Journal of Manufacturing Systems 45 (2017) 1–16

Contents lists available at ScienceDirect

Journal of Manufacturing Systems

journal homepage: www.elsevier.com/locate/jmansys

Implementing Information Technologies and Operational Excellence:

Planning, emergence and randomness in the survival of adaptive
manufacturing systems
Antonio Sartal a,b,∗ , Xosé H. Vázquez b
a
UNIDEMI, Department of Mechanical and Industrial Engineering, New University of Lisbon, Caparica 2829-516, Portugal
b
School of Economics and Business, Rúa Leonardo Da Vinci, s/n. Campus das Lagoas/Marcosende, University of Vigo, 36310, Spain

a r t i c l e i n f o a b s t r a c t

Article history: We explore the why and how of synergies and dysfunctions between Information Technologies (IT) and
Received 18 December 2016 Operational Excellence (OE). Moving away from the type of analyses that focus on the starting-point
Received in revised form 30 July 2017 and the final stage, we address the evolution of the complex factors involved in the joint deployment of
Accepted 30 July 2017
IT and OE that try to transform manufacturing systems. Evidence stems from a longitudinal case study
over a period of 11 years (2003–2014) in one of the world’s largest canning factories. Combining quasi-
Keywords:
experiments at the systems level with time series data, we compare the evolution of IT implementation in
Information Technologies
a Lean environment with 17 different assembly lines and 1,100 workers. Thus, building on an evolutionary
Operational Excellence
Lean routines
framework, the paper highlights the simultaneity of planned, random and emergent components in
Operational performance mutations; the influence of the rhythm and sequence of reforms in the construction of capabilities; and
Evolutionary Theory of the Firm the effect of environmental evolution on the type of mutations passed on over time.
© 2017 The Society of Manufacturing Engineers. Published by Elsevier Ltd. All rights reserved.

1. INTRODUCTION IT has nevertheless developed in its benefits and ease of use

and the digitalization of society has led the gradual incorporation
We explore the synergies and dysfunctions between Informa- of these technologies into manufacturing environments [4]. Today
tion Technologies (IT) and Operational Excellence (OE) in a Lean many young workers in any sector are better at using a tablet PC
environment, delving into how changes come about in a complex than writing with a pen. It is perhaps for these reasons that the man-
system through evolving interactions of individuals, technologies agement and operations literature has gradually moved toward
and departments. positions that show a greater acceptance of IT deployment. Liker [5],
In principle, the introduction of IT in business seems perfectly for example, claimed that using IT was justified considering it is a
in line with OE objectives. However, ever since the invention of the reliable, proven technology, whereas Riezebos et al. [6] suggested
transistor in 1947 and its consequent influence on the dissemina- that introducing IT could add great value to production planning
tion of IT in industrial environments, many authors have been wary or the maintenance management and business aspects, promoting
of how IT might affect the relevance of the human factor. Sugimori principles such as JIT. More recent studies go a step further and
et al. [1] and Ohno [2] stated that the use of IT implied a certain show that IT is a key factor for improving efficiency [7]. In fact,
dehumanization of processes and, more importantly, gave rise to although Lean Manufacturing (LM) can be implemented concep-
unnecessary costs associated with surplus information. Along the tually without IT, under current market conditions, the leaner an
same lines, Toyoda [3] stated that the fact that it was possible to organisation aims to be, the higher the level of IT required. Actually,
receive a huge amount of information “just by pressing a single but- there are more and more studies that analyze Lean-IT initiatives in
ton” could damage workers’ capacity to think and, therefore, their any area of organization: from the business area [8], to production
problem-solving skills. [9,10], maintenance [6] and even associated to the New Product
Development (NPD) process [11].
All of these analyses have undoubtedly been very useful for
considering the potential synergies between IT and organizational
tools for optimizing operations. Few doubts are raised today, in
∗ Corresponding author.
fact, about the capacity of IT to maximize the speed with which
E-mail addresses: antoniosartal@unidemi.com, antoniosartal@uvigo.es
(A. Sartal), xhvazquez@uvigo.es (X.H. Vázquez). information can be exchanged, for facilitating data processing and

http://dx.doi.org/10.1016/j.jmsy.2017.07.007
0278-6125/© 2017 The Society of Manufacturing Engineers. Published by Elsevier Ltd. All rights reserved.
2 A. Sartal, X.H. Vázquez / Journal of Manufacturing Systems 45 (2017) 1–16

analysis and for improving the quality of in-company commu- one of the most influential paradigms in operations management
nication [12,13]. It is less clear, however, how such benefits are [15]. However it seems that there is still no consensus on a single
produced, at what rhythm and in what sequence [14]. Thus, to definition of what “lean” is, what practices and tools it involves, or
understand the complex factors involved in the joint deployment how it should be measured [17,18].
of IT and OE, it is insufficient to solely analyze the starting-point or LM over the time has seen a lot of transformations (Lean Pro-
the final goal, as most of these studies suggest. We have to observe duction, Lean Management, Lean Thinking, etc.) which have been
evolution in the process of transformation, step-by-step. discussed by contributors such as Womack et al. [16], Womack and
In order to do so, the analyses cannot be based exclusively on Jones [19], Shah and Ward [20,9], Bhamu and Singh Sangwan [18],
descriptions of good practices from a strictly empirical point of etc. In fact, the eclectic nature of this concept has led to many
view and based on anecdotal evidence. In the absence of theory, different approaches in the literature [21,9]. Continuing on from
such approximations often lead to specific tools (sub-routines) to the summary drawn up by Bhamu and Singh Sangwan [18], two
be identified as the essence of OE, or many of these tools may be main streams can be identified, one practical and one philosoph-
presented as different alternatives when they are really just dif- ical. While the practical approach or “lean toolbox” considers LM
ferent packaging covering up a lack of novelty in many concepts. as a set of tools and techniques to reduce waste (e.g. [20,9]); the
The organizational solutions featured in academic analyses and in philosophical or “lean culture” approach describes LM as a bundle
much of the more general literature should not be confused with of overarching guiding principles for achieving total transformation
the development of a theory of the meaning of OE, the process of of the firm (e.g. [19,22]).
organizational transformation needed to pursue it, and the role IT Our aim is to build an analysis based on what is permanent,
can play in this context. The theory we need is therefore eclectic, so we address the study of LM from the two main organizational
at the intersection of numerous bodies of literature to clarify the routines of this philosophy: JIT and Jidoka [16,21] and recovering
nature of causal relations among variables necessarily related to IT a third pillar, Respect for People (RfP), very much present in the
implementation, Operations Management and routine dynamics. foundations of the TPS but not in the subsequent dissemination of
Our goal is to identify which elements come first and when. LM [2,23].
Against this background, the following section presents an evo- It is precisely this concept of routine, as a unit of analysis in the
lutionary interpretation of OE. Although different strategies exist study of organizational change, one of the most important ideas in
(Six Sigma, Agile Manufacturing, World Class Manufacturing, etc.), our framework: the Evolutionary Theory of the Firm [24]. Although
since LM can be considered the current dominant paradigm in oper- firms are subject to stochastic processes giving rise to irregular and
ations management [15], we take its principles as the basis for our unpredictable patterns of behavior —an idea that is essential in
analysis. Thus, we first identify three main groups of organizational the paper— the concept of routine reflects what is permanent in
routines existing in any organization and then consider how IT a firm’s behavior. Routines operate as protocols whose capacity for
can stimulate certain mutations giving rise to new (Lean-IT) rou- reproduction —like genes in biology— preserves and reproduces
tines that, depending on the traits of the specific manufacturing the information they contain over time and thus serves as a firm’s
environment, are converted into specific static and dynamic capa- “organizational memory” ().
bilities. Section 3 explains the analytical method used and justifies LM routines cannot therefore be confused with tools such as 5S,
its validity and reliability. Then Section 4, Results, describes the case Kanban or the 5 Whys. More broadly, they provide the warehouse
study evaluating the interaction between lean principles (JIT, Jidoka where the knowledge of each organization is stored (knowledge
and Respect for People −RfP−) and IT features: data reliability, on specific tools and on how to improve them), and simultane-
real-time information and data integration. In Section 5, Discus- ously allow the coordination of individuals with different interests,
sion, we interpret the lessons learned from Jealsaı́s experience. In know-how and information. We can thus identify Just in Time (JIT),
order to facilitate the line of reasoning from project evidence to Jidoka, and Respect for People (RfP) as the key sources of LM rou-
case study analysis and development of propositions, we draw a tines [1,2].
complete parallel between the Results and Discussion sections in In this way, routines allow organizations to coordinate the deci-
order to illustrate how concepts emerging from our data analysis sions and behaviors of a multitude of agents with different interests
provide a solid and grounded base for our conclusions. In Section and know-how. Three main areas can be distinguished: a) informa-
6 we describe the main implications for manufacturing industries. tion processing for coordinating the members of the organization,
While analyzing the synergies between IT and LM in the food indus- b) solving technical problems, and c) designing solutions to the con-
try, we also try to explain how the conclusions obtained here might flicts generated among stakeholders [24–26]. These three types of
be transferable to other manufacturing sectors and different pro- routine can be perfectly transferred to the field of OE according to
cess types. Finally, Section 7 concludes with the main findings and the three pillars of LM as identified by seminal authors [1,2,5]: Just
theoretical implications. in Time (JIT), Jidoka and Respect for People (RfP).
To understand how routines for processing information can be
reflected in JIT as a management philosophy, such operations have
2. AN EVOLUTIONARY INTERPRETATION OF OPERATIONAL to be considered as activities that transform, transfer and trans-
EXCELLENCE port information assets made up of materials and energy [27]. In
fact, information cannot exist without the materials or energy with
2.1. Toyota Production System, Lean Routines and Operational which it is associated. Hence, when we sustain that JIT seeks to
Excellence produce the products required at the right time [1], we are really
stating that information transmission routines must guarantee the
Lean Manufacturing (LM), often known simply as “Lean”, is coordination of operations by generating the right information at
a systematic approach for eliminating the waste or Muda (non- the right time.
value-added activities) by means of on-going improvement. LM is However perfect the initial design of information transmission
based on the Toyota Production System (TPS), which was largely routines might seem, deploying them is subject to constant pres-
grounded on the work done by Taichi Ohno in Toyota [2,5]. The sures that obstruct coordination. Complementary routines thus
generic term “LM”, coined by Womack et al. [16], appeared for the have to be designed to guarantee the quality of processes so that
first time in their seminal book “The machine that changed the information is incorporated in products in the way planned. The
World”. Thirty years later LM is considered by many authors to be problem-solving routines identified by Evolutionary Theory of the
 A. Sartal, X.H. Vázquez / Journal of Manufacturing Systems 45 (2017) 1–16 3

