Implementing Lean Construction

Abstract

Lean thinking is a new way to manage construction. Born in manufacturing, the goals demand a new way to coordinate action, one that is applicable to industries far removed from manufacturing. Implementation requires action be shaped by a deeper understanding of the goals and techniques. This paper explains the implications of the goals and key production principles, and how when taken together they result in a different way to manage construction. Implementing lean in construction then becomes a matter of developing and acting on this new knowledge. Advice on implementation is offered.

Full text

Implementing Lean Construction: Understanding and Action
Proceedings IGLC ‘98
IMPLEMENTING LEAN CONSTRUCTION:
UNDERSTANDING AND ACTION
Greg Howell1 and Glenn Ballard2
ABSTRACT
Lean thinking is a new way to manage construction. Born in manufacturing, the goals
demand a new way to coordinate action, one that is applicable to industries far removed from
manufacturing. Implementation requires action be shaped by a deeper understanding of the
goals and techniques. This paper explains the implications of the goals and key production
principles, and how when taken together they result in a different way to manage
construction. Implementing lean in construction then becomes a matter of developing and
acting on this new knowledge. Advice on implementation is offered.
KEY WORDS
Lean construction, implementation, project management.
1 Adjunct Professor at Boise State and Virginia Tech.; Executive Director of the Lean Construction Institute,
ghowell@micron.net
2 Lecturer at U.C. Berkeley and Research Director for the Lean Construction Institute,
ballard@ce.berkeley.edu

Howell and Ballard
Guaruja, Brazil
INTRODUCTION
Lean thinking is a new way to manage construction. Many people object on first exposure
because lean thinking appears to be the application of a manufacturing technique to
construction. One response to the arguments that “construction is different” is to make
construction more like manufacturing through greater standardization. We take the opposite
view (Ballard and Howell 1998b) as we believe the goals of lean thinking describe the
management of dynamic projects. But objections to lean in construction are hardly a surprise
as lean was indeed developed in manufacturing, and individual tenets appear either already in
practice or incomprehensible.
The goals of lean thinking redefine performance against three dimensions of perfection:
(1) a uniquely custom product, (2) delivered instantly, with (3) nothing in stores. This is an
ideal that maximizes value and minimizes waste. The goals demand a new way to coordinate
action, one that is applicable to industries far removed from manufacturing.
The principles of lean thinking and production: (1) Stopping the Line, (2) Pulling Product
Forward, (3) One-Piece Flow, (4) Synchronize and Align, and (5) Transparency, are
techniques which support the goal. Implementation requires a deeper understanding of the
goals and techniques. Some go astray by comforting themselves that they are already doing
some or most of it, while those who believed going in that lean is about manufacturing want
their money back. In this paper we explain the implications of the goals and key production
principles, and how when taken together they result in a different way to manage
construction. Implementing lean in construction then becomes a matter of developing and
acting on this new knowledge. Simply put, but still only partly comprehensible in current
thinking, lean is a value seeking process that maximizes value and continually redefines
perfection as described above. Moving toward this form of perfection, requires more than a
change in procedure, it requires changing the way we think about and do construction.
This paper first explains lean construction by exploring the essential differences between
lean and current practice, then explores the underlying nature and implications of lean
thinking. The paper closes with implementation advice based on this new understanding.
COMPARING CURRENT PRACTICE WITH LEAN
CURRENT PRACTICE
Construction management is activity or contract centered, with transactional contracts or
assignments defining and balancing the objectives of various participants. Coordination
between organizations or crews is primarily controlled from a central plan that establishes
sequence and determines when an activity will start. Costs, errors, and learning occur within
activities. Cost reduction results from improving productivity, and project duration is
shortened by accelerating activities, or by changing logic to allow concurrent work. Waste is
cost that could have been avoided within the activities, such as rework, or cost due to
extended activity duration along the critical path.

