employees and more than

60 assembly factories on six continents. The deductive part of our study tests the research

hypothesis using GEM-internal scorecard data. The inductive part of our study analyzes

interview and observation data collected during field visits and seeks to provide detailed

insights into the theoretical explanation for the quantitative results. Thus, our research design

first uses quantitative data to test our hypothesis and then leverages our qualitative data to shed

additional light on why the quantitative results are present in the data.

Our research setting offers several important advantages. Our sample consists of

factories with distinctly different production rhythms, ranging from a few minutes to several

hours between consecutive products. This variation provides an opportunity to examine lean

implementation across factories with significantly different rhythms while holding constant the

characteristics of the lean program itself. Further, the GEM lean program is a highly

standardized and typical lean program. It is based on the classic groups of lean practices found

in the literature: just in time, total quality management, total productive maintenance, human

resource management, and continuous improvement (e.g., Shah and Ward 2003; Anand et al.

2009; Furlan et al. 2011). The program is supported by significant internal resources, resulting

in detailed scorecards that measure the extent of implementation of lean practices at the factory-

level. This is particularly important because measuring the extent of lean implementation in a

factory is difficult. Most important, unlike many quantitative studies of lean (and other

improvement programs) that are based on cross-sectional data from many firms and industries

Page 9: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

8

(a few examples include Shah and Ward 2003; Shah et al. 2008; Fullerton et al. 2014; Jacobs

et al. 2015; Onofrei and Prester 2019), our sample is from one company, and all the factories

operate in a narrow segment of one industry. Hence, our analysis implicitly controls for a

variety of contextual variables. These advantages provide an opportunity to bring both focus

and depth to our mixed-method study.

4.1 Measuring the dependent variable: Lean Implementation

For our quantitative analysis, we measure the dependent variable with GEM-internal

scorecard data of lean implementation in each factory. GEM assesses the extent of lean

implementation in each of its factories regularly. Our sample is based on the intersection of the

factories that had been assessed by GEM’s lean implementation auditors from 2009-2012 and

for which we have data about the production rhythm. This intersection consists of 49 factories,

37 of which have had more than one lean assessment. In total, there have been 95 lean

assessments of these 49 factories.

GEM divides its lean program into five “groups” of practices, with each group

consisting of three to five “principles,” and each principle consisting of a number of

“practices.” Appendix A describes GEM’s lean program, its constructs, and its measurements.

GEM’s target is to assess the implementation of lean in each factory once every two years, but

this varies based on individual situations and logistical concerns. The headquarters usually

suggests when to assess each factory three to six months in advance. A team of experts,

consisting of two or three experienced lean auditors from headquarters and two to four certified

or in-training assessors from other GEM factories, spend four or five days in the factory. This

team assigns a rating based on the factory’s adherence to each of the nearly 100 practices in

the GEM’s lean program. These ratings are then used to compute, using simple averages, the

scores for each principle, each group, and the entire factory. For example, the score for the

“TQM group” is the average score for “zero defects,” “quality assurance,” and “product and

Page 10: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

9

process quality planning.” Table A-1 reports high values of Cronbach’s alpha in every group,

indicating within-set homogeneity of the principles GEM uses in its lean assessments and

suggests the GEM measurement process is internally consistent. The computed score for the

entire factory, Lean Implementation, is our dependent variable.

While we follow GEM’s method and use simple averages to measure Lean

Implementation, we also examined an alternative method for aggregating the lean assessment

data to ensure our evidence is not sensitive to the use of averages. Specifically, we conducted

a factor analysis of all principles and found that the inferences remain unchanged—i.e., using

simple averages does not drive our empirical results. As another reliability check, following

the general consensus in the literature that a factory’s operational performance improves as it

implements lean more extensively (e.g., Shah and Ward 2003; Browning and Heath 2009;

Fullerton et al. 2014; Netland and Ferdows 2016; Onofrei and Prester 2019), we reviewed

internal company factory performance data to see if there were any anomalies; we found none.

Overall, we believe GEM’s Lean Implementation assessment scores are reliable measures to

gauge the extent of lean implementation in a factory.

4.2 Measuring the independent variable of interest: Production rhythm

Our independent variable of interest is the production rhythm of the factory. We use the

average time in minutes between the production of two consecutive units on the main assembly

line in the factory. Because production rhythm is the key variable of our study, we triangulate

three sources to obtain these average times: consultation with GEM managers, direct

observation during factory visits, and reviews of GEM internal documents. We define Rhythm

to be the natural logarithm of these average times, as using the logarithm delineates large

differences.

While the rhythm of production within a factory might vary slightly day-to-day, GEM

managers indicate that this variation is not dramatic; for instance, the day-to-day rhythm does

Page 11: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

10

not vary from minutes to hours. This point is important for our analysis. As mentioned

previously, we examine the impact of large differences in production rhythms (e.g., minutes

versus tens of minutes versus hours) and not small variations (e.g., plus or minus a few

minutes). We also checked to see if there had been any major changes in Rhythm of any factory

in GEM’s network during our sample period; GEM’s records do not show any. GEM managers

explained that this was partly because a factory’s physical layout, machinery, and material

handling systems often constrain the ability to make major changes in production rhythm.

When GEM needs to make a major change in the factory’s production volume, it is usually

done by adjusting employee overtime, number of shifts, or number of working days instead of

changing in the speed of the assembly line.

4.3 Control variables

4.3.1 Management focus on the lean program

Studies suggest that the management’s focus on the lean program relates favorably to

lean implementation (Beer 2003; Emiliani and Stec 2005; Bortolotti et al. 2015; Netland 2016).

We develop two control variables that reflect factory-level management focus on the lean

program using data from GEM internal documents. For the first measure, we examine the

extent to which plant managers engage in “gemba walks” (Gemba Walks), which require

managers to visit the “place where work happens” and to gain first-hand experience regarding

day-to-day work practices and improvement efforts (Bicheno 2004). We measure Gemba

Walks using evidence collected during interviews with senior plant management regarding the

extent to which the factory had institutionalized gemba walks.

The second variable we use to capture management focus is the number of formal lean

assessment audits in a factory (Formal Assessments). Formal assessments require significant

attention from factory managers and operators, and we argue that recurrent assessments foster

a sustained focus across time. Although GEM expects all factories to undergo a formal

Page 12: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

11

assessment approximately every two years, the factories in our sample have had different

numbers of assessments during our sample period, ranging from 1 to 4 (see Panel A of Table

1). This is partly due to scheduling (e.g., the first assessment happened early in the sample

period) and partly due to idiosyncratic reasons (e.g., travel schedules). It is important to note

that assessments do not affect resource allocations—that is, there is neither an explicit link

between the number of formal assessments nor the lean assessment score and the allocation of

factory-level resources. Hence, it is reasonable to assume that management in factories with

multiple assessments has focused more on lean program implementation rather than requesting

assessments to obtain additional resources.

The observation that some factories have had more than one lean assessment suggests

the potential presence of selection issues. We examine this possibility by developing tests of

whether assessment frequency—measured either as (1) an indicator variable that captures

whether a factory has had multiple assessments or (2) as a variable that measures the total

number of assessments—is driven by three factory-level factors: proximity to GEM’s Europe-

based corporate headquarters, the potential strategic importance of the factory (i.e., factory

resides in GEM’s largest division), and factory size (i.e., the number of factory employees).

We also test for associations between Rhythm and each measure of assessment frequency. In

untabulated analyses, we estimate a series of regressions and find no relation between any of

these factors and either the indicator variable that captures whether a factory has had multiple

assessments or the variable that measures the total number of assessments. Most critically, we

do not find a relation between Rhythm and either measure of assessment frequency. The

evidence suggests that selection issues regarding assessment frequency are unlikely to

influence our analyses.