Firm are thus reflected perfectly in the second main pillar of Lean: lar characteristic is adaptive per se; it all depends on the capabilities
Jidoka, or “automation with a human touch”. This allows the pro- it generates to confront the specific selection mechanisms of the
cess to have its own quality self-control (“zero defects”) so that, in environment.
case of an anomaly, production can be stopped and any defective It is at this point that we understand that IT can play a key role in
parts prevented from moving on to the following workstation [2]. the mutation of OE routines —Lean routines— to successfully adapt
Finally, even in those industrial environments where the groups to the new customer requirements related to response times, flex-
of routines described (JIT and Jidoka) are properly established, pres- ibility and customization [32]. Although the mistrust expressed by
sure groups often have different interests and a stable social pact traditional lean authors regarding the effect of IT on the human
is therefore needed to avoid constant conflicts in the organization factor continues to limit the widespread adoption of IT in lean envi-
of operations. From this point of view, Coriat [26] emphasize the ronments [2,5], the digitalization of society has made it easier for
traditional conflict between employers and employees over the dis- IT to be gradually incorporated into industrial environments [7,14].
tribution of surpluses and over work organization as the third key Perhaps for this reason, although IT and LM have traditionally been
challenge that a specific type of routine must pay attention to. The considered antagonistic, a new trend is growing within Lean, the
objective is to achieve a stable truce so that information transmis- “Lean-IT partnering”, that not only favors their joint deployment
sion and problem-solving routines can be deployed correctly. This but also emphasizes their complementarity and inter-dependence
is the argument behind the third pillar, called “Respect for People” [6]. In fact, ample evidence suggests nowadays that use of IT is
(RfP) [1,2,5]. Avoiding a romantic view of participation, even Ohno indispensable in current lean environments for managing the huge
[2] reflected on the “JIT + Jidoka” system as a type of management amount of required data and supporting new lean e-tools [10,33].
by stress. This is a type of management that, compared to Taylorism But to understand the complex factors involved in the joint deploy-
or Fordism, stimulates a high participation of workers in problem- ment of Lean-IT, it is insufficient to solely analyze the starting-point
solving. However, it can simultaneously generate even more stress or the final goal. It is necessary to observe the origin of each IT-
than that stemming from systematic repetition of a few simple mutation, its evolution, and even the possibility that organizational
tasks. We can therefore observe here how this third pillar comple- routines might develop that are different but adaptively efficient in
ments the empowerment inherent in JIT and Jidoka with an explicit the environment for which they were designed.
concern for improving working conditions and worker motivation, The concepts of mutation and selection mechanism are there-
balancing out the job burnout generated by tight flows and con- fore fundamental for understanding the different paths that can
tinuous improvement processes among different stakeholders [23] lead a firm toward Operational Excellence (OE) and how the three
(Fig. 1). socio-technical routines identified previously (JIT, Jidoka and RfP)
turn into certain capabilities in a specific manufacturing environ-
2.2. Generating capabilities from Lean routines: mutations, ment [27]. Following Grant [34], we define the firmı́s capabilities as
selection environments and IT the collective organizational skills or competencies —routines— for
performing a specific task or activity. While much of the literature
In addition to defining and identifying the routines involved in has only focused on the competitive consequences −the outputs−,
operations management, it is important to understand why the we aim to explicitly evaluate the influence —the why & how— of
content of such routines varies. Bearing in mind that firms vary manufacturing environments. We consider not only how IT causes
greatly in strategic and organizational terms, the organization of mutations in organizational routines (JIT, Jidoka and RfP) giving rise
operations may undergo a phenomenon equivalent to the genetic to new Lean-IT routines, but also how these new routines (JIT-IT,
recombination and mutations in biological species [28]. At the root Jidoka-IT and RfP-IT) might lead to specific capabilities depending
of this phenomenon, one can find two irreconcilable perspectives in on the specific traits of manufacturing environments. With this goal
biology but compatible in a social environment: the standard ran- in mind we differentiate between two types of capabilities, static
dom variations of the Darwinian approach, on the one hand, and and dynamic, and use several plant KPIs to determine whether new
the variations planned —that sought the greatest possible returns— Lean-IT routines actually give rise to new (static or dynamic) capa-
associated with the Lamarckian view of evolution, on the other. bilities, that is, whether target indicators rose significantly in the
Since the Lamarckian perspective is probably the most period analyzed (2003–2014).
widespread among analysts studying social systems and, in par- Static capabilities are the skills with which firms roll out rou-
ticular, business management, an increasing number of authors tines in order to respond to recurring stimuli. They are therefore
emphasize the need to complement the intentionality of man- mainly linked to performance indicators. Such capabilities affect
agers in organizational change with the stochastic—and even the level of productivity, quality, or customer service at any given
emergent—origin of many other mutations [27,28]. From this point time and are difficult to alter in the short term because the routines
of view mutations could have a random origin and unpredictable that sustain them cannot substantially change how and how much
results, but they could also arise out of self-organization of the a firm produces. On the other hand, dynamic capabilities reflect the
elements in the system without the need for a central planning firm’s potential for creating, developing and recombining resources
authority (as with the biosphere, cities or the stock market). The with a view to deploying new strategies to create value [30]. Thus,
organization of operations would thus change both abruptly and while static capabilities relate to the firm’s day-to-day activities
radically (through mutations in routines), and slowly and incre- and can be perceived through fixed efficiency indicators (such as
mentally (by recombining sub-routines), guaranteeing the transfer costs and productivity ratios), dynamic capabilities (e.g. continu-
of apparently adaptive solutions over time. ous improvement) reflect the firm’s skill at adapting its routines to
Evolutionary Theory of the Firm stresses the term ‘adaptive’ changes in the environment [35].
rather than ‘efficiency’ because, in the same way that biological
evolution does not guarantee ‘fair’ results, the evolution of oper-
ations in a firm will not necessarily guarantee optimal results in 3. METHODOLOGY
terms of efficiency in comparison with previous situations (see, e.g.
[24,29,30]). Thus, a new organizational routine is adaptive to the 3.1. Research design
extent that it gives rise to a capability that increases survival against
the selection mechanism of the environment in which it operates To analyze the interaction between IT and OE, we conducted a
[24,31,28]. It cannot therefore be anticipated a priori that a particu- longitudinal case study for 11 years (2003–2014) in a specific sec-
4 A. Sartal, X.H. Vázquez / Journal of Manufacturing Systems 45 (2017) 1–16

Fig. 1. Analysis of Lean routines from an evolutionary perspective [69].

tor, the food and beverage industry. We follow a mixed-methods traits we needed to obtain the insights that other firms could not
strategy that combines time series data with quasi-experiments provide.
to compare the ecosystem before, during and after the Lean-IT.
The rationale is threefold. First, our study aimed to investigate
3.1.2. Reliability
how changes come about in a complex system through evolving
The field study followed a plan in stages. It included a protocol
interactions of individuals, technologies and different departments,
with working rules, data collection tools and timing, and proce-
which are normally difficult to quantify. Second, we examined why
dures for analyzing the evidence [41]. The protocol is useful as a
and how synergies and dysfunctions exist by adopting a holis-
working guide (internal reliability), and it allows other researchers
tic perspective of the phenomenon [36]. Third, we used various
to reproduce the study (external reliability). It covers the motiva-
data collection and analysis techniques so that the information
tion of the research, the field procedure (arranging for interviews,
could be triangulated. All of these tasks required an open and iter-
experiments, etc.) and the guide for reporting the case. Its valida-
ative approach to data collection and analysis to conceptualize the
tion started from the pilot case with the canning lines and gradually
evolutionary processes involved [37], anticipate when plausible
improved with each visit and interview [41]. All of the informa-
propositions were raised and, in such cases, develop a theory with a
tion obtained from different sources was arranged by category and
few specific guidelines that would be perfectly transferable to any
recorded in a single research database using specific analytical soft-
type of process or manufacturing environment different from the
ware to guarantee the chain of evidence.
analyzed one [38]. Qualitative studies such as ours are nevertheless
subject to selection bias and potential prejudices of researchers
throughout, from the initial planning of the research to the final 3.1.3. Validity
stage of publication. We therefore took due care to select the right The accuracy of our results was supported by the triangulation of
case and to ensure a strict chain of evidence so that our findings data, researchers and methodologies [42]. The data were collected
could be analyzed correctly [39]. from secondary sources of a documentary nature and primary
sources such as direct observation (on-site visits, direct participa-
tion in groups for improvement, conversations with workers. . .)
and 56 semi-structured interviews (45 min. −2 hours). These inter-
3.1.1. Case selection
views took place with different agents in the chain of command
We needed to find a firm that would allow us to examine paral-
(workers, middle managers and senior managers), tapping into
lel adoption of lean tools and IT from the start and to do so over a
potential differences in perspectives to mitigate biases and lapses
sufficiently long period in order to be able (a) to observe the influ-
[43]. The interviews were conducted by at least two researchers to
ence of this interaction on the construction of capabilities, and (b)
pick up possible points missed by each interviewer and to avoid
to study the influence of the environment on the types of muta-
‘observer bias’ [44]. Table 1 describes the number of interviews per
tions that survive over time. On the one hand, Jealsa Rianxeira
round and respondent category.
allowed this research question to be posed in a “green field” with
Finally, in addition to semi-structured interviews on the rela-
limited experience in the development of OE methodologies and
tions of interest, we carried out three different experiments
IT implementation. On the other hand, the industrial ecosystem
in accordance with the case study objectives. The diversity of
chosen (1,100 workers and 17 canning lines, 23 autoclaves and 13
information sources and participants, and the use of multiple
different packaging lines) works with “High Volume-High Variety”
methodologies not only provided an enhanced understanding of
processes: 700 million cans per year, with 1,500 different living ref-
the interaction between IT and OE but also helped to prevent
erences. This allowed us to test IT mutations where lean principles
overoptimistic assessments on our part and to verify the findings.
are apparently more difficult to implement [5].
Finally, the opportunity to link the lean project to a newly grad-
uated engineer allowed us to participate directly in the design and 3.2. Data Analysis and chronology
implementation of the Lean-IT initiatives from the beginning. The
experiential nature of our analysis—with strict procedures to min- Our data analysis combines time series data with quasi-
imize observer bias—allowed us to experience and interpret the experiments to compare the ecosystem before, during and after the
events first hand and in real time, facilitating the study of processes Lean-IT implementation from a longitudinal case study perspective.
in all of their richness, compared, for instance, with retrospective Using the qualitative data analysis software, NVivo, and follow-
research [40]. Hence, regardless of whether Jealsa Rianxeira reaches ing Yin [41], we created a case archive providing a chronological
the distinctiveness of a “talking pig” [38], it definitely reflected the overview of events that is summarized in Fig. 2.
 A. Sartal, X.H. Vázquez / Journal of Manufacturing Systems 45 (2017) 1–16 5

Table 1
Description of interviews conducted.

