Professional Documents
Culture Documents
Patrick DUPIN
Advisors:
Pr Christine PASQUIRE
Dr Roy STRATTON
Autumn 2016
ABSTRACT Commenté [SR1]: It helps to use a stucturd abstract
The purpose of this research is
The purpose of this research is to improve production flow stability of real estate development sites by The method/approach
complementing the existing planning systems. The research hence investigates the application of pull The findings
systems from the manufacturing industries that have not been fully tested in the construction industry. The impact
Limitations
An action circles approach has been used to conduct this research. Two action circles have been conducted
so far, the third has been initiated. The method used in action circle is an ascending spiral that consists of
four major phrases: ‘planning, acting, observing and reflecting’, Zuber-Skerrit(1991). The initial cycle of
these four activities lead to the next cycle in which the reflections of the previous cycle inform the plan of
the next. The cyclic process alternates between action and critical reflection (Dick, 2002). As the cycles
progress a greater understanding is developed through the continuous refining of methods, data and
interpretation (Dick, 2002). A three block real estate development project in Brussels has been used to
ground the research. The sequential development of the three developments blocks, of similar techniques
and made by the same teams, eased the process of capturing the learning from one phase another and
building up an optimized scheduling process. Default of works visibility has been found as a major issue
when the research started on the ongoing works of the first block. “5S” and Last planner® System were
tested in the as a mean to improve works visibility from both physical and managerial facets.
On site measurements showed that despite a high Last Planner® PPC1 the weekly individual progresses of
each apartment did not follow a smooth line. The works were still highly volatile and unstable. It has been
found that the volatility came from a lack visibility in the scheduling system.
Despite that the progresses of each week were determined by what and should be done based on the Last
Planner® System, a deeper structuration of the scheduling system and a more stable forecast of the works
were needed to help stabilize the works. Given that takt time has been successfully used in the
manufacturing industry to address variability in the demand and that large similarities exist between a
construction site and a manufacturing line, it has been decided to test the application of takt time as a
complement of the Last Planner® System in the second action of the second bloc. Encouraging results were
measured, on site works were smoothened in the second block.
It was found that the discipline to respect the system, the rules and sequence of the works demanded a high
level and constant surveillance from the site management. In the long run, this could put the whole system
at risk. In order to be as efficient, sustainable and duplicable as possible, the scheduling system should be
visual, need no interpretation and lead works in a pull flow. Manufacturing successfully used a simple but
efficient system named Kanban to level production and ensure a fully pulled system in a variable demand
environment while limiting the sources of errors. It is currently tested in action research of the third bloc
that Kanban can improve the scheduling system developed in the first two blocs by favoring a self-pull
system.
The impact of the scheduling system developed by the action circles in block 2 has been a reduction of
nearly 20% of the lead time measured and a significant increase of quality (number of snagging works
decreased 95%). As a consequence the stress on the site management team mechanically decreased
(captured by interviews).
The early findings and encouraging results are signs of the importance of the research undertaken.
Limitations
The research has been conducted in sites of highly classical typology, size, techniques, management
structure, and contractors so the findings are addressed to a large audiauce from the construction industry.
The limitation of the research comes from the domain of the construction observed: real estate in Belgium.
Despite that it is highly probable that the issued detected on site by the research and the mechanisms used
to address those issued are duplicable on other domains of the construction industry, there is, at this point of
the research, no clear evidence.
1
Percentage of Promise Completed, by the foremen on site from one week another.
Page 2 / 86
TABLE OF CONTENT
ABSTRACT .............................................................................................................................. 2
INTRODUCTION ....................................................................................................................5
1.1 BACKGROUND ............................................................................................................................... 5
1.1.1 Complex environment. ................................................................................................. 5
1.1.2 Low Productivity in the construction industry. .......................................................... 5
1.2 RATIONALE .................................................................................................................................... 7
1.3 RESEARCH AIM, OBJECTIVES AND QUESTIONS .................................................................... 8
Page 3 / 86
3.1.2 Determining the type of Case Study .......................................................................... 37
3.1.3 How to generalize the findings? ................................................................................ 38
3.2 Action Research (differentiated from Consultancy) ........................................................................ 33
3.3 Proposed methodology: Experimental Multiple Case Study from Action Research........................ 47
REFERENCES .......................................................................................................................75
Page 4 / 86
1 INTRODUCTION
1.1 BACKGROUND
As Ballard describes in 1993 after he has been working closely with several contractors during
the construction phase of EPC industrial projects, disruption and waste caused by poor quality
and delivery of offsite resources, specifically design documents and permanent plant materials
and equipment are common. Ballard goes further and states that the construction industry is in
some ways not completely shifted from craft to mass production- much less to lean production.
He argues that the absence of industrial engineers from project sites and the lack of standard
work methods is one sign of the dominance of the craft production model--no little assisted by
the contention that each facility is unique. What Ballard was describing in 1993, at the very
initial stages of the birth of Lean Construction, is still widely valid in most of the construction
sites, the real estate development site used to conduct this research made no exception.
The research conducted here ambitions to improve project management of real estate
developments by providing a more robust system of works coordination with the eventual aim
to achieve “on time” delivery projects.
While chaos and desynchronization can be found at all phases of a construction phase
(Engineering/Design, Procurement, Construction) and desynchronization gets amplified from
the previous stages (domino effect), the research investigates the ultimate process of
coordinating the works on site. Any construction project needs planning before, during and
sometimes after construction. Planning is also a complex process that involves coordination
between many stakeholders. The complexity of construction projects has constantly increased in
ages with the use of new materials, new trades, new owners and environmental requirements,
new technologies… such as the number of parties involved (Consultants, Quantity Surveyors,
Contracts Administrators, Planners, Specialists…).
1.1.2 Motivation of the research: Low Productivity in the construction industry. Commenté [SR2]: The figure needs explaining especially your
addition.
Meanwhile production cycles have been compressed in many industries sectors that have
Commenté [PDN3]: completed
leveraged technology in unprecedented ways compared to the Architectural Engineering
Construction (AEC) industry. Studies in the past have indicated productivity and performance
of the AEC industry performance is poor compared to that of manufacturing and or other
industries (Teicholz, 2013). In his 2013 book “Labor-Productivity Declines in the Construction
Industry: Causes and Remedies (Another Look)” (page 12) Teicholz proposes the following
graph. It describes the evolution of value added per employee in various sectors from 1950 to
2010 and illustrates the productivity gap that the construction industry created in comparison to
other industries. At a closer look, we notice that, over ages, value added per employee has even
decreased in the construction industry.
Page 5 / 86
Fig 1.1. Real Productivity by industry in the US (value added by employee). Teicholz, 2013.
The problem of the current planning systems as currently designed is that they do not shield
direct production from upstream variation and uncertainty. The result is longer durations and
higher costs (Ballard 1993).
Also, high level trade specialization eased work flexibility and benefits to the industry but also
resulted in high costs in the form of fragmented decision-making (Howard et al. 1989). More
and more technologically complex environment requires specialized skills within the specialty
contractors. Meanwhile, general contractors have adopted a contract-brokering role rather than
the role of production coordination (Tommelein and Ballard 1997a) that has exacerbated the
fragmentation among parties and the tensions in there relations.
The above explains why coordinating the different trades and specialists to deliver a project “on
time” has become a difficult challenge. People and systems need to work in a synergistic and
self-reinforcing fashion to achieve true coordination (Collaborative Process Institute, 1997).
Developments of internet and IT tools were envisaged as magic solutions to address
coordination. In search for promoting team-work relationships and improving communication
and coordination, partnering sessions or similar team-building programs have arose in the
early 2000: ProjectNet, Team Builder, ProjectPoint, Constructw@re, ProjectTalk… but these
programs, despite huge promises, have only provided a partial solution. O’Brien (2000a) states
that processes of all participants need to be investigated in order to best design and use these
tools to support daily operations. Coordination at the production level needs detailed
information about all aspects of the site requirements (logistics, assignments, specification of
labor, equipment, materials, space use…).
Also, the need for coordination is twofold: first in physical production resources, e.g., labor,
equipment, material, and space (where material should not only be read as raw material but also
as work-in-process) and second in information. Architects and Engineers now use powerful
information treatment system to draw, add specifications, and realize shop drawings in form of
3D models that clearly show the final product. Construction Managers and Site Foremen
usually use the elements they have on hand (the ones they were given, not necessarily the ones
they need) to erect the product in the right place, in the right sequence and at the right pace of
works but there is a large gap between the way a site is organized and the way the design and
procurement phases are prepared, coordinated and monitored.
Page 6 / 86
Improving construction planning requires a clear understanding of the limits of the current
practice at different levels. In a classical system a Project Planner builds a project planning by
determining an array of works sequences, tasks durations and interdependencies that fit the
Client requirements. This document is then sent to the contractors who must respect it.
Despite that the planning made as such could be advocated about its theoretical validity, its
actual ability to keep the budget and delay asked by the Client can be questioned. When delays
occur, the immediate answer stands in the load of extra resources which create many peaks on
site and unplanned reallocation of resources causing an imbalance of resources (Kenley et al.,
2009). This imbalance fathers more works desynchronizations.
A planning alone is hence not sufficient to successfully manage a complex construction project.
Thus, it is important to review the current role project management assumes during a
construction project.
The traditional approach to production planning is twofold. The works are planned at the site
level by the Site Manager who organizes and coordinate the works on site to provide resources
where “appropriate”. The relevance of the site organization can be questioned as mainly based
on experience at this level. At the SME level, the production planning should also meet an
internal constraint: resource allocation. Any company has limited resources and the ones
available are allocated to the most urgent tasks (based on the Site Managers’ requests).
distributed in the different sites that the company has contracted. Then, if we consider that the
skills in a company are many, such as the ones to erect a building, the likelihood of getting the
right resource at the right location at the right time (not to mention the right cost) is very
limited. This difficulty fathers tensions from both the Site Manager who must organize the
construction and the companies that must provide the ad hoc resources, that does not go in favor
of sound site planning.
Whereas the works on site should obey a logical and robust sequence between trade to
maximize efficiency, if no Planner or Trade Coordinator is involved, works sequencing is
mainly made “by experience” upon the situation as observed on site (Choo, 2003).
It is often argued that the flow principles and theories developed in the manufacturing industry
cannot be imported in the construction industry that produces only prototypes with no Commenté [SR4]: I now get a hint at where you are going but it
repetitively in the processes: each building being different from the next while series production is still not clear.
is widely used in the manufacturing industry (Takim, Akintoye and Kelly, 2003). The author Commenté [PDN5]: Now?
claims and supports the contrary, that the construction industry is a highly repetitive process.
The research conducted here is grounded in previous works and experiences by Lean
Construction pioneers who successfully developed methods directly inspired from flow systems
developed by and for the manufacturing industry. The most notable is Ballard (2000) who
developed a new way of approaching scheduling, by tackling flow reliability through human
commitment rather that by complex computed tools, namely the Last Planner® System.
The research investigates further in practice how flow systems can be imported into the
construction industry and contribute to close the productivity gap with the manufacturing
industry.
1.2 RATIONALE
The construction industry environment evolves rapidly in search for operational performance:
use of new IT tools on site (eg iPad and increased virtual reality), 7D modelling drawings, new
materials (eg high performance concrete) and technics (eg 3D printing). Hence, erecting a
building in a more and more complex environment needs more and more parties involved and
flow management became after years a key success factor for on time and on budget delivery
(Suermann,2009)
The argument developed in this report is that on site productivity can be increased (ie decrease
Page 7 / 86
lead time and costs) by creating a continuous and stable workflow. As Olawale, Y., and Sun M.
(2010) point, during the last few decades, numerous project control methods, such as Gantt Bar
Chart, Program Evaluation and Review Technique (PERT) and Critical Path Method (CPM),
have been developed (Nicholas 2001, Lester 2000). A variety of software packages have
become available to support the application of these project control methods, for example
Microsoft Project, Asta Power Project, Primavera, etc. Despite the wide use of these methods
and software packages in practice, many construction projects still suffer time and cost
overruns. Takim, R, Akintoye, A and Kelly, J (2003) recalls that performance measurement is
the regular collecting and reporting of information about the inputs, efficiency and effectiveness
of construction projects. They further develop by arguing that performance measurement is
made by those who organize the works to judge their project performances, both in terms of the
financial and non-financial aspects and to compare and contrast the performance with others,
with the aim to improve program efficiency and effectiveness in their organizations. Moreover,
according to Steven et al. (1996), measurements are needed to tract, forecast, and ultimately
control those variables and this has been agreed by many researchers and practitioners (Sinclair
and Zairi, 1995; Mbugua et al., 1999; Love et al., 2000 and Chan, 2001). Takim, R, Akintoye,
A and Kelly, J (2003) further argue that these classical indicators give a limited view on the
actual situation as they narrow the investigation scope too immediate and restricted measures.
Hence, the argument developed here is that, despite being a fundamental data (as a source of
other indicators’ performance) workflow reliability is rarely measured on a construction site;
the main measures used to monitor performance being the physical progresses and financial
expenses at a certain date.
Page 8 / 86
not permit any fat. From the 1920’s to the 1990’s, both approaches and systems have been
successfully developed in parallel. Line of balance, Material Requirement Planning, Total
Quality Management, 6 Sigma, ISO… have been developed in America. Autonomation, Just In
Time, Poke Yoke, Kanban, production level, 5S... have meanwhile been developed in Japan.
This shows that the manufacturing industry has spent large efforts in the development of its
operating and managerial tool to address an unstable demand.
As argued and later developed, large similarities can be found between the construction industry
and the manufacturing industry. Similarly to the manufacturing industry, the construction
industry must also cope with a variable demand (each building is different) and is based on a
sequenced operating system. Therefore, in both industries, managing the operating flow is an
important stake in search for productivity.
However, while many approaches have been developed from both sides of the Pacific in the
manufacturing industry as seen above, the construction industry lacks such developments. Gant
charts and Critical Path Method date from the 1910’s and 1930’s and are still widely, not to say
almost exclusively, currently in use in the construction industry. The recent emergence of the
use of Critical Chain Method and the development of Last Planner ® System in the 1990’s gave
new perspectives. Nevertheless, a long path remains in the construction industry to catch up the
manufacturing industry.
For Ballard (2005) flow reliability is the proper concept and reliable work flow impacts the
productivity of downstream players. Ballard further develops and argues that understanding and
managing the construction process as a flow have been a key issue for IGLC since its very first
meeting in 1993. Shingo (1989) and the Toyota Production System (Shiomi and Wada, 1995; Commenté [SR11]: You are not being inspired to write a novel!
Ohno, 1988) along with its Western interpretation as Lean Production (Womack et al, 1990; – more objectivity.
Womack and Jones, 1996) and even earlier sources such as Aristotle (app. 330 B.C.) Koskela Commenté [PDN12]: Erased
and Kagligou (2005 and 2006) have been influencing the IGLC thinking.
This chapter helps the reader understand the developments of flow management systems, the
gap between both construction and manufacturing industries and how new management
approaches of flow can be further developed in the first based on existing systems in the
second.
Walter Shewhart invented the process control chart (1924), a modern study at this time that
Shewhart's further expanded out from industrial quality to wider concerns in science and
statistical inference to later (1939) issue his new book “Statistical Methods from the Viewpoint
of Quality Control”. This book introduces the notion of a continuous quality improvement
cycle: Plan-Do-Study-Act, more widely known from its refining alteration (Deming 1939) Plan-
Do-Check-Act. This continuous improvement often represented by a wheel is still extensively
Page 9 / 86
used in the current days. Commenté [SR13]: OK but why is this relevant?
Commenté [PDN14]: Repositionned
The focus of Frank G. Woollard 1954 book “Principles of Mass and Flow Production” was
already on achieving flow in processes upstream.
“the virtue of flow production lies in the fact that it brings all inconsistencies into the light of
day and so provides the opportunity for correcting them,” and “[the] high visibility conferred on
the company’s activities by flow production will lead to unceasing and continuous
improvement” (Woollard, 1954, p. 87)
The goal of production scheduling is to strike a profitable balance among conflicting objectives
(Hopp and Spearman 2008). When dealing with controlling the production to achieve profit and
profitability through meeting the demand, two approaches apply. A push based and a pull
based. In a puch system, a part is released in the production process when asked by an
exogenous schedule, independently of what is happening in the system itself. In a pull system, a
part is only released when a signal is generated as an indication to proceed. This signal, flag,
card or any type of upstream information is called by a change in the line status.
According to Hopp and Spearman (2008), the distinguishing feature between push and pull
systems is how the movement of work is triggered. In a push system, work orders are scheduled
based on actual or forecasted demand by a central system. In a pull system, work is authorized
based on the current system status.
Push Pull
Exogenous
(schedule) Endogenous
(stock)
Production process
Job
Production process
Job
Fig. 2.1 Push and pull mechanics (Adapted from Hopp and Spearman 2008)
Hopp and Spearman (2008) refine this definition based on the distinguishing effect of the two
principles and state “A pull system establishes an a priori limit on work-in-process, while a
push system does not.”
In brief, in a pull system enables the production based on customer demand; the downstream Commenté [SR15]: Where does this definition come from – too
process/customer takes the product/service they need and 'pulls' it from the producer. In his loose – look at Hopp 2008.
book “Creating Level Pull” (2004) Smalley further classifies the pull as replenishment pull, Commenté [PDN16]: Completed
sequential pull, and mixed pull system. . The successful pull system depends on flowing Commenté [SR17]: Says who?
product in small batches (approaching one piece flow where possible), pacing the processes to
Commenté [PDN18]: Completed
takt time (to stop overproduction), and Signaling replenishment via a Kanban signal and
leveling of product mix and quantity over time.
The above production approaches have marked a large step in search for productivity. Before,
Page 10 / 86
the vast majority of production flow and inventory were controlled using reorder-point/reorder-
quantity (ROP/ROQ) type methods. The ROP/ROQ systems were based on process sequences
that needed a large inventory of standard pieces, special purpose machines and Go/ NoGo
gauges. Hopp and Spearman (2002). ROP / ROQ were typical push systems, inventory and
flow were forced into the system to ensure enough production and meet the Customer demand.
The demand, until the 1930’s was dictated of the offer, and the offer was limited in a few
models available of each product. The best example is certainly the Ford T Model that was only
available in one reference in one black color. Painting in black all their models not only enabled
Ford to realize large economies of scale but also significantly reduced the drying time. The
manufacturing industry interests dictated the offer and hence tried to control the demand.
From the 1930’s, the demand became more exigent and asked for product diversity as a driver
for mass consumption. This forced the manufacturing industry to adapt the offer to a more and
more variable demand.
The development of computers and software (initially the ability to compile a large number of
information) eased the processes of managing variability in the demand by inventory. The mass
production system in the 1930’s was based on large throughput of part, large inventory sizes
and optimized but oversized tools. Time lag between steps increased and required but also
promoted the use of sophisticated information system. A concrete answer has been brought in
the 1960’s by Orlicky and Wight Plossl who developed a new system, which they termed
Material Requirements Planning (MRP). MRP was mainly a sophisticated information
management tool, one of the earliest production scheduling approaches based on the application
of computers. Although it started slowly, MRP got an extensive development in 1972 because
the American Production and Inventory Control Society (APICS) launched its "MRP Crusade"
to promote its use (Hopp and Spearman 2008). MRP is called a push system because it
computes schedules of what should be started (or pushed) into production based on demand
(Hopp and Spearman 2008).
The development of Total Quality Management (TQM) and later of 6 Sigma based on standards
and statistical analysis in the 1970’s lead to the development of MRP II, that incorporated
capacity planning, capacity requirement, master planning and input/output control and required,
of course, an even more complex and powerful software.
We can say that, from the 1930’s to the late 1990’s, the American production system relied
mainly on the development and use of more and more complex softwares.
While MRP was steadily dominating the American production control scene, history was taking
a different course in Japan. There, several Japanese companies, most notably Toyota, developed
the older ROP/ROQ methods to a high level. Starting in the 1940’s, Taiichi Ohno began
evolving a system that would enable Toyota to compete with American automaker but would
not depend on efficiencies resulting from long production runs that Toyota did not have the
volumes to support. This approach, now known as the “Toyota Production System,” was
designed to “make goods, as much as possible, in a continuous flow” (Ohno 1988).
According to Ohno, the Toyota Production System rests on two “pillars”; 1. “Autonomation”
2.Just-in-time production (JIT). Autonomation, or “automation with a human touch,” is the
practice of determining the optimal way to perform a given task and then making this the “best
practice” standard method. Autonomation also involved “fool proofing” or “poke yoke,” which
involved using devices to quickly check dimensions and other quality attributes to allow
workers to be responsible for their own quality. If problems were found, the production line
stopped until the problems were corrected. This eliminated the need for rework lines and,
eventually, eliminated most scrap. The Toyota Production System also promoted “5S,” Seiri,
Seiton, Seiso, Seiketsu, and Shitsuke which are organization and housekeeping techniques
aimed at achieving Autonomation and Visual Control.