Page 13: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

12

4.3.2 Process complexity

Studies argue that process complexity influences lean implementation (MacDuffie et

al. 1996). GEM management considers that the complexity of production processes does not

differ widely between its assembly factories. However, to ensure that we capture the effect of

broad differences in process complexity, we introduce an indicator variable for the factories

that assemble off-road products, which come in less variety, are less customized, and often

require assembly of fewer components than GEM’s other products. We define Process

Complexity to be an indicator variable equal to one when a factory belongs to the off-road line

of business and zero otherwise.

4.3.3 Other control variables

Based on our discussions with GEM personnel, we also control for factory size,

unionization, and geographic location. Studies suggest that these variables have a direct

influence on the extent of improvement program implementation (e.g., White et al. 1999; Shah

and Ward 2003; Jacobs et al. 2015; Wiengarten et al. 2015). Larger factories can have greater

flexibility in allocating resources and time to the lean program but are also likely to face greater

challenges due to the scale of the implementation effort. We define Factory Size as the natural

logarithm of the number of full-time factory employees. Unionization is sometimes included

as a control variable in the process improvement literature (e.g., Shah and Ward, 2003; Jayaram

et al., 2010; Fullerton et al., 2013). Some authors argue that unionization can slow down the

implementation of change initiatives (Sim and Rogers 2008), while others suggest that a

cooperative union can create change-ready organizations (Fast 2014). We define Union as the

percentage of factory-level employees who belong to a union. Finally, to account for the

potential influence of geographic location, we define Location as an indicator variable equal to

one for factories located in Europe and zero otherwise; the reason for this choice is that Europe

is the home continent of GEM.

Page 14: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

13

Table 1 Sample characteristics Panel A: Descriptive statistics

Variable Name N Mean Standard Deviation Min Max

Lean Implementation 95 1.69 0.79 0.50 3.39 Rhythm (minutes) 95 51.9 88.3 0.30 480 Formal Assessments 95 2.37 0.80 1.00 4.00 Gemba Walks 95 0.64 0.48 0.00 1.00 Process Complexity 95 0.38 0.49 0.00 1.00 Factory Size (number of employees) 95 893.5 608.5 100 2400 Union 95 0.76 0.40 0.00 1.00 Location 95 0.54 0.50 0.00 1.00

Panel B: Correlations

1 2 3 4 5 6 7 1 Lean Implementation 1.00 2 Rhythm -0.40** 1.00 3 Formal Assessments 0.27** 0.11 1.00 4 Gemba Walks 0.39** 0.12 0.19^ 1.00 5 Process Complexity -0.16 0.49** 0.31** 0.04 1.00 6 Factory Size 0.04 -0.17^ 0.02 -0.13 -0.21* 1.00 7 Union -0.02 -0.09 -0.16 0.11 -0.34** -0.02 1.00 8 Location -0.19^ -0.01 -0.02 -0.06 0.16 -0.04 0.01

Notes: ** = p<0.01, * = p<0.05, and ^ = p<0.10, two-tailed tests of statistical significaence.

Panels A and B of Table 1 present descriptive statistics and correlations for the

respective variables used in our multivariate analysis. Note that the sample size is 95 factory

observations, representing all lean assessments for each factory in the sample. Consistent with

our prediction, Panel B reports a negative association between Rhythm and Lean

Implementation (ρ=-0.40, p<0.01), suggesting that factories with smaller Rhythms (that is,

factories with faster rhythms of production) have larger scores for lean implementation. Post-

estimation review of variance inflation factors confirms that the magnitudes of the correlations

reported in Panel B do not generate multicollinearity concerns (untabulated).

Page 15: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

14

4.4 Conducting the qualitative analysis

In our mixed-method design, we complement the quantitative evidence with a

qualitative analysis (Jick 1979). The call for such deductive-inductive research designs is

longstanding in the operations management literature (Swamidass 1991). Adding detailed

contextual data enriches both the theoretical contribution (Meredith 1998; Barratt et al. 2011)

and the practical importance (Swamidass 1991; Toffel 2016). We follow the advice in the

literature for structured qualitative analysis (Eisenhardt 1989; Voss et al. 2002; Barratt et al.

2011). Specifically, we follow the eight-step procedure proposed by Eisenhardt (1989) because

it is compatible with our research objective, which seeks to explain an a priori identified

relationship in a complex process (Gehman et al. 2018). Eisenhardt’s eight steps are: (1)

Getting started, (2) selecting cases, (3) crafting instruments, (4) entering the field, (5) analyzing

data, (6) shaping hypotheses, (7) enfolding literature, and (8) reaching closure.

In the first step, we specified our research question (based on the results of our

quantitative analysis): Why are faster production rhythms related to faster lean

implementation? In the second step, we selected the factories to visit based on several criteria

that ensured a balanced sample (cf. Barratt et al. 2011): a) factories with low and high lean

assessment scores (in their latest assessment), b) factories of different sizes (small and large),

and c) factories on five different continents. Visiting these factories helped us to develop an

on-the-ground understanding of GEM’s lean implementation context.

The third step involves specifying our research instruments. We developed a detailed

research protocol for factory visits. It specified what information we requested prior to our

visit, the schedule for our one-day visits, the number and positions of employees we would like

to interview, information regarding anonymity and confidentiality, and the link to an online

survey we asked factory employees to complete. We communicated a shorter version of the

research protocol with our factory contact persons approximately one to two months before

Page 16: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

15

each visit. We developed an interview guide and pre-tested it with GEM employees. As it is

often usual in qualitative research, the interview guide evolved throughout the research when

some of the initial questions were answered or dismissed (Eisenhardt 1989).

Before entering the field (step 4), we reviewed factory internal presentations and lean

assessment scores that had been made available to us by GEM. The factory visits followed

roughly the same routine, starting with an introduction and a general presentation of the factory,

followed by a one-to-two hours tour of the shop-floor. We then had three to eight interviews

(depending on factory size) with key employees, including representatives from top

management, the lean implementation team, factory line management, and union

representatives. During the tour of the shop floor, we could ask questions directly to shop floor

workers and take photographs. These visits allowed us to make our own observations about

lean implementation in the factory. Immediately after the factory visits, all notes were carefully

typed up in detailed case reports (resulting in a qualitative database of approx. 300 pages).

We used a structured approach to analyze the data in step 5, following the suggestions

by Barratt et al. (2011) to compare cases against the constructs and each other. We first compare

factories against a list of parameters that we had either adopted from the literature or noted as

potentially interesting during the factory visits and interviews. We then map the factories along

two axes—the rhythm of production and the extent of lean implementation in the factory—and

compare different groups of fast-rhythm and slow-rhythm factories, drawing on our database

and quotations from the interviews. This process allows us to “search evidence for [the] ‘why’

between relationships” (Eisenhardt 1989, p. 533), and we use this evidence to shape new

propositions for our emerging theory (step 6). In step 7, we anchor our findings in relevant

literature to sharpen our theoretical contribution. Finally, in step 8, we stopped our field visits

after reaching a “theoretical saturation” point (Eisenhardt 1989), which happened almost

naturally in our research. We had visited 37 of the 49 factories in our sample and had reached

the point of diminishing new insights from additional visits.

Page 17: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

16

5. HYPOTHESIS TESTING

5.1 Evidence from consecutive factory assessments

Our first set of tests exploits the fact that 37 factories have multiple lean assessments.

This allows us to investigate the association between the speed of change in factories’ lean

scores between different assessments and their production rhythms. Our hypothesis posits that

factories with faster production rhythms should have a faster pace of lean implementation (i.e.,

a faster rate of improvement between each lean assessment). To test this prediction, we estimate

the following ordinary least squares regression:

Lean Implementationi,t = α0 + β1Rhythmi,t + β2Gemba Walksi,t + β3Formal Assessmentsi,t +

β4Process Complexityi,t + β5Factory Sizei,t + β6Unioni,t + β7Locationi,t + β8PLASi,t + β9Time Betweeni,t

+ λYEAR + εi,t.