Respondent category 1st round (2004-05) 2nd round (2009-10) 3rd round (2012-13) Total Respondents

Department Manager 1 2 2 5 (9%)

Area Manager 2 2 2 6 (11%)
Project Team 2 2 3 7 (13%)
Maintenance Staff 2 4 5 11 (20%)
Shop-floor worker 2 3 8 13 (23%)
Office Staff 2 3 4 9 (16%)
External consultant 1 2 2 5 (9%)
Total 12 (21%) 18 (32%) 26 (46%) 56 (100%)

Fig. 2. Lean-IT project (below) and longitudinal case study and methodological tools used (above).

This structure has a dual function. It allows us to identify over random and emergent IT events can provoke mutations in OE rou-
time the sequence of landmarks in the Lean-IT project (below) tines that are as adaptive as those with a planned origin.
while simultaneously considering the different methodologies in
our research (interviews, quasi-experiments and time series data). 3.2.2. Step 2. Emergence of new Lean-IT routines: Mutations in
Each stage of the project closed with a round of interviews (Fig. 2) to organizational routines provoked by IT (2006–09)
assess the evolutionary nature of the relations studied. Then, from In the second stage we evaluated, qualitatively and quanti-
this file, we performed our analysis in three sequential steps as tatively, how IT characteristics identified by Moschella [45] and
described below. As can be seen in each of the following steps, the Dewett and Jones [46] (data reliability, real time information and
mixed-methods strategy followed allow us to establish a complete knowledge integration) affect each of the three main Lean rou-
parallel between the Lean-IT project development, the case study tines separately: JIT, Jidoka and RfP giving rise to new Lean-IT
analysis and the development of propositions in order to illustrate routines (JIT-IT, Jidoka-IT and RfP-IT). We adopted a replicating
clearly how concepts emerging from our data analysis provide a logic equivalent to multiple-case studies and used three different
solid and grounded base for our contributions. quasi-experiments to compare two identical production lines, one
of which maintains the old organizational routines (control line)
3.2.1. Step 1. The origin of mutations: Implementation of IT while the other applies different Lean-IT initiatives (Table 2). Our
systems in a lean environment (2003–08) purpose is to analyze how each of the IT-mutated routines seize
We start our analysis by describing the implementation of the in different IT traits: technical problem-solving hinges specially
main IT systems applied in the lean project: first, the acquisition on data reliability, information transmission on real-time informa-
of Enterprise Resource Planning (ERP), then the implementation tion, and conflict management on knowledge integration. In order
of the Manufacturing Execution System (MES) and, finally, their to control as far as possible the conditions of the experiment, the
integration (ERP-MES). methodology in step 2 reflects the logic of a typical healthcare
In this first stage, we focus the discussion on the different trig- experiment, where one group receives treatment and another (the
gers for each IT implementation. While IT projects may all be clinical control group) does not, in order to determine the effec-
intentional, their trigger may have a planned origin, or may be tiveness of the drug or treatment being tested. We nevertheless
caused by random or emergent events. The methodology used performed a Factorial Analysis of Variance (ANOVA) to rule out
here comprised semi-structured, multi-interviewer interviews. alternative explanations based on differences between shifts.
Two sequential rounds were held at different times (2004 and
2009) where the same questions were asked to compare the initial 3.2.3. Step 3 Conversion of new Lean-IT routines into specific
and final status of the implementation. The information garnered, static and dynamic capabilities (2009–14)
together with that from 2012, complements the results obtained Finally, using a time series data analysis and different plant indi-
from the quasi-experiments and the time series data (Fig. 2). With cators based on Hill et al. [47], we checked how the new Lean-IT
the data collected, we can thus evaluate, in this initial stage, how routines affected different areas of industrial performance and the
6 A. Sartal, X.H. Vázquez / Journal of Manufacturing Systems 45 (2017) 1–16

Table 2
Description of the quasi-experiments performed.

New Lean-IT routines evaluated IT feature Description of quasi-experiment Date

Jidoka-IT Reliability Comparative analysis using a t-test and Oct–Dec 2006

factorial analysis: Manual recording vs.
automatic identification and data capture
(AIDC) systems
JIT-IT Real-time information Comparative analysis using a t-test and Feb-April 2008
factorial analysis: “e-kanban” vs. conventional
(manual) kanban
RfP-IT Data integration Qualitative analysis: IT vs. non-IT, Toyota’s Jan-Oct 2009
7-step problem-solving method

generation of specific static and dynamic capabilities according to vious background in IT (that the firm did not actually search for)
project goals. On the one hand, this analysis attempts to assess how had an unexpected domino effect: when lean initiatives began to
the specific traits of manufacturing environments make groupthink demand huge amounts of real-time data from the shopfloor, his
biases a key determinant of how Lean-IT routines are converted influence over the Industrial Managerı́s decisions became decisive
into specific static and dynamic capabilities. On the other hand, we for the adoption of MES.
address how the rhythm and sequence of implementations influ- Initially, due to the high costs and complex implementation of
ence the timing of the conversion of new Lean-IT routines into the MES, only two options had been contemplated to deal with
capabilities. data management in the plant: extending ERP downstream, or rein-
forcing manual data recording. However, based on his previous
4. Case findings: THE LEAN-IT PROJECT IN JEALSA knowledge, the new engineer suggested dividing the implemen-
tation of MES into various phases and to start with a single module
4.1. Implementation and integration of IT to test its applicability. Furthermore, in parallel and on his own
initiative, he obtained a public subsidy to support this possibility,
4.1.1. Acquisition of Enterprise Resource Planning (ERP) allowing the firm to perform “a little pilot experiment” almost for
Jealsa acquired its corporate ERP (All-in-One@ConservasTM by free. This was undoubtedly the final trigger that ended up con-
SAP) in 2003. With advice from the IT department, the management vincing the top management team. Thus, a random event (hiring
decided that the firm’s internationalization process begun one year someone with an educational background that the firm did not
earlier required the adoption of a global solution to standardize actually search for) changed the decision. So the introduction of
the management system and integrate all of the information from the MES, in itself a planned initiative, had nevertheless a random
plants all over the world [48]. The IBM AS/400 system that the origin.
firm already had was no longer meeting the requirements of the Jealsa acquired this first module, the productivity module
group, which had doubled its sales over recent years. Moreover, IT (TraksysTM from Parsec), in 2005. The choice of the productivity
quick fixes or made-to-measure applications were common prac- module implicitly established the order for the industrial man-
tice. Therefore, after a thorough comparative study, the group chose ager’s goals (efficiency and quality) in the subsequent deployment
to adopt the sectorial solution developed by SAP and supported of the project. After successfully completing this first pilot stage, the
with Hewlett-Packard hardware. This new technological platform project was completed into successive stages for both software and
was expected to meet the firm’s specific needs and its potential hardware. After a while, the facilities were also adapted for Auto-
future outsourcing requirements, and to integrate the business matic identification and data capture (AIDC), and different modules
systems efficiently without affecting productivity [49]. Given the were installed progressively to monitor quality or traceability data,
variety and extent of the group’s activities and facilities, the migra- among other functionalities (Fig. 2).
tion process took place in two stages. In 2003, a pilot project was
run in one of the firms of the group. Then, starting in 2004, adoption 4.1.3. Integration of ERP and MES
went ahead in the group’s other 21 firms around the world. The two initiatives—industrial and corporate—came together in
Although the whole group today uses the same information 2009. On the one hand, the ERP adopted a few years before was,
system, the degree of integration and utilization of SAP differs as usual in these cases, under-used in the production area in com-
substantially from area to area. Although the process started out parison with other business modules [50]. It had been gradually
in a planned way, the actual adoption process varied in different displaced by the MES, so it was only receiving information from
areas of the company. Both the nature of the ERP and the guidance the old paper-based data collection system. On the other hand, the
provided by the IT department caused the focus to be placed on lean project initiated in the industrial area in 2005, which found
‘standard’ business areas (finance, sourcing or sales), while other a perfect ally in the MES, needed greater visibility in the plant as
areas such as production or quality, with much more specific pro- well as information from the whole organization. This could only
cesses, were addressed in a second stage and with less vigor [50]. be done by integrating the two knowledge models —industrial and
corporate— or, in other words, by means of two-way integration of
4.1.2. Implementation of Manufacturing Execution System (MES) information between the MES and ERP platforms.
While the ERP acquisition was a strategic management decision, We observe here that a gradual evolution took place to cover
the MES implementation arose in 2004 from a bottom-up decision the gap between automatic data collection in the plant (MES) and
sparked by a random trigger: after decades of taking on middle production planning (ERP). With both systems running on a regular
management with qualifications in food sciences the firm accepted, basis, managers in each area (planning, logistics, human resources,
for the first time in this area, a systems engineering student intern etc.) began to see the usefulness of integrating information from
to prepare his bachelor thesis on autoclaves monitoring. On com- “both sides”. Disconnected initiatives began then to emerge, not
pletion of this academic requirement, this engineer was taken onto only demanding new functionalities but also incorporating new
the staff of Jealsa (October 2004) with the general (and unspecific) utilities to the ERP-MES integration: statistical process control,
aim of “improving production processes”. It was here that his pre- alarm management, traceability, etc.
 A. Sartal, X.H. Vázquez / Journal of Manufacturing Systems 45 (2017) 1–16 7