Implementing Lean Construction: Understanding and Action
Proceedings IGLC ‘98
LEAN PRODUCTION
Lean production presents a very different model. Production is managed so that actions are
aligned to produce unique value for the customer. Project duration and cost are considered in
“project-as-production system” terms making concern for project total cost and duration
more important than the cost or duration of any activity. Coordination is accomplished in
general by the central schedule while the details of work flow are managed throughout the
organization by people who are aware of and support project goals (as opposed to activity or
local) performance. Value to the customer and throughput, the movement of information or
materials to completion are the primary objectives. Improvement results from reducing
waste, that is the difference between the current situation and perfection, i.e., meeting
customer unique requirements in zero time with nothing in stores.
Lean thinking forces attention on how value is generated rather than how any one activity
is managed. Where current project management views a project as the combination of
activities, lean thinking views the entire project in production system terms, that is, as if the
project were one large operation. It is difficult to optimize a large production system in
construction (a project) because of the complex interaction between the parts. Lean
production principles, rules for organizing and managing production, certainly work in
manufacturing. Once understood, they also apply in construction, if not directly then in
principle. For example, lacking a production line in the manufacturing sense, we believe that
planning at the assignment level is the place to “stop the line” in construction to assure a
reliable flow of work through assembly on site. Stopping the line in manufacturing prevents
the release of defective work downstream. Construction is directive driven in contrast to
manufacturing which is routing driven. Planning is the place to “stop the line” by assuring no
defective assignments are released downstream (Ballard and Howell 1998a). Stopping the
line reduces uncertainty transmitted downstream making coordination much less difficult.
With this sort of expanded understanding lean appears to apply in any industry segment
or situation. It can even be argued that manufacturing is a special case of construction
because it alone is characterized by multiple copies of the same product. Both construction
and manufacturing require prototyping, that is the design of both product and delivery
process. Manufacturing is the special case because it alone moves beyond prototype to
routine production.
Thus implementing lean production does not require making construction manufacturing
by standardizing products, rather implementation starts by accepting the ideal of perfection
offered by lean and understanding the application of each principle and technique to
construction. Beyond these initial steps lies the development of new principles guiding the
pursuit of the lean goal in the specific conditions of the construction industry. Implementing
lean means adopting a “project-as-production-system” approach to construction, defining the
objective in customer terms, and decentralizing management to maximize throughput and
reduce inventories.
We also believe that implementation of lean in construction should start on projects. Lean
developed on the operating floor and its implications spread throughout the organization and
to suppliers. While processes may be improved anywhere with lean thinking, the project is
always in the value stream of the client and the home office accounting systems rarely are.

Howell and Ballard
Guaruja, Brazil
IMPLICATIONS OF LEAN THINKING AND PRODUCTION
The tenets of lean thinking and lean production are drawn from Lean Thinking (Womack and
Jones 1996).
Lean Thinking
Specify Value By Product
Identify the Value Stream
Make Product Flow
At the Pull of the Customer
While Pursuing Perfection -
Custom product, Zero time delivery, nothing in stores.
Lean Production
1. Stopping the Line
2. Pulling product
3. One Piece Flow
4. Synchronize and Align
5. Transparency
The first tenet, “Specify Value” appears to be part of current practice (and to an extent it is)
but “stopping the line,” controlling work flow by “pull” and one piece flow” do not appear to
be about construction. Designers can claim that their programming and design services
produce specifications of value, just as those working in partnering relations can claim they
are attending to project system issues instead of narrower interests. But these partial (and
usually less complete than supposed) implementations come short as they lack the “pursuit of
perfection” goal and the system focus brought by lean production techniques. On the plus
side, partnering does establish a base level of trust which allows people within a system to
shift their attention to improving at the system level instead of simply defending their
interests. But trust is hard to maintain in the absence of reliable work flow. We believe trust
between people requires system reliability.
It is worth reflecting on how lean thinking coordinates action. Specifying value by
product to the customer shapes all actions around customer requirements. In construction,
specifying value comes before design.
Identifying the value stream, the way value will be realized, establishes when and how
decisions should be made. Mapping the value stream shows when the information necessary
to meet owner requirements will be available and when it is required.
A value stream map is a comprehensive model of the project that reveals issues hidden in
current approaches. Value stream maps can be understood as process flow charts that identify
what action releases work to the next operation. Mapping brings choices to the surface and
raises the possibility of maximizing performance at the project level. Normally maps are
prepared at the project level and then decomposed to better understand how the design of
planning, logistics and operations systems work together to support customer value.
The next principle, “Make it (value) flow” says that value development and therefore
product components should be in constant motion, that is without stopping. This tenet
supports the complimentary goals of zero stores and maximum throughput. Current practice
ignores or accepts large inventories or backlogs as the natural consequence of the
commercial situation. Lean works to eliminate those places where value adding work on
material or information is interrupted. In construction this may mean repackaging work so
that parts of the project can proceed without completion of others and/or assure that resources
are delivered in order required directly to the installation location. We argue that current