By the end of the 1980’s, JIT began being eclipsed by Enterprise Resources Planning (ERP). As
ERP began its rise, it appeared that the JIT movement had run its course. But in 1990, a
Page 11 / 86
landmark case study conducted by MIT was published in The Machine That Changed the World
by Womack, Jones and Roos (1990). This study compared American, European, and Japanese
automobile manufacturing techniques and concluded in no uncertain terms that the Japanese
methods, particularly those of Toyota, were vastly superior. In addition, the authors freshened
JIT by recasting it as “Lean Manufacturing.”
After year 2000’bug proved to be a non-event, ERP faded as a term but was quickly replaced by
SCM (Supply Chain Management) but both trace their roots back to quantitative production and
inventory control.
The following paragraphs aim at recalling the reader with the main flow systems and theories
that arose since the 1910’s when Henri Ford’s Flow Production system was based on unique
model (outputs) to the current state of manufacturing where demand varies rapidly both in
quantities and in the references to be produced and delivered.
2.1.1 1910’s in the US: Emergence of the GANTT & PERT charts
American mechanical engineers Henry Gantt and Frederick Taylor designed a graphic method
for planning and controlling a work schedule and recording its progress. The Gantt chart is
represents horizontal bars that vary in lengths to represent each activity's time span. The vertical
axis separates activities upon a configured time line. Gantt charts are still in use today because
they are easily understood by the whole project workforce. The shortcomings of a Gantt chart
are that it does not show the interrelationships between the activities within a work sequence, or
the constraints they may impose on one another. PERT was essentially a project road map
identifying major activities and their interrelationships. Unlike Dupont’s CPM, it only showed
time constraints and did not deal with the cost and quality issues faced by commercial
organizations. T
Ohno’s insight that leveling production was key to “strategic” pull and that the way to
accomplish this was to use a “takt time” or “takt-paced production.” For example, Boeing
defined “takt-paced production” as :
Takt-Paced Production describes the rate of assembly in a factory. Lean does not mean
doing things faster; it means doing things at the right pace. Essentially, the customer’s
rate of demand establishes the pace, or takt time. So, rather than simply maximizing the
rate of work, lean sets the pace in the factory, ensuring that the customer’s needs are
met on time.
Taken at face value, it would appear that demand must be extremely regular because otherwise
following customer demands would be hugely inefficient. In reality, however, setting a pace
(instead of chasing demand) is exactly what Boeing (and Toyota) do to smooth the demand
that is seen by the plant.
This means that they set the “takt time” based on a current backlog of orders and then adjust it
from time to time. Since most releases can be connected to a customer order, these systems are,
Page 12 / 86
in an overall sense, make-to-order.
In 2002, Hopp and Spearman eventually propose a clear definition of a pull system:
A pull production system is one that explicitly limits the amount of work in process that
can be in the system. By default, this implies that a push production system is one that
has no explicit limit on the amount of work in process that can be in the system.
Page 13 / 86
2.1.4 1950’s in the US: Line balancing
Monden et al. (1983) suggested that U-line balancing problem are fathered by task time
variability and human factors or various disruptions. Task time variability and respect to work
rate are mainly due to the instability of human factors. Becker et al. (2006) and Chiang et al.
(2006) suggested that the task itself is a source of variability; such as the worker performing the
task, and the environment where the task is performed. These sources of variability are
controlled by minimizing the moving cost of men and machine. The operator walking time and
fluctuation of man and machine cycle time leads to line imbalance. Also the change over time
creates imbalance in the line for mixed model line which is necessary for lean. Based on
demand, the number of worker and machine within the workstation are increased or decreased
in order to overcome the line imbalance. Man-machine flexibility is achieved through free flow
of material and information in the manufacturing process, Monden et al. (1983).
Miltenburg et al. (2001) suggested that the break-through or tedious process flow can be
balanced by introducing the customized machine in the workstation in order to balance the
machine with the workstation cycle time. The mixed model flow is smoothened by designing
the workstation with Quick changeover and Small lot size.
Short after having validated the utility of Heijunka as a mean of production leveling, Taiichi
Onho introduced a new management strategy as a means of competitive advantage, named JIT
(Just In Time). The post-World War II Japanese automobile industry faced a crisis of existence,
and companies such as Toyota looked to benchmark their thriving American counterparts. The
productivity of an American car worker was nine times that of a Japanese car worker at that
time, and Taiichi Onho had to reach up such levels. However, the markets and production
strategies were quite different and made it almost impossible to copy what was made in the US.
Indeed, American car manufacturers were making “lots" or a “batch" of a model or a
component before switching over to a new model or component, such a system was not suited
to the Japanese conditions where a small market required manufacturing in small quantities.
Also, the car pricing policy of US manufacturers was to charge a mark-up on the cost price, the
low demand in Japan led to price resistance. Taiichi Onho identified waste as the primary evil
to reach the US productivity levels in the Japanese environment. Taiichi Onho then sought to
eliminate waste through the just-in-time philosophy, where items moved through the production
system only as and when needed.
De Treville argued that the application of JIT flow control alone would not lead to the ideal rate
of learning and that disruption as means for learning should primarily be applied in situations
where production inefficiencies could not be found analytically. De Treville (1987) later
Page 14 / 86
proposed embedding flow control and either of the two elements: Flexible resources for
improved line balance, and/or learning through disruptions occurring at less than optimal buffer
levels. Learning is key to the improvement process, and this is particularly true when comes
investigating the JIT approach. Long term productivity gains are attained by investing in
learning and switching from short term productivity for information collection.
Idle workers should be considered as resources (not as an issue), an opportunity to improve line
balance by training in search for buffer removal.
However, De treville (1987) points that the usage of disruptions in a learning process requires
careful management of improvement activities and buffers to avoid unnecessary productivity
losses and organizational stress. “A separate decision as to the “best” type of JIT must be made
for every intermediate buffer.” Designing a JIT system hence requires a large understanding of
the context and of the specific design choices.
Later (2006) De Treville emphasizes respect for people, the establishment of good job
characteristics and motivation as main subsystem of TMS, in line with lean flow management
fundamental principles.
2.1.6 1960’s in Japan: Emergence of Kanban and the origin of “PULL” system
The first academic paper describing kanban was published by Japanese researchers in 1977 by
Sugimori, et al. “Toyota Production System and Kanban System: Materialization of Just-In-
Time and Respect-For-Human System.” Like Orlicky’s “New Way of Life,” the Toyota
Production System is something big that goes beyond production control. It was followed four
years later by another paper by two other Japanese researchers, Kimura and Terada. Both papers
described the mechanics of kanban and the requirements for its implementation. As such, they
set the stage for the books of Schonberger, Hall and Monden.
Hence, by the early 1980’s the mechanics of kanban had been widely described. Because it
represented the first system to be termed a “pull” system, I briefly describe these mechanics
below.
Karmarkar appears to have been first to note the similarities of kanban to earlier base stock
systems in two simple and insightful papers (1986, 1991). The system he discusses is very
similar to a traditional base stock system as described by Simpson (1958). Simpson attributes
this system to George Kimball in an unpublished Arthur D. Little report, which Karmarkar had
published in 1988. Simpson describes the base stock system as follows:
“When an order is placed, it is filled from inventory if the inventory is not zero. If the
inventory is zero, the order is placed in a backorder file, to be filled when an item arrives. In
any event, a manufacturing order is immediately placed with the preceding manufacturing
operation to produce an item to replace the item that has been consumed. The
manufacturing operator, in turn, immediately places an order for the required raw materials
against the preceding inventory, and as soon as this order is filled [i.e., he has the needed
inventory], proceeds to ‘operate’ on it to produce the required item. In this way an order
against the last inventory for a finished item is immediately transmitted all the way back
along the line to all the manufacturing operations, each of which is galvanized into
production.”
But, while kanban and the base stock systems are similar, they are not identical. The difference
was described by Spearman (1992) as follows:
“Kanban would not place an order for more parts if a demand (in the form of a move card)
had arrived when there was no stock in the outbound stock point. Instead, there would be
one or more production cards already in process and whenever one of these was completed,
the waiting move card would be immediately attached to the recently completed container of
parts and the production card would be sent back into production. In this way a kanban
system bounds the amount of WIP there can be in the system while the base stock system
does not.”
Page 15 / 86
Spearman et al. (1990) and then Spearman and Zazanis (1992) found that while specific
environmental improvements are certainly influential (e.g., setup reduction, production
smoothing), there are three primary logistical reasons for the improved performance of pull
systems:
1. Less Congestion: Comparison of open queueing network with an “equivalent” closed
one shows that the average WIP is lower in the closed network than the open network given
the same throughput. The effect is relatively minor and is due to the fact that queue
lengths have no correlation in an open system but are negatively correlated in a closed
queueing network, an observation made earlier by Whitt.
2. Easier Control. This fundamental benefit results from several observations:
a. WIP is easier to control than throughput since it can be observed directly.
b. Throughput is typically controlled with respect to capacity. Because it cannot be
observed directly, capacity must be estimated by considering process time, setup time,
random outages, worker efficiency, rework and other factors that affect the potential
rate of production.
c. Throughput is controlled by specifying an input rate. If the input rate is less than the
capacity of the line then throughput is equal to input. If not, throughput is equal to
capacity and WIP builds without bound. By incorrectly estimating capacity, input can
easily exceed the true capacity. This is particularly true when seeking high utilization
rates. As a result, systems that control WIP are substantially more robust to control
errors than are systems that control throughput.
3. WIP Cap: The benefits of a pull environment owe more to the fact that WIP is bounded
than to the practice of “pulling” everywhere. This was argued by observing that a simple
overall bound on the WIP (i.e., as in a closed queueing network without blocking) will
promote the same benefits as those cited for kanban and by showing that the throughput of a
closed queueing network without blocking is greater than that of a closed queueing network
with blocking (i.e., a kanban system). , Spearman (2003) Commenté [SR19]: Reference? Spearman?
Commenté [PDN20]: Yes
Spearman and Zazanis (1992) proposed a hybrid push/pull system known as CONWIP that
possesses the benefits of kanban but can be applied to more general man- ufacturing settings.
Independent work by Veatch and Wein (1994) also concluded that the WIP constraint is a major
source of the benefit of kanban.
However, at this point, the understanding of “pull” varied from authors and advocators. One
reason for this, from Hopp and Spearman (2002), is that Ohno and the other early practitioners
of JIT discussed pull only in very general, high-level terms.
Manufacturers and workplaces can no longer base production on desktop planning alone
and then distribute, or push, them onto the market. It has become a matter of course
for customers, or users, each with a different value system, to stand in the front line
of the marketplace and, so to speak, pull the goods they need, in the amount and at the
time they need them. (Ohno, 1988, p. xiv)
Although Ohno did outline the elements needed to make pull to work, namely (1) standard
work methods (autonomation) and (2) level production, his writings fell short of providing a
working description of pull and hence, at the shop floor level, pull quickly became equated with
its first manifestation, kanban. Symmetrically, push became nearly synonymous with MRP.
Cheng (1993) stated that the pull mode of manufacturing only allows parts to be moved
from the previous operation to the next when the subsequent operation is ready to process.
One of the advantages of pull was described as follows:
Pull systems by far outreach the responsiveness of a push system. The responsiveness of
the system to changes and problems which arise in upstream processes allows the
downstream processes to be shut down. This prevents the accumulation of inventory on
the plant floor.
Page 16 / 86
A key catalyst for this change was the 1996 book by Womack and Jones, Lean Thinking,
which was a follow-on to their highly successful The Machine That Changed the World. But
still no clean definition of pull was provided. The book stars as follows:
Pull in simplest terms means that no one upstream should produce a good or service
until the customer downstream asks for it, but actually following this rule in practice is
a bit more complicated.
As seen in the above paragraph, managing inventory levels became an important skate in search for
resources and quality optimization. The Kanban, a subsystem of the Lean manufacturing system
was created in Japan by Taiichi Ohno to control the production and supply of components and hence
limit inventory needs. Sipper et al. (1997) classifies the Kanban system into the dual card Kanban
system for signaling production and transportation Kanban system for signaling. During demand
uncertainty the buffer maintenance is necessary for smoothening production flow and reconfigures
the Kanban System in order to lower the inventory. Thus Kanban system provide mixed model
production along with optimal inventory level which results in less lead time in product delivery and
effective utilization of resources such as man, machine etc. Junior et al. (2010) suggest that with the
knowledge in the creation and accumulation of Kanban system, the implementer can classify, and
analyze the variations of the Kanban.
While Ohno was speaking at the strategic level about the basic connection between production and
demand, Womack and Jones were talking about the tactics of implementing lean. At the tactical
level, waiting for customers to ask for goods or services does indeed cause things to get “a bit more
complicated.” To describe the Toyta kanban system, Hopp and Spearman (2008) distinguish
between push and pull production control systems as follows: In a push system, such as MRP, work
releases are scheduled, and in a pull system, releases are authorized. The difference is that a
schedule must be prepared in advance, while an authorization is decided by the status of the plant.
It has become common for managers to state that pull is make-to-order while push is make-to-
stock. Using this definition, a make-to-order MRP system would be an ideal pull system, which
clearly contradicts the historical intent of pull.
Page 17 / 86
specific work. From many experiences of this sort, the person gradually learns to generalize
how to design his / her activity. The Toyota Production System has so far been transferred
successfully only when managers have been able and willing to engage in a similar process of
questioning to facilitate learning by doing.
The experience conducted in the research came up to the same conclusion, showing that
the willingness to learn is a main prerequisite to the adoption of lean in on a construction
site.
The overall TPS approach has been summarized within a 4 rule approach proposed by Spear:
1 The person or machine can do the The activity is not Determine the true skill level of the person
activity as specified. If the activity is done as specified. or the true capability of the machine and
done as specified, the good or service The outcome is train or modify as appropriate. Modify the
will be defect free. defective. design activity.
2 Customers’ requests will be for goods Responses don’t keep Determine the true mix and volume of
and services in a specific mix and pace with requests. demand and the true capability of the
volume. supplier; retrain, modify activities, or
The supplier is idle,
reassign customer-supplier pairs as
The supplier can respond to customers’ waiting for requests.
appropriate.
requests.
3 Every supplier that is connected to the A person or machine Determine why the supplier was unnecessary,
flow path is required. is not actually and redesign the flow path.
needed. A supplier
Any supplier not connected to the flow Learn why the non specified supplier was
provides a good or
path is not needed. actually required, and redesign the flow path.
service.
4 A specific change in an activity, The actual result is Learn how the activity was actually
connection, or flow path will improve different from the performed or the connection or flow path
cost, quality, lead time, batch size, or expected result. was actually operated. Determine the true
safety by a specific amount. effects of the change. Redesign the change.
Spear argues that constant application of rules-in-use (specify, build in tests and improve) to
system design constitute meta routines that create “highly situated learning that is both broadly
distributed” and where “learning occurs through frequent practice that allows for repeated
failure”. Spear also state that “The point of process improvement is to improve the participants’
process- improvement capabilities by coaching them as they try to improve the process.”.
Toyota furthermore takes significant care to extract and disseminate learning from this process
where process improvement is thus a mean for improving individual processes as well as a
mean for accumulating organizational resources in terms of individual skills for process
improvement and in terms of widely distributed deep process and design knowledge.
Spear reverses the roles of organization and production: “The factory was not only a place to
produce physical products; it was also a place to learn how to produce those products and [...]
keep learning how to produce those products.” Despite that the system should work to reduce
variability and edict clear specifications to allow detection of and learning about process
inefficiencies and that production should be thought of as repetitive activities, permutations
provide learning opportunities for the organization. Those learning opportunities should be
exploited to create better process designs and build resources in terms of process improvement
skills in the individual, and in terms of accumulated process knowledge (Spear 2009).
Page 18 / 86
The organization not only operates production but also continuously redesigns production.
Highly situated learning implies highly situated design which requires design knowledge, skills
and motivation, in even the smallest units Employee motivation and respect are important
subsystems of TPS. Workers’ motivation is especially dependent on the experience of skill
variety and responsible autonomy; hence a dedicated application of HR practices.
The limit of applicability of CPM in the construction industry comes by the fact that
construction projects mainly differ from manufacturing plants by highly variable production
rates. The impact of variability is demonstrated by the simulation “Parade Of Trades” by
Tommelein et al. (1999) and revealed in study fields; Thomas et al. (2002), Sack and Goldin
(2007). For Sack and Brodetskaia (2007), a major issue in accurately and reliably managing
flow progress in the construction industry is the application of the Critical Path Method (CPM)
and software. CPM mainly focusses on determining the critical path from an array of tasks
linked by interpedencies, which is valid in a linear process such as in the Manufacturing
industry but the processes in the construction industry are very different. In the transformation
concept that manages construction management approaches, Koskela and Vrijhoef (2000) recall
that total transformation can only be achieved by realizing all parts of it, introducing
decompositions into parts and further into tasks. Those tasks are then assigned to workstations
and resources are added to the system to accomplish progresses.
Page 19 / 86
2.2.2 1994: The Last Planner® System (LPS)
The Last Planner System™ (LPS) aim to address flow stability and reliability. It is a well-
developed lean-production tool for project planning and management (Ballard 2000).
Implementation of the LPS demonstrates continuous improvement of project flow and increased
levels of PPC. However, even where the LPS has been applied well, PPC levels of 100% have
not been achieved (Bortolazza et al. 2005). Many techniques have been developed to run the 5
levels of discussions contained in the LPS: 1. Milestone Planning, 2. Phase Scheduling (macro
collaborative schedule of the next phase), 3. Make all tasks Ready to start in 6 weeks, 4.
Production Planning (works synchronization) and 5. Feedback and checks the completion of
commitments. A 6th conversation has been added to meet the TPS essence, learning. Despite its
unquestionable progress from CPM, LPS should be improved. LPS itself is not capable of
reacting to unpredicted conditions emerging through a working day and has no mechanism to
prioritize the work packages already filtered through a make-ready process.
Each practitioner should hence adapt the use of LPS to his own environment and constraints.
Another limit of the LPS is that is not sufficient by itself to ensure both works stability and
continuity as shown in the site research conducted.
Page 20 / 86
continuous flow. Lean production emphasizes smooth and small batch flow with little inventory
of work-in-process, short cycle-times and the ability to respond to changes in the products a
market demands. Partouche (2008). One of the main characteristics of construction is that work
flows through temporary production systems from the start to the end of a project// Once the
pull signal has been released, the same amount of work flows through the production system,
meaning the WIP levels are the same for each signal. We need to close the gap by increasing
work flow predictability beyond 1 day and by reducing delivery times 2006 Arbulu. Client
changes delivered during construction adversely affect the stability and flow of work at the
construction site. Esquenazi (2006) argues that reduced batch size appears to be the key to
improving work flow, and has a positive impact on WIP levels, cash flow and project
completion time. Partouche, Sacks &Bertelsen (2008) remind that the importance of smooth
product flow and reducing variation (at least in product) were well understood in mass
production systems; they were not new in lean production.
The batching effect is a big contributor to lead time. A lean production perspective on any given
situation suggests that in the best of circumstances the batch size should be 1, so that the flow is
continuous and incurs the least delay. Arbulu Tommelein (2002)
Small batch sizes allow for quick detection and correction of problems in production, and
consequently allow for a smoother flow of work through the various workstations in a
production system. 2003,Alves Tommelein. Hopp & Spearman’s Factory Physics uses batch
flow systems on which to develop their ‘science of manufacturing’, looking to make them more
like line flow or continuous flow processes, and neglecting job shops altogether. 2005 Ballard.
Small batches assembly strategy, standardizing and controlling the assembly cycle, and process
stabilization led to a 105% productivity increase, in relation to the budget structure 2006
Bulhoes. The transfer and production batches were defined as equal to a base-unit (two rooms),
making it easier to establish a synchronized workflow 2006 Schramm.