Each variable in the regression is measured for factory i at time t. In addition to the

independent variables discussed previously, we introduce Prior Lean Assessment Score

(PLAS), which is the lagged value of Lean Implementation, thereby enabling us to examine the

relation between the pace of lean implementation between assessments and the rhythm of

production. The introduction of PLAS might create serial auto-correlation in the model. Hence,

we report Driscoll-Kraay standard errors, which assumes errors are heteroskedastic and

autocorrelated (see, for example, Driscoll and Kraay 1998; Hoechle 2007). To account for the

presence of multiple observations for some factories, we also check and confirm that all

inferences are robust if we cluster standard errors by factory. We introduce Time Between,

defined as the natural logarithm of the number of months between consecutive lean assessments

in a factory, to capture differences in when factory assessments occurred (we obtain statistically

similar results when we do not use the natural logarithm). The model includes indicator

variables for the year in which an assessment was completed to account for any impact that the

passage of time may have on lean assessment scores.

Page 18: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

17

Table 2 Production rhythm and lean implementation: Evidence from consecutive assessments

Lean

Implementation Rhythm -0.093**

[0.009] Gemba Walks -0.034

[0.089] Formal Assessments 0.070*

[0.031] Product Complexity 0.090

[0.092] Factory Size 0.014

[0.054] Union 0.271**

[0.070] Location -0.027

[0.057] Time Between 0.057

[0.040] PLAS 0.725**

[0.068] Intercept 2.825**

[0.380] Fixed Effects YEAR Observations 46 R2 79%

Notes: This table reports coefficient estimates and standard errors [in brackets] from an ordinary least squares regression that investigates the relation between factory-level production rhythm and the pace of lean implementation. The sample size is 46 factory observations because the analysis uses all lean assessments for each factory that had more than one assessment. ** = p<0.01, * = p<0.05, and ^ = p<0.10, two-tailed tests of statistical significance.

Table 2 presents the results of this analysis. The negative and statistically significant

coefficient between Rhythm and Lean Implementation indicates that factories with faster

(slower) rhythms have a faster (slower) pace of lean implementation between each assessment

(β=-0.093, p<0.01). This result is consistent with our hypothesis.

An important feature of our research setting is that GEM launched the lean program

throughout its global factory network at the same time. While GEM managers confirmed this

situation with us, it is nonetheless possible that individual factories had different experiences

with some lean practices when the program was launched; for instance, local managers may

Page 19: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

18

have been experimenting with aspects of lean indepdently. If so, then some factories may have

had different “starting points” at the launch of the lean program. We conduct two additional

tests to assess whether different “starting points” affect our results. First, we reestimate the

model after removing the two factories with the highest and the two factories with the lowest

lean assessment scores at the initial assessment (i.e., we remove four out of 47 factories)—all

results continue to hold (untabulated). Second, we reestimate the model after removing the

three factories with the highest and the three factories with the lowest lean assessment scores

at the initial assessment (i.e., we remove six out of 47 factories)—all results continue to hold

(untabulated). We conclude that the potential for different “starting points” does not unduly

affect the relation between Rhythm and Lean Implementation.

5.2 Additional evidence

We conduct four additional sets of tests to assess the strength of the evidence presented

in Table 2. First, we use only the first lean assessment completed by a factory, which captures

the pace of lean implementation from the beginning of the GEM lean program through the first

factory-level lean assessment. This analysis sheds light on whether factories with faster

(slower) production rhythms implement lean at a faster (slower) pace than other factories after

GEM’s worldwide launch of the lean program. To conduct this test, we estimate the following

ordinary least squares regression:

Lean Implementationi,t = α0 + β1Rhythmi,t + β2Gemba Walksi,t + β3Process Complexityi,t + β4Factory

Sizei,t + β5Unioni,t + β6Locationi,t + λYEAR + εi,t.

The variables are measured for factory i at time t, and we control for the year in which

the assessment was completed (λYEAR). Evidence consistent with our hypothesis would show a

negative relation between Lean Implementation and Rhythm. We report robust,

heteroskedasticity-consistent standard errors after confirming the errors are not smaller than

Page 20: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

19

conventional standard errors in our sample (see Imbens and Kolesar (2016) and Angrist and

Pischke (2009) for details about the properties of robust standard errors in small samples).

Table 3 presents the results of this regression. Columns 1-7 present evidence consistent

with our hypothesis—faster (slower) production rhythm has a favorable (unfavorable) relation

with the pace of lean implementation. This can be seen, for instance, in column 1 by the

negative and statistically significant coefficient estimate on Rhythm (β=-0.273, p<0.01). We

next add the control variables Gemba Walks (column 2), Process Complexity (column 3),

Factory Size (column 4), Union (column 5), and Location (column 6) and show that the

favorable (unfavorable) relation between faster (slower) production rhythm and Lean

Implementation persists (p<0.01 in all columns). Column 7 reports the results of the full model;

the favorable (unfavorable) relation between faster (slower) production rhythm and Lean

Implementation remains after including all control variables (β=-0.269, p<0.01).

For our second set of tests, we repeat the analysis in Table 3 using only the assessment

scores of factories that received multiple assessments (N=83; results reported in Panel A of

Table 4). This test addresses concerns that production rhythm is only influential for the initial

factory lean assessment and not as the lean journey continues. For our third set of tests, we

repeat the analysis using all lean assessment scores across the entire factory network to assess

whether the relation between production rhythm and lean implementation persists over time

(N=95; results reported in Panel B of Table 4). Finally, we follow the advice put forth in Angrist

and Pischke (2009) that in the presence of clustering, it is important to show that the evidence

is consistent with the inferences that arise from an analysis of group averages. Hence, we

construct a new sample that uses the average of all lean assessment scores for each factory in

the sample, thereby forming group averages for those factories that have more than one lean

assessment (N=49; results reported in Panel C of Table 4). Panels A and B report standard

errors clustered by factory to account for the presence of multiple observations for some

factories, while Panel C reports robust, heteroskedasticity-consistent standard errors.

Page 21: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

1

Table 3 Production rhythm and lean implementation: Evidence from initial factory assessments

(1) (2) (3) (4) (5) (6) (7)

Lean

Implementation Lean

Implementation Lean

Implementation Lean

Implementation Lean

Implementation Lean

Implementation Lean

Implementation Rhythm -0.273** -0.253** -0.298*** -0.264** -0.281** -0.272** -0.269**

[0.074] [0.065] [0.078] [0.075] [0.081] [0.074] [0.069] Gemba Walks 0.457**

[0.147] 0.497**

[0.167] Process Complexity 0.162

[0.187] 0.114 [0.174]

Factory Size 0.103 0.125 [0.101] [0.107]

Union -0.105 -0.152 [0.171] [0.156] Location -0.041 -0.029

[0.175] [0.152] Intercept 3.201** 2.873** 3.269*** 2.507** 3.344** 3.215** 2.270**

[0.243] [0.225] [0.245] [0.733] [0.396] [0.260] [0.850] Fixed Effects YEAR YEAR YEAR YEAR YEAR YEAR YEAR

Observations 49 49 49 49 49 49 49 Adj R2 46% 54% 47% 47% 47% 47% 57%

Notes: This table reports coefficient estimates and standard errors [in brackets] from ordinary least squares regressions that investigate the relation between factory-level production rhythm and the pace of lean implementation. The sample size is 49 factory observations because the analysis uses only the first lean assessment for each factory. ** = p<0.01, * = p<0.05, and ^ = p<0.10, two-tailed tests of statistical significance.

