Content uploaded by Edward D. Walker
Author content
All content in this area was uploaded by Edward D. Walker on Aug 23, 2016
Content may be subject to copyright.
ADDRESSING ILL-STRUCTURED PROBLEMS USING GOLDRATT’S THINKING
PROCESSES: A WHITE COLLAR EXAMPLE
By
Edward D. Walker II
Department of Management
College of Business Administration
1500 North Patterson Street
Valdosta State University
Valdosta, GA 31698-0076
(229) 245-2236
eddwalker@valdosta.edu
James F. Cox III
Department of Management
Terry College of Business
University of Georgia
Abstract: Problem-solving techniques for poorly structured problems have been the subject of
recent academic research and popular press texts. In this paper, we explain the use of one of the
Theory of Constraints thinking process tools — the Current Reality Tree (CRT). The purpose of
the tool is to clearly identify the core problem or problems that cause the surface problems or
undesirable effects occurring in an organization. One must fully understand the core problems in
the environment before proposing a system solution to these core problems. Without this
systems perspective, a proposed solution may create more surface problems than it solves.
Through the use of an actual white-collar service case, we explain how the CRT is created. The
overall objective of this research is not to propose solutions to the case but to demonstrate how
the CRT provides structure by identifying and logically linking the surface problems to their
cause, the core problems. We demonstrate the power of the CRT to address poorly structured
problems.
Key words: Theory of Constraints thinking processes, current reality tree, decision-making,
service operations
1
ADDRESSING ILL-STRUCTURED PROBLEMS USING GOLDRATT’S THINKING
PROCESSES: A WHITE COLLAR EXAMPLE
Abstract: Problem-solving techniques for poorly structured problems have been the subject of
recent academic research and popular press texts. In this paper, we explain the use of one of the
Theory of Constraints thinking process tools — the Current Reality Tree (CRT). The purpose of
the tool is to clearly identify the core problem or problems that cause the surface problems or
undesirable effects occurring in an organization. One must fully understand the core problems in
the environment before proposing a system solution to these core problems. Without this
systems perspective, a proposed solution may create more surface problems than it solves.
Through the use of an actual white-collar service case, we explain how the CRT is created. The
overall objective of this research is not to propose solutions to the case but to demonstrate how
the CRT provides structure by identifying and logically linking the surface problems to their
cause, the core problems. We demonstrate the power of the CRT to address poorly structured
problems.
Key words: Theory of Constraints thinking processes, current reality tree, decision-making,
service operations
2
ADDRESSING ILL-STRUCTURED PROBLEMS USING GOLDRATT’S THINKING
PROCESSES: A WHITE COLLAR EXAMPLE
WHY IS DECISION-MAKING SO TOUGH?
Billions of dollars have been expended on new technologies to improve white-collar
productivity. In most cases little to no improvement has been realized. In some cases, the
organization actually lost money with the “improvements”. Manufacturing went through two
decades of improvements – MRP, MRPII, JIT, and TQM. Service organizations have recently
joined manufacturers in the newest improvement tools—six sigma, lean, BPR, and ERP. Can it
be that most of these tools and technologies do not address the core (underlying) problems of the
organization? In many cases, organization policies and procedures are what block effective
organizational improvement.
In this paper, we present one of the Theory of Constraints (TOC) thinking processes (TP),
the current reality tree (CRT), which has been used effectively by many organizations (e.g. Israni,
1995; Kwolek & Cox, 1996; Moon, 1996; Mordoch, 1996; Murphy, 1996; Roadman, 1996;
Gattiker & Boyd, 1999) to identify the core problems of the organization. We discuss the CRT,
and provide an application of the CRT to a white-collar environment. We conclude with a
discussion of the use of the CRT in ill-structured problem identification.
3
STRUCTURED, STRUCTURABLE, AND UNSTRUCTURED PROBLEMS
Many managerial decisions are ill-structured and thus do not lend themselves to
quantitative methods – managerial decision making seems to be as much an art as a science,
particularly with respect to ill-structured problems. Problems are rarely defined in a way that
leads the decision-maker to appropriate data, the right tool, and the correct answer. Our
traditional academic texts provide little preparation for understanding the complexities of real
world problems.
