Content uploaded by Bryant C. Mitchell
Author content
All content in this area was uploaded by Bryant C. Mitchell on Jun 15, 2017
Content may be subject to copyright.
What Should Be Changed?
A comparison of cause and effect diagrams
and current reality trees shows which will bring
optimum results when making improvements
by
Lawrence D. Fredendall, J. Wayne Patterson,
Christoph Lenhartz and Bryant C. Mitchell
THEORY OF CONSTRAINTS
50
I
JANUARY 2002
I
WWW.ASQ.ORG
HEN MAKING IMPROVEMENTS,
a manager must answer three fun-
damental questions:
1. What should I change?
2. What should I change to?
3. How do I implement this change?
One management system many managers use to
answer these questions and facilitate change is total
quality management (TQM). Yet TQM programs
sometimes fail to achieve the desired benefits
because TQM does not explicitly identify the most
crucial area to change.
The theory of constraints (TOC), which is also
called constraints management, is a set of tools to
examine the entire system for continuous improve-
ment.1, 2 The TOC uses five tools (the current reality
tree, conflict resolution diagram, future reality tree,
prerequisite tree and transition tree) in its ongoing
improvement process. These tools answer those
three fundamental questions.3
One of the five tools, the current reality tree
(CRT), helps identify what to change. We will com-
pare CRTs to cause and effect diagrams (CEDs), a
common TQM tool.
Some comparisons between TQM and TOC
already exist. Cohen4explains the latter’s similari-
ties to Deming’s problem solving approach.
Gardiner and his co-authors5provide case examples
in which combining TOC with TQM produced the
focus needed to improve performance rapidly.
Dettmer6uses the constraints tools to identify
which TQM tool should be used.
Pasquarella, Mitchell and Suerken7briefly exam-
ine how the five constraints tools could replace the
common TQM tools. They compare CEDs to CRTs
and demonstrate some of the differences. However,
they do not give a detailed example of how to con-
struct the CRT or completely explain the advan-
tages or disadvantages of either tool.
We offer a simple example comparing the CED to
the CRT, which illustrates the differences in the con-
struction of each. We then apply each tool to an
actual business problem and explain the advan-
tages and disadvantages of each tool.
W
The CED
Constructing a CED is an exercise in structured
brainstorming. The objective is to ensure the CED
contains the true cause(s) of a problem. Ishikawa
lists three steps for making a CED:8
1. Determine the quality characteristic to be
improved and controlled, write it in a block and
draw an arrow pointing to the block.
2. Write the main factors that may be causing the
quality characteristic at the end of a branch arrow
drawn to the main arrow.
3. Next to each branch arrow, write the factors that
contribute to the main factor.
Brainstorming for the CED ensures all potential
root causes of a problem are captured in a rough
causal sequence. As part of this brainstorming, the
facilitator asks “why” five times. The answer to
each “why” is attached to a subbranch.
For example, in Figure 1 the cause and effect logic
starts with Ishikawa’s step one, in this case “1. The
car will not start.”
An example of step 2 (writing the main factors on
a branch arrow) is “2. The driver.” This is written
on the large arrow to explain why “1. The car will
not start.” An example of step 3 (recording the
detailed factors) is shown in “3. Left lights on previ-
ous night,” which explains why the driver was
responsible for the car’s not starting.
“4. Not yet dark when parking” explains why (3)
the lights were left on the previous night. The state-
ment “5. Lights on due to light drizzle” answers
why the lights were left on if it was not yet dark
when parking (4).
Because the use of a CED is well understood by
most readers, it is not explained in detail here.
Instead, we will focus on the use of a current reality
tree to analyze this same problem.
The CRT
The CRT employs cause and effect logic to deter-
mine what to change by identifying the root causes
or core problems.
Another purpose of the CRT, whether developed
by an individual or by a team, is to create a consensus
QUALIT Y PROGRESS
I
JANUARY 2002
I
51
Cause and Effect Diagram of Car Not Starting
FIGURE 1
Out of gas.
Running low and
did not fill.
Had too little time.
Starter broken.
Bendix spring broken.
Solenoid shorted out.
Battery too weak.
Engine will not turn over.
Radio will not turn on.