Page 22: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

1

Table 4, Panels A through C, present the results of these tests. Of note, the negative and

statistically significant coefficient estimates on Rhythm reported in column 1 of Panels A

through C provide further evidence in support of our hypothesis—faster (slower) production

rhythm has a favorable (unfavorable) relation with the pace of lean implementation (Panel A:

β=-0.132, p<0.01; Panel B: β= -0.221, p<0.01; Panel C: β=-0.266, p<0.01).

Table 4 Production rhythm and pace of overall lean implementation

Panel A: Evidence from factories with multiple assessments

(1) (2) (3) (4)

Lean Implementation

Lean Implementation (50 replications)

Lean Implementation (200 replications)

Lean Implementation

(Robust Regression)

Rhythm -0.132** -0.132** -0.132** -0.136^ [0.067] [0.066] [0.060] [0.072]

Gemba Walks 0.659** [0.191]

0.659** [0.154]

0.659** [0.168]

0.641** [0.178]

Formal Assessments 0.302 0.302^ 0.302^ 0.338^ [0.206] [0.170] [0.175] [0.191] Process Complexity -0.278 -0.278 -0.278 -0.311 [0.257] [0.275] [0.238] [0.265] Factory Size 0.008 0.008 0.008 0.003

[0.134] [0.091] [0.127] [0.140] Union -0.053 -0.053 -0.053 -0.074

[0.166] [0.198] [0.167] [0.194] Location -0.330^ -0.330* -0.330* -0.325^

[0.179] [0.145] [0.164] [0.175] Intercept 1.315 1.315 1.315 1.228

[1.061] [0.945] [0.956] [1.184] Fixed Effects YEAR YEAR YEAR YEAR Observations 83 83 83 83 Adj R2 41% 41% 41% NA

Panel B: Evidence from all factory assessments

(1) (2) (3) (4)

Lean Implementation

Lean Implementation (50 replications)

Lean Implementation (200 replications)

Lean Implementation

(Robust Regression)

Rhythm -0.221** -0.221** -0.221** -0.223** [0.065] [0.053] [0.064] [0.069]

Gemba Walks 0.590** [0.158]

0.590** [0.099]

0.590** [0.133]

0.585** [0.159]

Formal Assessments 0.153 0.153 0.153 0.149 [0.107] [0.135] [0.100] [0.111]

Process Complexity 0.105 0.105 0.105 0.110 [0.213] [0.215] [0.189] [0.206]

Page 23: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

2

Factory Size 0.062 0.062 0.062 0.067 [0.108] [0.116] [0.098] [0.116]

Union -0.057 -0.057 -0.057 -0.067 [0.136] [0.186] [0.178] [0.168]

Location -0.146 -0.146 -0.146 -0.134 [0.154] [0.144] [0.159] [0.151]

Intercept 1.934* 1.934* 1.934* 1.935* [0.886] [0.898] [0.803] [0.948]

Fixed Effects YEAR YEAR YEAR YEAR Observations 95 95 95 95 Adj R2 39% 39% 39% NA

Panel C: Evidence from average factory assessments

(1) (2) (3) (4)

Lean Implementation

Lean Implementation (50 replications)

Lean Implementation (200 replications)

Lean Implementation

(Robust Regression) Rhythm -0.266** -0.266** -0.266** -0.269**

[0.073] [0.087] [0.081] [0.082] Gemba Walks 0.485**

[0.170] 0.485* [0.206]

0.485* [0.195]

0.474* [0.198]

Formal 0.119 0.119 0.119 0.111 Assessments [0.102] [0.123] [0.119] [0.129] Process 0.251 0.251 0.251 0.281 Complexity [0.212] [0.257] [0.254] [0.246] Factory Size 0.062 0.062 0.062 0.052

[0.104] [0.133] [0.125] [0.135] Union -0.073 -0.073 -0.073 -0.076

[0.150] [0.190] [0.200] [0.204] Location -0.039 -0.039 -0.039 -0.014

[0.162] [0.184] [0.190] [0.193] Intercept 2.271** 2.271* 2.271* 2.358*

[0.854] [0.996] [1.030] [1.086] Fixed Effects YEAR YEAR YEAR YEAR Observations 49 49 49 49 Adj R2 38% 38% 38% NA

Notes: This table reports coefficient estimates and standard errors [in brackets] from ordinary least squares regressions that investigate the relation between factory-level production rhythm and the pace of lean implementation. The sample size in Panel A is 83 factory observations because the analysis uses all lean assessments for each factory that had more than one assessment. The sample size in Panel B is 95 factory observations because the analysis uses all lean assessments for each factory in the sample. The sample size in Panel C is 49 factory observations because the analysis uses the average of all lean assessments for each factory in the sample. ** = p<0.01, * = p<0.05, and ^ = p<0.10, two-tailed tests of statistical significance.

Page 24: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

3

Out of an abundance of caution, we next assess the strength of the evidence presented

in Table 4, Panels A through C, using an alternative approach to calculate standard errors in

small samples. We follow Mooney and Duval (1993) and repeat our analyses using

bootstrapped standard errors based on 50 replications (column 2) and 200 replications (column

3). We use serial numbers from randomly drawn U.S. $1 bills (after removal of any letters) to

set the seed for each bootstrap, which ensures that our procedures begin with random seeds.

Consistent with our prior evidence, columns 2 and 3 of Panels A through C show that the

favorable (unfavorable) relation between faster (slower) production rhythm and Lean

Implementation persists. We also use robust regression to assess whether outliers unduly

influence our results, which first screens the sample to eliminate high-leverage outliers (using

Cook’s D) and then performs Huber iterations followed by biweight iterations. Column 4 in

Panels A through C presents the results; our inferences remain unchanged.

6. QUALITATIVE ANALYSIS AND DISCUSSION

Our quantitative results provide robust evidence that faster (slower) production rhythm

has a favorable (unfavorable) relation with the pace of lean implementation. In this section, we

support this finding using our qualitative analysis. In particular, we use our qualitative data to

deepen our understanding of why this relation occurs, which provides new insights into how

managers can mitigate the adverse effects of slow production rhythms and maximize the

benefits of fast ones.

We begin our analysis with Figure 1, which presents a regression of lean

implementation on production rhythms for all factories from which we have collected

qualitative data (N=37). As expected from our quantitative analysis, the slope of the regression

line is downward: lean implementation slows as production rhythm slows. The factories that

are above the regression line implemented the lean program faster (and those below the line

slower) than expected, given their production rhythms. Factories that are farthest from the

Page 25: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

4

regression line in each group are the most intriguing ones: Why are they so much better or

worse in implementing the lean program than expected? We use our qualitative analysis to

investigate this question.

Notes: This figure presents a scatter plot of the relation between production rhythm and lean implementation for the 37 assembly factories from which we collected qualitative data. The solid line represents the OLS regression of the lean assessment score on production rhythm. The vertical dashed line represents the median production rhythm in this sample (16.5), and the downward-sloping dashed lines represent the cut-offs for the comparative analysis.

Figure 1 Relation between production rhythm and lean implementation in visited factories

For this analysis, we choose factories that are outside 0.75 standard deviations from the

regression line. First, we consider factories in the top right corner of Figure 1, which have slow

production rhythms yet have managed to implement the lean program noticeably faster than

factories in the lower right corner with similar rhythms. Second, we consider factories in the

bottom left corner of Figure 1, which have fast production rhythms yet are conspicuously

slower to implement lean than their peers in the upper left corner. We label these groups slow

rhythm leaders and –laggards and fast rhythm laggards and –leaders, respectively.