Page 21 / 86
2.3 THE CONCEPT OF FLOW IN THE CONSTRUCTION INDUSTRY
The concept of flow has been investigated from many facets since Koskela’s seminal work
(1999). Several studies have focused on the subject of flow tackling the concept from different
angles:
- Creating flow in construction (Santos 1999; Santos et al. 2002; Alves& Tommelein 2003),
Shen_Chua (2008), :
- and on tools to support flow, Bulhoes_Picchi_Granja 2005, Alves Tommelein Ballard2006,
Gafy_Abdelhamid 2008), Tuholski_Tommelein, 2008), Senthilkumar_Venkatachalam 2009,
Kerosuo&Koskela&Miettinen&Maki&Codinhoto 2012
- Mechanisms to pull production (Caldas Soibelman, 2002; Sacks_Goldin_Derin,2005;
Arbullu 2006; RookeSeymorKoskelaBertelsenOwenCleary 2008; Brodetskaia et al 2010,
Tiwari&Sarathy 2012),
- Visual devices to increase transparency (Formoso and Santos 2002), Santos Moser 2003,
Tan Horman Messner Riley 2003, Nakagawa 2005, Schramm Formoso 2007,
Lima_Formoso_Echeveste 2008, Chavada et al 2010, Koskela & Patricia et al 2011,
Brady&Tzortopoulos&Rooke 2012,
- Reduction of the cycle time (Ballard 2001; Elfving et al. 2002), Walsh Sawhney Bashford
2003, Chin Yoon Jung et al 2004, Sacks_Goldin_Derin 2005, Jang Ballard 2007,
Partouche_Sacks_Bertelsen 2008, Chin 2010,
- Batch size reduction, Alves Tommelein 2003, Maturana Alarcon Deprez 2003, Thomassen
Nielsen2004, Al-Sasi Brown 2006, Esquenazi Sacks2006, Ward McElwee2007,
Carneiro_Filho_Alves_Nascimento_Carneiro_Neto 2009
Picchi and Granja (2012) argue that despite the many researches undertook on flow the
application of continuous flow in construction is still understood in a fragmented way.
Bertelsen and Sacks (2007) argue that the general understanding of the construction process
within the industry and on the part of public administrators is too simple, as it sees the process
from a transformation perspective only, entirely leaving out the important flow perspective.
They develop in advocating that the construction process is seen as a basically ordered system
where order should be improved as much as possible and where unforeseen events
unfortunately happen from time to time, as opposite to understanding construction as a complex
and dynamic system where the complexity and its effects should be managed directly. They
conclude by stating that, therefore, a new mental model is needed.
As suggested by Hopp and Spearman (1996), the flow processes can advantageously be
modeled through the application of queuing theory. It reveals the central significance of
variability for the performance of flow processes, as well as the role of the process duration as a
metrics. The common conceptualization of business process serves as an example where the
flow processes and the value generation processes are considered as thing based processes, in
essence just as additional characteristics of transformation processes. Unfortunately, this thing
based approach misses the intrinsic features of the flow and value generation processes, and
cannot be recommended for generic use. Koskela& Rooke (2007)
Whilst Brodetskaia (2007) argue that a theoretical understanding is needed of the flow of
operations on the micro-level of Project Management – at the level of daily resource utilization
– in order to develop appropriate systems, Alsehaimi & Koskela (2008) further develop and
argue that Project Management theory focuses on the transformation concept without
considering both the value and flow concepts.
Sacks (2009) agrees and challenges the meaning of “process”. Sachs reminds that Koskela
(2004) showed that popular accounts like Womack & Jones (2003) may confound the two
Page 22 / 86
involved concepts, namely flow and value generation, and thus blur the existence of two
conceptualizations from which principles are being derived. Historically, lean was initiated
based on the flow concept, and the value concept, cultivated by the quality movement, was later
merged into lean.
It clearly appears from the above that understanding the concept of flow is a long lasting and
complex investigation that feeds many debates; this paper reflects the review of the IGLC
arguments and how the concept of flow has been understood.
Kalaas & Bolviken (2010) argue in paper “The flow of work in construction: a conceptual
discussion“ that the lean literature has (still) an unclear and imprecise approach to the concept
of flow. Shingo (1988, 232, 305 and 308) uses flow in the sense a chain of events, and he sees
the distinction between process and operation as one of his main contributions to production
theory and as fundamental for the understanding of production. He explains process and
operation through a two-dimensional figure with the processes flowing vertically (parallel to the
y axis) and the operations flowing horizontally (parallel to the x axis). Processes (the y axis) are
the chains of events during which raw materials are converted into products (p. 305). Processes
are object flows, flows of goods (p.78). Operations (the x axis) are the chains of events during
which workers and machines work on items (p. 305). Operations refer to ―a human temporal
and spatial flow that consistently centres around the worker‖ (p.5), and are subject flows, flows
of people (p. 78). Both processes and operations consist of four phenomena, however, which are
partly different in terms of content: Processing, inspection, transport, and delay (pp. 79–80).
When working to improve production, processes should have priority over operations (pp. 310–
311). Shingo‘s main point is understood as being that maximising output from operations
implies sub-optimisation, and operations should be coordinated to maximise overall throughput
/ throughput time, the flow of objects. This is central in lean thinking and represents a
philosophy that differs from for example the Western thinking underpinning methods
addressing the economic order quantity (Wilson 1934) on the operations level.
When applying Shingo‘s terms to construction, the process can be conceived of as the progress
of the project, while the work undertaken by the different trades constitutes the operations. A
construction project is seen as a process of aggregated sub-processes; however, not primarily
comprising sequential but also reciprocal interdependencies. Operations in constructions can be
split into work-packages with internal flow (processing, inspection, transport/movement and
delay). When work-packages are aggregated, with all the sub-flows involved, each package or
task becomes a process in Shingo‘s concept, influencing the progress of the whole project.
Kalsass & Bolviken
Kalaas and Bolviken developed a preliminary concept of flow that contains the following
properties:
Page 23 / 86
2.4 LINKING MANUFACTURING AND CONSTRUCTION FLOWS
Flow of components
The material flow of components to the site is comparable to that of car production. The
building frame proceeds through the different assembly phases (referring to processing of all
locations by a workstation), like a car proceeds through different workstations. However, due to
the size of a construction product, there is an additional workflow where all installation
locations proceed through the workstations, called ‘location flow’ (Koskela 1999).
Flow of trades
Flow of inputs (components and materials, labour, equipment, information, auxiliary tools)
While a construction building is static, a car body is (inherently) dynamic and while trades are
dynamic in a construction site, the work stations in a manufacturing plant are static.
We can hence quality a construction site ‘assembly line in multiple locations’. Any given
activity can only be performed when all required input flows with minimal required volumes
appear simultaneously in a given location. Brodetskaia (2007)
Just like in a manufacturing plant the whole construction process productivity is impacted by
the performance of as little as one trade. If the productivity of a single trade slows down, the
weakest productivity trade becomes the weakest link that determines the productivity of the
whole project chain (bottleneck). Acceleration of production of only a single trade leads to
accumulation of work locations in front of the successive trades.
In search for high pace continuous workflow, construction managers are often temped of
pushing new locations into production with the aim of creating a pull flow (large surfaces of
trade production) without necessarily understanding that it mainly increases work- in-process
(WIP) and production cycle times (CT).
Page 24 / 86
time to develop and deliver a new model. In opposition to Henry Ford’s ‘flow based’ strategy
Ohno started focusing on customer orders; a ‘value based’ strategy (Howell, 1999). Thus
‘Value can only be defined for the critical costumer and is only meaningful when expressed in
connection with a specific product, Bejder (2005). Theoretically Ohno wanted to be able to
adjust each car according to customer needs. A central and fruitful concern of lean construction
thinking has been the analysis of flows, leading to the elimination of waste. Shingo (1988)
identifies two types of flow: “Process refers to the flow of products from one worker to another,
that is, the stages through which raw materials gradually move to become finished products.
Operation refers to the discrete stage at which a worker may work on different products, i.e., a
human temporal and spatial flow that consistently centers on the worker.” (Shingo 1988) and
Rooke (2008)
Improvement in construction can either be generated by adapting conceptual and theoretical
theories from other industries (e.g. Hopp and Spearman 1996; Womack et al. 1991), or develop
new theories (for example, Koskela 2000; O’Brien et al. 2002).
Many studies that address production management in construction are application from the
thinking developed in Womack and Jones “Lean Thinking” that is based on the Toyota
Production System (TPS). Continuous flow is the underpinning goal that all these studies aim at
that can be reached by “producing and moving one item at a time, or a small and consistent bath
of items, through a series of processing steps as continuously as possible, with each step making
just what is requested by next step” (LEI, 2003, p.9). Hopp and Spearman (2008) highlight two
types of variability in a manufacturing production setting: (1) process time for a task executed
at a workstation and (2) the rate of task arrival at a workstation. In construction, various types
of variability, including those mentioned by Hopp and Spearman, impact task workflow and
impose a challenge to project teams in managing
workflow especially during design which comprises iterative processes fluctuating between the
owner’s value preposition and various design alternatives (Ballard 2000b and Ballard 2002).
The differences between construction and manufacturing are well known, and the translation of
concepts and tools requires generalization and adaptation (Lillrank, 1995). Translating
continuous flow concept to construction environment is a quite difficult challenge. On one hand
lack of flow is clear: construction production systems are characterized by many interruptions,
generating all kinds of wastes and sub-use of resources (labour, equipment and materials)
(Koskela 1992).
According to Shingo (1988 p. 4), process and operations are two different phenomena of the
structure of production and, as such, should be treated differently when one is analyzing
production systems: “Process analysis examines the flow of material or product; operations
analysis examines the work performed on products by worker and machine.
Construction as flow is a main concept in lean construction. There are three types of flow on a
construction site, namely, material flow, location flow and assembly flow, in contrast totwo
types in the factory (material flow and assembly flow) that Koskela (2000) developed in three
theoretical models of production: production is a transformation of production factors into the
product, production is a flow of material through the production system and production is value
generation, fulfilling the customers’ needs and wishes.
Page 25 / 86
According to Brodetskaia (2007), construction production processes have a 'floating' bottleneck.
At any stage of project progress, any trade that slows down its production rate or lingers
at any location may become a flow bottleneck. Therefore, under the bottleneck scheduling
approach, effective stabilization of construction flow depends on the ability of the works
manager to identify or predict which activity is the bottleneck process at any time, and schedule
or manage it properly in order to ensure it continuous workflow.
2.4.4 Takt Time Planning Commenté [SR22]: You need to conclude the chapter and there
should be a logic to the conclusion from what has been covered.
Takt time is the division of available work time per shift by the customer demand rate per shift In your case takt planning simply appears. Having reviewed the
(Rother and Shook 1998). It is the rate at which the customer demands the product and hence literature you are still attached to this focus but I am still not clear
about your justification.
dictating the rate at which production should take place to meet those demands exactly on time
without generating unnecessary inventory (Liker, 2004). The application of Takt is to make sure
that the customer demand rate is met Takt, “is the heart beat of one piece flow” (Liker, 2004).
In the construction industry, it is the overall progress rate at which all construction activities are
ideally supposed to move (Kenley et al.,2009) to reduces variability, decrease overall duration
and minimize the cost of the project (Kenley et al., 2009). Takt time requires balancing the
production rates of different workstations to ensure that product does not accumulate between
workstations and workstations do not starve waiting for work (Hopp and Spearman 2008).
The concept of Takt applied to construction has not yet been widely investigated.
It has first been presented at IGLC conferences on production system design using takt time
(Fiallo and Howell 2012, Frandson et al. 2013, Linnik et al. 2013). In building construction
applications, takt time has been defined as the maximum number of days allowed to complete
work in each location (Frandson et al. 2013).
Findings from the Sutter Health Women’s and Children’s Center (WCC) show that when takt
time and production was managed at daily level, the project was completed in 5 months instead
of the planned 11 (Linnik et al., 2013). Another case compared the productivity and production
rate of floor completion time within the same building (i.e.: same site conditions) between
Location Based Management System (LBMS) following the flow line visualization
management style and traditional ASAP management style. Results show that under LBMS,
productivity and production rates were enhanced by 18% and 10% respectively (Evinger et. al,
2013).
Those encouraging early findings and the limited research conducted so far open new
opportunities of further research in the application of takt time in the construction
industry .
3 OVERVIEW OF THE WORK DONE AND FINISHED TO DATE
The research has been started on a three block development project in Brussels (Belgium) and
conducted on the first two blocs. The following paragraphs present the general path followed to
date.
Page 26 / 86
- Planning Action: start of the first S of the “5S” approach
- Action taken: stopping the works for 30 minutes and evacuate all that is not needed on site.
- Evaluation of the action: visual impact and PPC
Action Research 1.3:
- Diagnostic: PPC still fluctuates and should be stabilized to ensure works stability
- Planning Action: Analysis the root causes for no keeping the weekly site promises
- Action taken: systematic questioning of the site foremen when a promise is not completed
- Evaluation of the action: impact on the measure of the weekly PPC
Action Research 1.4:
- Diagnostic: Despite that PPC is stabilized, a large proportion of the works desynchronization
originates from a lack of activity preparation
- Planning Action: Last Planner System (LPS) at the Make Ready level to improve site
preparation and decrease hazards probability
- Action taken: training and initiating LPS Make Ready at the site manager level
- Evaluation of the action: impact on the measure of the weekly PPC
4 LIMITATIONS
The research limits the scope to residential units made under a traditional form of EPC
(Engineering/ Procurement/ Construction) contract. Further application of Takt Planning on
other construction type and sectors can be envisaged (hospital, schools, warehouses,
housing…). The same approach can be tested even in civil works (bridges, roads, tunnels…).
The impact of Takt Planning will mainly be measured upon flow stability and continuity and its
consequence on lead time and snag works reduction. Stress reduction and financial
improvement will be checked as an expected consequence of the system but not analyzed in
deep.
Finally, duplicability of the findings of each phase will be tested over next phases, conclusions
on the conditions of duplicability will be investigated. Duplication to other sectors will be
discussed but not tested in this research.
Page 27 / 86
5 SCOPE & METHODOLOGY OF THE RESEARCH Commenté [SR23]: The lit review normally precedes the
method section. This is more than an overview in this document!
Commenté [PDN24]: I understood from our meetings and
conversations that the lit review needed to be as comprehensive as
5.1 Introduction possible to legitimate the roots, the understanding and the
contribution to knowledge.
Buckley et al., (1975) insist on the importance of meeting certain criteria to conduct good
quality research:
• it be an orderly investigation of a defined problem;
• appropriate scientific methods are used;
• adequate and representative evidence is gathered;
• logical reasoning, uncolored by bias, is employed in drawing conclusions on the basis
of the evidence;
• the researcher is able to demonstrate or prove the validity or reasonableness of their
conclusions;
• the cumulative results of research in a given area yield general principles or laws
that may be applied with confidence under similar conditions in the future.
There are many factors to consider when choosing an appropriate research method, with the
topic to be researched and the specific research question being primary drivers (Remenyi et
al., 1998). The choice of research method varies depending on the aim of the research and
the scientific discipline. Van Aken (2004) determines 3 categories of scientific disciplines
and what their missions are:
• The formal sciences, such as philosophy and mathematics: the formal sciences are
‘empirically void’. Their mission is to build systems of propositions whose main
test is their internal logical consistency;
• The explanatory sciences, such as the natural sciences and major sections of the
social sciences. The mission of explanatory science is to describe, explain and
possibly predict observable phenomena within its field. Research should lead to
‘true’ propositions, i.e. propositions which are accepted by the scientific forum as
true on the basis of the evidence provided. The typical research product of
explanatory science is the causal model, preferably expressed in quantitative
terms; and
• The design sciences, such as the engineering sciences, medical sciences and
modern psychotherapy. The mission of design science is to develop knowledge
for the design and realisation of artefacts, i.e. to solve construction problems, or
to be used in the improvement of the performance of existing entities, i.e. to solve
improvement problems (Van Aken, 2004, p. 224). Architects and civil engineers
deal predominantly with construction problems while medical doctors and
psychotherapists deal mainly with improvement problems.
A large proportion of the research conducted so far in the field of Lean Construction addresses
large sites made by very large contractors made by a design supported by BIM (Building
Information Modelling). Sutter Health and DPR construction were pioneers in the use of Lean
integrated with lean in the Camino Medical Centre project in 2003, a process further developed
in the Castro Valley project. In the UK, the Heathrow Terminal 5 project, Network Rail’s
Borough Viaduct project and the Highways Agency’s Bidston Moss (Costain) and M4/M5
Automation (Balfour Beatty) are other large scale construction projects made using Lean
Construction principle. La Laing O’Rourke (UK), Morgan & Sindall (UK), Skanska (Finland
Page 28 / 86
and UK), and BAM Nuttall (UK) are notable companies that have already at least experimented
or integrated Lean. Examples of small or medium projects conducted by SMEs under Lean
construction principles are still limited.
The scope of my research has been narrowed to address a gap in the body of knowledge:
understanding the effects of the application of Takt Planning on medium size sites (2 to 50
million € work turnover) made by medium size companies (20 to 200 employees), that
represents the vast majority of the construction sites in Europe. Determining an appropriate
research methodology is considered as an important element in a research study; especially in a
doctoral research study. It involves approach to the entire process of a research study, starting
from theoretical underpinnings and spanning to data collection and analysis, and extending to
developing the solutions for the problems investigated. Research methodology in essence is
focused around the problems to be investigated in a research study and therefore varies
according to the problems investigated. Thus, identifying the research methodology that best
suits a research in hand is important, not only as it will benefit achieving the set objectives of a
research, but also as it will serve establishing the credibility of the work (Wedawatta 2013).
The basic purpose of scientific research is theory, i.e. to understand and explain phenomena
(Kerlinger, 1977). In Kerlinger, (1973) Theory relates to "A set of interrelated constructs
(concepts), definitions, and propositions that present a systematic view of phenomena
specifying relations among variables, with the purpose of explaining and predicting the
phenomena.". A Scientific Research is “systematic, controlled, empirical, and critical
investigation of hypothetical propositions about the presumed relations among natural
phenomena.". The hypothesis is a “conjectural statement, a tentative proposition, about the
relation between two or more phenomena or variables." In short, "If X occurs ...then Y will
follow."
This thesis has been built following a combination of Design Science and Action Research
approaches. Next section introduces Design Science and Action Research and presents a
justification for their application in the present research context.
Brady (2014) points that in construction, some examples of Design Science studies that have
been carried out are by da Rocha (2011); for developing a framework to be used in defining
customization strategies for housing and Rooke (2012); for the development of guidelines for
improving way finding in hospital environments. Other examples include those carried out by
Page 29 / 86
Barker et al., (2004) who focuses on the development of a time compression model for
construction projects and a study carried out by Oyegoke (2011) focusing on highlighting the
need for constructive research and illustrating this need through the development of a
Specialist Task Organization (STO) procurement approach. This is used as a demonstrative
example to show the rigor and application of the Design Science research approach. As is the
case in these examples, rather than producing general theoretical knowledge, design scientists
produce and apply knowledge of tasks or situations in order to create effective artefacts (Van
Aken, 2004).
The core features of the Design Science approach require that it (Lukka, 2003):
A Design Science study is experimental by nature: the developed and implemented new
artefact should be regarded as a test instrument in an attempt to illustrate, test, or refine a
theory, or develop an entirely new one (Lukka, 2003). The Design Science research approach
is based on the belief, brought from the pragmatist philosophy of science that by a profound
analysis of what works (or does not work) in practice, one can make a significant contribution
to theory (Lukka, 2003).
Design Science outcomes are of four types (March & Smith, 1995): Constructs or concepts,
Models, Methods and Instantiations. The four outcomes and the relationship between each are
defined as follows (March & Smith, 1995):
Construct or concept
A construct or concept forms the vocabulary of a domain. They constitute a conceptualization
used to describe problems within the domain and to specify their solutions. Such constructs
may be highly formalized as in semantic data modelling formalisms (having constructs such
as entities, attributes, relationships, identifiers, constraints, as proposed by Hull & King,
1987) or informal as in cooperative work (e.g. consensus, participation, satisfaction, as
proposed by Kraemer & King, 1988);
Model
A model is a set of propositions or statements expressing relationships among constructs. A
model can be viewed as either a description or a prescription, that is, as a representation of
how things are, or a statement of how things should be;
Method
A method is a set of steps, e.g. an algorithm or guideline, used to perform a task. Methods are
based on a set of underlying constructs (language) and a representation (model) of the solution
space (March & Smith, 1995). Methods can be tied to particular models in that the steps take
parts of the model as input. Furthermore, methods are often used to translate from one model
or representation to another in the course of solving a problem;
Page 30 / 86
Instantiation
Instantiating means realizing an artefact in its environment where the methods are
operationalized. Vaishnavi & Kuechler (2007) develop the concept of instantiation and argue
that Design Science research supports better theories in two ways: 1) since the artefact
developed is an object for theorizing for many communities (e.g. how to build more
maintainable software), the construction phase of a Design Science research effort can be an
experimental proof of method or an experimental exploration of method, or both and 2) the
artefact can expose relationships between its elements, through better understanding and
making them more visible, thus potentially falsifying or elaborating on previously theorized
relationships (Vaishnavi & Kuechler, 2007). Van Aken (2004) further develops the outcome of
Design Science as a technological rule : “a chunk of general knowledge” (Van Aken, 2004, P.