From the industrial point of view, the integration of both sys- working days, see Table 3), we compared data collected automat-
tems made it possible, for example, to manage different batches ically (with the TraksysTM ) and manual data recording from two
better in real time, reducing their size and generating global identical processing lines —making the same product— with and
production efficiency indicators by linking plant and business infor- without automated data collection. We took Availability (percent-
mation. Additionally, users continued working in their usual ERP age of scheduled time in which it is possible to work with the
environment but with more and better information, and in real equipment) as the key measure to evaluate the importance of reli-
time. So, both initiatives started out separately but allowed two tra- ability. This is one of the three underlying parameters yielding the
ditionally isolated areas (industrial and corporate) to join together OEE indicator, and it is definitely where the human factor has the
even though there was no explicit plan to address this challenge. greatest impact.
We use Factorial Analysis of Variance (ANOVA) to test for differ-
4.2. How did IT affect the Operational Excellence organizational ences in the AIDC systems and manual recording (Table 3). Factorial
routines? ANOVA is an inferential statistical test that can be used to evaluate
the individual and joint effect of two or more factors (indepen-
4.2.1. The start of the lean project and the need for IT dent categorical variables) on a dependent quantitative variable
Taking the “canning–sterilization-packaging area” as the main (the main effect). In our experiment, the dependent variable was
field of concern, the top management formed a project team com- the recorded availability and dependent variables were the type of
posed of the main staff with production-related responsibilities record used (AIDC/manual) and the shift (morning/afternoon). Our
(Production, Quality, R&D and Maintenance), coordinated for oper- purpose is to contrast the null hypothesis that there are no signif-
ational purposes by the intern and led by the Industrial Manager. icant differences between both methods of data collection, while
With this group composition, the project was implemented in controlling simultaneously for possible differences between shifts
three sequential stages. First, the industrial quality and productiv- that could reflect variations in the initial conditions. Thus, if we
ity indicators (associated with strategic objectives) and the scope find that “shifts” make a difference in the dependent variable, we
(product family) were established area by area. To measure the could not rule out alternative explanations related, for instance, to
processes, we used, among others, Overall Equipment Effectiveness worse maintenance or staff with different training or experience.
(OEE) and Total Equipment Effectiveness Performance (TEEP). Both Conversely, alternative explanations seem unlikely if there are no
these measures are designed to identify losses of manufacturing differences between the manufacturing shifts. To be sure, further-
effectiveness (availability, performance, and quality rate) consid- more, the extensive time frame analyzed (three months) should
ering even planned downtime [51,52]. In the second stage, we absorb any small differences occasionally arising between the lines.
identified the flow of information and the value processes by means Table 3 describes the main statistics of our sample and then
of Value Stream Mapping (VSM). Finally, together with the indus- shows the Factorial ANOVA results. There is a null hypothesis for
trial staff, we designed the action plan based on “Toyotaı́s 7-step each factor (Capture type and Shift) and for each possible combi-
problem-solving method” [5]. This last stage led to the adoption of nation of factors (Shift*Capture type). The null hypothesis referred
different lean tools according to the specific quality or cost reduc- to a factor states that the means are the same and the critical level
tion purpose: 5S, SMED, Kanban, and others. (Sig) associated with each F statistic allows us to decide whether
The methodology worked well for triggering an improvement we can maintain or discard this hypothesis. Since the probability of
process that helped overcome the main organizational barriers. At this statistic for “Capture type” (F = 4.812, p = 0.031) is lower than
the same time, however, some important shortcomings became 0.05, we can reject the null hypothesis of equality of means, and
clear that were preventing full benefit from being obtained from the hence conclude that availability obtained through manual and AIDC
lean transformation. On the one hand, the production data were not capture is different. However, the values associated with the Shift
available in real time because they were being recorded manually; variable (F = 0.780, p = 0.380) confirm that there are no differences
on the other, the use of a paper-based method of data collection between shifts. We can therefore suggest that potential differences
was labor-intensive and resulted in many mistakes. The lack of in recorded availability are not due to the potential particularities
information was therefore often considered the main reason for the of each shift.
projects not achieving the expected success. It was in this context Results at first glance would suggest that automatic data capture
that, as explained above, the company acquired in 2005 an initial reduces line availability by 5.47% compared to manual recording.
MES prototype for the pilot area (canning lines), which was subse- Obviously, however, the information record does not affect the
quently extended to the rest of the plant (2005–2008). While many availability of the line, so what these findings really indicate is
of this tool’s utilities were not used initially (e.g. integration with that the automatic data record is more reliable than the manual
the ERP or planning), the IT contributions led to improvements in record. Many micro-downtimes and line slowdowns passed unno-
the total plant capacity or the Work-in-process (WIP) —semi-finished ticed, but, since control mechanisms and workers’ incentives are
buffered products to proceed to the next stage—, among other plant mainly focused on production volume, others were simply ignored.
indicators. Undoubtedly, the data reliability, real-time informa- As a result, because availability ratios were higher than the real
tion and knowledge integration changed the old Lean routines and ones, there was little pressure to achieve constant improvement
ended up improving the industrial indicators. and, even worse; too much emphasis was placed on actions on
We assess each of these relations in depth in the following sec- work lines that were not really of prime importance. In fact, real
tions. We first explain any mutations taking place in each of the knowledge of the situation allowed actions by the improvement
Lean routines using three different quasi-experiments (Table 2). group to be realigned and stepped up.
Then, in subsection 4.3., we assess the effect of organizational But this situation is not exclusive to the food industry. Recent
changes on Jealsa’s manufacturing strengths, and establish the basis industrial evidence highlights the increasing importance of tech-
for evaluating, in Section 5, to what extent new Lean-IT routines are nology for managing the huge amount of real-time information.
transformed into specific distinctive capabilities. There are various examples in the literature to support this asser-
tion. Ward and Zhou [8], for instance, found that when lean
4.2.2. The first change achieved by IT in Operational Excellence practices are adopted, the need to implement information sys-
routines: data reliability tems such as ERP arises. Similarly, Moyano-Fuentes et al. [10] show
One of the most important effects of IT in OE routines was asso- empirically that there is a direct link between the internal IT-level
ciated with data reliability. From October to December 2006 (42 and the degree of implementation of lean initiatives. This is espe-
8 A. Sartal, X.H. Vázquez / Journal of Manufacturing Systems 45 (2017) 1–16

Table 3
ANOVA test. Evaluating availability recorded: AIDC vs. Manual data recording.

Dependent variable Sub-groups N (days) Mean Diff. in means Std

Availability recorded Capture type AIDC 42 78,36% −5.47% 10,81%

Manual 42 83,83% 11,94%
Shift Morning 42 82,20% 2.20% 10,44%
Afternoon 42 80,00% 12,69%

Dependent variable: Availability recorded

Source Type III Sum of Squares df Mean Square F Sig

3a
Corrected Model 730,22 3 243,408 1,865 0,142
Intercept 552485,191 1 552485,191 4233,276 0,000
Capture type 627,966 1 627,966 4,812 0,031*
Shift 101,739 1 101,739 0,780 0,380
Shift* Capture type 0,519 1 0,519 0,004 0,950
Error 10440,807 80 130,510
Total 563656,221 84
Corrected Total 11171,030 83
a
R Squared = 0.065 (Adjusted R Squared = 0.030).
*
Statistical significance (p < 0.05).

cially important in certain engineering fields such as maintenance. for the same shift, stoppages (in red) on the traditional line with
Here the emergence over recent years of information-intensive those on the line where the new lean-IT organizational routines
Total Productive Maintenance (TPM) programs has made it essential were generated.
to implement management information systems [6]. Applications As in the previous case we have used a Factorial ANOVA to
such as CMMS (Computer Maintenance Management Software) are test the null hypothesis that there were no significant differences
increasingly common in the maintenance monitoring and for spare between the means of both samples. That is, we aimed to detect
parts management in order to meet customer demands in terms of significant differences between the average values for the level of
response times and flexibility. capacity achieved on the e-kanban line and on the traditional line
(Table 4). Just as we explained above, this type of analysis allows
to determine whether there can exist differences between shifts
4.2.3. Real-time information management: e-kanban in order to control as far as possible the conditions of the experi-
A second effect of IT was detected in real-time information ment. Since we compared two duplicate production lines, making
management. Upon the installation of an RFID system in 2008 identical products and during the same period of time; if we can
in the sterilization area, we performed a quasi-experiment over verify in addition that there are no differences between manufac-
three months (February–April 2008) to evaluate how this IT feature turing shifts (workers with similar experience and training, the
affected the old organizational routines. By adopting a replicat- same maintenance, etc.) alternative explanations to the influence
ing logic equivalent to multiple-case studies, we compared two of Lean-IT routines, seem unlikely. Furthermore, as in the previ-
identical production lines making the same product, one of which ous case, the extensive time frame analyzed (three months) should
maintained the old organizational routines (control line) while the absorb any small differences occasionally arising between the lines.
other adopted certain IT-resources (e-kanban system) supporting We verify firstly that the probability of the F statistic (F = 89.621,
lean practices. p = 0.000) regarding the “Type of line” is less than 0.05, so we can
With this goal in mind, an electronic chip was incorporated reject the assumption of equality of means and confirm that there
in every ‘sterilization basket’. In this way each basket became a is a significant difference between the levels of capacity achieved
decision-making unit that could manage the flow of materials and by each area (e-kanban vs. traditional kanban). Thus, the new Lean-
stock in real-time between the two working areas (canning and IT routines may improve the sterilization capacity by 19.6% (from
packaging). All production and quality information that was pre- 67.35% to 80.55%). Analogously to the previous experiment, we
viously collected manually started to be obtained automatically by can also suggest that this difference is not due to divergent Shift
each chip and managed by the MES. It was therefore possible to ana- characteristics (F = 0.259, p = 0.612).
lyze data on the production process in real time (type of product,
occupation, line speed, efficiency, downtime), as well as informa-
tion associated with quality assurance and food safety (traceability, 4.2.4. Knowledge integration: the reduction of information
waiting times, etc.). asymmetries through IT
Automatic data capture thus provided the user with the infor- A third effect of IT in OE routines was seen in knowledge
mation needed in the right format, at the right time and in the right integration and the reduction in information asymmetries. While
place [53]. Any incidents could then be resolved in real time and lean practices lead to a degree of empowerment in that workers
proposals for improvement adopted immediately without having have greater responsibility and more decision-making capacity,
to stop the line or re-plan production. In addition, inclusion of this it also means that they might find their work more stressful
information in the MES programming modules, with the appropri- [2,16]. To test this idea, we carried out a third quasi-experiment in
ate restrictions, allowed the system to learn and quickly respond 2009 (January–October) to see how the different IT systems (MES,
to internal fluctuations. This helped optimize work in process (WIP), ERP, digital screens) affected workers based on “Toyotaı́s 7-step
smooth production and, consequently, improve the capacity of the problem-solving method” [5].
area. In fact, while the other characteristics of the line remained Although the Liker method is the plan of action used through-
unchanged, the assimilation of new lean-IT routines resulted in an out the Lean-IT project, we decided to focus this third experiment
increase of 19.6% in the e-kanban area sterilization capacity, in com- on the last stage of our pilot ecosystem (packaging) and once the
parison with the traditional production area. In order to illustrate implementation was broadly introduced (2009). We did so because
this situation, Fig. 3 shows a qualitative example that compares, the preliminary areas (canning and sterilization) had to have con-
 A. Sartal, X.H. Vázquez / Journal of Manufacturing Systems 45 (2017) 1–16 9

Fig. 3. Downtimes on sterilization lines.