Equipment
2. The driver.
1. The car
will not start.
Did not follow start procedure
(transmission not in park).
3. Lights left on previous night.
4. Not yet dark when parking.
5. Lights turned on due to light drizzle.
Gasoline
Lights left on and no
automatic shutoff.
among those involved with a problem. To do
this, the CRT uses a set of clearly defined rules for its
construction, validation and interpretation.9These
rules ensure the CRT increases communication and
understanding of the problem. The construction and
validation of the CRT are illustrated using the exam-
ple of the car’s not starting.
The steps to create a CRT are shown in Table 1
(p. 53). The first step for any group solving a problem
is to identify the group’s span of control. The second
step is to list the symptoms or undesirable effects
(UDEs) of the problem (in other words, why the
group perceives a problem).
A set of UDEs of the failure of a car to start in the
morning is shown in Figure 2. Similar to the CED, this
initial list of UDEs may be generated through brain-
storming. To simplify discussion of the UDEs, each
undesirable effect is always numbered. For example,
the UDE in Figure 2 that states, “The car will not
start,” is labeled with a 500.
The number of the UDE is for identification and is
assigned arbitrarily. But there are some useful num-
bering guidelines in Dettmer.10 In examining a UDE,
we must agree it is real. Then we begin to address
whether it is a problem or simply an effect arising
from one or more underlying causes.
The third step is to illustrate the relationships
between UDEs by connecting them with arrows, as
shown in Figure 3 (p. 54). For example, an arrow runs
from 600 to 500 to illustrate 600 has caused 500.
The CRT is always read using “if ... then ...” state-
ments. In Figure 3, the CRT would be read in the fol-
lowing way: “If 600, the engine will not turn over,
then 500, the car will not start.”
Initially, there may not be a clear connection
between all the UDEs (for example, 400 and 300 in
Figure 2), so they are left without connecting arrows.
The fourth step validates the UDEs by adding other
entities to the original list of UDEs so there is a valid,
logical relationship, as shown in Figure 4 (p. 54). For
example, the entity “250. Battery is dead” was added
to explain UDEs 400, 600 and 300. Entities 200, 120 and
110 were added to explain why the battery is dead.
Additional entities are added as necessary or until the
span of control is reached.
At each step, it is important to allow everyone
involved the opportunity to critique the tree using the
eight questions shown in Table 2. These eight ques-
tions create a communications protocol that focuses
everyone’s attention on the process and helps avoid
wasted time and counterproductive behaviors such as
finger pointing or killing the messenger.
Question number 1 in Table 2 is very common.
Someone reading the tree asks for clarity about one
statement. The author then explains or clarifies the
statement. For example, if someone asks for clarity
about 200 to 250 in Figure 4 (p. 54), the author of these
entities may explain, “If 200, the lights were left on all
night, then the power was drained from the battery
and then 250, the battery is dead.”
There might be consensus to add another entity,
such as “the power was drained from the battery,” or
the decision may be to leave the tree unchanged. In a
similar fashion the other questions may be addressed
to remove any uncertainty about what the tree or any
portion of it actually means.
This validation step for the CRT helps ensure con-
sensus and promotes agreement and buy-in to deal
with the underlying problem. Dettmer11 offers a good
discussion of this process.
Step 5 identifies the root cause of the car’s not start-
ing. Typically, the root cause is at the bottom or near
52
I
JANUARY 2002
I
WWW.ASQ.ORG
WHAT SHOULD BE CHANGED?
Initial Undesirable Effects (UDEs)
FIGURE 2
500. The car will
not start.
400. The lights will
not come on.
600. The engine
will not turn over.
300. The radio will
not work.
the bottom of the CRT. In this case entities 120 and 110
would be logical candidates for the root cause because
they are at the bottom. After examining these entities,
you may state the root cause is “110. The driver did
not turn the car lights off before exiting the car.” This
root cause should, of course, be the answer to the first
fundamental question of management: What should
we change?
A business application
The application of the CRT and CED to the car
example illustrates how each functions. Lenhartz pro-
vides an example of a CRT created to examine the
problems a company was experiencing with business
process reengineering (BPR).12
The managers of the company performed this
analysis before beginning another BPR project. The
goal of the analysis was to identify the root causes of
BPR failures—what to change in BPR projects. We use
that example to contrast the CRT and CED.