Implementing Lean Construction: Understanding and Action
Proceedings IGLC ‘98
construction planning systems are unable to produce a reliable flow of work (Ballard and
Howell 1998a). Buffers or wasted capacity are the natural consequence of unreliable flow.
The current practice of urging speed at every turn results in large buffers because of the
uncertainty associated with supply and use rates. Despite complaints about unreliable
performance by other project participants, pressing for rapid completion is a common
approach because people believe improvements within activities or changing the logical
relationship between operations is the only way to reduce project cost and duration. Under
lean thinking, improvement is possible by reducing uncertainty in work flow thus eliminating
the need for intermediate backlogs. Redesigning the planning system at the assignment level
is the key to assuring reliable work flow and must be an early implementation step.
“At the pull of the customer” means “make only what is sold.” In construction, except in
speculative situations, this is already the norm as regards the facility as a whole. That is not
the case, however, as regards facility components. Extending the concept of customer
backwards from the ultimate user, the rule says that each production unit should make only
what its ‘customer’ production unit needs. The intent is to avoid the waste of overproduction;
i.e., doing work that does not release other work. Consequently, it becomes apparent that
there are three types of inventories that need to be minimized, namely, (1) materials and
design information, (2) labor and its tools, and (3) intermediate work product that is not
being exploited. Attempting to exploit all open work faces is a partially conscious effort to
minimize the latter type of inventory. However, the lack of conscious understanding allows
that technique to be coupled with striving to start each activity as soon as possible. Early start
in CPM terms reflects the urge to provide the most time for each activity, but is not governed
by the intent to select work that releases other work. The concept of pull has different
implications. The idea is that upstream activities should not start sooner than needed to
assure the continuous release of downstream work. Thus “pull” assures flow in contrast to
current construction practice which relies on the schedule to “push” resources. The
application of Just-in-Time supply in construction requires activities to be coordinated by
pulling. Push techniques controlled by the central plan, even in the more stable world of
manufacturing, are unable to time the arrival of resources at the work face with enough
precision to assure a reliable flow. In a sense, the CPM can be understood as a technique for
pushing long lead items into the project parking lot or lay down yard. The resources are then
pulled to their final location by an order from the field.
Pushing resources under even modest conditions of uncertainty means that resources
must arrive on site before needed (often to support a degenerative form of flexibility properly
called “slack resources”), or work may be delayed waiting arrival. Many in construction do
not believe that a just-in-time (or more accurately, justified in time) (Daneshgari 1997)
delivery is possible because they know close coordination is not possible using centralized
scheduling, and cannot conceive any other means of coordinating action. Manufacturers have
faced the same problem and now realize that pull is the only technique that will assure just-
in-time delivery.
“Pulling” has been discussed above. “One piece flow” is the logical extension of pulling
and flow. But what is a piece in construction? This principle is challenging as we tend to
think and package work by trade while ignoring the way work is released. Projects are not