Table 4
Factorial ANOVA test. Evaluating level of capacity achieved: traditional kanban vs. e-kanban.

Dependent variable Sub-groups N (days) Mean Diff. in means Std

Level of capacity Line type Traditional kanban 61 67,35% 13.20% 8,09%

achieved e-kanban 51 80,55% 7,31%
Shift Morning 61 74,19% 0.49% 10,63%
Afternoon 61 73,70% 9,71%

Dependent variable: Level of capacity achieved

Source Type III sum of Squares df Mean Square F Sig

Corrected Model 0.544a 3 0.181 30.548 0.000

Intercept 66.722 1 66.722 11234.376 0.000
Type of line 0.532 1 0.532 89.621 0.000*
Shift 0.002 1 0.002 0.259 0.612
Type of line*Shift 0.011 1 0.011 1.897 0.171
Error 0.701 118 0.006
Total 67.957 122
Corrected Total 1.245 121
a
R Squared = 0.437 (Adjusted R Squared = 0.423).
*
Statistical significance (p < 0.05).

Table 5
Third quasi-experiment: comparison between old Lean routines and new Lean-IT routines according to “Toyotaı́s 7-step problem-solving method”.

7-STEPS LEAN ROUTINES LEAN-IT ROUTINES

1. Initial perception of the problem Lean Manufacturing recommends going to the ERP-MES integration allows synthesizing both knowledge models —industrial
workplace, seeing the problem, gathering all and corporate— into an only platform and decision-making responsibilities to
possible information and comparing the initial be delegated to the work-teams in shopfloor [16]. Automatic data capture
situation with the objective. enables this information to be obtained in real time so that problems can be
2. Clear definition of the problem The objective is to convert a complicated prioritized automatically and possible corrective measures assessed and
problem into small projects (workshops) that compared with historical analysis of incidents.
can be dealt with successfully.
3. Analysis of causes Once the problem has been described, it is The MES allows data to be stratified and clustered in real time and, if
necessary to focus on the causes. Success will necessary, to be analyzed objectively and interpreted immediately using
depend on the skill with which the method is Pareto charts and other graphical tools integrated into the software.
implemented. Furthermore, IT frees workers from data collection tasks, allowing them to
4. Analysis of root cause The root cause is determined using tools such analyze available information directly and often to determine the potential
as the Ishikawa diagram and the 5 whys. causes and, finally, the “root cause” in a single step.
5. Drawing up counter-measures Once the root cause has been identified, the ERP-MES allows workers themselves to propose actions immediately and to
corrective plan can be adopted and monitoring quickly (in hours or days) determine the validity of counter-measures [14].
meetings planned. MES is accordingly considered the best tool for measuring performance
6. Verification of results and the process Lean projects are monitored in periodic indicators in real time, even by workers, using mobile plant devices and digital
meetings of improvement groups. screens.
7. Standardization If the counter-measure is effective, it will form MES-ERP integration generates standardized time series of data, promoting
part of the new standard and will gradually be the transfer of knowledge among all levels of the firm without additional cost
transferred to the plant. or effort. System scalability also allows for reduction in implementation times.

solidated the Lean-IT routines so that we could focus on knowledge this latter experiment showed how the use of IT had an immediate
integration. Second, we knew that in the packaging area there were effect on the reduction of information asymmetry, both verti-
still some lines that were not being monitored, which allowed us cal and horizontal. This is possible because ERP-MES integration
again to use a replicating logic to study this third relation. The com- allow synthesize the main two knowledge models —industrial
parison between the areas with IT and without IT is summarized in and corporate—, traditionally separate, into an only common plat-
Table 5. form. On the one hand, automatic data capture allowed managers
Therefore, in addition to positive effects on reliability and speed to supervise process information and monitor quality in real
in decision-making (already noted in the other quasi-experiments), time, which in turn increased the transparency of the process
10 A. Sartal, X.H. Vázquez / Journal of Manufacturing Systems 45 (2017) 1–16

and reduced opportunistic behavior by workers. On the other 5. DISCUSSION: interpreting THE JEALSA EXPERIENCE
hand, the use of plant devices such as Display Screen Equipment INTERPRETING
(DSE), very common in sectors such as automotive or electronic
manufacturing, allow efficiency ratios be transferred to the work- The case of Jealsa Rianxeira reflects the key intellectual chal-
place to link the variable part of wages to compliance with the lenge in any enterprise: how to manage technological change
OEE standard agreed on. At the same time, daily information is and organizational processes consistently and simultaneously with
fairly distributed among the members of the organization, avoid- constant requirements for adaptation [24]. Thus, having identified
ing conflict at every level by reducing the privileges that stem the main landmarks in the process of joint implementation of the
from asymmetric access to information. As in many manufacturing lean and IT tools, we now have to carry out an abstract exercise, cod-
firms with functional structures, the Jealsa departments developed ifying the knowledge to organize and structure the facts described
their own—sometimes divergent—goals, leading to organizational (Fig. 5).
entropy. In this sense, the Knowledge integration from ERP and MES In order to facilitate the line of reasoning from project evi-
into a common model had a synergic effect because they both work dence to case study analysis and development of propositions, we
with the same information, cooperation among departments was establish a complete parallel between the Results (case study) and
encouraged. Integration between the industrial and the corporate Discussion sections. This is summarized in Table 6.
area not only helped improve the OEE (reaching values of 67%) but We begin our analysis, from the evolutionary point of view,
also helped cover the gap between information in the plant and pro- with the adoption of ERP (planned mutation) and MES (stochas-
duction planning made viable by the ERP. The increase observed, tic mutation), moving on subsequently to their integration in the
for instance, in indicators such as TEEP (16%) and the reduction in ERP (emergent mutation). In this way, the evidence obtained from
the WIP inventory (27%) reflected this positive trend. Jealsa thus the 11-year case study allows us to assess how random and emer-
gained in ambidexterity by reducing its lead times while benefit- gent IT events can provoke mutations in OE routines that are as
ing from lower costs due, among other things, to the improvement adaptive as those with a planned origin. Then, in a second stage,
of processes capacity [54]. we interpret the results of the three quasi-experiments performed
and how the different IT features (data reliability, real-time infor-
mation and knowledge integration) influence specifically each of
the three Lean routines: technical problem-solving hinges espe-
cially on data reliability, information transmission on real-time
information and conflict management on knowledge integration.
4.3. The effect of the organizational changes on Jealsa’s Finally, we describe how the new Lean-IT routines convert into
manufacturing strengths certain capabilities according to specific traits of (lean) manufactur-
ing environments: standardization, centralized decision-making or
Beyond the improvements observed in the different areas of the “management by stress”. On the other hand, the case study also
plant, this last section aims to determine whether the increase by shows that the rhythm and sequence of Lean-IT implementations
new Lean-IT routines in these ‘local’ indicators really meant the determine the timing of conversion of new routines into capabili-
generation of manufacturing strengths according to project goals. ties. We describe below how the case study results give rise to four
Based on Hill et al. [47], we summarize these initiatives to address propositions.
the specific traits of manufacturing environments in four specific
fields: (1) Efficiency, measured in terms of Plant Labor Productivity 5.1. The mutations caused by IT
(total sales in millions of euros/total shopfloor workers), (2) Qual-
ity Improvement in terms of First-Time-Yield (FTY), i.e., good units The evolutionary framework proposed suggests that while IT
produced/total units in millions, (3) Innovation (no. of patents and projects may all be intentional, the origin of the mutations caused
utility models/year), and finally, (4) Customer satisfaction (no. of by IT can be interpreted from a Lamarckian or Darwinist perspec-
product complaints/total units sold in millions). tive. In the basis of the ERP implementation, for example, there
Fig. 4 shows that the only growing factors are related to variables was a clearly planned intention on the part of the top manage-
associated with efficiency and continuous improvement. On the ment team to face the changes they perceived in the environment.
one hand, Plant Labor productivity increased 68% (from 0.25 to 0.42) However, other factors, such as the hiring of unexpectedly quali-
for the period 2003-2013. On the other hand, Quality Improve- fied human capital in IT, were a clear example of random triggers of
ment —measured by FTY— ranged from 99.85% to 99.94%, which in ‘IT mutations’. Planned or random origins of mutations are not the
terms of non-conformity units entails an improvement of 60% (from only possibilities, however; the case study also shows that muta-
1,510 nonconformities per million in 2003, to 698 in 2013). Fig. 4 tions can follow the patterns of an emergent system. With ERP-MES
includes these values in parentheses and Table A1 in the Appendix integration, for example, we observed that there may be certain
A describes the calculations. Conversely, as can also be seen in Fig. 4, mutations resulting from the self-organization of manufacturing
other possible manufacturing strengths hardly changed (Innovation, systems without a central planning authority [55].
0%), or even showed a negative trend (Customer satisfaction, −14%). Whatever their origin, the specific mutations that prevail are
Clearly, the influence of the project team (mostly process- those that appear to be adaptive over time. As explained in the
oriented) and, especially, the autocratic leadership of the Industrial Section 2.2, Evolutionary Theory of the Firm stresses the word
Manager, who was particularly interested in improving the effi- ‘adaptive’ rather than ‘efficiency’ because, just as biological evolu-
ciency of “his” plant, entailed a clear bias toward cost saving tion cannot guarantee ‘fair’ results, the evolution of routines does
and process improvement to the detriment of Innovation or Cus- not necessarily reflect optimal results in terms of efficiency. Thus,
tomer Satisfaction. Even the human capital skills of the Lean-IT a new organizational routine is adaptive as long as it gives rise to a
project coordinator or the suppliers of IT tools, focusing on process capability that increases survival against the selection mechanism
engineering, also explain why priority was placed on continuous of the environment in which it operates [24,31,28]. It is not easy
improvement indicators. In fact, the design and development of the therefore to anticipate a priori that a particular quality is adaptive
ERP-MES platform can be interpreted as an attempt by the organi- per se; it all depends on the capabilities generated, on the differ-
zation to create a tool designed exclusively to improve operational ent selection mechanisms, and the evolution of the environment
efficiency ratios (Cost and Quality). for which they were designed. In fact, the Jealsa experience shows
 A. Sartal, X.H. Vázquez / Journal of Manufacturing Systems 45 (2017) 1–16 11

Fig. 4. Evolution of the main industrial performance indicators (2003–2013).