228), linking an intervention or artefact with a desired outcome or performance in a certain
field of application.
Page 31 / 86
Step 1: Problem identification and motivation
This first step aims to define the research problem and validate the significance of the
proposed solution. The readers better understand the origin, the context, the reasons and
reasoning of the researcher in researching a solution to the proposed problem. Peffers et al.
(2007) argue that a thorough definition of the problem and justification of the value of a
solution is essential for the development of effective artefacts. The links between the problem,
its justification, its value and context help capture the complexity of the problem and helps
build a justification for the value of the solution. Justifying the value of the solution is
important for motivating the researcher and the audience of the research to pursue the solution
and to accept the results and helps to understand the reasoning associated with the researcher’s
understanding of the problem (Peffers et al., 2007).
Identifying the problem is a critical step to understand what the problem is and therefore being
able to validate the value of the solution to the problem.
Step 4: Demonstration
In Step 4, experimentation, simulation, case study or other appropriate activity may be
conducted to demonstrate the use of the artefact to solve the problem with the help of effective
knowledge of the use of the artefact to solve the problem. The researcher hence demonstrates
how effectively the created artefact solves the research problem.
Step 5: Evaluation
Step 5 evaluates how well the created artefact supports the solution to the research problem by
observation and measurements. This activity is conducted by comparing the objectives of the
solution to the actual monitored results from use of the artefact in the demonstration. Evaluation
can be of various forms: surveys, user feedback and satisfaction questionnaires or comparing
the artefacts functionality with the solution objectives or include quantifiable measures of
system performance, such as response time or availability (Peffers et al., 2007). Another
iteration of the Design Science Research cycle may prove necessary from the analysis of the
results of the evaluation phase to solve the identified research problem.
Evaluating the artefact requires knowledge of relevant metrics and analysis techniques (Peffers
et al., 2007).
Page 32 / 86
Step 5: Communication
The final step closes the cycle when it has been confirmed from the above that the artefact
solved the research problem. The utility and novelty of the artefact the rigor of its design, and
its effectiveness are communicated in the findings to relevant audiences such as researchers and
practicing professionals (Peffers el al., 2007). Writing papers commonly achieve this final step,
giving lectures or conferences, sharing new knowledge to both the academic and the industry
worlds. The research problem and its significance are exposed, the artefact (innovativeness,
effectiveness, design objectivity and usefulness) is commutated to be challenged, criticized and
if possible further developed by the research community.
In 1955, for Blum, action research is a two stage process. The stage, diagnostic, involves a
collaborative analysis of the social situation and the researcher formulates hypotheses on the
nature of the research domain. Then, to “care”, collaborative change experiments are conducted
by introducing changes and the effects are studied. Scientific rigor has been introduced by
Susman and Evered in 1978 by detailing a five-phase cyclical process. The approach first
requires the establishment of a client-system infrastructure or research environment. Then, five
identifiable phases are iterated: (1) diagnosing, (2) action planning, (3) action taking, (4)
evaluating and (5) specifying learning. The figure below is a diagram of this action research
structural cycle.
Page 33 / 86
Fig 3.3: Adapted from Coghlan & Brannick (2001), p 19
The client-system infrastructure is the specification and agreement that constitutes the research
environment. For Baskerville (1997), a key aspect of the infrastructure is the collaborative
nature of the undertaking. The research scientists work closely with practitioners who are
located within the client-system. These individuals provide the subject system knowledge and
insight necessary to understand the anomalies being studied. For convenience it is useful to
think of the practitioner as part of a set of actors who are oriented to solution of practical
problems, who are essentially organizational scientists rather than academic scientists.
In the case of my research, the client-system infrastructure is given and the boundaries of the
actions that can be conducted determined by the contract that links my Client (JM Belgium) and
my company (Delta Partners). The contract and the scope of actions have been conjointly
decided with JM Belgium to plan to work on the three development block consecutively and to
build knowledge and experience on the previous findings to explore more improvement
opportunities at each stage. Hence, Block 1 would be used to initiate a lean culture supported by
fundamental tools such as Last Planner® System and 5S (2 first S). Block 2 would be the
opportunity the capitalize on the experience and learning from block 1 and decrease tolerance in
the respect of the fundamental tools (increase Pourcentage of Promise Completed) and secure
the three first S from the 5S approach. Block 3 would be a consolidation of the good practices
gained in the first two blocks and an initiation of the Integrated Project Delivery (collaborative
design).
Given the time allocated to the research (three years), the research made on each of the three
blocs would be made itself of successive action research cycles.
The research would therefore comprise of three distinct action projects within which cycles
would be undertaken.
Page 34 / 86
One difficulty (or trap) in this situation stands in my external position, not as a Researcher but
as a Consultant. The common view of organizational consulting holds that this activity is a
process of temporarily introducing external expertise in order to obtain objective analysis,
specialist knowledge, or the benefits of cross-organizational experience, without permanently
acquiring additional organizational members (Steele, 1975). Most scholarly work in the
consulting process is generally considered part of the organizational development stream. The
work is highly cross-disciplinary because the principles and techniques are widely adopted by
specialists in many management fields, e.g. marketing consultants, computer consultants,
manufacturing consultants, etc. (Kubr, 1986). The process of organizational consultation is
projected in various stages or phases. For Lippit and Lippit (1978) consulting is a succession of
phases such as engagement, analysis, action, and disengagement. Engagement is the process by
which the consultant is introduced into the organization under a formal contractual agreement
that links the organization and the consultant in search for the deliverable, that is the critical
output. During the analysis phase the consultant studies the organizational elements to identify
problem areas and possible solutions to conclude by a presentation of a consultancy report to
the management of the organization. This report typically proposes various actions to be
undertaken to bring concrete answers in seach for the deliverable. During the action phase, the
consultant works within the organizational structures to implement the actions that have been
selected by the Management. The disengagement phase is determined either if the deliverable is
delivered or if the mission (temporally fixed) has come to an end.
Bakersville (1997) argues that Consultancy is a form of participative case study, but a case
study in which the authors make no strong claims to being objective or uninvolved observers.
Rather, the case describers very often have a pronounced and understandable bias in promoting
their claims to having identified actions leading to successful outcomes.
By essence, a Consultant is not necessarily interested in maintaining a strong academic rigor as
the result (deliverable on which the fees can be claimed) is of more importance than the path
followed to attain the objective. A consultancy is managed as any profit based company: profit
is generated by the difference between the price sold for the mission and the operational costs to
deliver the mission. The price sold is most generally a lump sum (fixed) so in order to
maximize profits, a consultant must decrease operational costs by short cutting outputs,
generating immediate biased results and be result focused, that is mostly incompatible with
academic rigor.
Seashore (1976) first concerned and warned that without some rational linkage of action and
research, we will be increasingly embarrassed and impeded by professional services
masquerading as research and by scientific inquiry unjustifiably proffered as a client service.
Eric Trist (1976) pointed that "there is a great difference between simply acting as a consultant
and acting as a researcher in a role where both professional and scientific responsibility are
accepted. In the first case, there is no commitment to the advancement of scientific knowledge,
either on the part of the consultant or on the part of those for whom the inquiry is being made.
In the second case, this commitment is fundamental and must be explicitly accepted by both
sides. It is this that makes the relationship collaborative." Warmington (1980) adds that the
Action Researcher must not present himself as the expert who has some definitive techniques or
packaged methods, able to make his own recommendations to the organization concerned. He
must, they say, act throughout as a fellow student, collaborator and investigator, of equal status
with his managerial colleagues. But his position may be more difficult and ambiguous than this.
Gummesson (1988) proposed four factors of rigor to clearly differentiate action research and
consulting (a) researchers require more rigorous documentary records than consultants; (b)
researchers require rigorous theoretical justifications and consultants require less-rigorous
(often undetailed) empirical justifications; (c) consultants are subject to client-imposed rigor,
i.e., they operate under tight time and budget constraints, while researchers must often impose
their own rigor on the projects; (d) the consultation is usually linear--engage, analyze, action,
disengage--while the action research process is cyclical requiring additional rigor since each
Page 35 / 86
phase must be re- entrant. Baskerville (1997) proposes to summarize the key distinctions
between consulting and action research as follows:
From all the above, the consultant researcher (or researcher consultant should pay an extra care
and focus on focusing on the research part and leave the consultancy apart. The research
conducted here should hence, for the scientific stake, follow the academic rule and rigor while
allowing generating “some” profit to ensure the consultancy viability. The motivation is clearly
based on the quest of scientific knowledge that is not incompatible with profit, but profit as a
consequence of making a Client happy by having improved his operations and captured the
learning, not as a priority. The knowledge gained and experience will then be share with the
academic world to be disseminated, a fundamental difference with a classical consultancy
approach.
I am hence committed to both the academic and practice world, my approach is based on
collaboration with my Client who is deeply involved in the research path. The recommendations
come mainly from the theoretical framework despite that some empirical experience can also
help choosing the action to be tested. The finding their explanations will be constructed within
the action research framework, their validity and possible generalization will be assessed by the
scientific world and it is most probable that more research on the topic will be necessary,
capitalizing on the learning gained in this case, before dissemination.
From the above, it clearly appears that, despite that I am consultant and that JM Belgium first
contracted me as such, changing hat to a researcher is necessary to guarantee academic rigor
and scientific validity of the findings.
From the above, it also appears that one can be a researcher and a consultant; but a consultant
cannot necessarily be a researcher.
Page 36 / 86
In their book Architectural Research Methods, Linda Groat and David Wang (2002), illustrate
the relation between different research strategies in the field of architecture. They argue that the
strategies that are close to each other on their diagram have more similarities than those that are
further apart. Qualitative and interpretive researches have in common a holistic approach to the
research subject, but with differing time perspectives. Correlational research, on the other hand,
shares with qualitative research a focus on naturally occurring circumstances, but is dependent
on quantitative data. Experimentation is also dependent on quantitative data, but with the
requirement that the researcher must be able to manipulate isolated variables. Likewise,
simulation requires control and manipulation. Logical argumentation — which includes, for
instance, space syntax analysis — shares with simulation an emphasis on abstraction. And
interpretive-historical research is dependent on a constructed logic of interpretation. This
completes the circle. (from Johanson 2003)
Page 37 / 86
Type Definition Published Study
Explanatory Answering a question that sought to explain the presumed causal Joia (2002). Analysing
Example a web- based e-
links in real-life interventions that are too complex for the survey commerce learning community: A case study
or experimental strategies. In evaluation language, the in Brazil. Internet Research, 12, 305-317.
explanations would link program implementation with program
effects (Yin, 2003).
Exploratory Exploring those situations in which the intervention being Lotzkar & Bottorff (2001). An observational
evaluated has no clear, single set of outcomes (Yin, 2003). study of the development of a nurse-patient
relationship. Clinical Nursing Research, 10,
275-294.
Descriptive Describing an intervention or phenomenon and the real-life Tolson, Fleming, & Schartau (2002). Coping
context in which it occurred (Yin, 2003). with menstruation: Understanding the needs
of women with Parkinson’s disease. Journal
of Advanced Nursing, 40
Multiple- Exploring differences within and between cases. The goal is to Campbell & Ahrens (1998). Innovative
replicate findings across cases. Because comparisons will be community services for rape victims: An
case studies drawn, it is imperative that the cases are chosen carefully so that application of multiple case study
the researcher can predict similar results across cases, or predict methodology. American Journal of
contrasting results based on a theory (Yin, 2003). Community Psychology, 26, 537-571.
Intrinsic Stake (1995) uses the term intrinsic and suggests that Hellström, Nolan, & Lundh (2005). “We do
researchers who have a genuine interest in the case should use things together” A case study of “couplehood”
this approach when the intent is to better understand the case. It in dementia. Dementia, 4(1), 7-22.
is not undertaken primarily because the case represents other
cases or because it illustrates a particular trait or problem, but
because in all its particularity and ordinariness, the case itself is
of interest. The purpose is NOT to come to understand some
abstract construct or generic phenomenon. The purpose is NOT
to build theory (although that is an option; Stake, 1995).
Instrumental Accomplishing something other than understanding a particular Luck, Jackson, & Usher (2007). STAMP:
situation. It provides insight into an issue or helps to refine a Components of observable behaviour that
theory. The case is of secondary interest; it plays a supportive indicate potential for patient violence in
role, facilitating our understanding of something else. The case is emergency departments. Journal of Advanced
often looked at in depth, its contexts scrutinized, its ordinary Nursing, 59, 11-19.
activities detailed, and because it helps the researcher pursue
the external interest. The case may or may not be seen as typical
of other cases (Stake, 1995).
Collective Collective case studies are similar in nature and description to Scheib (2003). Role stress in the professional
multiple case studies (Yin, 2003) life of the school music teacher: A collective
case study. Journal of Research in Music
Education, 51.
Table 3.1: Definitions and published examples of case studies (Baxter 2008)
The research will address real life construction sites in which the researcher is professionally
engaged, and has a genuine interest to better understand the case. This tends to advocate that
appropriate type of case study “Intrinsic”. However, if we consider that the case will comprise
of all three apartment blocks, exploration differences within and between sub-cases could be
possible (each apartment block) and replication of the findings across cases made, the ad hoc
type of case study would rather be “Multiple Case Studies”. In a multiple case study, we are
examining several cases to understand the similarities and differences between the cases. Yin
(2003) describes how multiple case studies can be used to either, “(a) predicts similar results (a
literal replication) or (b) predicts contrasting results but for predictable reasons (a theoretical
replication)”. The types “Explanatory”, “Exploratory”, “Descriptive”, “Instrumental” and
“Collective” do not seem appropriate here.
Page 38 / 86
relevance. The audiences themselves can validate and generalize finding of the case (cases
selected within a representational sample strategy used in correlational research, Stake 1995,
Patton 1990) or the generalization can be intrinsically generated when a case is purposefully
selected in virtue of being information-rich, critical, revelatory, unique, or extreme.
The case observed here has been purposefully been selected as the property blocks to be built
are most common: typical property development of standard apartments in a European capital
suburb made by classically selected SMEs with no particular technique or architectural fantasy.
The generalization is hence expected validated through the intrinsic characteristic of the
selected case. Generalization can then be built by deductive principle (a hypothesis is
formulated an testable consequences are derived by deduction) or through induction (theory
generation based on the data of the case) or by abduction (particular cases). Ginzburg refers to
these kinds of generalization as occurring within the “evidential paradigm” (Ginzburg 1989).
Table 3.2: Modes of generalization and reasoning within case study methodology (Johansson 2012)
From the above, two main generalization approach seems valid: deductive and inductive.
“Deductive” as the research here aim at testing the applicability of takt approach to the
organization (scheduling) of construction sites and “inductive” as the approach that is proposed
to be tested is not yet a theory. By essence the conclusion is not known at the early stages of the
research and in a case study, the different modes of generalization are often combined.
Layder argues that theory testing and theory generating are combined in practice. He names this
combined approach “adaptive theory approach”. I shall retain the latter as it best fit the duality
and complexity of my case study research.
Interest from scholars in the field of Action Research (AR) and Design Science (DS), two
main approaches in information systems (IS) have been growing in the academic world.
Despite that Kock & Lau, 2001; Baskerville & Myers, 2004; Kock, 2006 have provided
legitimacy to AR, March & Smith, 1995; Hevner et al, 2004 (and others) proposed
articulations of design research in IS hence the development of an other (yet complementary)
research approach. While AR have been the one research approach when dealing with
observing, modifying and learning from concrete actions within reality through iterative steps,
DS emerged and there is nowadays a growing interest in DS and perhaps a declining one in
AR.
Action research and design research originate from different scientific traditions. AR has
Page 39 / 86
mainly initially been developed in social science (Dewey ,1938 and later Kurt Lewin, 1947).
DS on the other hand introduced the concept of “sciences of the artificial” taken from
engineering science (Simon, 1996). Both traditions have found their ways into the information
systems discipline. May authors proposed attempts to understand the relations between AR
and DS, as Burstein & Gregor (1999), van Aken (2004), Cole et al (2005), Andriessen (2007),
Lee (2007), Järvinen (2007), Figueiredo & Cunha (2007).Baskerville et al (2009), Iivari &
Venable (2009), Sein et al (2011), Papas et al (2012), Wieringa & Morali (2012), Alturki et al
(2012), and more recently Goldkuhl (2013) proposed using practice research as a lens for
comparison and integration between action research vs. design research.
Summary of the paradigmatic assumptions of Action Research and Design Science Research.
Similarities?
On the basis of a conceptual analysis, Järvinen (2007) argue that AR and DR are similar
approaches. He compared the previous work on AR from Susman & Evered (1978) and
Nunamaker et al (1991), March & Smith (1995) and van Aken (2004) for DR, searched for
similar characteristics and found similarities: seven fundamental similarities. Goldkuhl (2013)
proposed a categorization of the similarities based on Järvinen (2007): 1) striving for utility,
2) production of useful knowledge, 3) combination of building/acting and evaluation, 4)
collaboration between researchers and practitioners, 5) aiming for development and
improvement, 6) intervention in a local practice and 7) knowledge creation and testing during
the process.
The above table, adapted from Järvinen (2005) presents Similarities of the fundamental
Page 40 / 86
characteristics of action research and design science
Action research means both action taking and evaluating. Building and evaluation are the two main activities of
design science.
Design science research is initiated by the researcher(s)
interested in developing technological rules for a certain
Action research is carried out in collaboration between
type of issue.
action researcher and the client system.
Each individual case is primarily oriented at solving the
local problem in close collaboration with the local people;
Design science solves construction problems (producing
Action research modifies a given reality or develops new
new innovations) and improvement problems (improving
system
the performance of existing entities).
Design science research is initiated by the researcher(s)
interested in developing technological rules for a certain
The researcher intervenes in the problem setting.
type of issue. Each individual case is primarily oriented at
solving the local problem in close collaboration with the
local people.
Knowledge is generated, used, tested and modified in the Knowledge is generated, used and evaluated through the
course of the action research project. building action.
After comparing important characteristics (concepts and paths) of both AR DS there seem to
be important overlaps and similarities in both approaches. On this base, Järvinen, (2007) claim
that action research and design science should be considered as similar research approaches.
Dissimilarities?
Unlike Järvinen who conducted a conceptual analysis, Iivari & Venable (2009) conducted a
paradigm analysis. Despite that they found some similarities, they argue that the similarities
they found are “superficial” and that the two approaches are “decisively dissimilar”.
Similarities and differences were found from their comparisons, where based on ontological,
epistemological, methodological and ethical assomptions. They claim that there is no
similarities between AR and DR in the cases of an AR case made with no technical design or a
DR case with pure technical problem solving or a DR case without any local practice
intervention (Goldkuhl, 2013).
The aim of AR and DR can also divert from studies to studies. While the major contribution of
DR is to create “new means for achieving some general (unsituated) goal and demands
innovation and novel technology”, AR searches “a general practice contribution and not for
any local practice contribution linked towards “normal design practice”. Iivari & Venable
(2009) conclude by orienting research toward DS in the field of cutting-edge technologies and
toward AR in the field of safe solutions based on robust technology (Goldkuhl, 2013).
Combination?
Burstein & Gregor (1999) proposed an analysis of the two approaches in relation to each other.
Their case was not conducted under AR or DS but under general scientific criteria. They
introduces a new term of system development (SD) based on the terminology from
Nunamaker et al, (1991) and conclude that “The iterative nature of SD research, with the
intention of generating new knowledge and improving social acceptance of the system under
construction, makes it also quite distinct from computer science research where the main
purpose is the creation of new methods or programs – not the application of the methods in
real-life.”. While some authors claim that AR and DS are similar approaches (eg Järvinen
Page 41 / 86
(2007)), others claim that they are quite different (eg Iivari & Venable (2009)). Others, such as
Cole et al (2005) are more in favour of combining the two approaches rather than opposing
them. Based on a cross-criteria analysis, they compared DS criteria (from Hevner et al, 2004)
and AR criteria (from Davison et al, 2004), based on the other’s approach case. They
discovered that despite the approaches claimed to be different, the cases fulfilled criteria from
the other research approach. Cole et al (2009) used a down to earth paradigmatic analysis and
outlined an integrated research approach. They found that a common area could be found since
“the process models of both approaches are similar to a degree that we can form a common
process model for them”. Papas et al (2012) made a first of a kind experiment: they conducted
their case both from AR and DS analysis after that the paradigm analysis of the two
approaches has revealed many similarities, but also some differences. In light of this
experience, Papas et al (2012) claim that it is possible and valuable to combine AR and DS
and used both criteria.