0.00 1.00 2.00 3.00 4.00 5.00 6.00

LOW

40314855207,42,7Minutes

HIGH

Impl

emen

tatio

n of

lean

(Lea

n im

plem

enta

tion

scor

e)

0

Production rhythm(logarithmic scale)

FAST SLOW

Slow rhythm leaders

Fast rhythm laggards

Slow rhythm laggards

Fast rhythm leaders

Page 26: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

5

We then ask the following questions: Why are the slow rhythm leaders implementing

lean faster than the slow rhythm laggards? And, why are the fast rhythm laggards

implementing lean slower than fast rhythm leaders? We use the structured analysis of the

qualitative material to search for patterns that could answer these questions. (cf. Eisenhardt

1989; Barratt et al. 2011). Drawing on the framework suggested by Anand et al. (2009)—

categorizing the efforts for continuous improvement into “purpose,” “process,” and “people”—

we compare practices in the factories in each group. Table 5 reports the result of the aggregate

cross-case analysis.

Table 5 Aggregated summary of qualitative cross-case analysis

Variables

Fast Rhythm Slow Rhythm

Clear Laggards

(N=4)

Clear Leaders (N=6)

Clear Laggards

(N=3)

Clear Leaders (N=3)

Purpose Lean program recognized as a strategic goal Mixed Present Absent Present Top-management team owns the lean program Mixed Present Absent Present

Area managers held responsible for lean program Present Present Absent Present

Program communication Mixed Mixed Absent Present Process Codification of tacit knowledge Present Present Absent Present Standard improvement method Mixed Present Absent Present Whole factory involved Mixed Present Present Present Internal audits Mixed Mixed Mixed Mixed Benchmarking other factories Mixed Present Absent Mixed Pilot/project-based approach Present Mixed Absent Mixed Physical space for kaizen* Mixed Present Absent Mixed Physical space for training* Mixed Present Mixed Mixed People Evidence of scientific mindset Absent Mixed Absent Present Dedicated lean program manager/team/champions Present Present Present Present

Structured lean training program Mixed Mixed Mixed Present Use of external consultants* Mixed Mixed Mixed Mixed Hired external lean expert* Absent Absent Absent Mixed

Notes: This table reports the aggregated summary of the qualitative cross-case analysis of the four groups of factories identified in Figure 1. Words in bold-italics show the strongest patterns in the cross-case analysis. *Variables included based on our field observations, not mentioned in Anand et al. (2009).

Page 27: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

6

6.1 Comparing slow rhythm leaders versus slow rhythm laggards

Three policies stand out as differentiating factors between the three factories that were

slow rhythm leaders with three factories that were slow rhythm laggards: (1) strong

management commitment to the lean program, manifested in strategy, engagement, and

communication, (2) proactive codification of tacit know-how, and (3) frequent scientific

experimentation on the factory floor (see Table 5).

6.1.1 Strong management commitment to the lean program

Our qualitative analysis confirms that management commitment is important for lean

implementation, which is consistent with several studies in the lean literature (e.g., Emiliani

and Stec 2005; Netland 2016) and other improvement programs (e.g., Rodgers et al. 1993; Beer

2003). From our field notes, we discerned an appreciable difference in the level of management

commitment between the leaders and laggards in this group. As we observed during our visits

and interviews, managers in the leader group went beyond slogans and declarations that the

lean program is a strategic objective. They took ownership and responsibility for the lean

implementation. Importantly, we found no evidence that there was systematically more (or

less) management commitment in factories with different production rhythms (for example,

management did not apply more attention on average to factories with fast (or slow) production

rhythms). This point is also confirmed in our quantitative data (cf. Table 1, Panel B). We

conclude that a strong management commitment is essential—independent of production

rhythm.

6.1.2 Proactive codification of tacit production know-how

Next, we observe that slow rhythm leaders tended to emphasize the standardization of

improvement methods and production methods, while slow rhythm laggards did not. In the

slow rhythm leaders, we observed teams of operators, technicians, and managers routinely

analyzing longer production tasks at each assembly station by dividing them into shorter

Page 28: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

7

segments and developing the best method to perform them. The new methods were then

recorded into new standard procedures (i.e., codified), which, as discussed previously, made it

is easier to learn, share, improve, and act on the methods. This is a proactive codification of

tacit knowledge and, as such, likely expedites the adoption of lean practices in a factory. This

finding resonates with studies that argue for the importance of managing forms of knowledge

in improvement programs (e.g., Choo et al. 2007).

Relatedly, our factory visits and discussions with GEM personnel highlighted how slow

rhythm leaders were keen to make the operations on the factory floor visible to all employees,

especially operators and supervisors. By doing so, they mitigate one of the inherent problems

of a slow production rhythm—that of obscuring variations in production methods and systemic

problems in the day-to-day operations (as explained in Section 2). These factories are not only

fervent users of visualization, but they also facilitate rapid communication between the shop

floor and managers and supervisors. In stark contrast, the absence of such policies was palpable

in the slow rhythm laggards during our visits.

A GEM factory in South America provides a clear example of systemic and rapid

codification of know-how. Despite having one of the slowest production rhythms in the GEM

network, this factory has one of the highest lean assessment scores. Visitors cannot miss

colored, eye-catching descriptions and photographs of standard procedures at each assembly

station. Shop-floor operators help update the contents of these displays. Putting their best

standards on display is both a source of pride and an effective way to codify the tacit knowledge

on the shop floor.

6.1.3 Frequent use of scientific experimentation

Spear and Bowen (1999), among others, posit that the fundamental element of a lean

manufacturing system is the creation of “a community of scientists… [who] establish sets of

hypotheses that can then be tested” (p. 97). Consistent with this idea, we observed a higher

Page 29: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

8

inclination to conduct formal scientific experiments in all factories in the slow rhythm leader

group compared to the slow rhythm laggards.

As an example, in one GEM factory, a team of employees was conducting an

experiment on the assembly line during our visit. Using a typical “plan-do-study-act” method,

the team compared existing assembly processes, redesigned the procedures, and then tried the

new methods on the assembly line. The employees later recorded the results and discussed

what they had learned. While this particular experiment’s results were not perfect, the act of

experimentation itself appealed to many employees, convinced them of its benefits, and

increased the number of Kaizen suggestions and formal experiments to test them. As a lean

manager in this factory explained, “It [experimentation] does not have to be perfect right away;

just starting to improve and taking small steps shortens the learning cycle.” A focus on

scientific experimentation appears to build a culture of continuous improvement that fosters a

faster pace of lean implementation (see also Anand et al. 2009; Ballé et al. 2016).

6.2 Comparing fast rhythm laggards versus fast rhythm leaders

We compare four factories that were fast rhythm laggards with four factories that were

fast rhythm leaders (See Figure 1). The primary reason for the slower pace of lean

implementation in the former group, we find, is the presence of a culture that dismisses the

importance of the lean program. While the fast rhythm leaders followed the practices described

in the previous section (i.e., strong management commitment to the lean program, the proactive

codification of tacit production know-how, and frequent use of scientific experimentation),

which were actually easier to implement in their environments, the fast rhythm laggards

seemed to question the relevance or potential benefits of a lean program for their factories and

did not promote these practices.

During our factory visits and interviews in the fast rhythm laggard factories, it was

apparent that key employees remained unconvinced about the benefits of the lean program. A

Page 30: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

9

senior manager in a factory located in an advanced country described the sentiment we

observed in several such factories: “We have much more complicated processes than other

assembly factories,” implying that the lean program did not fit such factories. The feeling was

sometimes stronger. In a factory in another advanced country, we heard: “We have produced

at this site for almost a century now; we know what it takes to be successful,” suggesting that

the factory did not need to be told how to run its operations. There were also questions about

the relevance of the measures used for assessing the implementation of the lean program. “We

are in a poor developing country, very different from Japan,” stated a plant manager in one

such factory.