Fig. 5. Evolutionary interpretation of the case study.

Table 6
Evolutionary rationalization of the case study results.

Lean-IT Project stages Stylized facts (from case study) An evolutionary rationalization of the facts

1. The origin of mutations: Acquisition of IT - Planned acquisition of ERP by IT manager P1: Random and emergent IT events can
(ERP and MES) within a Lean environment (semi-structured interviews). provoke mutations on Operational Excellence
- MES implementation, stemming from the Lean project routines that are as adaptive as those with a
manager (semi-structured interviews). planned origin.
- ERP and MES integration through ex post synergies
(semi-structured interviews).

2. Mutations in organizational routines - Data reliability enhances “0 defects” in information P2: Each of the IT-mutated routines seize in
provoked by IT:Influence of IT on Lean routines process (interviews & 1st quasi-experiment: manual different IT traits: technical problem-solving
(JIT, Jidoka and RfP) recording vs. AIDC). hinges specially on data reliability, information
- Real-time information improves flows in JIT systems transmission on real-time information and
(interviews & 2nd quasi-experiment: “e-kanban” vs. conflict management on knowledge
conventional kanban). integration.
- Knowledge Integration stimulates the reduction of
information asymmetries (interviews & 3rd
quasi-experiment: IT vs. non-IT, Toyota’s 7-step
problem-solving method).

3. Conversion of Lean-IT mutations into - The impact of organizational changes (new Lean-IT P3: The specific traits of manufacturing
specific static and dynamic capabilities: How routines) on creation of capabilities: cost, quality, environments make groupthink a key
the rhythm and sequence of reforms influence innovation and customer satisfaction (time series data for determinant of how Lean-IT routines convert
timing of transformation different operational performance indicators from into specific static and dynamic capabilities.
2003–2013). P4: The rhythm and sequence of Lean-IT
implementations determines the timing of the
conversion of new routines into capabilities.
12 A. Sartal, X.H. Vázquez / Journal of Manufacturing Systems 45 (2017) 1–16

Table 7
Generating new capabilities and latent capabilities in Jealsa.

Static capabilities Dynamic capabilities

New capabilities (FULFILLED) • Labor productivity • Quality improvement

Latent capabilities (NOT FULFILLED) • Customer satisfaction • Innovation

Note: See Fig. 4 with the evolution of the main industrial performance indicators (2003–2013).

that new lean-IT routines raised from random or emerging IT muta- distribution of information among all members of the organization,
tions (the three quasi-experiments are based on MES or ERP-MES avoiding the type of conflict that arises where there is asymmetric
initiatives) can significantly improve plant indicators. However, we access to information and, in addition, allowed for the merger of the
also noted that the specific traits of manufacturing environments industrial and business areas, which had traditionally been sepa-
will actually define which capabilities arise (Table 7). rate. On the other hand, the adoption of certain IT solutions on the
A comprehensive perspective of the case study suggests, there- shopfloor (e.g. digital screens and personal RFID cards) increased
fore, that while there may be different triggers of mutation, once the transparency of processes in the plant. This helped managers
the routines have been mutated, their adaptability does not depend supervise tasks as an initiative inherent in the job, avoiding the
on whether their origin was planned, random or emergent. Both typical conflicts that can arise from a direct supervision. Thus, in
random and emergent mutations show that there can exist syner- view of these three issues addressed in the quasi-experiments, we
gies in the joint deployment of IT and lean routines even without propose the following proposition:
an initial specific direction or time frame. Against this background,
we propose our first proposition: P2. Each of the IT-mutated routines seize in different IT traits:
technical problem-solving hinges specially on data reliability,
P1. Random and emergent IT events can provoke mutations on
information transmission on real-time information and conflict
Operational Excellence routines that are as adaptive as those with
management on knowledge integration.
a planned origin.

5.2. Mutations in each of the organizational routines provoked by

IT-features: data reliability, real time information and knowledge 5.3. Conversion of IT mutations into specific static and dynamic
integration capabilities

The Jealsa case shows that IT is an important factor for mutation The new Lean-IT routines in Jealsa resulted primarily in better
in the three groups of routines we have identified as bastions of static and dynamic capabilities related to efficiency and continuous
OE: a) the solving of technical problems, b) the processing of infor- improvement. Capabilities such as Plant Labor productivity (+68%)
mation assets, and c) the management of conflict. Based on the −static− or Quality Improvement (+60%) −dynamic− emerged,
three quasi-experiments designed, our study explores how each of while others such as Customer satisfaction (−14%) or Innovation
the different IT features (data reliability, real-time information and (0%) remained dormant (Table 7).
knowledge integration) mainly affect each type of routine, giving Out interpretation is that both the composition of the project
rise to new Lean-IT routines (Jidoka-IT, JIT-IT and RfP-IT) (Table 2). team (mostly manufacturing personnel) and the leadership of the
Firstly, thanks to the 1 st quasi-experiment, we observed that industrial manager played a determinant role in these results. The
AIDC allowed reliable line data to be obtained, showing the true common interests shared by the majority of the Lean-IT project
availability of the facilities (Table 3). Previously, with manual data team (focused on cost reduction and continuous improvement)
recording, many micro-downtimes and line slowdowns passed stirred a long term coalition (Pfeffer, 1981; Pettigrew, 2001) that
unnoticed or were simply ignored. As a result, because the old avail- suffered from groupthink: a psychological phenomenon that occurs
ability ratios hid the real problems, there was little pressure for when a group values harmony and coherence over accurate anal-
improvement and, worse still, the available resources were usu- ysis and critical evaluation, which not only makes individuals
ally focused on other activities that were actually less critical. This unquestioningly follow the word of the leader, but also encourages
experiment allowed us to highlight the importance of data relia- conformity with the group mainstream ideology. In the particular
bility —one of the main contributions of IT— in problem-solving case of Jealsa, the coalition built mainly around the manufacturing
routines. According to the principles of continuous improvement area determined the project goals [56,57], leaning towards biased
and elimination of non-value added activities, Jidoka routines hinge initiatives that basically supported process optimization.
especially on data reliability, since “zero defects” is essential in Other equilibriums in typical dilemmas, such as cost leadership
information processes. vs. differentiation or process innovation vs. product innovation,
Similarly, in the 2nd quasi-experiment, e-kanban vs. conven- could have probably been achieved if the coalition changed to
tional kanban, we assess whether organizational techniques can, by include more individuals from outside the manufacturing area (for
themselves and in real time, manage the increasingly huge amount instance, R&D or marketing). What is relevant to observe in the light
of data required today in manufacturing systems. In this case, the of Jealsa’s experience, however, is that by contrast with other firm
e-kanban system provided the user —as one more JIT tool— with areas in which decentralization of responsibilities for day-to-day
the information needed “in the right format, at the right time and management is common, manufacturing environments show more
in the right place” [53], improving flows and quickly responding standardization of procedures and centralized decision-making
to internal fluctuations (Table 4). We can therefore conclude that [58,59]. This creates the perfect breeding ground for groupthink,
JIT routines hinge especially on real-time data, since immediacy is since bureaucratic procedures foster group isolation, whereas cen-
essential in the information processes. tralized decision-making encourages autocratic leadership [60].
Finally, the 3rd quasi-experiment shows that knowledge inte- Furthermore, the typical “management by stress” inherent to LM
gration supported by IT helped resolve information asymmetries [27] creates pressure to follow the leaders’ opinion and to conform
(vertical and horizontal) and, therefore, improved conflict man- to the majority view [61,62]. We therefore propose the following
agement. On the one hand, the ERP-MES integration led to a fairer proposition:
 A. Sartal, X.H. Vázquez / Journal of Manufacturing Systems 45 (2017) 1–16 13