Integration?
The above sections describes important similarities that have been found between AR and
DSR and how several several integrative approaches combining AR and DSR have been
created. In search for a combined approach, as remined by Goldkuhl (2013); Cole et al (2005)
proposed a four stage process: 1) problem identification, 2) intervention, 3) evaluation and 4)
reflection and learning; where intervention (step 2) is a combination of action planning and
action taking (from AR) and building (from DSR).
Later, in 2007, Lee (2007) claim that “action research and design science have the potential to
bring about greater rigor and greater relevance by acting together than by acting alone” (ibid p
45). Capitalizing on the framework from March & Smith (1995), Lee combined the four types
of activities and the four types of artefact; respectively: build, evaluate, theorize, justify and
construct, model, method, instantiation. The dialogical action research from Mårtensson & Lee
(2004) helps him map on this DR framework and make a step forward from the previous
publications by including “theorize” and “justify” in DSR, hence completing March & Smith
(1995) and Hevner et al (2004) work. Lee then separates the activities conducted by the
researcher and by the practitioner and claims that the researcher should not be pa participating
in the building of the instantiation
Lee (2007) defines, following the ideas of dialogical action research, what activities will be
conducted by the researcher and what activities will be conducted by the practitioner. It is
notable that the researcher is not participating in the building of the instantiation. This is due to
the researcher role of being an “expert advisor” … [taking] the scientific attitude” (ibid p 53).
Baskerville et al (2009) in their “Soft design science methodology”. They integrate AR and
DR through the use of a well-known AR approach; Soft Systems Methodology – SSM
(Checkland, 1981). They adapt here SSM to a DR context. Within this integrated approach,
they combine working 1) with specific problems and solutions and 2) with generalized
problems and solutions. The problem solving logic is from 1) specific problem via 2)
generalized problem and 3) general design requirements and solutions to 4) adaptation of
specific solution.
A later development (2011) of Cole et al (2005) has been proposed by Sein et al, an integrated
approach emerged: Action Design Research (ADR) that further integrates intervention and
evaluation. Inception of ADR is the formulation of an organizational problem to be treated by
a research process that leads to the generation of design principles.
Wieringa & Morali (2012) have presented a combined AR and DR approach called Technical
Action Research (TAR). This is defined as a researcher-driven approach based on ideas of
Page 42 / 86
idealized design contrary to problem-driven design. TAR is working with two related
“engineering cycles” and a complementary empirical research cycle. One engineering cycle is
the artefact cycle where a researcher aims at improving a class of problems through artefact
design. The other engineering cycle is helping a client improving a particular problem. The
empirical research cycle aims at answering research questions through making investigations.
Wieringa & Morali (2012) explicitly contrast their approach to ADR, which is labelled as
“problem-driven”. TAR does not start with particular problems. Its start is rather that “the
researcher has identified a class of problems, and aims to develop an artefact to mitigate those
problems” (ibid p 234).
As noted in Section 2 considering our conceptions of AR and DSR, by definition the research
interest of DSR is to construct new and innovative ways to solve a class or classes of
problems, thus creating new reality. AR does not necessarily share any such purpose. Much of
AR is conducted to understand existing reality, such as the complex workings of
organisational situations and human behaviour. These research interests are decidedly
different.
We now turn our analysis to the activities of AR and DSR. To facilitate our analysis, we
conceptualize of two broad classes of problems and solutions: purely technical problems and
innovations and socio- technical problems and innovations. Furthermore, we assume that any
solutions or innovations to purely technical problems do not have direct and complex
implications for the socio-technical systems within which they will be embedded (other than
monetary costs, which are always taken into consideration in any innovation). Therefore, there
is little or no interest in such problems and innovations in the context of AR, although there
can be considerable interest in the context of DSR. Rather, AR is exclusively interested in
innovations with significant impacts on socio-technical systems and the human context. This
distinction between purely technical and socio-technical problems and innovations is reflected
in the potential activities or AR and DSR, which we will now analyse.
We base our analysis of AR and DSR activities on the model of DSR activities presented in
Venable (2006), which extended the multi-methodological framework of Nunamaker (1990)
and Venable and Travis (1999) to differentiate between Naturalistic and Artificial evaluation
in DSR, as in Figure 1.
Theory
Building
Solution
Technology
Invention
Page 43 / 86
Figure 1. Framework and Context for DSR (Venable, 2006).
Page 44 / 86
1c. No None Solving a socio-technical Theory building,
overlap problem in a non-AR Solution technology
context by developing a invention, and
new solution technology, Artificial and/or
but evaluating it by natu- ralistic
2. Slight Evaluating an existing Evaluation
means otherof a solution
than AR Naturalistic
evaluation
overlap solu- tion technology tech- nology developed evaluation only
in an orga- nizational separately
3. Solving
context a socio- Solving a socio-technical Theory building,
Significa technical problem by problem by developing Solution technology
nt developing and a new and a new solution invention, and
overlap solution technology technology and Naturalistic
and evaluating it in an evaluating it in an evaluation
Table 2. Overlaps in activities
organiza- between
tional AR and DSR.
organiza- tional context
context
It appears that no clear consensus has emerged yet. This section aims to develop the debate
between AR and DS and propose a justification of the approach(s) retained.
“A problem-centered approach is the basis of the nominal sequence, starting with activity
one. Researchers might proceed in this sequence if the idea for the research resulted from
observation of the problem or from suggested future research in a paper from a prior
project.
Page 45 / 86
that worked; it starts with activity four, resulting in a Design Science solution if researchers
work backwards to apply rigor to the process retroactively. This could be the by- product of
a consulting experience”.
In the case of this research, Design science approaches have been applied at different stages of
the project: block 1, block 2 and block 3.
AlSehaimi et al. (2013) criticize research in construction management as a too descriptive and
explanatory approach (AlSehaimi et al., 2013) using quantitative surveys or case studies
(Azhar et al., 2009). Managerial problems experienced in construction would hence not be
addressed with tangible and applicable answers. They further argue that the main problem with
existing approaches to research in construction is that research findings tend to fall short of
providing clear recommendations for the improvement of Construction Management practice
(AlSehaimi et al., 2013). Most recommendations are general and not devoted to solving the
problems observed in practice. Studies often find planning and control to be ineffective, yet
solutions to this problem are not recommended and while some studies do recommend
solutions, they do not identify the necessary tools needed to facilitate those (AlSehaimi et al.,
2013). While theory building, theory testing and explanation undoubtedly remain
indispensable, a deeper focus on discovery and problem solving in the form of Design Science
research to complement existing approaches is needed (Simon,1969 & 1996; Simon, 1973;
Klahr & Simon, 1999; Holmström 2009). For Peffers et al., (2007), Design Science can be
adopted for a research study in various ways depending on the circumstances of the research;
ie whether it is a problem-, an objective- or a design centered research project or if the
research study was initiated from a client-context situation (Section 3.5.5).
Thus, such as Brady’s doctoral research “Using visual management to improve transparency in
planning and control in construction”, Salford 2014, Design Science was also adopted for this
research to fit the unique circumstances of how this work was initiated in practice in a client-
context situation. Alike Brady, my role as a consultant with a company specialized in the aera
of lean management applied to the construction industry with the aim to improve delays and
costs through creating reliable and stable production flows on site, the researcher developed a
first version of improvement tool. This initial experiment aimed resolve issues experienced in
the daily planning and control of construction work onsite.
According to Peffers et al., (2007) a typical Design Science researcher would start with a
problem and then work from there, but a research study could also be initiated by the
development and application of a solution that worked in practice as is the case with this
research.
Iterative improvements in weekly and daily planning were noted after the first implementation
of LPS in practice, which initiated its further study and development within the context of this
PhD work. Similarly to Brady (2014), the researcher is in a sense “working backwards” to
apply rigor to the process. Design science can result from the existence of an artefact that had
not yet been formally thought through (Peffers et al., 2007, p.14), as is the case with the
model developed in this research. Through the three-case study and developments contained in
this research, the development process of the model is presented, analyzed, evaluated and
further developed to determine its contribution to practice and theory using the Design Science
approach. This need to apply rigor to the initial design process of the model is also an
important reason for applying Design Science.
As mentioned at the beginning of this section, research in construction tends to fall short of
providing clear recommendations and solutions to improve construction management practice
Page 46 / 86
(AlSehaimi et al., 2013). A further important reason for adopting Design Science is due to
nature of the aim of the research, which is to provide a clear solution that can be applied to
improve current practice in production planning and contol and to clarify the theoretical
significance of this solution (March & Smith, 1995; Lukka, 2003; Hevner et al., 2004; Van
Aken, 2004; Venable, 2006). Design Science is prescriptive in nature, creating artefacts that
embody those prescriptions (March & Smith, 1995) and this research intends to change the
real world by proposing a solution to solve a practical problem and to add to existing
knowledge about the world by establishing the theoretical significance of the problem and its
solution.
Through the application of Design Science, this work also intends to bridge practice to theory,
rather than theory to practice as proposed by Holmström & Ketokivi (2009). They present four
phases of research, the first two being exploratory (solution incubation and solution
refinement) followed by two phases of explanatory research (establishing theoretical
relevance and development of formal theory). According to the authors Holmström & Ketokivi
(2009), typically, problem solvers in industry will likely stop at phase II once the solution is
refined and the solution design has met its goal. However, solving a managerial problem does
not constitute a scientific contribution. Through the course of the research, the initial phases I
and II as described by Holmström & Ketokivi (2009) are linked to phase III by generalizing
the findings and demonstrating a theoretical contribution. Holmström & Ketokivi (2009)
conclude that the successful bridging of managerial relevance and theoretical contribution lies
in the ability to bridge Phase II and Phase III types of research. Design Science intends to
make academic research more relevant to practitioners and extends practical solutions by
seeking more thorough theoretical understanding and contribution. Design Science better
aligns the theoretical and research interests of this work with the interests of managerial
practice (Holmström & Ketokivi, 2009; Lukka, 2003). Despite ambitious efforts in various
fields of research over the years, the goal of making academic research relevant to the
practitioner remains elusive (Holmström & Ketokivi, 2009).
Furthermore, a key part of Design Science is the evaluation and testing of the solution
(Kasanen, 1993; Lukka, 2003; Vaishnavi & Kuechler, 2007 & Peffers et al., 2007) and an
important focus of this research work is to determine whether the solution provided utility for
the task and the market intended (Kasanen, 1993). Design Science can be combined well with
other traditional approaches to research such as case studies, which are an important method
of investigation in this research.
Sections XXX presents the outcomes of Design Science research and Section 3.5.1 discusses
the most important steps involved when adopting this approach.
Page 47 / 86
Data were regularly collected the research, from different sources: workflow indicators,
documentation, semi- structured interviews and direct and participant observation.
Page 48 / 86
5.8.1 workflow indicators
In measuring the workflow stability or reliability in a construction site, the PPC
(Percentage of Promise Completed) is widely used, amongst whom Ballard 2002,
Hamzeh 2006, Gehbauer 2008.
5.8.2 Documentation
Documents are important sources of data in research and their range might include
diaries, letters, agendas, minutes of meetings, personal notes, field notes and reports,
images, sounds and objects and computerized records (Bryman, 2001; Yin, 2003;
Finnegan, 2006). According to Yin (2003), documents can provide other specific details
to corroborate information from other sources and additionally, related inferences can
be made from documents.
5.8.4 Observation
Hancock (1998) describes observation as a technique that can be used when data
collected through other means can be of limited value or it is difficult to validate.
Observation can be defined as a systematic method of data collection that relies on a
researcher’s ability to gather data through his or her sense (O’Leary, 2004). Hussey et
al. (1997) and late Saunders et al. (2007) divide observation into two categories: direct
(non- participant) and participant. In direct observation, people or experiments are
observed and recorded without that the researcher is involved (i.e. a field visit to the
case study). In a participant observation, the researcher is not just a passive observer,
instead, he/she is involved and may play a variety of roles in the event being researched
(Yin, 2003; Easterby-Smith et al., 2002; Bryman, 2001).
A detailed description of the data collected during each part of the research is presented
in each of the case study Chapters XXXXX
6 RESEARCH PROCESSES
Research building/ mind process consisted in defining the problem identified in practice
Page 49 / 86
and deepening the knowledge of this problem through a synthesis of the literature, then
test the solution and measure the results; before getting back to step one as an iterative
improvement process.
Diagnostic
The researcher, was given the task of conducting an analysis at the building site in
question, to determine any potential to reduce waste as defined in the Toyota
Production System and to optimize the building process.
The research process began with an initial solution to a problem perceived in practice:
the site works are highly instable and desynchronized, there was a low visibility of the
structural works schedule. Deeper understanding the practical problem has been the first
step of in case study 1, so to say of the research. As a Consultant, the researcher is
valued to provide immediate solutions to immediate issues. As a Researcher, the switch
from consulting had to be made to increase clarity, actuality and causes of the problem
from a theoretical perspective. The theoretical part was to be investigated within the
body of knowledge, in the existing literature. Most of the literature and cases available
described situations of large sites made by large and mature (lean-wise) companies. A
very limited amount was found in the residential development field made by SMEs to
gain this deeper understanding of the problem and of the theoretical background that
could provide basis for addressing it. The issues, problematic and solutions applicable
on a large hospital project in California erected by multinational lean leader
construction companies are limitedly applicable to a small 35 apartment bloc in
Brussels suburb made by small SMEs who have never heard about lean.
The literature review focused on the principle of flow stability, scheduling, Lean
Production, Lean Construction and the deficiencies of classical Project Management.
Once the problem been duely identified and its effects measured, step 2 consisted in the
definition of the objectives of a solution that would be implemented, or to say, the
expected results and/ or effects to step 1. The data gathered in step 1 were then
reviewed and analyzed (Chapter XXX, Table XXX). This step is necessary to detail
how the objectives for a solution were expected to bring a concrete answer to the
problems observed in practice. Reviewing the important concepts in the literature and
the researchers practical experience on the application of Lean tools and flow
Page 50 / 86
production tools developed prior to the current research helped refine the possible
objectives and solutions.
The main objective of the research was to stabilize the production flow on site, with
expected effects or impacts on reducing the delay, the perceived stress, the costs while
increasing motivation, quality and health and safety levels. (Chapter XXX, section
XXX.).
Action planned
After that this operations audit or “gemba walk” had been conducted, that the main
wastes and origins of these wastes have hence been identified and measured, it was time
to initiate the research. With a knowledge of possible production management tools to
be applied in practice in the aim to stabilize the production flow (LPS, Takt Planning,
Kanban systems…), the researcher in his role as a consultant, applied the most common
and widely accepted in the construction industry: Last Planner. The experience of the
researcher as a consultant on various construction projects prior to the research
conducted here helped choose the most appropriate “level of conversation” of the Last
Planner as starting point: the Weekly Work Plan (WWP). Unlike the other
conversations of the LPS (except from the daily review, that needs feeding from the
latter and could not hence be tested first), the Weekly Work Plan brings concrete
discussions and impacts that helps motivate and embark site teams that could, by
nature, be reluctant in new managerial tools. Hence WWP has been been chosen to start
the research.
Action taken
In evaluating what a solution has achieved, a difficulty often lies in determining when a
solution is complete. Henver et al., (2004) states that a solution is complete and
effective when it satisfies the requirements and constraints of the problem it was meant
Page 51 / 86
to solve. An important part of the evaluation of the LCM model is to establish whether
it contributed to an improvement in transparency in the planning and control of daily
operations and to clarify its contribution to knowledge and practice. In order to carry
out the summative evaluation, it was first of all necessary to develop an evaluation
framework. The development of the evaluation framework and the evaluation process is
a significant part of the overall research design.
Three important steps were taken to develop a suitable framework for the evaluation:
The first step involved the definition of evaluation criteria. In the literature, a number of
important criteria can be found that should be considered when evaluating an artefact
using the Design Science approach. Hevner et al., (2004) point out that utility,
quality, and efficacy are parameters for evaluating a solution. March & Smith et al.,
(1997), state that research in the build activity should be judged based on value or
utility to a community of users. Likewise, Kasanen argues that artefacts should be
validated based on their utility and applicability in the market (Kasenen, 1993). Van
Aken (2004) refers to evaluating the effectiveness of a certain rule in the original
context while Lukka (2000) emphasises the need to “ponder the scope of applicability
of the solution” or translating the rule to other contexts. Lukka (2003) also emphasises
the need for the researcher to explicate the theoretical contribution of the artefact during
the evaluation by reflecting the findings back to prior theory. From these ideas, high
level criteria were identified as basis for the evaluation framework. They are:
usefulness, applicability and theoretical importance (Figures 3.4 and 3.5).
While the applicability of the model could be evaluated based on its adaptability to
different types of construction scenarios, in order to establish criteria for evaluating
utility a deeper understanding of the aims of the model and its outcomes according to
Design Science was needed.
The LCM model was analysed and its elements classified according to the outcomes of
Design Science: concepts, models, methods and instantiations (Section 3.5). This is an
important step, to fully understand the artefact and its aims to determine what is
understood by utility of the LCM model. The foundational concepts of the LCM model
are based on the idea behind the Lean principles of value, value stream, flow, pull and
perfection (Womack & Jones, 1996) and a need to improve transparency in the
construction process onsite.
A total of 16 different elements as part of the LCM model were identified which are
based on these concepts (see Figure 3.3). Five of the elements (1. Overall Process Map
2. Process Planning tool 3. Overall Process Analysis action list 4. Process Planning
action list and 5. The Stability of Process Planning Metric PP metric) aim to focus on
the principle of value by ensuring that the optimal process is fully understood and that
Page 52 / 86
there is a common understanding of what the value of the process is. Six further
elements of the model (6. The Planning board 7. The construction cards 8. The problem
cards 9. The logistic board 10. Logistic cards 11. Visualised colour-coded site layout
and 12. Material database (power plant) aim to implement flow and pull in the
construction process onsite and in the wider processes. Finally, four elements of the
model (13. Apartment clock / colour- coded plans 14. Visualised action plans in LCM
area 15. Metric for On-Time- Performance and 16 metric for quality) aim promote
continuous improvement and the principle of perfection in the process. How these
elements are applied together in a specified way is the method. Each application of
LCM represents an instantiation.
To link each bloc and grow the chain of evidences, I have added
- 5. Learning and 6. Question to initiate next AR cycle.
The tables below present the logical path and steps undertaken from the diagnostic of the first action
research in bloc 1 (AR 1.1) to the question that introduces the third action research (AR3.1)
Block 1A:
AR 1.1 AR 1.2 AR 1.3 AR 1.4
YSAYE
The site works are highly
Delays (and root for lower PPC) PPC fluctuates and should be Lack of activity fathers works
instable and desynchronized,
Diagnostic low visibility of the structural
mainly due to labor congestion stabilized to ensure works desynchronization despite high
on site. stability PPC
works schedule
Start the implementation of Last
Action Continue the implementation of
Planner System by Weekly Work Initiate 5S on site Analysis the root causes
planned LPS to the next step
Plan (WWP) meetings
Introduction and implementation
of “Make Ready” to the trade
2hour training in Lean project managers and
Systematic questioning of the
Action Construction with the foremen, Start 5S by the first “S”: determined the theoretical works
site foremen, 5 whys analysis
taken setting of the rules of the WWP evacuate the wastes sequence and unit time.(Note:
when possible / appropriate
meetings. inception of the development
path to the application of Takt
Time latter developed)
Works commitment on site
Evaluation PPC continued to rise, though PPC kept high, works stability
(PPC) doubled in the first month PPC levelled
of the action instable from one week another improved
continued to grow after.
A clear understanding of the A clear working area helps the Immediate answers hide real The works can be
“Why” and favorable meeting foremen be aware of their own difficulties faced by the foremen.
Learning issues and that of the others that
desynchronized despite a high
conditions are key to first lure A natural tendency to not open PPC
the foremen who will then come eases mutual help. the constraints and bad news
Page 53 / 86
naturally at the meetings, kept early. The most accurate and
short and useful for their daily usable answers (in a continuous
organization. improvement goal) came from
the foremen, not from the project
managers.
Question to Why are the works are running What can be improved in the
What prevents site commitment
initiate next (PPC) to raise more ?
Why is PPC not stable? late vs schedule despite a high next bloc (Chopin) that has
AR cycle and levelled PPC? slowed the works in this block?
Page 54 / 86
Block 2:
AR 2.1 AR 2.2 AR 2.3 AR 2.4
Chopin
Accelerations and decelerations The pull system based on takt
Works in the raisers have been a
The works are still not stable in the pace of the works are time requires a solid discipline.
bottleneck in the start of the
Diagnostic works and delays the whole
despite a high PPC and measured from an apartment Despite that flow constancy
technical anticipations another, from a week another, improves, site congestion
progress of block 1.