We find most of these arguments unconvincing. Several fast rhythm leaders are in

similar conditions as those that consider themselves to be unique or superior. Some are located

in “poor developing countries” or produce similar products using similar production processes

as the fast rhythm laggards, yet they have excelled in implementing the corporate lean program.

It appears that the negative perception of local managers about the need for the lean program,

its relevance to their operating context, and the manner in which lean is measured is a primary

obstacle that slows the pace of lean implementation in fast rhythm laggards factories.

6.3 Learning lean

The leaders in both groups seem to follow practices and policies that promote learning,

but a faster production rhythm seems to ease implementing such practices and policies.

Therefore, we theorize that a fast production rhythm accelerates lean implementation because

it promotes learning. In fast rhythm assembly factories, the practices that expedite learning

come more naturally than in slow rhythm assembly factories. When the time for assembly tasks

is shorter, it is easier to codify the tacit knowledge, standardize, and simplify—all of which

enhance learning (Choo et al. 2007). Moreover, tasks can be learned faster when they are

repeated more frequently (Adler and Clark 1991). This is a fundamental insight that contributes

Page 31: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

10

to the explanation of why the pace of lean implementation is slower in some settings than in

others.

7. CONCLUSIONS AND IMPLICATIONS

In this paper, we uncover a relation between a factory’s production rhythm and the pace

of implementation of a lean program in that factory: the faster the production rhythm, the faster

the implementation. To our knowledge, this is the first time that this relation has been identified

and empirically tested. Our qualitative analysis reveals a theoretical proposition: a fast

production rhythm supports the pace of lean implementation because a fast rhythm promotes

policies that expedite learning about lean practices in the factory. This finding has important

implications for both practitioners and scholars.

As a senior manager at GEM who collaborated closely with us on this project,

explained: “An open question for us has always been whether the differences in the speed of

production lines in our factories affect their lean implementation.” Our results provide evidence

that they do. If the company has factories with distinctly different production rhythms,

managers should calibrate their expectations about the pace of lean implementation in different

factories based on their production rhythms.

Practitioners should also try to mitigate the inherent disadvantages of a slow rhythm by

promoting policies that facilitate learning about lean practices, which come more naturally in

fast rhythm factories. For instance, factories can be more proactive in making problems on the

line visible quickly, running more scientific experiments on the shop floor, and, in general,

promote policies that help codify tacit production knowledge quickly. These practices help

factory employees learn lean and support a learning cycle such that the benefits of incremental

improvements motivate employees to improve continuously. While these policies help every

factory, they are particularly helpful in overcoming the adverse effects of slow production

rhythms.

Page 32: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

11

While we have studied only assembly lines, logical extrapolation of our findings raises

intriguing practical questions. If lean implementation is slower in factories with slower rhythms

of production, then should we expect lean implementation to take longer in low-volume

production settings, such as construction, shipbuilding, or aerospace (e.g., producing a unit

every week, month, or year) than in high-volume production settings, such as automobiles,

furniture, and consumer products (e.g., producing a unit every few minutes)? If that is indeed

the case, our findings suggest that these slow-rhythm industries can mitigate the unfavorable

conditions in their environments by adopting policies that promote learning.

For academics, we confirm and shed additional light on the critical role of learning in

explaining the pace of lean implementation. We also show how theories related to the learning

curve, forms of knowledge (codified or tacit) and organizational routines provide support for

how factories can accelerate learning about the lean practices and spreading them throughout

the organization. Our theoretical proposition can be modeled and tested using survey studies

or further explored in qualitative field research.

Our findings also raise interesting questions for future research. For example, do the

inherent structural characteristics in a production system, such as production rhythm, override

the more transient managerial actions, such as management commitment and resource

allocation, in setting the pace of implementation of a broad improvement program such as lean?

Is it possible to reduce the potential adverse effects of such inherent structural characteristics?

If slow rhythm factories are at a disadvantage when implementing lean, are there ways to

mitigate the adverse effects of slow rhythms other than the tactics we have suggested? Delving

deeper into the underlying reasons for our findings can potentially provide new insights into

the enduring puzzle of the factors that affect lean implementation.

Page 33: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

12

REFERENCES

Adler, P.S., Clark, K.B., (1991). Behind the learning curve: A sketch of the learning process. Management Science, Vol. 37, No. 3, pp. 267-281

Anand, G., Ward, P.T., Tatikonda, M.V., Schilling, D.A., (2009). Dynamic capabilities through continuous improvement infrastructure. Journal of Operations Management, Vol. 27, No. 6, pp. 444-461

Angrist, J.D., Pischke, J.S., (2009). Mostly harmless econometrics: An empiricist's companion. Princeton University Press, Princeton, NJ

Ballé, M., Morgan, J., Sobek, D.K., (2016). Why learning is central to sustained innovation. MIT Sloan Management Review, Vol. 57, No. 3, pp. 63

Barratt, M., Choi, T.Y., Li, M., (2011). Qualitative case studies in operations management: Trends, research outcomes, and future research implications. Journal of Operations Management, Vol. 29, No. 4, pp. 329-342

Bateman, N., (2005). Sustainability: the elusive element of process improvement. International Journal of Operations & Production Management, Vol. 25, No. 3/4, pp. 261-276

Baudin, M., (2002). Lean assembly: the nuts and bolts of making assembly operations flow. Productivity Press, New York, NY

Beer, M., (2003). Why total quality management programs do not persist: the role of management quality and implications for leading a TQM transformation. Decision Sciences, Vol. 34, No. 4, pp. 623-642

Bicheno, J., (2004). The new lean toolbox: towards fast, flexible flow. Production and Inventory Control, Systems and Industrial Engineering Books, Buckingham

Bortolotti, T., Boscari, S., Danese, P., (2015). Successful lean implementation: Organizational culture and soft lean practices. International Journal of Production Economics, Vol. 160, No. 0, pp. 182-201

Bortolotti, T., Danese, P., Romano, P., (2013). Assessing the impact of just-in-time on operational performance at varying degrees of repetitiveness. International Journal of Production Research, Vol. 51, No. 4, pp. 1117-1130

Browning, T.R., Heath, R.D., (2009). Reconceptualizing the effects of lean on production costs with evidence from the F-22 program. Journal of Operations Management, Vol. 27, No. 1, pp. 23-44

Chavez, R., Yu, W., Jacobs, M., Fynes, B., Wiengarten, F., Lecuna, A., (2015). Internal lean practices and performance: The role of technological turbulence. International Journal of Production Economics, Vol. 160, pp. 157-171

Choo, A.S., Linderman, K.W., Schroeder, R.G., (2007). Method and context perspectives on learning and knowledge creation in quality management. Journal of Operations Management, Vol. 25, No. 4, pp. 918-931

Deming, W.E., (1986). Out of the crisis: Quality, productivity and competitive position. Cambridge University Press, Cambridge, MS

Derfus, P.J., Maggitti, P.G., Grimm, C.M., Smith, K.G., (2008). The Red Queen Effect: Competitive actions and firm performance. The Academy of Management Journal, Vol. 51, No. 1, pp. 61

Distelhorst, G., Hainmueller, J., Locke, R.M., (2016). Does lean improve labor standards? Management and social performance in the Nike supply chain. Management Science, Vol. 63, No. 3, pp. 707-728

Driscoll, J.C., Kraay, A.C., (1998). Consistent covariance matrix estimation with spatially dependent panel data. The Review of Economics and Statistics, Vol. 80, No. 4, pp. 549-560

Page 34: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

13

Eisenhardt, K.M., (1989). Building theories from case study research. Academy of Management Review, Vol. 14, No. 4, pp. 532-550

Emiliani, M.L., Stec, D.J., (2005). Leaders lost in transformation. Leadership and Organization Development Journal, Vol. 26, No. 5, pp. 370-387