P3. The specific traits of manufacturing environments make of Jealsa would probably have been transferred to the plant’s overall
groupthink a key determinant of how Lean-IT routines are con- indicators immediately and with a greater impact.
verted into specific static and dynamic capabilities. The debate, therefore, should not focus on Lean vs. IT but on
establishing ex ante the sequence and rhythm. Although simultane-
ity is positive in an initial stage because IT speeds up the appearance
5.4. How the rhythm and sequence of implementation influence of new routines and their impact on local indicators (I0 → I1 ), the
the timing of organizational transformation Jealsa experience shows that an under-developed lean environ-
ment, especially in ecosystems with a low level of organization,
The case study also shows that the rhythm and sequence of will lead to a sharp slowdown in the subsequent phase (I1 → I2 ).
implementation in the Lean-IT project will affect the timing of We therefore propose the following proposition:
transformations. Although there can be little doubt about the
capacity of IT to facilitate data management or to improve the qual- P4. The rhythm and sequence of Lean-IT implementations deter-
ity of communication in a firm, there is less certainty about how mines the timing of conversion of new routines into capabilities.
such benefits take place and at what rhythm and in what sequence
[14]. To unravel the factors involved in the joint deployment of IT 6. IMPLICATIONS FOR MANUFACTURING INDUSTRIES
and OE, it is insufficient to analyze the starting-point or the final
goal, we also need to observe the transformation process. For this Although the case study is addressed in the food and beverage
purpose, we drew up Fig. 6, in which, on a sequence of landmarks sector, the synergies and dysfunctions found between Information
related to the Lean-IT project (below), we show the evolution of Technologies (IT) and Operational Excellence (OE) offer valuable
the two capabilities that were built: Plant Labor Productivity (total lessons for other industries or types of processes, with particular
sales/total shopfloor workers) and Quality Improvement in terms of emphasis on lean practitioners and IT managers.
FTY, i.e., good units produced/total units in millions. To begin with, while real-time information management was
Considering three instants in time (I0 , I1 e I2 ) and with the important to enhance the peformance of LM in Jealsa’s Make-
results of the quasi-experiments, we suggest that while the new to-Stock (MTS) processes, it can become critical in the lean
Lean-IT routines (I0 → I1 ) arose rapidly in a matter of weeks, the transformation of Assemble-to-Order (ATO) or Make-to-Order
conversion of these mutations into capabilities (I1 → I2 ) took much (MTO) processes. In MTS environments, product variety is compar-
longer (years). In fact, it was only with the ‘mature’ lean project atively low, manufacturing is decoupled from customer orders and
in 2009–2012 that we observed real upturns in these two new demand is seasonable or easy to predict. However, MTO (e.g., air-
capabilities: Quality Improvement (+56%) in 2010 and Plant Labor craft, vessels) and ATO (e.g., automotive, computers) shop-floors
Productivity (+45%) in 2012. We examine this fact as well as the work often with a greater degree of product complexity (large
process of transformation in terms of sequence and rhythm. While variety of components and technical characteristics), manufactur-
sequence indicates the order or arrangement in which the lean and ing begins when the production order is launched, and demand
IT initiatives should take place, the rhythm reflects the intensity of volatility is higher. For instance, the extreme volatility in indus-
the different events. tries such as fast- fashion requires not only much faster processes,
Regarding the sequence, we note that when the lean project but above all a more flexible supply chain −beyond conventional
began to stagnate in 2005 because of increasing information lean supply– to assimilate the constant changes in demand [63,64].
requirements (I0 ), the simultaneous incorporation of IT became an Thus, the greater flexibility required seems to involve solutions that
important catalyst for both improving the results and shortening go beyond isolated LM initiatives. It is here that IT becomes crucial
the timing. These achievements, which point to a successful pro- for managing the huge amount of data, and enhance the absorptive
cess of mutation in organizational routines (I0 → I1 ), also helped capacity of the company to respond systematically to the variation
meet certain lean requirements such as standardization and level- in client requirements [33,65].
ing. However, in retrospect, it might have been more efficient to Secondly, we have observed that IT resources offer real value
initially focus efforts only on the lean project to avoid “automating when they are embedded within organisational processes. The
the Muda”. Simultaneous Lean-IT deployment, without a strategy Jealsa case suggests, in fact, that IT features enable lean principles
planned by the top management, gave rise on some occasions to ini- and tools to evolve and adapt successfully to any manufactur-
tiatives that were mistaken or produced few results. As discussed ing environment [33]. The case study shows that IT features
above, the attempts to adopt programs for improvement on all lines may constitute powerful mechanisms through which lean rou-
simultaneously and the sudden replacement of written records tines evolve and reinforce their contributions to manufacturing
with automatic capture are examples of initiatives that were not strengths (specifically productivity and quality improvement). The
as successful as expected. In the same way, other initiatives such second quasi-experiment shows, for instance, that classic lean tools
as improving productivity or rotating stock went practically unno- such as Kanban cards can become unmanageable nowadays if one’s
ticed until they were perfectly aligned with strategic objectives. goal is to work with real-time information. It is here that IT fea-
Something similar happened with regard to rhythm. From the tures appear to be essential for managing massive amounts of data
start, possibly because of the team’s professional bias, the priority on a daily basis. Different examples in the literature endorse this
was placed on IT to the detriment of lean initiatives in the belief that perspective. Chen [66] describes, for example, the emergence of
IT alone would solve all of the organizational problems. This was a web-based kanban system that, based on the traditional con-
perhaps the reason why the purchased tool went far beyond the ini- cept, expands monitoring and real-time performance indicators. In
tial needs and why many of its functionalities were only used years the same vein, different RFID-based applications (similar to those
later. The Jealsa experience suggests that at the start of the project, described in the case study), are being increasingly used in the dis-
a basic system offering reliable, real-time information would have tribution industry supporting JIT principles and tools in a faster and
been the best option. It would also have entailed a mutation fac- more accurate manner [67].
tor and would have directed the use of limited human resources Thirdly, another valuable lesson learned from Jealsaı́s experi-
toward the utilization of lean tools while postponing the adoption ence has to do with how manufacturing environments determine
of other IT functionalities (statistical process control, batch sys- the generation of specific static and dynamic capabilities. Partic-
tems, traceability, etc.) until they were really necessary. Without ularly, managers should take into account that the absence of
these miscalculations, the benefits obtained in the different areas groupthink facilitates an efficient decision-making process and
14 A. Sartal, X.H. Vázquez / Journal of Manufacturing Systems 45 (2017) 1–16

Fig. 6. Influence of rhythm and sequence in the transformation of Lean-IT routines into capabilities.
*Note: Data from the quasi-experiments (Section 4.2 “How IT affected organizational routines in Operational Excellence”).

goal congruence during the Lean-IT implementation. This fact has Firstly, building on an evolutionary framework, the paper high-
proved to be relevant in industrial environments, such as that of lights an aspect not previously addressed in the lean-IT literature:
Jealsa, with limited experience in LM. Since we justified that “man- the simultaneity of planned, random and emergent triggers in IT
agement by stress” puts pressure on operators to follow the leaders’ mutations. The possibility of carrying out the analysis from inside
opinion and to conform with the majority view, it may well become and for such a long period, allowed us to show that certain random
a more critical factor in more developed lean environments such as and emergent events complement planned behavior and actually
automotive, aerospace or machinery [18]. Our results suggest the have an important role in the transformation of organisational rou-
need for an early detection and correction of cognitive biases such tines. Although this does not contradict the idea that the integration
as groupthink, accordingly, to avoid automating Muda (non-value- of IT projects in manufacturing environments should be planned, it
added activities). indicates that random and emergent triggers can be as adaptive for
Finally, our findings also show that the rhythm and sequence of the implementation of Lean-IT projects as planned initiatives. The
Lean-IT implementation are two key issues for practitioners regard- fact that activities and processes are constantly being challenged
less of sector. If continuous improvement initiatives never end by unexpectedly, therefore, calls for the need of learning systems sup-
definition, then excessively early adoption of IT —as in the case ported by IT and organizational routines that push the flexibility of
discussed— may represent at best an investment with low returns manufacturing systems to yield a higher level of performance.
and a high opportunity cost. Hence, by contrast with authors such as In the same vein, the analysis of the Lean-IT process of trans-
Cottyn et al. [14], we posit that lean projects do not necessarily ben- formation, step-by-step and from an evolutionary point of view,
efit from the outset from IT support for data collection and analysis. allowed us to address how the different IT features (data reliability,
Even when static and dynamic capabilities are generated, IT muta- real-time information and knowledge integration) influence specif-
tions in Lean routines will not necessarily adapt to the selection ically each of the three Lean routines. Technical problem-solving
mechanisms in the environment. hinges especially on data reliability, information transmission on
real-time information and conflict management on knowledge
integration. Hence, based on three quasi-experiments carried out
7. CONCLUSIONS on Jealsa’s shop floor, our findings contribute to the LM literature by
qualitatively analyzing how each of the IT-mutated routines seize
This paper explores the synergies and dysfunctions between in different IT traits.
Information Technologies (IT) and Operational Excellence (OE). The Thirdly, the evidence obtained suggests how the specific
Lean-IT topic has received increasing attention by the scientific traits of manufacturing environments (standardization, centralized
community [33], and few doubts are raised today about the capacity decision-making or management by stress, amongst others) can be
of IT for facilitating data processing and analysis [13]. The litera- a key factor in the conversion of Lean-IT routines into specific static
ture shows some major and relevant research gaps, however. It is and dynamic capabilities. Specifically, our paper addresses qualita-
not clear, for instance, how IT benefits are produced. Thus, mov- tively the moderation role of groupthink between Lean-IT routines
ing away from the usual analyses that focus on the starting-point and specific static and dynamic capabilities. Based on time series
and the final stage, we address the evolution of the complex factors data for different plant performance indicators, the study reveals
involved in the joint deployment of IT and lean practices step-by- that groupthink in Jealsa biased the transformation of Lean-IT rou-
step, and by means of a longitudinal case study (2003–2014). We tines towards efficiency and continuous improvement capabilities
describe below the main findings of our research. (Plant Labor productivity and Quality Improvement), while other
 A. Sartal, X.H. Vázquez / Journal of Manufacturing Systems 45 (2017) 1–16 15

capabilities (such as customer satisfaction or innovation) remained [15] Forrester PL, Kazumi Shimizu U, Soriano-Meier H, Arturo Garza-Reyes J,
neglected. Fernando Cruz Basso L. Lean production, market share and value creation in
the agricultural machinery sector in Brazil. Journal of Manufacturing
Finally, the paper emphasizes that the full potential of IT can Technology Management 2010;21(7):853–71.
only be achieved when it is used to support (as a complementary [16] Womack JP, Jones DT, Ross D. The machine that changed the world. New
resource) organizational routines such as those linked to OE. Lean York: MacMillan/Rawson Associates; 1990.
[17] Pettersen J. Defining lean production: some conceptual and practical issues.
and IT are complementary, and while they are distinct, they are also The TQM Journal 2009;21(2):127–42.
interdependent. This is the reason why, rather than discussing at [18] Bhamu J, Singh Sangwan K. Lean Manufacturing: literature review and
this point in time whether LM and IT generate synergies, it is impor- research issues. International Journal of Operations & Production
Management 2014;34(7):876–940.
tant to delve for each firm into the rhythm and sequence of Lean-IT
[19] Womack JP, Jones DT. Lean Thinking: Banish Waste and Create Wealth in Your
implementations. The case study has showed, for instance, that an Corporation. New York: Simon & Schuster; 1996.
under-developed lean environment delays the realization of new [20] Shah R, Ward PT. Lean manufacturing: context, practice bundles, and
performance. Journal of Operations Management 2003;21(2):129–49.
capabilities, therefore slowing down or even hindering the pro-
[21] Womack J, Jones D. Lean Thinking. London: Simon & Schuster; 2003.
cess of Lean-IT transformation. From a broader perspective, Jealsa’s [22] Spear S, Bowen HK. Decoding the DNA of the Toyota Production System.
experience suggests that the deployment of IT in a firm is not just Harvard Business Review 1999:97–106.
a technological endeavor [10,33]. It is above all an organizational [23] Emiliani B. The Equally Important Respect for People Principle. Real Lean: The
Keys to Sustaining Lean Management, vol. 3. Wethersfied: The Center for Lean
challenge that determines the timing of conversion of new routines Business Management; 2008.
into capabilities. [24] Nelson RR, Winter SG. An evolutionary theory of economic behavior and
capabilities. Cambridge: Harvard University Press; 1982. p. 195–307.
[25] Coriat B, Dosi G. Learning how to govern and learning how to solve problems:
Appendix A. on the co-evolution of competences, conflicts, and organizational routines.
CNR Press; 1998.