The first step in creating the CRT was to list the
UDEs of prior BPR projects. This list is shown in Table
3. The second step was to create the CRT shown in
Figures 5a (p. 56) and 5b (p. 58). The original UDEs are
shown with shadows. Beginning with entity 110 in
Figure 5a, the CRT is read, “If 110, senior management
is measured on cost reduction, then 120, the business
strategy emphasizes cutting costs, not increasing
throughput. If 120, then 130, senior management
views cost cutting as the fastest means to improve.”
The logical “and” is shown as an ellipse that joins
two arrows in Figure 5a. This is read, “If 125, BPR
strategy is aimed at cutting costs, and if 130, senior
management views cost cutting as the fastest means to
improve, then 135, senior management expects fast
results from BPR projects.”
Often a CRT cannot fit on one page. When a causal
connection exists between entities on different pages,
the entity on the next page is shown with a circle. For
example, in Figure 5a, entity 125 leads to entity 220 in
Figure 5b.
For convenience, the originating entity is repeated
on the next page. For example, in Figure 5b the causal
connection is read, “If 125, BPR strategy is aimed at
cutting costs, then 220, BPR often is not integrated
with the organization’s strategy and planning.” Notice
entity 220 was shown in Figure 5a in a circle, but in
Figure 5b, entity 125 is repeated in full since it is the
originating entity.
Identifying the root causes
After the CRT is constructed, as in the simpler
example shown earlier, it is clarified by asking the
eight questions in Table 2. When these questions are
QUALIT Y PROGRESS
I
JANUARY 2002
I
53
Steps To Create a
Current Reality Tree
TABLE 1
Step
1.
2.
3.
4.
5.
Description
Identify the span of control and sphere of influence.
Create a list of undesirable effects (UDEs).
Develop causal connections between the UDEs.
Build a cause and effect chain to validate UDEs.
Identify root causes and the core problem.
Adapted from William H. Dettmer, Goldratt’s Theory of Constraints: A Systems
Approach to Continuous Improvement (Milwaukee: ASQ Quality Press, 1997).
Categories of Legitimate
Reservation
TABLE 2
Question
1.
2.
3.
4.
5.
6.
7.
8.
Description
Clarity—Is the meaning of the statement clear?
Entity existence—Is this a valid, complete and structural-
ly sound statement? Questions about this are often
resolved by ensuring that it is a single idea expressed in
a grammatically correct sentence.
Causality—Does the cause actually lead to the stated
effect?
Cause insufficiency—Is stated cause sufficient to
produce effect?
Additional cause—Are there additional causes that will
produce the effect?
Cause-effect reversal—Does the stated cause create the
effect, or vice versa?
Predicted effect existence—If the stated cause and effect
relationship is true, are there other effects that should
also be present?
Tautology—Is this circular logic?
Adapted from William H. Dettmer, Goldratt’s Theory of Constraints: A Systems Approach
to Continuous Improvement (Milwaukee: ASQ Quality Press, 1997).
Undesirable Effects of Business
Process Reengineering (BPR)
Projects
TABLE 3
170.
175.
180.
190.
275.
290.
Our BPR projects don’t yield lasting systems improvements.
BPR projects often take longer than expected.
Many BPR projects do not result in overall system
improvement.
Top management tends to lose interest in BPR projects.
Some employees resent the way they are treated in BPR
projects.
Many employees don’t support BPR projects.
54
I
JANUARY 2002
I
WWW.ASQ.ORG
WHAT SHOULD BE CHANGED?
Initial Connection for the Current Reality Tree for Starting the Car
FIGURE 3
500. The car will
not start.
400. The lights will
not come on.
600. The engine
will not turn over.
300. The radio will
not work.
Complete Current Reality Tree
FIGURE 4
500. The car will
not start.
400. The lights will
not come on.
600. The engine
will not turn over.
250. The battery
is dead.
120. The car lights
did not shut
off automatically.
110. The driver did
not turn the car
lights off before
exiting the car.
200. The lights
were left on
all night.
300. The radio will
not work.