Howell and Ballard
Guaruja, Brazil
one piece—at least the interesting ones aren’t—because the sub-assemblies have different
work content, duration and types of connections with other units.
The rule to synchronize and align applies more easily and directly to construction.
Consider two broad types of application. The first is the flow of materials from suppliers and
fabricators to the site. Synchronizing the sequence and rate of delivery with the sequence and
rate of installation is obviously valuable. Failing to do so causes both installation inefficiency
and large inventories of materials on site. A second application is to a sequence of trades;
e.g., finishing rooms of a hotel. Again, the need is apparent to synchronize the rate and
sequence of work done by each trade. Otherwise, there are rooms ready to work lying empty,
or crews are standing idle waiting on those ahead of them in the chain.
“Transparency” means that the state of the system is made visible to people making
decisions throughout the production system so that they will take decisions locally in support
of systems objectives. Transparency implies decentralized decision making which in turn,
allows people to coordinate through mutual adjustment.
NATURE OF THE CHANGE
We believe that lean production is a new way to coordinate action that rests on a new mental
model. Aspects of this model appear to be both sensible—Specify Value, and strange—Stop
the Line. Problems solved by lean such as unreliable flow have been recognized as problems
in construction, but no solutions are offered by common sense. Partial implementation of
lean thinking, particularly organizational or relational aspects, have been developed and used
on complex, uncertain and quick projects.
These dynamic projects are the ones most likely to fail when only traditional approaches
are used. Lacking a comprehensive underlying theory, efforts such as partnering are little
more than patches. Of course these patches are helpful when they lay the ground work for
larger steps, but alone they do not fix the underlying problem, the design of the production
system.
Lean thinking takes a project-as-production-system view as opposed to the current
activity or contract centered perspective (Ballard and Howell 1998a). In this way lean
embraces the uncertainty and complexity of construction and aims to assure that local
improvements actually lead to results at the project level. But the shift from activity to
system is a change at the mental model level that has important implications for
implementation.
Peter Senge (Senge et al. 1994) says there are four levels of change: events, behavior,
system (here used as “cost control system” or “logistic system”), and mental model. The
industry is more or less adept at changes at the system or procedure level, and spends great
energy trying to change behavior. Improvements at these levels are can be characterized as
problem/solution, that is, there is some problem with the current procedure so we solve it.
Change at the mental model level is different because each opens a new opportunity.
Companies beginning to implement lean report an unexpected phenomenon; each change
creates the opportunity for more and often larger improvements. Thus for a time it appears
that the amount of change possible increases with each step of implementation.

Implementing Lean Construction: Understanding and Action
Proceedings IGLC ‘98
Peter Senge offers general advice for managing a change at the mental model level in The
Fifth Discipline Fieldbook. Dr. James Womack makes specific recommendations in Lean
Thinking (Womack and Jones 1996) which are named and considered below.
Find a change agent: This is a person who makes things happen. Action is a must
because it develops a new mind, and gives people confidence that changes can be
made. Above all the change agent must have the courage to cause principle driven
action, and to stand against attempts to dilute lean. For example, lean requires
decentralized decision making. Decentralized decision making will be fiercely
resisted by those who equate it to a loss in their power. Similar problems arise when a
pull mechanism is installed to manage materials.
Get the knowledge: This means get enough knowledge to start and then continue to
gain knowledge through implementation. A focus on system thinking is in order
because it is uncommon in construction. Do not ignore the need to understand basic
production management. Few universities offer courses in production management in
construction and instead still teach courses in productivity improvement. Be cautious
of such courses as a focus on productivity at the activity level and productivity
reporting means you are missing the point of system performance. We join Dr.
Womack in his advice that both change agents and senior managers must master lean
thinking.
Find a lever by seizing a crisis or creating one: In some sense, every construction
project is one crisis after another so this should not be difficult. We have had two
different experiences worth considering. In one we worked on a large project with a
series of subcontractors selected in the order of closest to be terminated. This is in
line with one approach suggested by Dr. Womack of starting with a sub organization
already in crisis. This approach can be successful but it makes the experience difficult
to generalize to areas not in crisis. Worse, improvements which occur in crisis
situations are often attributed to a failure to do it right on the part of one or more
people. If people see the crisis in activity terms they will assume the problem is due
to someone not doing their job, and not to a failure in the underlying systems. As a
result, people learn little.
We propose an opposite approach: Start on the best projects. Application of lean thinking
from the start with a good project team, or in the middle of a well run project underway
reveals the weakness of current systems and the power of lean thinking. Mapping the project
value stream usually provokes remarkable improvements as people see redundancy and
waste as the map unfolds. In a sense, starting on the best is similar to Dr. Womack’s
suggestion to cause a crisis.
Forget grand strategy for the moment: While we agree that worrying why you
were born is not useful, there are fundamental strategic implications in lean thinking.
However, we believe that construction companies will want to start with work on
projects. Because lean is a systems approach, finding opportunities and bottlenecks
on projects first will reveal inadequacies in other areas which must be addressed to