Table A1
Evolution of non-conforming units, good units and annual FTY ratio.

Years 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013

Non-conforming units per million 1,544 1,480 1,400 1,480 1,320 1,360 880 600 640 592 610
Good units per million 998.456 998,520 998,600 998,520 998,680 998,640 999,120 999,400 999,360 999.408 999,390
First-Time-Yield (FTY)a 99.85% 99.85% 99.86% 99.85% 99.87% 99.86% 99.91% 99.94% 99.94% 99.94% 99.94%
a
Percentage of good units produced per million.
[26] Coriat B. The abominable Ohno Production System. Competences, monitoring,
and routines in Japanese production systems. In: The nature and dynamics of
organizational capabilities; 2000. p. 213–43.
Appendix B. Supplementary data [27] Fujimoto T. The evolution of a manufacturing system at Toyota. Oxford
University Press; 1999.
Supplementary data associated with this article can be found, [28] Hodgson GM, Knudsen T. Why we need a generalized Darwinism, and why
generalized Darwinism is not enough. Journal of Economic Behavior &
in the online version, at http://dx.doi.org/10.1016/j.jmsy.2017.07. Organization 2006;61(1):1–19.
007. [29] Dosi G, Marengo L. Some Elements of an Evolutionary Theory of
Organizational Competences. In: England RW, editor. Evolutionary Concepts
in Contemporary Economics. 1994. p. 157–78. USA.
References [30] Teece DJ, Pisano G, Shuen A. Dynamic capabilities and strategic management.
Strategic Management Journal 1997;18(7):509–33.
[1] Sugimori Y, Kusunoki K, Cho F, Uchikawa S. TPS and kanban system [31] Hodgson GM. Evolutionary and competence-based theories of the firm.
materialization of JIT and respect-for-human system. Int. Journal of Journal of Economic Studies 1998;25(1):25–56.
Production Research 1977;15(6):553–64. [32] Dubey R, Gunasekaran A, Childe SJ. The design of a responsive sustainable
[2] Ohno T. Toyota production system: beyond large-scale production. supply chain network under uncertainty. The International Journal of
Productivity Press; 1988. Advanced Manufacturing Technology 2015;80(1-4):427–45.
[3] Toyoda E. Creativity, Challenge and Courage. Toyota Motor Corporation; 1983. [33] Ghobakhloo M, Hong TS. IT investments and business performance
[4] Esmaeilian B, Behdad S, Wang B. The evolution and future of manufacturing: a improvement: the mediating role of Lean Manufacturing implementation.
review. Journal of Manufacturing Systems 2016;39:79–100. International Journal of Production Research 2014;52(18):5367–84.
[5] Liker JK. The Toyota way. NY: McGraw-Hill; 2004. [34] Grant RM. Prospering in dynamically-competitive environments:
[6] Riezebos J, Klingenberg W, Hicks C. Lean Production and information Organizational capability as knowledge integration. Organization science
technology: connection or contradiction? Computers in Industry 1996;7(4):375–87.
2009;60(4):237–47. [35] Eisenhardt KM, Martin JA. Dynamic capabilities: what are they? Strategic
[7] Bell SC, Orzen MA. Lean IT: Enabling and sustaining your lean transformation. management journal 2000;21(10-11):1105–21.
CRC Press; 2016. [36] Gummesson E. Qualitative methods in management research. Sage; 2000.
[8] Ward P, Zhou H. Impact of information technology integration and [37] Strauss AL. Qualitative analysis for social scientists. Cambridge University
Lean/just-in-time practices on lead-time performance. Decision Sciences Press; 1987.
2006;37(2):177–203. [38] Siggelkow N. Persuasion with case studies. Academy of Management Journal
[9] Shah R, Ward PT. Defining and developing measures of lean production. 2007;50(1):20–4.
Journal of Operations Management 2007;25(4):785–805. [39] Sutton RI. Crossroads-the virtues of closet qualitative research. Organization
[10] Moyano-Fuentes J, Martínez Jurado PJ, Maqueira Marín JM, Bruque Cámara S. Science 1997;8(1):97–106.
El papel de las TIC en la búsqueda de la eficiencia: un análisis desde Lean [40] Halinen A, Medlin CJ, Törnroos J. Time and process in business network
Production y la integración electrónica de la cadena de suministro. CEDE research. Industrial Marketing Management 2012;41(2):215–23.
2012;15(3):105–16. [41] Yin RK. Case study research: Design and methods. Applied social research
[11] Ghobakhloo M, Tang SH, Sabouri MS, Zulkifli N. The Impact of Information methods series, 5. Sage Publications Inc; 2003.
System-Enabled Supply Chain Process Integration on Business Performance: [42] Eisenhardt KM, Graebner ME. Theory building from cases: opportunities and
A Resource-Based Analysis. Int. Journal of Information Technology & Decision challenges. Academy of Management Journal 2007;50(1):25–32.
Making 2013;13(05):1075–113. [43] Pratt MG. For the lack of a boilerplate: tips on writing up (and reviewing)
[12] Deveraj S, Krajewski L, Wei JC. Impact of e-business technologies on qualitative research. Academy of Management Journal 2009;52:856–62.
operational performance: the role of product information integration in the [44] Eisenhardt KM. Making fast strategic decisions in high-velocity environments.
supply chain. Journal of Operations Management 2007;25:1199–216. Academy of Management Journal 1989;32(3):543–76.
[13] Heim G, Peng D. The impact of information technology use on plant structure, [45] Moschella DC. Waves of Power: The Dynamics of Global Technology
practices, and performance: an Exploratory study. Journal of Operations Leadership, 1964-2010. American Management Assoc., Inc; 1997.
Management 2010;28(2):144–62. [46] Dewett T, Jones GR. The role of information technology in the organization: a
[14] Cottyn J, Van Landeghem H, Stockman K, Derammelaere S. A method to align review, model, and assessment. Journal of Management 2001;27(3):313–46.
a manufacturing execution system with Lean objectives. International Journal [47] Hill C, Jones G, Schilling M. Strategic Management. Theory: an integrated
of Production Research 2011;49(14):4397–413. approach. Cengage Learning; 2014.
16 A. Sartal, X.H. Vázquez / Journal of Manufacturing Systems 45 (2017) 1–16

[48] Fernández-Olmos M, Gargallo-Castel A, Giner-Bagües E. Internationalisation [59] Vajjhala NR. Communities of practice in transition economies: Innovation in
and performance in Spanish family SMES: the W-curve. BRQ Business small-and medium-sized enterprises. In: Organizational Culture and
Research Quarterly 2016;19(2):122–36. Behavior: Concepts, Methodologies, Tools, and Applications. IGI Global; 2017.
[49] Farinas JC, López A, Martín-Marcos A. Sourcing strategies and productivity: p. 134–48.
evidence for Spanish manufacturing firms. BRQ Business Research Quarterly [60] Janis IL. Groupthink: Psychological studies of policy decisions and fiascoes,
2016;19(2):90–106. vol. 349. Boston: Houghton Mifflin; 1982.
[50] Olhager J, Selldin E. Enterprise resource planning survey of Swedish [61] Horrobin DF. The philosophical basis of peer review and the suppression of
manufacturing firms. European Journal of Operational Research innovation. Journal of the American Medical Association
2003;146(2):365–73. 1990;263(10):1438–41.
[51] Muchiri P, Pintelon L. Performance measurement using overall equipment [62] Turner ME, Pratkanis AR. Twenty-five years of groupthink theory and
effectiveness (OEE): literature review and practical application discussion. Int research: lessons from the evaluation of a theory. Organizational behavior
Journal of Production Research 2008;46(13):3517–35. and human decision processes 1998;73(2-3):105–15.
[52] Sartal A, Armesto JI, López-González MF. OEE: una propuesta para definir la [63] Vázquez XH, Sartal A, Lozano-Lozano LM. Watch the working capital of
productividad real de la planta. Automática e Instrumentación 2008;396:76. tier-two suppliers: a financial perspective of supply chain collaboration in the
[53] Hicks B. Lean information management: understanding and eliminating automotive industry. Supply Chain Management: An International Journal
waste. International Journal of Information Management 2016;21(3):321–33.
2007;27(1-2):233–49. [64] Sartal A, Martínez-Senra AI, García-Vázquez JM. Balancing offshoring and
[54] Diaz-Fernandez M, Pasamar-Reyes S, Valle-Cabrera R. Human capital and agility in apparel industry: lessons from Benetton and Inditex. FIBRES &
human resource management to achieve ambidextrous learning: a structural TEXTILES in Eastern Europe 2017;25(2):16–23.
perspective. BRQ Business Research Quarterly 2017;20(1):63–77. [65] Martínez-Senra AI, Quintás MA, Sartal A, Vázquez XH. ¿Es rentable «pensar
[55] Fujimoto T. Evolution of Manufacturing Systems and Ex-post Dynamic por pensar»? Evidencia sobre innovación en España. Cuadernos de Economía
Capabilities. In: Dosi G, Nelson RR, Winter SG, editors. The Nature and y Dirección de la Empresa 2013;16(2):142–53.
Dynamics of Org. Capabilities. 2000. p. 244–80. [66] Chen FF. A Web-based Kanban system for job dispatching, tracking, and
[56] Nadler DA, Tushman ML. A model for diagnosing organizational behavior. performance monitoring. The International Journal of Advanced
Organizational Dynamics 1980;9(2):35–51. Manufacturing Technology 2008;38(9-10):995–1005.
[57] Kakabadse A, Parker C. Power, politics, and organizations: A behavioural [67] Chen JC, Cheng CH, Huang PB, Wang KJ, Huang CJ, Ting TC. Warehouse
science view. John Wiley & Sons Inc; 1984. management with lean and RFID application: a case study. The International
[58] Shalley CE, Gilson LL. Creativity and the management of technology: Journal of Advanced Manufacturing Technology 2013;69(1-4):531–42.
balancing creativity and standardization. Production and Operations [69] Monden Y. Toyota production system: practical approach to production
Management 2017;26(4):605–16. management. Industrial Engineering & Management Press; 1983.