QUALIT Y PROGRESS
I
JANUARY 2002
I
55
answered for each entity, the CRT is examined to iden-
tify the root causes of the problems.
In this example, the potential core problems are
identified as entities 200, 215 and 265. Entity 200
reads, “BPR is not constraints focused,” which means
the BPR does not ensure the business process being
changed is the entity keeping the system from achiev-
ing a higher level of throughput.
Entity 215, which reads, “There is no systems
approach to BPR,” means BPR will be directed only at
improvements in the business process being reengi-
neered and ignore interdependencies with other com-
pany systems. Entity 265 reads, “BPR is not an
ongoing process of improvement,” and means BPR is
usually a one-time effort that does not create a system
to change the process as the environment changes.
After examining this limited set of potential core
problems, the company decided two themes were pre-
sent. First, entities 215 and 200 both lead to UDE 180,
that BPR projects do not result in overall system
improvement. Entity 265 leads to the UDE 170, that
BPR projects don’t yield lasting improvement.
Changing entity 200 would require limited change
to the BPR system but would ensure the BPR project
resulted in an improvement to the system. This is seen
as achievable. Further, it is believed achieving real
improvement would keep top management interested
in system improvement (see UDE 190). With the selec-
tion of UDE 200 as the root cause of this problem, the
answer to the fundamental question of what to change
is the fact that BPR is not constraints focused.
Showing the differences
To illustrate the differences between the CED and the
CRT, a CED to analyze why 70% plus of BPR projects
fail is illustrated in Figure 6 (p. 57). Each entry on the
cause and effect diagram includes the number that the
same entity has in the CRT in either Figure 5a or 5b.
Those entities on the CED more than once are in bold
type.
It is clear the cause and effect diagram maintains
much of the causal logic given in the CRT. For exam-
ple, entity 190 in Figure 5a is shown on the first
branch (management) of the CED in Figure 6. The
sequence of entities below 190 is the same on both the
CED and CRT.
The CRT shows the “if ... then ...” logic more pre-
cisely than the CED. In Figure 5a, the logic would
read: “If 160, some middle managers fear loss of influ-
ence, then 165, middle management does not support
BPR projects enthusiastically.” Likewise, “if 165 and
150, middle managers play an important role in man-
aging change, then 175, BPR projects often take longer
than expected.”
The CED does not easily allow the use of “and,” but
both the CED and CRT allow the presentation of mul-
tiple causes of an entity. In this example, in addition to
175 causing 190, it is possible for 180 to cause 190.
Lenhartz13 reports the presentation and discussion
using the categories of legitimate reservation of this
CRT forced this company to carefully develop a BRP
plan that examined the entire system, with focus on
the constraints.
By addressing the root cause of failure, the compa-
ny was able to successfully implement its BPR project.
The creation of a CRT by the team that was to lead the
BPR effort improved communication between all the
different individuals involved and helped convince
top management to avoid isolated BPR effort, and
instead create a process of continuous improvement
focused on the constraints.
In addition, the generic CRT and that of the compa-
ny’s current situation were used to communicate with
workers and middle managers. This helped to prevent
morale threatening rumors and ensure middle man-
agement support.
The two differences
There are two evident differences between CEDs
and CRTs. First, the physical layout of the data is dif-
ferent. The CRT is a vertical tree reading from bottom
to top, while the CED is a horizontal diagram reading
from right to left (for example, “The car will not start
because of the driver.”). The second difference is the
CRT has fewer entities than the CED.
These two differences simplify the determination of
what to change and what to change to. Because the
CRT allows loops in its process, one entity can rein-
force the action of a second entity.
Turner14 points out the CED does not consider non-
linear relationships. Also, while the cause and effect
diagram may contain the root cause, its structure does
not help the user easily identify the root cause.
To be highly effective, the CED probably requires
more skill from the facilitator. The art of facilitating
the development of a CED is being able to ask “why”
enough times that the root cause of the problem is on
the board. The CRT is more mechanical than the CED
and requires less art. It builds the structure in a logical
manner so the root cause is either on or near the bot-
tom of the tree.
While the CRT provides a more systematic, logical
analysis of a problem than the CED does, and this
analysis serves as a communication protocol that
assists in implementing changes, many find strict
application of the “if ... then ...” logic intimidating.