Last Planner is not enough. becomes again an issue
Develop further the pull system
Technical optimization and micro Determine a pace to align all
Action Compare the theoretical and by integrating a Kanban system
planning of the works in the trades and develop a
planned actual works progresses (discipline by visual
raisers complementary pull system
management).
Determined a model to calculate
Calculated a Takt time that Developed a table displayed on
the theoretical weekly
Special collaborative meeting should allow each trade to the wall of the site meeting room
Action progresses for each apartment
with the M&E+ Structural trade complete its works in a single with sticky vignettes (Kanban)
taken in a stable flow vs weekly
foremen and project managers. apartment. Initiate Takt Planning for the next trade in queue,
progresses measurements on
following Frandson’s 5 steps. mainly a communication tool.
site
Individual and collective
A truly collaborative spirit has
Permitted to find distortions improvement on the state of the
Evaluation allowed tradeoffs. The raisers Measured works stability,
between theory and actual site and self-commitment to
of the action works were optimized and doubled
physical progress follow the rules. Works stability
prepared well in advance.
doubled.
The Project Managers are not
It is somehow possible to apply
used to addressing issues well in Discipline can be eased by a
A high and stable PPC does not Takt Time to the construction
Learning advance, a trade dedicated
necessarily stable works industry, early key success
simple effective visual
preparation could have eased management system.
factors identified.
the process.
Block 3:
AR 3.1 AR 3.2 AR 3.3 AR 3.4
Vivaldi
To be started.
Diagnostic (Concrete works only started. No
relevant diagnostic yet.)
Action
planned
Action
taken
Evaluation
of the action
Learning
Question to
initiate next
AR cycle
Table 4.3: Action research cycles in bloc 3 (Vivaldi)
Page 55 / 86
7.2 ACTION RESEARCH CONDUCTED ON BLOCK 1 (YSAYE) Commenté [SR29]: You need to provide a structure that reflects
the AR cycles and the teams involved in these cycles etc need to
have been introduced in the methods section. The AR account could
be more clearly structured around the cycle headings where the
linkage between cycles is clear. You are expected to be keeping a
journal and quoting from it is good practice.
Commenté [PDN30]: Modified
7.2.1 Action Research 1.1: Weekly Work Plan (Last Planner® System)
- Diagnostic: the works on site are highly instable and desynchronized, the existing planning
of the concrete trade lacks visibility The first three measures of the promised completed by
the site foremen, before the instalment of the Last Planner® System, showed levels
between 10 to 20%. Such a level indicates that the vast majority (80 to 90%) of the works
undertaken in a day were not planned for that day a week ago. The overall planning is
followed using critical path method but the planning is 4 month old and need to be updated
to take into account the actual constraints of the project and their impact on the
synchronization of the works. It is to be noted that, with no surprise, the site was largely
congestioned by materials, tools and over storage and that safety was to be improved. The
lack of visibility on the planning forced the project management team to focus on the
physical progresses made weekly, under the well disseminated sentence in the construction
industry “all what is done, is done”, a typical push site management system.
- Action planned: Last Planner System (LPS) at the Weekly Work Plan (WWP) level to
Page 56 / 86
improve weekly works reliability. The WWP, through the enhanced communication it
promotes, facilitates synchronization among the trades at the site level. Very often, the
foremen have very limited knowledge and interest in the works and constraints of the other
foremen on site. Each foreman focusses on the works of his trade under the milestones
given by the project manager in charge of coordinating the works. The WWP, by the
openness and communication it imposes was planned first, to produce and follow
immediately the PPC that is a robust indicator of the short term planning reliability.
Communication is a key to assure impeccable works synchronization, one of the main
pillars of LPS. The aim was to double the reliability (raising from 20 to 40% PPC) in a
month. . It is to be noted that, right from the first meeting, a heavy concern has been put on
offering a very propitious environment to the meeting. The meeting room was clean,
heated, ventilated, well lighted, vast, with comfortable chairs, biscuits and good coffee to
lure the site foremen and maximize the chances of their coming before they actually
understand the deeper questions of why and for what they come.
- Action taken: the 5 site foremen have been made aware that their lean journey would
collectively start. The 2 hour on site training followed the golden circles path described by Commenté [SR31]: Detail – how long what material who was
Sineck (2009). After an brief introduction of the presence of the researcher on site, the involved (remember this is research).
“Why” question has been addressed, emphasizing the opportunity to change the way site Commenté [PDN32]: Detailed
managers assure works scheduling and sharing the vision that it was possible to deliver
quicker, cheaper and better. Examples from previous cases from the lean literature were
briefly presented, creating a sense of urgency. The “How” question was then addressed.
The context and general mechanisms of the Last Planner System was presented to provide
the site managers with an understanding of the 5 different steps that would take place
(Master Planning, Milestone Planning, Make Ready, Weekly Work Plan and Learning).
The “What” question logically ended the session. The first rules to follow during the WWP
sessions were shared: 1) discipline and 2) commitment. 1) “if you can commit, engage
yourself. If you cannot commit, do not. Do not promise if you are not convinced.” and 2)
“arrive on time and prepared and the meeting will stop on time after just an hour and a half,
at the latest”. The first rule was made to make all aware that the aim was not to get nice
promises but commitments that can actually be kept and measured. The second rule was
made to emphasize the important of respect for people. I asked an engagement to each of
the foremen to commit to the rules, which they verbally agreed. The promise I made was to
keep the meeting short, provided that the foremen kept their engagement to arrive prepared
and on time. The 3 first meetings have not started on time (20 minutes late). No site
foremen were present on time (counted at the minute) but the meeting closed on time. This
put the foremen in a position in which they were the ones who did not keep their
engagement while I did keep mine. The number of engagement increased after the
discovery phase (first two weeks), the PPC went up very rapidly (from 30 to 60+ in the
first month as planned) as the site foremen understood there were really no point in
announcing over optimistic promises.
The tendering process in the construction industry can be long and fastidious. The initial
group of 5 foremen representing the companies already chosen were soon joined by other
site managers from complementary trades.
On the fifth week, as the Client had appointed two more trade contractors to join the team, Commenté [SR33]: How are you defining the AR cycles – this
two newly landed site foremen joined the site and the WWP meeting. Like the firsts, they needs to be clear and justified.
followed the 2 hour on site training, made aware of the rules to follow and engaged Commenté [PDN34]: Completed. This is an external constraint.
themselves to respect them. They came on time, this put the late ones in an uncomfortable
situation; the group started to be constituted. Using sociological drivers from the group’s
theories and human bias from the games theory and / or the prisoners’ dilemma are
powerful helps in achieving true commitments. The sociological studies showed that a
group is composed of seven. From seven on, the sum of individuals is not a sum of
individualities but one single group. It is then each’s stake to stay in the group, by
Page 57 / 86
respecting the simple enacted rules. After six weeks, the majority of the site foremen
arrived on time for the meeting (most often five minutes in advance) the meeting could
then start on time, the promises of the last week checked and the reasons for not keeping a
promise analyzed first before getting on planning the following week.
- Evaluation of the action: The result has overcome the expectations as the PPC has indeed
doubled in the first month continued to grow after.
Page 58 / 86
Fig 4.3: “5S” on the slab
- Evaluated action: the visual impact was immediate. PPC continued to rise to 70+%,
which is considered a very good score. Also, these actions consolidated and enlarged the
group, provided coherence/ respect between trades. The first mutual aid and care of the
other foremen where noticed, which unfeasible even 2 months ago was when each site
foreman only cared for his portion of works, regardless of the other trades.
The site foremen have been interviewed on their perception of the action (plus, delta, why, for
what), the results were very encouraging as they all showed interest in the approach and
committed to keep it.
Page 59 / 86
7.2.3 Action Research 1.3: Root Cause Analysis of PPC (Last Planner® System)
- Diagnostic: PPC is still high but fluctuates and should be stabilized to ensure works
stability
- Action planned: Fluctuation of PPC Analysis the root causes for no keeping the weekly
site promises
- Action taken: systematic questioning of the site foremen when a promise is not completed
- Evaluation of the action: impact on the measure of the weekly Percentage of promise
completed (PPC)
- Action planned: Last Planner System (LPS) at the Make Ready level to improve site
preparation and decrease hazards probability
- Action taken: training and initiating LPS Make Ready at the site manager level
- Evaluation of the action: impact on the measure of the weekly Percentage of promise
completed (PPC)
Page 60 / 86
7.3 ACTION RESEARCH CONDUCTED ON BLOCK 2 (CHOPIN)
Page 61 / 86
✓ Commit to follow the rules
✓ Do not start works in a zone before ready
✓ Start works in a zone only when your turn
✓ Share issues asap, during the weekly meeting (at the latest)
✓ Accept possibility of rethinking the sequences (within or amongst) areas,
redesigning the zones: collaboratively to reach tradeoff. (…)
✓ Engage foremen in search for the best system
Page 62 / 86
Fig 5.1: Illustration of production flow in the manufacturing industry
In a construction site, the production does not flow. The whole production tool does flow
(workmen, materials and materiel), passing from an area (block, floor, area, room) to another.
Coupled with “Micro Zoning” that aims at dividing the project into “manageable” areas and
find relevant sequences amongst & into them. The resources (Manpower, Materials, Tools)
loaded on site independently from the (actual) needs of the site vs Resources delivered on site
“on (anticipated) demand ”Static production tool & dynamic product vs dynamic production
tool & static product The respect of the sequences agreed in the Last Planner® System phase is
essential to a sound synchronization of the works on site. There is a heavy tendency from the
site foremen to not respect a sequence and improvise new sequence after that the first have not
been kept. The consequences of over taking the next trade in the sequence (not waiting that all
the preliminary works are completed for a given task) can be many: the next trade work proves
more difficult (de-synchronization, under-preparation), working without all reliable information
or installation on site, improvising (technically)…
Extracted from the tasks identified in the phase scheduling workshop (LPS), the entire work
sequence of the works has been entered in a spreadsheet that indicates the state of the works on
each apartment. A “1” was entered to indicate that the works were completed.
In the next continuous improvement circle (step 2), each site foreman will enter their input in
the IT system (or glue a sticker on the table printed) to indicate that their works for a particular
apartment are completed (and checked), ready for the next trade in queue.
Each site foreman hence had a global vision of the state of the works and had a precise
knowledge of what apartment was ready for the next trade. This table was also used to feed the
discussion at the “Weekly Work Plan” level.
Page 63 / 86
A log of the progresses of the works was kept over a 6 week period to illustrate the search for a
continuous and constant work flow. Each week, the global progresses made were measured in
each apartment and showed in a graph. Given that only one trade was to work in a given
apartment, the graph represents the works progresses of the trades in each apartment.
General program of the works: 42 main works were identified with a planned duration of half a
week (105 days). Each apartment was to be started half a week after its predecessor. The total
theoretical lead time was hence 105d +2,5dx34apt = 190days. The difference of progresses (in
percentage) for a given week between two consecutive apartments should be 100%/ 42/ 2 =
1/84%. The difference in the progresses from one week another for a given apartment should be
100%/42x2 = 1/21%.
And if P(1,1) represents the progresses of Apartment 1 in week one, P(1,2) the progresses of
Apartment 1 in week 2, P(2,1) the progresses of Apartment 2 in week 1 and P(2,2) the
progresses of Apartment 2 in week 2, etc…
And if “Ew” is the number of Elementary works to be achieved in sequence, then the progresses
from a week another can be defined as in our example as:
To complete this equation, if TT is the time to complete the works of a given task in an
apartment and L the Lag between two apartments then:
26 27 28 29 30 31
A01 88 93 98 100 100 100
Were: A02 87 92 96 100 100 100
A03 86 90 95 100 100 100
- L is the Lag between two volume start (to be expressed A04 84 89 94 99 100 100
in relation with TT ie L=3TT). A11 83 88 93 98 100 100
- P is the period between two progresses A12 82 87 92 96 100 100
A13 81 86 90 95 100 100
A14 80 84 89 94 99 100
In our example, the theoretical data are shown in the A21 78 83 88 93 98 100
following table aside, the linearity of this equation illustrates A22 77 82 87 92 96 100
the continuous and constant flow that is searched, illustrated A23 76 81 86 90 95 100
A24 75 80 84 89 94 99
in the following graph A31 74 78 83 88 93 98
- (1) Apartment A03 should be completed in week 28. A32 73 77 82 87 92 96
- (2) Apartment A14 has been completed at 80% in A33 71 76 81 86 90 95
A34 70 75 80 84 89 94
week 26 and should be completed in week 31 A41 69 74 78 83 88 93
- (3) Apartment A34 has been completed at 70% in A42 68 73 77 82 87 92
A43 67 71 76 81 86 90
week 26 and should be at 80 in week 28. A44 65 70 75 80 84 89
A51 64 69 74 78 83 88
A52 63 68 73 77 82 87
Page 64 / 86
(1)
(2) CP week
31
CP week
30
CP week
29
CP week
28
CP week
27
(3)
Cumulative
progress (CP)
at week 26
26 27 28 29 30 31
Fig 5.1: Measurements of works progresses (%) in Part A of Ysaye Building (Block 1)
The variations are important, mounts and valleys in the graph show respectively low activities
Page 65 / 86
and high activities, illustrating the volatility of the works on site. Some week, even no works
at all were made in some apartments; see Apartment 13 & 14 in week 29 for example (cf the
red circle in the graph). The analysis of the root causes of the PPC helped limit this happening
in search for a constant work flow.
From the data collected the weekly differences of progresses made gave a good indication of
the how level the works. The calculation of the standard deviation of the progresses of each
week gave a good indication of the “volatility” of the works.
For each apartment, the differences (Δ) between the progresses measured between two weeks
were calculated. This measure (Actual) represents the progresses made on each apartment on a
given week. The calculation of the standard deviation of the differences in those progresses
gave another good indication of the level of works volatility.
The value of the box has been taken out from the calculation if the works of an apartment was
completed (100%) to avoid biasing the results (yellow boxes in the tables of the follow pages).
Initial Measures YSAYE - VARIATION OF WEEKLY PROGRESSES
26 27 28 29 30 31
Actual Actual Δ Actual Δ Actual Δ Actual Δ Actual Δ
A01 88 98 10 100 2 100 - 100 - 100 -
A02 93 100 7 100 0 100 - 100 - 100 -
A03 90 95 5 98 2 100 2 100 - 100 -
A04 88 98 10 98 0 98 0 98 0 100 -
A11 85 93 7 98 5 98 0 98 0 100 -
A12 78 85 7 95 10 100 - 100 - 100 -
A13 90 93 2 95 3 95 0 95 0 100 5
A14 80 88 7 95 7 95 0 98 2 100 2
A21 78 78 0 88 10 95 7 98 2 100 2
A22 73 73 0 88 15 100 12 100 - 100 -
A23 80 80 0 90 10 98 8 98 0 100 2
A24 78 78 0 88 10 93 5 95 2 100 5
A31 66 76 10 93 17 93 0 98 5 98 0
A32 66 78 12 88 10 95 7 100 5 100 -
A33 61 76 15 86 10 88 2 95 7 95 0
A34 61 78 17 86 8 88 2 100 12 100 -
A41 59 68 10 75 7 90 15 100 10 100 -
A42 56 66 10 81 15 86 5 95 10 100 5
A43 56 66 10 79 13 85 6 93 8 94 1
A44 59 66 7 79 13 83 4 93 10 98 5
A51 49 53 4 62 9 74 12 80 6 92 12
A52 45 55 10 62 7 71 10 79 7 93 14
Standard Deviation: 4,73 4,78 4,60 4,00 4,44
Table 5.1: Measurements the weekly works (%) variations in Part A of Ysaye Building (Block 1)
As the standard deviation indicates the distribution of the progresses, the larger the standard
deviation, the larger the volatility of the work. The mean of the standard deviations of the
progresses over the 6 week period observed of Action Research 1 above is 4.5, oscillating
between 4 and 5.
As this measure only gives an indication of the “volatility” of the works, it is important at this
stage to calculate the differences in the progresses, planned (or theoretical) and actual. The
table below presents this calculation for each apartment, and presents both mean and standard
deviation.
Page 66 / 86
YSAYE (A) - DIFFERENCE OF WEEKLY PROGRESSES (THEORY AND ACTUAL)
26 27 28 29 30 31
Theory Actual Δ Theory Actual Δ Theory Actual Δ Theory Actual Δ Theory Actual Δ Theory Actual Δ
A01 88 88 0 92,76 98 5 98 100 2 100 100 0 100 100 - 100 100 -
A02 86,81 93 6 91,57 100 8 96 100 4 100 100 0 100 100 - 100 100 -
A03 86 90 5 90 95 5 95 98 2 100 100 0 100 100 - 100 100 -
A04 84 88 3 89 98 8 94 98 4 99 98 -1 100 98 -2 100 100 0
A11 83 85 2 88 93 5 93 98 5 98 98 0 100 98 -2 100 100 0
A12 82 78 -4 87 85 -2 92 95 3 96 100 4 100 100 - 100 100 -
A13 81 90 9 86 93 7 90 95 5 95 95 0 100 95 -5 100 100 0
A14 80 80 1 84 88 3 89 95 6 94 95 1 99 98 -1 100 100 0
A21 78 78 0 83 78 -5 88 88 0 93 95 2 98 98 0 100 100 0
A22 77 73 -4 82 73 -9 87 88 1 92 100 8 96 100 4 100 100 0
A23 76 80 4 81 80 0 86 90 4 90 98 7 95 98 2 100 100 0
A24 75 78 3 80 78 -2 84 88 4 89 93 4 94 95 1 99 100 1
A31 74 66 -8 78 76 -3 83 93 10 88 93 5 93 98 5 98 98 0
A32 73 66 -7 77 78 1 82 88 6 87 95 8 92 100 8 96 100 4
A33 71 61 -10 76 76 0 81 86 5 86 88 2 90 95 5 95 95 0
A34 70 61 -9 75 78 3 80 86 6 84 88 4 89 100 11 94 100 6
A41 69 59 -10 74 68 -5 78 75 -3 83 90 7 88 100 12 93 100 7
A42 68 56 -12 73 66 -7 77 81 4 82 86 4 87 95 8 92 100 8
A43 67 56 -10 71 66 -5 76 79 2 81 85 4 86 93 7 90 94 4
A44 65 59 -7 70 66 -4 75 79 4 80 83 3 84 93 8 89 98 8
A51 64 49 -15 69 53 -16 74 62 -12 78 74 -5 83 80 -3 88 92 4
A52 63 45 -18 68 55 -13 73 62 -11 77 71 -6 82 79 -3 87 93 6
Mean: -3,72 -1,20 2,33 2,40 3,07 2,75
Standard Deviation: 7,49 6,58 5,05 3,72 5,33 3,23
Table 5.2: Mean and standard deviation of weekly progresses (%) in Part A of Ysaye Building (Block 1)
The average of the means is 0.9%, which indicates that, in average, 0.9% of the progresses
were gained from the forecast. But the global standard deviation of the Δ is 6.0. This means
that most of the weekly progresses were comprised between -5,1% and +6,9% from the target.
This scale gives an indication on both the level of volatility and how the weekly targets were
reached.
The same exercise has been undertaken on block B, where project management and works
scheduling were made “as usual”: the planning had been made by the project manager who
lead the sub-contractors under the critical path method. The results are presented below.
26 27 28 29 30 31
Fig 5.2: Measurements of works progresses (%) in Part B of Ysaye Building (Block 1)
Page 67 / 86
As noticed in part A of block A that was scheduled under the Last Planner® System, it
happens that no works are made in some apartments from a week another (red circles above).
26 27 28 29 30 31
Actual Actual Δ Actual Δ Actual Δ Actual Δ Actual Δ
B01 85 95 10 95 0 100 5 100 - 100 -
B02 88 93 5 93 0 98 5 100 - 100 -
B11/B12 85 88 3 90 2 95 5 95 0 95 0
B13 88 95 7 100 5 100 - 100 - 100 -
B21 61 68 7 76 8 76 0 81 5 98 17
B22 66 75 9 81 6 86 5 86 0 95 10
B23 76 78 2 83 5 88 5 90 2 95 5
B31 56 71 15 74 3 81 7 90 10 92 2
B32 51 66 15 69 3 71 2 79 8 83 4
B33 40 45 5 65 20 79 14 92 13 94 2
B41 55 71 16 79 8 83 5 83 0 88 5
B42 50 66 16 74 8 79 5 79 0 83 5
B43 45 60 15 71 11 76 5 76 0 79 3
B51 42 55 13 52 -3 69 17 76 7 76 0
Standard Deviation: 4,94 5,60 4,41 4,73 4,91
Table 5.3: Measurements the weekly works (%) variations in Part B of Ysaye Building (Block 1)
The mean of the standard deviations of the progresses over the 6 week period observed of
Action Research 1 above is 4,9; also oscillating between 4 and 5. Compared to the 4,5 of part
A that has been scheduled with LPS® it is not clear on whether LPS® helped lowering works
volatility. The numbers rather show that the LPS had no impact on how level the works.