Howell and Ballard
Guaruja, Brazil
assure maximum lean implementation at the project level. Projects are where we
serve customers and it is on projects that we must improve.
Having said that, there is no reason to exclude home office or other functions from lean
implementation. But we believe designing a supporting system to mesh with an inadequate
project planning system is starting at the wrong end.
Map your value streams: Dr. Womack correctly cautions against relying on
reengineering as it is often too narrow and lacks the sense of perfection that shapes
lean thinking. In construction, the value stream inevitably involves other
organizations and seeing them on the map is the first step to developing their
participation.
Begin as soon as possible with an important and visible activity: We believe the
production planning system and the way assignments are made is the place to start.
No other activity is as important because changes in planning performance have both
an immediate impact and reliable planning demands a lean response throughout the
organization. Measuring and improving planning performance is a new idea in
construction; it causes immediate improvement and reveals gaps in common sense.
Demand immediate results: We agree but with a caution. If you demand the wrong
results you will destroy the lean initiative. Reduced costs within an activity are the
wrong results to ask for, just as more rapid completion of tasks. Better to ask for
immediate results in planning performance and for how such improvements translate
into reduced backlogs between activities. Throughput measures are another good area
for consideration. How much time has been cut from the time of order to delivery and
delivery to installation?
Asking to review before and after value stream maps and/or changes in procedure are another
way to assure changes are happening.
We support Dr. Womack’s advice to avoid benchmarking, and are believe the recent
resurgence of work sampling is a mistake. Data gathered by work sampling is aimed at
improving performance of workers or crews by increasing the time working. But increasing
labor utilization is not the same as increasing throughput. The pressure to increase direct
work is likely to cause work to be released downstream in a less reliable flow as people
adjust their behavior to improve their numbers. Making the numbers look good adds no value
and provides measures of waste which are misleading. Such data is of no use in redesigning
planning and logistic systems to increase throughput. Finally, work sampling maintains
central control. In lean organizations, teams collect and act on their own data in support of
larger objectives.
As soon as you have momentum, expand your scope: This advice is always worth
remembering. The rate of change is the most important lean metric. If your effort is
not causing action in all corners of the organization, if people aren’t finding and
making changes on their own, you are off track. Implementing lean always brings
more opportunities to the surface.

Implementing Lean Construction: Understanding and Action
Proceedings IGLC ‘98
OUR ADVICE
Implementing lean thinking will lead to change in almost every aspect of project and
company management. No one step by step guide can be offered because change at the
mental model level is a developmental process. Each principle driven action will reveal new
opportunities hidden because people simply could not think in ways that made the change
possible. Thinking causes action, action causes deep learning, and learning causes new
thinking.
Take care to develop systems thinking, understand the difficulty of changing mental
models, expect deep resistance in yourself and others to decentralized decision making, and
learn about production management.
On the commercial side, begin to form long term alliances with like minds along the
value stream. As a citizen of the industry, spend time and effort developing lean thinking in
others, even competitors. More than one firm changing to lean has come up against the
inability of their suppliers to support lean projects. Once on the way to becoming lean, these
suppliers become almost schizophrenic as part of their customers work one way and the
others lean. Expand your view to consider the construction industry as a system and work to
improve performance everywhere. If this advice sounds wrong, it probably means you have
not confronted the depth of opportunity and change lean offers you and your organization.
REFERENCES
Ballard, G. and Howell, G. (1998a). “Shielding Production: An Essential Step in Production
Control” ASCE, J. Constr. Engrg. and Mgmt., 124 (1) 18-24.
Ballard, G. and Howell, G. (1998b). “What kind of Production is Construction?” Proc. 6th
Annual Conf. Int’l. Group for Lean Construction, Sao Paulo Brazil, August 13-15.
Daneshgari, P. (1997). The Chase. Unpublished manuscript. Flint, Michigan.
Senge, P., Kleiner, A., Roberts, C., Ross, R., and Smith, B. (1994). The Fifth Discipline
Fieldbook. Doubleday.
Womack, J. and Jones, D. (1996). Lean Thinking. Simon & Schuster.