And they may resent having to phrase their sugges-
tions and objections in terms of the eight questions in
56
I
JANUARY 2002
I
WWW.ASQ.ORG
WHAT SHOULD BE CHANGED?
Complete Current Reality Tree of Business Process Reengineering Failures
FIGURE 5a
190. Top management tends
to lose interest in BPR projects.
170. Our BPR projects
don't yield lasting systems
improvement. 175. BPR projects often
take longer than expected.
180. Many BPR projects
do not result in overall
system improvement.
150. Middle managers
play an important role
in managing change.
165. Middle management
does not support BPR
projects enthusiastically.
160. Some middle managers
fear loss of influence.
155. Planning
for change managment
is often neglected.
140. Many BPR projects
are conducted under
extreme time pressure.
145. Planning for change
management is time
consuming.
135. Senior management
expects fast results
from BPR projects.
125. BPR strategy is
aimed at cutting costs.
220
130. Senior management
views cost cutting as the
fastest means to improve.
120. The business strategy
emphasizes cutting costs,
not increasing throughput.
110. Senior management is
measured on cost reduction.
270
QUALIT Y PROGRESS
I
JANUARY 2002
I
57
Table 2 (p. 53). The role of the facilitator in the con-
struction of the CRT is to reduce the participants’ anx-
iety about their ability to contribute.
Still, most people probably view the CED as easy
to use. It can be constructed quickly by a facilitator
working with a team and probably requires less
training than the CRT. But the CED does not present
the underlying logic of the problem as clearly as the
CRT. Further, the CED structure does not quickly
identify the root cause, while the CRT visually
points to it.
Both the CRT and CED serve valuable functions in
the process of ongoing improvement. In fact, it is pos-
sible to use them in tandem. A cross functional team
meeting to solve a problem may start by brainstorm-
ing and creating a CED.
The team or some members of it could use input
from the CED as a list of undesirable effects when cre-
ating a CRT. In this way, everyone would have input
into the creation of the CRT through involvement in
the creation of the CED. Those on the team with a
higher level of comfort using the CRT could then con-
duct additional analysis and submit it back to the
team for action.
A major advantage of the CRT over the CED is
everything in the CRT is structured to identify what to
change. This structure can help focus team members
on identifying the root cause of a problem, which may
then lead to a breakthrough solution.
REFERENCES
1. Eliyahu M. Goldratt and Jeff Cox, The Goal, second edition
(Croton-on-Hudson, NY: North River Press Inc., 1992), p. 333.
2. William H. Dettmer, Goldratt’s Theory of Constraints: A
Systems Approach to Continuous Improvement (Milwaukee: ASQ
Quality Press, 1997), p. 11.
3. Eliyahu M. Goldratt, What is This Thing Called the Theory of
Constraints? APICS, The Performance Advantage, Alexandria,
VA, 1993, pp. 18-20.
4 Oded Cohen, “Ten Steps Ahead: Combining Deming and the
TOC,” APICS Constraints Management Technical Conference
Proceedings, Tampa, FL, March 13-14, 2000, pp. 14-23.
5. Stanley C. Gardiner, John H. Blackstone Jr. and Larraine R.
Gardiner, “The Evolution of the Theory of Constraints,”
Cause-and-Effect Diagram of Why Business Process Reengineering (BPR) Projects Fail
FIGURE 6
70% of BPR
projects fail.
125. BPR strategy is
aimed at cutting costs.
190. Top management tends
to lose interest in BPR projects.
190. Top management tends
to lose interest in BPR projects.
ManagementMeasurement
175. BPR projects often take
longer than expected.
170. Our BPR projects don't yield
lasting systems improvements.
130. Senior management
views cost cutting as the
fastest means to improve.
235. BPR is aimed at
local improvements.
225. Most BPR projects
examine only a core
business process. 245. The system
constraint has not
been improved.
155. Planning for change
management is often neglected.
140. Many BPR projects are
conducted under extreme
time pressure.
135. Senior management
expects fast results from
BPR projects.
215. There is no
systems approach
to BPR.
120. The business strategy
emphasizes cutting costs, not
increasing throughput.