Eroglu, C., Hofer, C., (2011). Lean, leaner, too lean? The inventory-performance link revisited. Journal of Operations Management, Vol. 29, No. 4, pp. 356-369

Fast, L., (2014). Can Union Facilities Support a Lean, Continuous Improvement Culture?, Industry Week

Feldman, M.S., Pentland, B.T., (2003). Reconceptualizing organizational routines as a source of flexibility and change. Administrative Science Quarterly, Vol. 48, No. 1, pp. 94-118

Fogliatto, F., Anzanello, M., (2011). Learning Curves: The State of the Art and Research Directions, In: Jaber, M.Y. (Ed), Learning Curves: Theory, Models, and Applications. CRC Press, pp. 3-21

Fullerton, R.R., Kennedy, F.A., Widener, S.K., (2014). Lean manufacturing and firm performance: The incremental contribution of lean management accounting practices. Journal of Operations Management, Vol. 32, No. 7–8, pp. 414-428

Furlan, A., Vinelli, A., Dal Pont, G., (2011). Complementarity and lean manufacturing bundles: an empirical analysis. International Journal of Operations & Production Management, Vol. 31, No. 8, pp. 835-850

Galeazzo, A., Furlan, A., (2018). Lean bundles and configurations: a fsQCA approach. International Journal of Operations & Production Management, Vol. 38, No. 2, pp. 513-533

Gehman, J., Glaser, V.L., Eisenhardt, K.M., Gioia, D., Langley, A., Corley, K.G., (2018). Finding theory–method fit: A comparison of three qualitative approaches to theory building. Journal of Management Inquiry, Vol. 27, No. 3, pp. 284-300

Hayes, R.H., Wheelwright, S.C., (1984). Restoring our competitive edge: Competing through manufacturing. John Wiley & Sons, New York

Hoechle, D., (2007). Robust standard errors for panel regressions with cross-sectional dependence. Stata Journal, Vol. 7, No. 3, pp. 281

Holweg, M., (2007). The genealogy of lean production. Journal of Operations Management, Vol. 25, No. 2, pp. 420-437

Hopp, W.J., Spearman, M.L., (2004). To pull or not to pull: What Is the question? Manufacturing & Service Operations Management, Vol. 6, No. 2, pp. 133-148

Hopp, W.J., Spearman, M.L., (2011). Factory physics. Waveland Press, Long Grove, Illinois Imbens, G.W., Kolesar, M., (2016). Robust standard errors in small samples: Some practical

advice. Review of Economics and Statistics, Vol. 98, No. 4, pp. 701-712 Jacobs, B.W., Swink, M., Linderman, K., (2015). Performance effects of early and late Six

Sigma adoptions. Journal of Operations Management, Vol. 36, pp. 244-257 Jick, T.D., (1979). Mixing qualitative and quantitative methods: Triangulation in action.

Administrative Science Quarterly, Vol. 24, No. 4, pp. 602-611 Knol, W.H., Slomp, J., Schouteten, R.L.J., Lauche, K., (2019). The relative importance of

improvement routines for implementing lean practices. International Journal of Operations & Production Management, Vol. 39, No. 2, pp. 214-237

Lapré, M.A., Mukherjee, A.S., Van Wassenhove, L.N., (2000). Behind the learning curve: Linking learning activities to waste reduction. Management Science, Vol. 46, No. 5, pp. 597-611

Linderman, K., Schroeder, R.G., Zaheer, S., Liedtke, C., Choo, A.S., (2004). Integrating quality management practices with knowledge creation processes. Journal of Operations Management, Vol. 22, No. 6, pp. 589-607

Page 35: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

14

MacDuffie, J.P., (1997). The road to “root cause”: Shop-floor problem-solving at three auto assembly plants. Management Science, Vol. 43, No. 4, pp. 479-502

MacDuffie, J.P., Sethuraman, K., Fisher, M.L., (1996). Product variety and manufacturing performance: Evidence from the International Automotive Assembly Plant Study. Management Science, Vol. 42, No. 3, pp. 350-369

Meredith, J., (1998). Building operations management theory through case and field research. Journal of Operations Management, Vol. 16, No. 4, pp. 441-454

Mooney, C.Z., Duval, R.D., (1993). Bootstrapping: A nonparametric approach to statistical inference. Sage, Newbury Park, CA

Netland, T.H., (2016). Critical success factors for implementing lean production: the effect of contingencies. International Journal of Production Research, Vol. 54, No. 8, pp. 2433-2448

Netland, T.H., Ferdows, K., (2016). The S-curve effect of lean implementation. Production and Operations Management, Vol. 25, No. 6, pp. 1106-1120

Netland, T.H., Powell, D.J., (2016). A lean world, In: Netland, T.H., Powell, D.J. (Eds), The Routledge Companion to Lean Management. Routledge, New York

Netland, T.H., Schloetzer, J.D., Ferdows, K., (2015). Implementing corporate lean programs: The effect of management control practices. Journal of Operations Management, Vol. 36, No. 0, pp. 90-102

Nonaka, I., (1994). A dynamic theory of organizational knowledge creation. Organization Science, Vol. 5, No. 1, pp. 14-37

Nonaka, I., von Krogh, G., (2009). Perspective - Tacit knowledge and knowledge conversion: Controversy and advancement in organizational knowledge creation theory. Organization Science, Vol. 20, No. 3, pp. 635-652

Ohno, T., (1988). Toyota Production System: beyond large-scale production. Productivity Press, New York, NY

Onofrei, G., Prester, J., (2019). The relationship between investments in lean practices and operational performance. International Journal of Operations & Production Management, Vol. 39, No. 3, pp. 406-428

Overy, R.J., (1994). War and economy in the Third Reich. Oxford University Press, Oxford Pay, R., (2008). Everybody’s jumping on the lean bandwagon, but many are being taken for a

ride, Industry Week Polanyi, M., (1967). The tacit dimension. Routledge & Kegan Paul, London Powell, D.J., Coughlan, P., (2020). Rethinking lean supplier development as a learning

system. International Journal of Operations & Production Management, Vol. forthcoming,

Rodgers, R., Hunter, J.E., Rogers, D.L., (1993). Influence of top management commitment on management program success. Journal of applied psychology, Vol. 78, No. 1, pp. 151-155

Sadun, R., Bloom, N., Reenen, J.v., (2017). Why do we undervalue competent management? Harvard Business Review, Vol. 95, No. 5, pp. 120-127

Shah, R., Chandrasekaran, A., Linderman, K., (2008). In pursuit of implementation patterns: the context of Lean and Six Sigma. International Journal of Production Research, Vol. 46, No. 23, pp. 6679-6699

Shah, R., Ward, P.T., (2003). Lean manufacturing: context, practice bundles, and performance. Journal of Operations Management, Vol. 21, No. 2, pp. 129-149

Sim, K.L., Rogers, J.W., (2008). Implementing lean production systems: barriers to change. Management Research News, Vol. 32, No. 1, pp. 37-49

Spear, S., Bowen, H.K., (1999). Decoding the DNA of the Toyota Production System. Harvard Business Review, Vol. 77, No. 5, pp. 96-106

Page 36: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

15

Su, H.-C., Dhanorkar, S., Linderman, K., (2015). A competitive advantage from the implementation timing of ISO management standards. Journal of Operations Management, Vol. 37, pp. 31-44

Swamidass, P.M., (1991). Empirical science: new frontier in operations management research. Academy of Management Review, Vol. 16, No. 4, pp. 793-814

Toffel, M.W., (2016). Enhancing the practical relevance of research. Production and Operations Management, Vol. 25, No. 9, pp. 1493-1505

Voss, C., Tsikriktsis, N., Frohlich, M., (2002). Case research in operations management. International Journal of Operations & Production Management, Vol. 22, No. 2, pp. 195-219