The same calculations as to the data of Part A of Ysaye Building have been applied to Part B
to further investigate the impact of LPS® vs traditional (CPM) scheduling system and shows
surprising results.
While the works are as volatile in the traditional scheduling system as under LPS, the weekly
target is less attained. The measure of the mean and standard deviation of the theoretical and
actual works presented below illustrate this finding.
26 27 28 29 30 31
Theory Actual Δ Theory Actual Δ Theory Actual Δ Theory Actual Δ Theory Actual Δ Theory Actual Δ
B01 85 95 10 90 95 5 95 95 0 99 100 1 100 100 0 100 100 -
B02 84 93 9 89 93 4 93 93 0 98 98 0 100 100 0 100 100 -
B11/B12 83 88 5 87 88 1 92 90 -2 97 95 -2 100 95 -5 100 95 -5
B13 81 95 14 86 95 9 91 100 9 96 100 4 100 100 0 100 100 -
B21 80 68 -12 85 68 -17 90 76 -14 95 76 -18 99 81 -18 100 98 -2
B22 79 75 -4 84 75 -9 89 81 -8 93 86 -8 98 86 -12 100 95 -5
B23 78 78 0 83 78 -5 87 83 -4 92 88 -4 97 90 -6 100 95 -5
B31 77 71 -6 81 71 -11 86 74 -12 91 81 -10 96 90 -5 100 92 -8
B32 75 66 -10 80 66 -14 85 69 -16 90 71 -18 95 79 -16 99 83 -16
B33 74 45 -29 79 45 -34 84 65 -19 89 79 -10 93 92 -1 98 94 -4
B41 73 71 -2 78 71 -7 83 79 -4 87 83 -4 92 83 -9 97 88 -9
B42 72 66 -6 77 66 -11 81 74 -8 86 79 -8 91 79 -12 96 83 -12
B43 71 60 -11 75 60 -15 80 71 -9 85 76 -9 90 76 -14 95 79 -16
B51 70 55 -15 74 55 -19 79 52 -27 84 69 -15 89 76 -13 93 76 -17
Table 5.4: Mean and standard deviation of weekly progresses (%) in Part B of Ysaye Building (Block 1)
Page 68 / 86
The average of the means is -7,5%, which indicates that, in average the works were late 7,5%
in comparison to theory. But the global standard deviations of the Δ is 8.7. This means that
most of the weekly progresses were comprised between -16,2% and -1,2% from theory. This
scale gives an indication on both the volatility and how the weekly targets were reached.
Page 69 / 86
In the contrary Bock 1 Ysaye Building (both Part A and B), it has not happened during the six
week observation period that no works are made in some apartments from a week another (red
circles above). In other words, works were being carried in each apartment, in each week.
Based on the same Takt Time (that determine the progresses possible in theory), the data were
compiled the same way as for block 1 and are presented in the below table.
44 45 46 47 48 49
Actual Actual Δ Actual Δ Actual Δ Actual Δ Actual Δ
C01 75 82 7 88 6 95 7 97 3 100 3
C02 73 82 9 90 8 97 7 97 1 100 3
C03 72 78 6 86 8 95 9 97 3 100 3
C04 68 79 11 94 15 101 7 97 -4 100 3
C05 68 75 7 82 7 90 8 97 8 100 3
C06 70 78 8 86 8 93 7 97 5 100 3
C07 67 75 8 82 7 88 6 95 8 100 5
C08 65 72 7 79 7 85 6 91 7 100 9
C09 68 72 4 79 7 86 6 92 7 100 8
C10 65 73 8 78 5 85 7 92 8 100 8
C11 68 74 6 77 3 78 1 85 8 92 7
C12 65 71 6 77 6 85 8 93 9 96 3
C13 60 62 2 66 4 70 4 77 8 88 11
C14 63 65 2 74 9 81 7 87 7 94 7
C15 60 66 6 74 8 79 5 84 6 91 7
C16 50 60 10 69 9 77 8 83 7 89 6
C17 49 62 13 69 7 70 1 75 5 82 7
C18 50 57 7 64 7 65 1 73 9 81 8
C19 51 58 7 62 4 70 8 77 8 86 9
C20 49 56 7 64 8 70 6 74 5 81 7
C21 44 58 14 64 6 72 8 76 5 83 7
C22 44 52 8 59 7 70 11 76 7 81 5
C23 40 45 5 54 9 62 8 67 6 76 9
C24 42 48 6 53 5 61 7 68 8 75 7
C25 45 50 5 55 5 63 8 68 5 73 5
C26 39 44 5 52 8 60 8 67 8 72 5
C27 36 48 12 54 6 55 1 63 9 68 5
C28 34 38 4 46 8 53 7 57 5 66 9
C29 39 42 3 48 6 55 7 56 2 62 6
C30 43 45 2 49 4 55 6 59 5 67 8
C31 39 48 9 58 10 63 5 70 7 74 4
C32 41 45 4 50 5 51 1 59 9 67 8
C33 38 44 6 52 8 59 7 63 5 69 6
C34 35 42 7 49 7 55 6 61 7 66 5
C35 33 40 7 46 6 52 6 55 4 64 9
Table 5.5: Measurements the weekly works (%) variations in Part B of Ysaye Building (Block 1)
The mean of the standard deviations of the progresses over the 6 week period observed of
Action Research 2 above is 2.4; also oscillating between 2 and 3. Compared to the 4,5 of part
A that has been scheduled with LPS® it is clear that takt helped lowering works volatility.
Page 70 / 86
Again, to investigate further, the data have been compiled to extract the works performance
(how far from the linear target), the figures are presented in the table of next page.
44 45 46 47 48 49
Theory Actual Δ Theory Actual Δ Theory Actual Δ Theory Actual Δ Theory Actual Δ Theory Actual Δ
C01 75 75 0 80 82 2 85 88 3 89 95 5 94 100 6 99 100 1
C02 74 73 -1 79 82 3 83 90 7 88 97 8 100 100 0 100 100 -
C03 73 72 -1 77 78 1 82 86 4 87 95 8 100 100 0 100 100 -
C04 71 68 -4 76 79 3 81 94 13 86 100 14 100 100 - 100 100 -
C05 70 68 -2 75 75 0 80 82 2 85 90 5 89 97 8 100 100 -
C06 69 70 1 74 78 4 79 86 7 83 93 9 88 100 12 100 100 -
C07 68 67 -1 73 75 2 77 82 4 82 88 5 87 95 8 100 100 0
C08 67 65 -2 71 72 1 76 79 3 81 85 4 86 91 5 90 100 10
C09 65 68 3 70 72 2 75 79 4 80 86 6 85 92 7 89 100 11
C10 64 65 1 69 73 4 74 78 4 79 85 6 83 92 9 88 100 12
C11 63 68 5 68 74 6 73 77 4 77 78 0 82 85 3 87 92 5
C12 62 65 3 67 71 4 71 77 6 76 85 8 81 93 12 86 96 10
C13 61 60 -1 65 62 -3 70 66 -4 75 70 -6 80 77 -3 85 88 3
C14 60 63 3 64 65 1 69 74 5 74 81 7 79 87 8 83 94 11
C15 58 60 2 63 66 3 68 74 6 73 79 6 77 84 7 82 91 9
C16 57 50 -7 62 60 -2 67 69 2 71 77 5 76 83 7 81 89 8
C17 56 49 -7 61 62 2 65 69 4 70 70 0 75 75 0 80 82 2
C18 55 50 -5 60 57 -3 64 64 0 69 65 -5 74 73 -1 79 81 2
C19 54 51 -3 58 58 -1 63 62 -1 68 70 2 73 77 4 77 86 9
C20 52 49 -3 57 56 -1 62 64 2 67 70 3 71 74 3 76 81 5
C21 51 44 -7 56 58 2 61 64 3 65 72 6 70 76 6 75 83 8
C22 50 44 -6 55 52 -3 60 59 -1 64 70 5 69 76 7 74 81 7
C23 49 40 -9 54 45 -9 58 54 -4 63 62 -2 68 67 -1 73 76 3
C24 48 42 -6 52 48 -5 57 53 -4 62 61 -1 67 68 2 71 75 4
C25 46 45 -1 51 50 -1 56 55 -1 61 63 2 65 68 3 70 73 3
C26 45 39 -7 50 44 -6 55 52 -3 60 60 0 64 67 3 69 72 3
C27 44 36 -8 49 48 -1 54 54 0 58 55 -4 63 63 0 68 68 0
C28 43 34 -9 48 38 -10 52 46 -6 57 53 -5 62 57 -5 67 66 -1
C29 42 39 -3 46 42 -4 51 48 -3 56 55 -1 61 56 -5 65 62 -3
C30 40 43 3 45 45 0 50 49 -1 55 55 0 60 59 -1 64 67 3
C31 39 39 -1 44 48 4 49 58 9 54 63 9 58 70 12 63 74 11
C32 38 41 3 43 45 2 48 50 2 52 51 -2 57 59 2 62 67 5
C33 37 38 1 42 44 2 46 52 6 51 59 7 56 63 7 61 69 8
C34 36 35 -1 40 42 2 45 49 4 50 55 4 55 61 6 60 66 6
C35 35 33 -2 39 40 1 44 46 2 49 52 3 54 55 1 58 64 6
Mean: -2,00 0,05 2,25 3,23 3,88 5,36
Standard Deviation: 3,81 3,60 4,13 4,65 4,52 3,98
The average of the means is 2,0%, which indicates that, in average the works were early 2,0%
in comparison to theory. But the global standard deviations of the Δ is 4.1. This means that
most of the weekly progresses were comprised between -2.1% and -6,1% from theory. This
scale gives an indication on both the volatility and how the weekly targets were reached.
From the above, in this case, it was found that the application of Last Planner® alone was not
sufficient to reduce works volatility and ensure an increase in the work flow stability.
Page 71 / 86
9 EARLY FINDINGS, DISCOVERY & POTENTIAL BENEFITS OF THE
RESEARCH
9.1 DATA SUMMARY AT THIS POINT
The data collected on Action Research on bloc 1 (Part A scheduled in LPS®, Part B scheduled
in CPM) and on bloc 2 (scheduled with takt) have been compiled and figured extracted.
The following table summarizes and emphasizes the figures obtained and scheduling systems
used:
Page 72 / 86
levels of details at both the scheduling time unit (e.g., months, weeks, or days) and the
scheduling window (e.g., one week, four weeks, two months, or whole project duration) under
which the measurements will be made. This constraint will probably raise the need for a
training of the contractors at the earliest stage of the project.
Some contractors with well-defined work packages and few interaction with other trades may
be able to look three months ahead whereas others with changing work packages and more
interaction may only be able to look two weeks ahead. It is then important for all parties to
understand that there is only a small if no benefit at all to look much farther than what their
scheduling ability (reliability) permits.
In the same approach, scheduling with tasks under 1 one day (exception for the milestones) is a
non-sense at this point. The tasks should first be dimensioned as factors of week given that the
site meetings are most often on a weekly basis. The scheduling window must also allow for
flexibility depending on the situation on site and zooms into some particular points or
coordination details can be useful.
Page 73 / 86
showed that there is a need to investigate beyond the classical works coordination tools and
methods. The application of the advanced lean coordination tools Last Planner® System
showed it could be complemented in search for works robustness. Adapting and applying the
takt principle to the construction industry made a new scheduling system arise, based on the
search and the measure of works stability and continuity. Those two concerns dramatically
changed the observation view. The search for the application of Takt Planning on site has
showed encouraging results so far, both in terms of lead time and snag works reduction. The
last phase (block 3) will be an unique opportunity to test the new system from the beginning of
the construction site and measure its impact. From the data collected on site (action research),
following case study generation, conclusions will be drawn and further developments proposed.
10 RESEARCH PLAN: Commenté [SR36]: Lacking content!
10.2 Exploitation of the data from the 3 blocks in Belgium + testing in Mauritius:
end of 2016
Toward the end of 2016, the data collected by the Action Research on the three blocks will be
compiled, sorted and the contours of the case will emerge from the multiple case studies
approach. The application of the system developed by the 3 first block of action research will be
tested on a new site in Mauritus. The same approach will be adopted on this fast track mall-
development project permit comparison after that the differences (both intrinsic and extrinsic)
have been clearly identified. This action research will conducted from March to end of 2016.
10.3 Case writing from the data and findings: 2017- mid 2018 / submitting: end of
2018.
The year 2017 will be devoted to writing the case generated by the 3 blocks in Brussels and
draw conclusions on the duplication of the system from the site in Mauritius to identify further
research opportunity. Also, the development of the cases, the demonstration of the logical path
undertook and the chain of evidence from the theory to practice and between the Action
Research will be further developed to build the PhD Thesis. The PhD should hence be
submitted by the end of 2018 after that the report has been checked and amended to meet all
academic standards.
Page 74 / 86
11 REFERENCES Commenté [SR37]: References!!! No bibliography
Commenté [PDN38]: Indeed.
Aier S, Fischer C (2011) Criteria of progress for information systems design, Journal of Information
Systems and e-Business Management, Vol 9, p 133–172
Alturki A, BandaraW, Gable G (2012) Design science research and the core of information systems,
Ballard, G. (2000) The Last Planner™ System of Production Control. PhD Dissertation, School
Ballard, G. (2000). “Lean Project Delivery System.” White Paper 8, Lean Construction Institute
Baskerville R (2008) What design science is not, European Journal of Information Systems, Vol 17,
p 441–443
Baskerville R, Myers M (2004) Special issue on action research in information systems: making IS
research relevant to practice – foreword, MIS Quarterly, Vol 28 (3), p 329-335
Baskerville R, Pries-Heje J (2010) Explanatory Design Theory, Business & Information Systems
Engineering, Vol 5, p 271-282
Baskerville R, Pries-Heje J, Venable J (2009) Soft design science methodology, in Proceedings of
DESRIST ’09, Malvern
Baskerville R, Wood-Harper T (1998) Diversity in information systems action research methods,
Bernstein (2009) “The Lean Years: A History of the American Worker, 1920-1933” Haymarket
Page 75 / 86
Books,
Berthon, P., Pitt, L., Ewing, M., and Carr, C.L. (2002) Potential research space in MIS: A framework
for envisioning and evaluating research replication, extension, and generation. Information
Systems Research, 13(4), 416-427.
Blum, F., H. (1955). Action Research-A Scientific Approach? Philosophy of Science, 22(1), 1-7
Carr, W. & Kemmis, S. (1986). Becoming critical: Education, knowledge and action
research.
Bohnen,F; Thomas Maschek, Jochen Deuse, Leveling of low volume and high mix production based
on a Group Technology approach, CIRP Journal of Manufacturing Science and Technology
4 (2011) 247–25.
Brodetskaia, I. & Sacks, R. (2007), “Understanding Flow and Micro-Variability in Construction:
Theory and Practice” 15th Annual Conference of the International Group for Lean
Construction. East Lansing, Michigan, USA, 18-20
Bruner J (1964) The course of cognitive growth, American Psychologist, Vol 19, p 1-15 Burstein F,
Gregor S (1999) The systems development or engineering approach to research in
Buchanan, D.A. (2003). Getting the story straight: Illusions and delusions in the organizational
change process. Tamara Journal of Critical Postmodern Organization Science, 2(4), 7-21.
Cassell C, Johnson P (2006) Action research: Explaining the diversity, Human Relations, Vol 59 (6),
p 783–814
Chan, A. (2001). A Quest for Better Construction Quality in Hong Kong. Construction Paper 131,
CIOB Construction Information Quarterly, 3(2): 9-16
Chandler D, Torbert B (2003) Transforming inquiry and action. Interweaving 27 flavors of action
research, Action Research, Vol 1 (2), p 133–152
Checkland, P. and Holwell, S. (1998) Information, systems, and information systems: making sense
of the field. Chichester: Wiley.
Chein I, Cook S W, Harding J (1948) The field of action research, American Psychologist, Vol 3, p
43-50
Cheng, T.C.E., and S. Podolsky, Just-in-Time Manufacturing: An Introduction, Chapman & Hall,
London,1993.
Christian Becker, Armin Scholl, A survey on problems and methods in generalized assembly line
balancing, European Journal of Operational Research 168 (2006) 694–715
Christiansson M-T, Granström K (2012) Sharpening the knowledge domain transfer in practice
research design: The BPM assessment, Systems, Signs & Actions, Vol 6 (1), p 22–45
Clark, P.A. (1972). Action research and organisational change. London: Harper and Row.
Cole R, Purao S, Rossi M, Sein M (2005) Being Proactive: Where Action Research meets Design
Research, Proceedings of the Twenty-Sixth International Conference on Information
Systems, Las Vegas, p 325-336
Colquitt, J.A., and Zapata-Phelan, C. P. (2007). Trends in theory building and theory testing: A five-
decade study of the Academy of Management Journal. Academy of Management Journal,
50(6), 1281- 1303.
Page 76 / 86
Cronen V (2001) Practical theory, practical art, and the pragmatic-systemic account of inquiry,
De Treville, S., Antonakis, J.: Could lean production job design be intrinsically motivating?
Contextual, configurational, and levels-of-analysis issues. Journal of Operations
Management 24, 99--123 (2006) Two distinct theories of production: Lean and Toyota
management system (2012)
De Treville, S.: Disruption, learning, and system improvement in just-in-time manufacturing.
Doctoral dissertation, Harvard University (1987)
Dewey J (1938) Logic: The theory of inquiry, Henry Holt, New York
Dix, A. Finlay, J., Abowd, G.D. & Beale, R. (2004). Human-Computer Interaction. 3rd ed.
Edmondson, A.C., and McManus, S.E. (2007). Methodological fit in management field research.
Acade- my of Management Review, 32(4), 1155-1179.
Elden M. & Chisholm, R.F. (1993). Features of emerging action research. Human Relations,
46(2):275-298.
England: Pearson Education.
Epstein I (2002) Using available clinical information in practice-based research, Social Work in
Health Care, Vol 33 (3-4), p 15-32
Fawaz A. Abdulmalek, Jayant Rajgopal, Analyzing the benefits of lean manufacturing and value
stream mapping via simulation: A process sector case study, Int. J. Production Economics
107 (2007) 223–236.
Feldman M, Orlikowski W (2011) Theorizing practice and practicing theory, Organization Science,
Fiallo, M. and Howell, G. (2012). “Using Production System Design and Takt Time to Improve
Project Performance.” Proceedings of the 20th Annual Conference of the International
Group for Lean Construction, San Diego, CA.
Figueiredo A, Cunha P (2007) Action research and design in information systems: Two faces of a
single coin, in Kock N, (Ed. 2007) Information Systems Action Research. An Applied View
of Emerging Concepts and Methods, Springer
Flood R.L. & Jackson, M.C. (1991). Total systems intervention: A practical face to critical systems
thinking. (In Flood, R.L. & Jackson, M.C., eds. Critical systems thinking: Directed readings.
Chichester: Wiley. p.19-40.)
Fondahl, John W. “The History of Modern Project Management – Precedence Diagramming
Methods: Origins and Early Development”. Project Management Journal. Volume XVIII.
No. 2. June 1987.
Fondahl, John W.(1961), “Non-Computer Approach to the Critical Path Method for the Construction
Industry”, Report #9, Stanford University
Page 77 / 86
Fondahl, John W., “Non-Computer Approach to the Critical Path Method for the Construction
Industry”, Report #9, Stanford University, 1961
Fook J (2002) Theorizing from practice: towards an inclusive approach for social work research,
Frandson, A, Berghede, K. & Tommelein, I. (2013). “Takt Time Planning for Construction of
Exterior Cladding”. Proceedings of the 21st Annual Conference of the International Group
for Lean Construction, Fortaleza, Brazil
Gefen, D., Straub, D.W., and Boudreau, M.C. (2000). Structurational equation modeling and
regression: Guidelines for research practice. Communications of the Association of
Information Systems, 4(7), 1-76.
Ginzburg, Carlo. (1989). “Clues: Roots of an Evidential Paradigm”. In Carlo Ginzburg. Clues,
Myths, and the Historical Method. Baltimore: John Hopkins University Press.
Glaser, B., and Strauss, A. (1967). The discovery of grounded theory: Strategies of qualitative
research. London: Wiedenfeld and Nicholson.