110. Senior management
is measured on cost reduction.
125. BPR strategy is
aimed at cutting costs.
280. BPR is a
one-time activity.
290. Many employees
don't support BPR projects.
275. Some employees
resent the way they are
treated during BPR projects.
165. Middle management
does not support BPR
projects enthusiastically.
160. Some middle
managers fear loss
of influence.
180. Many BPR projects do
not result in overall
system improvement.
180. Many BPR projects do
not result in overall
system improvement.
Process
270. Many employees
fear losing their jobs.
275. Some employees
resent the way they are
treated during BPR projects.
290. Many employees
don't support BPR projects.
Employees
58
I
JANUARY 2002
I
WWW.ASQ.ORG
WHAT SHOULD BE CHANGED?
Current Reality Tree of Business Process Reengineering Failures
FIGURE 5b
170. Our BPR projects
don't yield lasting
improvemnt.
285. BPR solutions
deteriorate over time.
280. BPR is a one-
time activity.
285. The system's
environment changes.
290. Many employees don't
support BPR projects.
295. Sustained improvement
requires employee support.
275. Some employees resent
the way they are treated
during BPR projects.
255. Employees are
viewed as costs.
125. BPR strategy is
aimed at cutting costs.
180. Many BPR projects
do not result in overall
system improvement.
240. A system's overall
performance is often
limited by a constraint.
245. The system contraint
has not been improved.
230. Often BPR does not
address the system constraint.
205. The constraint
is not identified.
200. BPR is not
constraints focused.
220. Often BPR is not integrated
with the organization's
strategy and planning.
215. There is no systems
approach to BPR.
235. BPR is aimed at
local improvements.
225. Most BPR projects
examine only a core
business process.
270. Many employees
fear losing their jobs.
265. BPR is not an ongoing
process of improvement.
190
190
Industrial Management, Vol. 36, No. 3, 1994, pp. 13-17.
6. William H. Dettmer, Goldratt’s Theory of Constraints: A
Systems Approach to Continuous Improvement, see reference 2,
p. 10-12.
7. Marcia Pasquarella, Bob Mitchell and Kathy Suerken, “A
Comparison of Thinking Process and Total Quality
Management Tools,” APICS Constraints Management
Symposium Proceedings, Denver, 1997, pp. 59-64.
8. Kaoru Ishikawa, Guide to Quality Control (White Plains,
NY: Quality Resources, 1982), pp. 18-20.
9. Marcia Pasquarella, et al., “A Comparison of Thinking
Process and Total Quality Management Tools,” see reference 7.
10. William H. Dettmer, Goldratt’s Theory of Constraints: A
Systems Approach to Continuous Improvement,see reference 2.
11. Ibid.
12. Christoph Lenhartz, “Using Theory of Constraints
Thinking Process Tools To Enhance the Effectiveness of
Business System Optimization,” APICS: Constraints
Management Technical Conference Proceedings, Tampa, FL,
March 13-14, 2000, pp. 42-46.
13. Ibid.
14. Ronald E. Turner, “CEDs Alone Don’t Tell the Whole
Story,” Quality Progress, January 1997, pp. 53-56.
LAWRENCE D. FREDENDALL is an associate professor of man-
agement at Clemson University. He earned a doctorate in opera-
tions management from Michigan State University. Fredendall is
a member of ASQ.
J. WAYNE PATTERSON is a professor of industrial management
at Clemson University. He received his doctorate in business
administration from the University of Arkansas.
CHRISTOPH LENHARTZ is a senior consultant with Logo IT
Consulting in Moers, Germany, where he focuses on business
system optimization and supply chain management issues. He
holds a master’s degree in business administration from Clemson
University and a Diplom-Kaufmann-degree from the University
of Essen, Germany. Lenhartz is certified as an assessor by the
European Foundation for Quality Management.
BRYANT C. MITCHELL is an instructor and faculty advisor in the
department of business, management and accounting at the
University of Maryland Eastern Shore and a founding partner of
the Ansar Group, a general management-consulting firm.
Mitchell earned a master’s degree in business administration
from Columbia University.
QUALIT Y PROGRESS
I
JANUARY 2002
I
59
QP