White, R.E., Pearson, J.N., Wilson, J.R., (1999). JIT manufacturing: A survey of implementations in small and large U.S. manufacturers. Management Science, Vol. 45, No. 1, pp. 1-15

Wiengarten, F., Gimenez, C., Fynes, B., Ferdows, K., (2015). Exploring the importance of cultural collectivism on the efficacy of lean practices: taking an organisational and national perspective. International Journal of Operations & Production Management, Vol. 35, No. 3, pp. 370-391

Womack, J.P., (2017). Jim Womack on where lean has failed and why not to give up. Lean Enterprise Institute, Planet Lean: The Lean Global Network Journal

Wright, T.P., (1936). Factors affecting the cost of airplanes. Journal of the Aeronautical Sciences (Institute of the Aeronautical Sciences), Vol. 3, No. 4, pp. 122-128

Yelle, L.E., (1979). The learning curve: Historical review and comprehensive survey. Decision Sciences, Vol. 10, No. 2, pp. 302-328

Page 37: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

16

APPENDIX A Description of the GEM Lean Program

Launced in 2007, the GEM lean program is based on the five traditional groups of practices: just-

in-time (JIT), total quality management (TQM), total productive maintenance (TPM), human resource

management (HRM), and continuous improvement (CI). Each group contains three to five principles, and

each principle has two to seven practices. In total, there are over 100 practices. Table A-1 summarizes the

groups and principles, while Table A-2 provides examples of the practices included in the principles. A

factory’s Lean Assessment Score is an equally-weighted sum of the extent to which a group of principles

has been implemented. GEM has defined five ratings of implementation for each practice: 1 (“Basic”), 2

(“Structured”), 3 (“Improving”), 4 (“Best-in-Industry”), and 5 (“World-Class”). These “anchor” points

describe what should be in place to qualify for the rating; see Table A-2 for an example. A factory will

receive a score of zero for an element if it has not reached the “Basic” level for that element. Using simple

averages, the scores of practices are aggregated into scores for the principles, and scores for the principles

are aggregated into scores for a group of principles, which, finally, are aggregated into the Lean Assessment

Score for the factory.

Table A-1 GEM Lean Program Groups of Lean Practices and Descriptive Statistics

Group Principles Cronbach’s alpha Mean Std.

Dev. Min Max

Just in time

• Flexible workforce • Pull system • Adherence to takt time • Continuous flow • Material supply

0.904 1.66 0.84 0.23 3.97

Total quality management

• Zero defects • Quality assurance • Product and process quality

planning

0.836 1.52 0.84 0.25 3.63

Total productive maintenance

• Standardized work • Production leveling • Maintenance system • Workplace organization

0.823 1.55 0.85 0.30 3.79

Human resource management

practices

• Goal-oriented teams • Cross-functional work • Organizational design

0.814 1.78 0.84 0.33 3.58

Continuous improvement

• Prioritization • Problem-solving methods • Improvement organization • Improvement approach

0.921 1.78 1.04 0.25 4.25

Page 38: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

17

Table A-2 Example of GEM’s Assessment of Lean Practice Implementation

Gro

up

Prin

cipl

e

Prac

tice

Stage 1. Basic

Stage 2. Structured

Stage 3. Improving

Stage 4. Best in industry

Stage 5. World Class

JIT

Pu

ll sy

stem

s

Sync

hron

izat

ion

A fixed production sequence is used to synchronize production of the main flows in the factory. Buffer areas and quantities between lines/areas of the main flow are clearly defined and visualized.

Synchronization between the main flow and sub-assembly areas, and between the factory and main component suppliers, with clearly defined buffer areas, clearly visualized with max and min levels.

Internal flows, between the main flows and supplying processes, synchronized with delivery of material JIT, just-in-sequence (JIS), and clearly defined buffer locations and quantities

Synchronization exists between all local suppliers and the main factory with material delivered JIT, JIS, and clearly defined buffer locations and quantities.

Synchronization exists in the end-to-end supply chain, with material delivered JIT, JIS, and clearly defined buffer locations and quantities.

TQ

M

Zer

o de

fect

s

Mis

take

avo

idan

ce

(Pok

a Y

oke)

In at least one process, the implementation of mistake avoidance can be demonstrated, to ensure that defects cannot occur.

The factory can show a structured approach to the application of mistake avoidance devices, by for example implementing for prioritized quality issues and/or safety-critical features/parameters.

Mistake avoidance devices are commonly used (where applicable) as part of the problem resolution process to avoid the recurrence of root causes.

The factory is committed to installing mistake avoidance in all areas of the factory and has succeeded in applying it to nearly all areas. The use of low-cost solutions developed in-house makes this possible.

Mistake avoidance devices and systems are in place in all processes, machines, equipment, etc to ensure zero defects.

TPM

M

aint

enan

ce sy

stem

s

Aut

onom

ous M

aint

enan

ce (A

M)

AM has been deployed to a pilot machine or equipment. Following the GEM standard method for AM. A significant reduction of AM related breakdowns can be shown over at least three months. Before and after photos are recommended. Tags are closed out within 3 months.

AM activates as described in Stage 1 deployed to all “AA equipment.” Visual management is introduced (e.g. safety, tools, manuals, oil levels, gauges, …) to deployed machines. Zero AM related breakdowns is achieved over a six month period on pilot machine.

AM activities as described in Stage 2 deployed to all “A machines.” Lubrication moved to AM activities for those points manageable by operators. Zero AM related breakdowns is achieved over 6 month period on “AA machines.”

AM activities as described in Stage 2 and 3 deployed to all “B machines.”

The factory is considered as a top benchmark in AM.

Page 39: Learning lean: Rhythm of production and the pace of lean … · knowledge, suggests that managers and operators in factories with faster rhythms production are likely to have more

18

HR

M

Goa

l-ori

ente

d te

ams

Polic

y de

ploy

men

t

Clear business plan with defined KPIs and target levels for safety, quality, and delivery is in place and communicated down through the whole organization for the factory/unit.

Structured yearly strategy and policy deployment process with cascading (with interactive catch ball process) to at least team leader level.

As Stage 2, but for all competitive priorities (SQDCEP). A dialogue between leaders and teams takes place at all levels in the organization to agree on objectives (tactical and operational levels).

All teams (including support functions and management teams) are constantly challenging themselves by setting 0/100 targets, e.g., 0 defects, 100% delivery performance, etc. randomly sampled.

The overall policy deployment is visualized in the team area, and there is a clear linkage shown to tactical and strategic factory goals. KPIs and broken-down metrics/PIs are continuously refined to ensure they are applicable & best practice.

CI

Prob

lem

-sol

ving

met

hodo

logy

Prob

lem

-sol

ving

cul

ture

The majority of key production support personnel (e.g., production, quality engineers, and team leaders) show that they are competent in use of PSM and basic tools. For example, the problem-solving activity takes place with the disciplined use of the methodology and correct use of problem-solving tools to understand the root causes.

Factory shows a structured approach to developing problem-solving competence in the organization, e.g., ongoing training of employees (blue collar and white collar), coaching in problem-solving, review of problem-solving discipline. Factory demonstrates ability to react to problems quickly and effectively.

Horizontal expansion is performed to relevant areas (e.g., similar processes and/or products) to spread solutions and develop competence in order to prevent problems.

All levels of the organization and all functions are able to react quickly and effectively to problems. The factory can demonstrate examples of solving problems at a high pace and, more importantly, can demonstrate that the learning is transferred quickly and effectively to applicable areas/teams.

The factory is considered as top benchmark in problem-solving methodology

Notes: This table provides examples of how GEM assesses the implementation of five of the nearly 100 practices in the GEM Lean Program. It gives one example for each of the five groups JIT, TQM, TPM, HRM, and CI. Stage descriptions are cumulative. Descriptions have been shortened.