Goldkuhl G (2008) Practical inquiry as action research and beyond, in Proceedings of the 16th
European Conference on Information Systems, Galway
Goldkuhl G (2011) The research practice of practice research: theorizing and situational inquiry,
Goldkuhl G (2012a) From action research to practice research, Australasian Journal of Information
Systems, Vol 17 (2), p 57-78
Goldkuhl G (2012b) Pragmatism vs. interpretivism in qualitative information systems research,
Greenwood, D.J. & Levin, M. (1998). Introduction to action research. Thousand Oaks, Calif.: Sage.
Gregor S, Hevner A (2011) Introduction to the special issue on design science, Journal of
Information Systems and e-Business Management, Vol 9, p 1–9
Gregor S, Jones D (2007) The Anatomy of a Design Theory, Journal of AIS, Vol 8 (5), p 312-335
Hevner A, Chatterjee S (2010) Design research in information systems. Theory and practice,
Springer,
Gregor, S. & Hevner, A.R. (2013). Positioning and Presenting Design Science Research for
Maximum Impact. MIS Quarterly, 37 (2):337-355, June.
Hevner A R, March S T, Park J, Ram S (2004) Design science in information systems research, MIS
Quarterly, Vol 28 (1), p 75-115
Page 78 / 86
Hevner, A.R. & Chatterjee, S. (2010). Design research in information systems. Vol. 22. New York:
Springer.
Hevner, A.R., March, S.T., Park, J. & Ram, S. (2004). Design science in information systems
research. MIS Quarterly, 28 (1):75-105.
Hopp, Spearman (2002) To pull or not to pull: what is the question?, McGraw-Hill, New York.
Hopp, W., and Spearman, M. (2004). "Commissioned Paper To Pull or Not to Pull: What Is the
Question?" Manufacturing & Service Operations Management, 6(2), 133- 148.
Hopp, W.J., Spearman (2000) M.L.: Factory Physics: Foundations of Manufacturing Management.
McGraw-Hill, New York
Hult M, Lennung S-Å (1980) Towards a definition of action research: a note and bibliography,
Iivari, Juhani and Venable, John R., "Action research and design science research - Seemingly
similar but decisively dissimilar" (2009). ECIS 2009 Proceedings. Paper 73. 1.
Jackson, M.C. (1991). Systems methodology for the management sciences. New York, N.Y.:
Plenum.
Järvinen P (2007) Action research is similar to design science, Quality & Quantity, Vol 41, p 37–54
Kock N, (Ed. 2007) Information Systems Action Research. An Applied View of Emerging
Concepts
Järvinen, P. (2004). On research methods. Tampere: Opinpajan kirja.
Jonsson K, Levén P (2012) A relevant issue: Establishing collaborations with multiple practitioners,
Jönsson, S. (1991) Action research. in H.-E. Nissen, H.K. Klein & R. Hirschheim, (Eds.)
Information Systems Research: Contemporary Approaches and Emergent Traditions.
Amsterdam: North-Holland, 371-396.
Junior Muris Lage, Moacir, Godinho, Filho, (2010) Variations of the Kanban system: Literature
review and classification, Int. J. Production Economics 125
Karmarkar, U.S. (1986) “Kanban Systems,” Working Paper Series No. QM8612: Center for Manuf.
and Oper. Mngmt., The Graduate School of Mngmt., The Univ. of Rochester, Jun. (1986).
R.L. (1999). Investigating information systems with action research.
Karmarkar, U.S. and S. Kekre (1989) “Batching policy in kanban systems.” Journal of
Page 79 / 86
Manufacturing Systems, 8:4, 317–328, 1989.
Karmarkar, U.S., (1991) “Push, Pull and Hybrid Control Schemes,” Tijdschrift voor Economie en
Man- agement 26, 345–363, 1991.
Kenley, R. (2005). “Dispelling the Complexity Myth: Founding Lean Construction on Location-
based Planning.” Proceedings of the 13th Annual Conference of the International Group for
Lean Construction, July 19-21, Sydney, Australia
Kimball, G.E.(1998) “General Principles of Inventory Control,” J. Manufacturing and Opns. Mgmt.
1 (1988):1, 119-130.
Kimura, O. and H. Terada, “Design and Analysis of Pull System, A Method of Multi-Stage Pro-
duction Control,” Intl. J. of Prod. Res. 19:3, 241–253, 1981.
Kock N, Lau F (2001) Information Systems Action Research: Serving Two Demanding Masters,
Koskela, L (2000): An exploration towards a production theory and its application to construction.
Doctoral dissertation, Helsinki University of Technology (2000)
Koskela, L., and Vrijhoef, R. "The Prevalent Theory of Construction is a Hindrance for Innovation."
8th Annual Conference of the International Group for Lean Construction, IGLC-8, Brighton,
UK.
Kubr, M. (Ed.) (1986) Management Consulting: A Guide to the Profession 2nd Ed. Geneva:
International Labour Office.
Kuechler B, Vaishnavi V (2012) A Framework for Theory Development in Design Science
Research: Multiple Perspectives, Journal of AIS, Vol 13 (6), pp 395-423
Lau F (1997) A review on the use of action research in information systems studies, In Lee A,
Liebenau J, DeGross J (Eds, 1997), Information Systems and Qualitative Research,
Chapman & Hall, London
Layder, Derek. (1998). Sociological Practice. Linking theory and social research. Thousand Oaks:
Sage.Baxter, P., & Jack, S. (2008). Qualitative case study methodology: Study design and
implementation for novice researchers. The Qualitative Report, 13(4), 544-559.
Lee A (2007) Action is an artifact: What action research and design science offers each other, in
Kock N, (Ed. 2007) Information Systems Action Research. An Applied View of Emerging
Concepts and Methods, Springer
Lester (2000)” Project Management, Planning and Control: Managing Engineering”, Butterworth-
Heinemann
Lewin K (1947) Frontiers in Group Dynamics II. Channels of group life; social planning and action
research, Human Relations, Vol 1 (2), pp. 143-153
Lewin, K. (1947). Frontiers in group dynamics. Human Relations, 1(1):5-41. Lewin, K. (1948).
Resolving social conflicts. New York, N.Y.: Harper.
Li, Y.L. Rong (2009) The reliable design of one-piece flow production system using fuzzy ant
Page 80 / 86
colony Optimization, Computers & Operations Research 36 (2009) 1656 – 1663
Lillrank, P (1995). “The transfer of management innovations from Japan”. Organization Studies,
16(6) 971–989. A set of principles for conducting Critical Research in
Lindgren R (2002) Competence systems, PhD diss, Gothenburg Studies in Informatics 23,
Gothenburg University
Linnik, M., Berghede, K., and Ballard, G. (2013). “An experiment in takt time planning applied to
non-repetitive work.” Proceedings of the 21st Annual Conference of the International Group
for Lean Construction, Fortaleza, Brazil.
Liu, M., Ballard, G., and Ibbs, W. (2011). "Work Flow Variation and Labor Productivity: Case
Study." J. Manage. Eng., 10.1061/(ASCE)ME.1943-5479.0000056, 236-242.
March S T, Smith G F (1995) Design and natural science research in information technology,
Decision Support Systems, Vol 15 (4), pp 251-266
March S, Storey V (2008) Design science in the information systems discipline: An introduction to
the special issue on design science research, MIS Quarterly, Vol 32 (4), p 725-730
Mårtensson P, Lee A (2004) Dialogical action research at Omega corporation, MIS Quarterly, Vol
28 (3), p 507-536
Mathiassen L (2002) Collaborative practice research, Information Technology & People, Vol 15 (4),
p 321-345
Mbugua, L. M., Harris, P., Holt, G. D., and Olomolaiye, P. O (1999). A framework for determining
critical success factors influencing construction business performance. In: Hughes, W. (ed)
Procs. 15Th Annual ARCOM Conference. September 5-7, Reading: ARCOM. 1: 255-264
McDonald T, Van Aken E.M, Rentes A.F, Utilizing simulation to enhance value stream mapping: a
manufacturing, case application. International Journal of Logistics, Research and
Applications 5 (2) 2002 213–232.
Miltenburg J, One-piece flow manufacturing on U-shaped production lines: a tutorial source. IIE
Transactions 33 (2001) 303--21. 2001
Monden Y, Toyota Production System. Industrial Engineering and Management Press, Norcross, GA
1983
Myers, M.D. (1997). Qualitative research in information systems. MIS Quarterly, 21(2):241-242.
Myers, M.D. & Klein H.K. (2011).
Naumann J, Jenkins M (1982) Prototyping: The new paradigm for systems development, MIS
Quarterly, Vol 6 (3), p 29-44
New York
Nicholas, J. (2001) Project Management for Business and Technology. Prentice Hall, New Jersey.
Page 81 / 86
Nordström M, Axelsson K (2011) Practitioners’ motives as a key issue in organizing practice
research collaboration, Systems, Signs & Actions, Vol 5 (1), p 133–146
O’Brien, W.J.; london, K. & Urijhoef, R. (2002).“Construction supply chain modeling: reseach
review and interdisciplinary research agenda”. Proceedings of the 10th Conference of the
International Group for Lean Construction, Gramado, Brazil.
Offerman P, Levina O, Schönherr M, Bub U (2009) Outline of a Design Science Research Process,
Ohno, T. Toyota Production System: Beyond Large Scale Production, Productivity Press, Cam-
bridge, Mass, 1988.
Olawale, Y., and Sun M. (2010). “Cost and time control of construction projects: Inhibiting factors
and mitigating measures in practice.” Construction Management and Economics, 28 (5), 509
– 526.
Organization Studies, Vol 30 (12), p 1309–1327
Orlecky J. (1975) Material Requirement Planning – The New Way of Life in Production & Inventory
Management. New York : Mc Graw Hill.
Orlecky J., Ploss G. and Wight O. (1972), “Structuring the bill of Material” Production & Inventory
Management, V.13, No 4
Orlikowski W J, Iacono C S (2001) Desperately seeking the “IT” in IT research – a call to theorizing
the IT artifact, Information Systems Research, Vol 12 (2), pp 121-134
Pain H (2011) Practice research: what it is and its place in the social work profession, European
Journal of Social Work, Vol 14 (4), p 545-562
Papas N, O’Keefe R, Seltsikas P (2012) The action research vs design science debate: reflections
from an intervention in eGovernment, European Journal of Information Systems, Vol 21 (2),
p 147–159
Patton, Michael Quinn. (1990). Qualitative Evaluation and Research Methods. Thousand Oaks,
London, New Delhi: Sage.
Peffers K, Tuunanen T, Rothenberger M A, Chatterjee S (2007) A design science research
methodology for information systems research, Journal of Management Information
Systems, Vol 24 (3), p 45–77
Peffers, K., Tuunanen, T., Gengler, C.E., Rossi, M., Hui, W., Virtanen, V. (2006). The design
science research process: a model for producing and presenting information systems
research. (In Proceedings of the first international conference on design science research in
information systems and technology (DESRIST 2006)
Pickering A (1995) The mangle of practice. Time, agency & science, University of Chicago Press,
Chicago
Qualitative Social Work, Vol 1 (1), p 79–95
Rahani AR, Muhammad al-Ashraf, Production Flow Analysis through Value Stream Mapping: A
Lean Manufacturing Process Case Study, Procedia Engineering 41 ( 2012 )
Rapoport R N (1970) Three dilemmas in action research, Human Relations, Vol 23 (6), p 499-513
Reckwitz A (2002) Toward a Theory of Social Practices. A Development in Culturalist
Page 82 / 86
Theorizing,
Rapoport, R. (1970) Three dilemmas of action research. Human Relations 23, 499-513. Sanford, N.
(1976) Whatever happened to action research? in A. Clark (Comp.),Experimenting
with Organizational Life: The Action Research Approach. New York: Plenum.
Rapoport, R. N. (1970). Three dilemmas in action research. Human Relations 23: 499–513. Schwarz,
G. (2007). Media Literacy, Graphic Novels and Social Issues. Studies in Media &
Richard L. Baskerville (1997) Dintinguishing research from participative case studies. Journal of
Systems and Information Technology.
Rother, M. & Shook, J. (1998). Learning to See: Value Stream Mapping to Add Value and eliminate
MUDA. Lean Enterprise Institute, Cambridge, MA.
Sacks, R., and Goldin, M. (2007). "Lean Management Model for Construction of High-Rise
Apartment Buildings." Journal of Construction Engineering and Management, 133(5), 374-
384.
Salisbury Forum Group (2011) The Salisbury Statement, Social Work & Society, Vol 9
Santiago, J., and Magallon, D. 2009. "Critical Path Method," in CEE320, Winter 2013, Stanford
University
Schatzki T R, Knorr Cetina K, von Savigny E (Eds, 2001) The practice turn in contemporary theory,
Routledge, London
Schein E (2001) Clinical inquiry/research, in Reason P, Bradbury H (Eds, 2001) Handbook of action
research, Sage, London
Schultze, U. and Leidner D.E. (2002). Studying knowledge management in information systems
research: Discourses and theoretical assumptions. MIS Quarterly, 26(3), 213-242.
Seashore, S. (1976) The design of action research. in A. Clark (Comp.) Experimenting with
Organizational Life: The Action Research Approach. New York: Plenum, 103-117.
Sein M, Henfridsson O, Purao S, Rossi M, Lindgren R (2011) Action design research, MIS
Quarterly, Vol 35 (1), p 37-56
Short, J. & Reeves, T. (2009). The Graphic Novel: a ''Cool'' Format for Communicating to
Generation Y. Business Communication Quarterly, 72 (4):417.
Simon, H.A. 1996. The sciences of the artificial. 3rd Ed. London: The MIT Press.
Simpson B (2009) Pragmatism, Mead and the Practice Turn, Organization Studies, Vol 30 (12), p
1329–1347
Sinclair, Zairi (1995) “Business Process Re-engineering & Management Journal” , ISSN: 1355-
2503, Business Process Management Journal
Sinek, S. (2009) You need to provide a structure that reflects the AR cycles and the teams involved
in these cycles etc need to have been introduced in the methods section.
Sipper, D, Bulfin Jr, R.L. Production: Planning, Control, and Integration, Mcgraw-Hill, New York,
1997.
Social Work & Society, Vol 9
Sonnenberg C, vom Brocke J (2012) Evaluation patterns for design science research artefacts, in
Helfert M, Donnellan B (Eds, 2012) EDSS 2011, CCIS-286, Springer, Berlin
Page 83 / 86
Spear, S., Bowen, H.K.: Decoding the DNA of the Toyota Production System. Harvard Business
Review 77, 97--106 (1999)
Spear, S.J.: Just-in-Time in practice at Toyota: Rules-in-Use for building selfdiagnostic, adaptive
work-systems. Working paper 02-043, Harvard Business School (2002)
Spear, S.J.: The High Velocity Edge, How market leaders leverage operational excellence to beat
the competition. McGraw-Hill, New York (2009)
Spearman, M.L., “Customer Service in Pull Production Systems,” Operations Research, 40 (1992),
948–958.
Spearman, Mark L. and Michael A Zazanis, "Push and Pull Production Systems: Issues and
Comparisons", Op. Res., Vol40(3), pp 521-532, May-June 1992.
Stake, Robert. (1995). The Art of case study Research. Thousand Oaks, London, New Delhi: Sage.
Stake, Robert. (1998). “Case Studies” in: Norman Denzin & Yvonna Lincoln. (eds.): Strategies of
Qualitative Inquiry. Thousand Oaks, London, New Delhi: Sage.
Stevens et al (1996) “Innovation and the Built Environment” in Proceedings of IBEA Conference, 7-
9th October, 2011, London South Bank University.
Stevenson C (2005) Practical inquiry/theory in nursing, Journal of Advanced Nursing, Vol 50 (2),
196–203
Stockton DJ, Ardon-Finch J, Khalil R. Walk cycle design for flexible manpower lines using genetic
algorithms, International Journal of Computer Integrated Manufacturing 18 (2005) 15--26.
Straczynski, J. M., Davis, S., Hope, S., Leigh, R., Siegel, J. and Shuster, J. (2010). Superman - Earth
One. New York: DC Comics
Suddaby, R. (2006). From the editors: What grounded theory is not. Academy of Management
Journal, 49(4), 633-642.
Suermann P, Issa R (2009) Evaluating industry perceptions of building information modelling(BIM)
impact on construction, Journal of Information Technology in Construction (ITcon), Vol. 14,
pg. 574-594,
Sugimori, Y., K. Kusunoki, F. Cho and S. Uchikawa, “Toyota Production System and Kanban
System: Materialization of Just-In-Time and Respect-For-Human System,” Internat. J.
Production
Susman G I, Evered R D (1978) An assessment of the scientific merits of action research,
Susman, G. (1983) Action research: a sociotechnical systems perspective. in G. Morgan, ed. Beyond
Method: Strategies for Social Research. Newbury Park: Sage.
Susman, G. and Evered, R. (1978) An assessment of the scientific merits of action research.
Administrative Science Quarterly 23 (Dec) 582-603.
Page 84 / 86
Susman, G.I. and Evered, R. D. (1978). An assessment of the scientific merits of action research.
Sutton, R.I., and Staw, B.M. (1995). What theory is not. Administrative Science Quarterly, 40(3),
371- 384.
Systems, Signs & Actions, Vol 5 (1), p 7-29
Takim, R, Akintoye, A and Kelly, J (2003) Performance measurement systems in construction. In:
Greenwood, D J (Ed.), 19th Annual ARCOM Conference, 3-5 September 2003, University
of Brighton. Association of Researchers in Construction Management, Vol. 1, 423-32.
Thomas, H. R., Horman, M. J., de Souza, U. E. L., and Zavrski, I. (2002). "Reducing Variability to
Improve Performance as a Lean Construction Principle." ASCE Journal of Construction
Engineering and Management, 128(2), 144-154.
Tommelein, I. D., Riley, D. R., and Howell, G. A. (1999). "Parade Game: Impact of Work Flow
Variability on Trade Performance." ASCE Journal of Construction Engineering and
Management, 125(5), 304-310.
Uggerhøj L (2011) What is Practice Research in Social Work - Definitions, Barriers and
Possibilities,
Ulrich, W. (1983). Critical heuristics of social planning: A new approach to practical philosophy.
Haupt. Bern.
Vaishnavi, V. & Kuechler, W. (2004). Design science research in information systems. Last updated:
October 23, 2013. URL: http://desrist.org/desrist. Date of access: 15 May 2015.
Vaishnavi, V. and Kuechler, W. (2004). Design Research in Information Systems. January 20, 2004,
last updated August 16, 2009. URL: http://desrist.org/design-research-in-information-
systems Accessed 26 April 2010.
van Aken J (2004) Management research based on the paradigm of the design sciences: The quest
for field-tested and grounded technological rules, Journal of Management Studies, Vol 41 (2)
p 221- 246
Veatch, M.H., and L.M. Wein, “Optimal Control of a Two-station Tandem Production/Inventory
System,” Operations Research, 42, 337–350, 1994.
Venable J (2006) The role of theory and theorising in design science research, in Proc of DESRIST
2006, Claremont
Venable J, Pries-Heje J, Baskerville R (2012) A comprehensive framework for evaluation in design
science research, in DESRIST 2012, LNCS 7286, Springer, Berlin
Walsham, G. (2006). Doing interpretive research. European Journal of Information Systems, 15(3),
Page 85 / 86
320- 330.
Walsham, G., and Sahay, S. (1999). GIS for District-Level Administration in India: Problems and
Oppor- tunities. MIS Quarterly, 23(1), 39-66.
Weber, Sandra C., Scheduling Construction Projects: Principles and Practices, Prentice Hall, 2005
Weick, K.E. (1995). What theory is not, theorizing is, Administrative Science Quarterly, 40(3), 385-
390.
Wen-Chyuan Chiang, Timothy L. Urban,The stochastic U-line balancing problem:A heuristic
procedure, European Journal of Operational Research 175 (2006) 1767–1781.
Whittington R (2006) Completing the practice turn in strategy research, Organization Studies, Vol 27
(5), p 613–634
Wieringa R, Morali A (2012) Technical action research as a validation method in information
systems design science, Proceedings DESRIST 2012, LNCS 7286, Springer, Berlin
Winter R (2008) Design science research in Europe, European Journal of Information Systems, Vol
17, p 470–475
Wolf, G. (1996). The wisdom of Saint Marshall, the holy fool. Wired, 4(1):124.
Womack, J.P., D.T. Jones and D. Roos, The Machine That Changed the World: The Story of Lean
Production, HarperCollins Publishers, New York, 1990.
Yin, Robert. (1984/1994). Case Study Research: Design and Methods. Thousand Oaks, London,
New Delhi: Sage.
Zuber-Skerritt (1991) Action research for change and development, Avebury
Page 86 / 86