Smith (1988) argues that problems are structured, structurable, or unstructured for a
particular problem solver at a given time. With a structured problem, the problem solver can
readily identify a solution strategy – the six-line problem in academic texts. For a structurable
problem, a solution strategy becomes apparent if additional information becomes available or the
problem solver can divide the problem into more manageable sub-problems. An unstructured
problem has no solution strategy because of limitations of the problem-solving tools.
Both researchers (e.g. Anderson, 1985; Bransford, Sherwood, & Sturdevant, 1987;
Nickerson, 1994; and, Sacerdoti, 1977) and the popular press (e.g. Altier, 1999; Palady & Snabb,
2000; Rasiel & Friga, 2001; Sproull, 2001) have addressed the nature of complex problem-
solving. Tacit knowledge of the environment (Balconi, 2002) and a clear definition of the
problem (Campbell, 1960; Getzels & Csikszentmihalyi, 1976) are required. An accurate
identification of the core problem provides a platform "to analyze the roots, rather than the
symptoms, of success and failure" (Lamont & Friedman, 1997). Investigations into solving ill-
4
structured problems (e.g. Chen, 1999, and Fazlollahi & Vahidov, 2001) reveal that a structured
analysis is helpful.
Spence & Brucks (1997) defined an expert as one “who has acquired domain specific
knowledge through experience and training” (p. 234) and concluded that experts perform much
better than novices on structurable problems. Eli Goldratt (1994) developed a simple logic
method based on causality, called the Theory of Constraints Thinking Processes (TP), to impress
structure on most structurable problems and on many unstructured problems. The Thinking
Processes are based on the premise of systems thinking. Dettmer (1995) states that systems
thinking is comprised of four principles:
1. The performance of the system is affected by each of its components.
2. The parts of the system are interdependent.
3. If the parts of the system are grouped in any way, they form subgroups that
are subject to the first two principles.
4. If the performance of each part of the system is individually maximized,
the system, as a whole, will not behave as well as it could.
Novices, those without specific domain knowledge, can in our view make decisions comparable
to experts through the careful application of the thinking process tools. Using only the data we
gained through interviews with company employees (managers and workers in various functional
areas), we will attempt to support this assertion.
THEORY OF CONSTRAINTS THINKING PROCESSES
Goldratt’s TP are logic-based tools provide the user the ability to:
5
1. Identify the core problem of a system.
2. Identify and test a win-win solution (before implementation).
3. Create an implementation plan (that is almost fool-proof).
4. Communicate the above without creating resistance.
Goldratt (1994) describes six basic tools (referred to as thinking processes) that allow
managers to answer the three most important questions when confronted with an apparently
intractable problem, whether it may be structurable or unstructured:
1. What to change?
2. To what to change?
3. How to implement change?
We will focus on the use of the Current Reality Tree in this research. The TP have a set of rules,
called categories of legitimate reservation (CLR), for constructing and testing the logical
arguments.
The CRT is a logic-based tool for using cause-and-effect relationships to determine core
problems that cause the undesirable effects of the system (Cox, Blackstone, & Schleier, 2003) It
answers the first managerial question – what to change? Of all the organizational problems, what
problem is blocking the organization from real improvement?
A CRT begins with the identification of several surface problems or undesirable effects
(UDEs) through interviews with the parties involved in the situation. A single statement
summarizes each UDE. Two approaches are used to build a CRT. The traditional approach is to
build the causal logic between two UDEs by using if-then logic or by selecting an UDE and
asking, “Why does this UDE exist?” The second approach, the three cloud approach, is to build
6
an evaporating cloud for each of three of the UDEs. From these three clouds, a generic cloud is
constructed; next, all of the UDEs are connected to this generic cloud base using causal logic as
described above in the traditional approach. For a detailed discussion of the traditional approach,
see Scheinkopf (1996) or Dettmer (1997); for a detailed discussion of the traditional and the
three cloud approaches, see Cox, Blackstone, & Schleier (2003). We use the traditional approach
to demonstrate the creation of a CRT; the 10-step procedure is presented in Table 1. [Though
only four of the CLR are explained here, Dettmer (op. cit., pp. 30-61) explains each of the seven
CLR.]
<Place Table 1 here>
AN APPLICATION OF THE CRT
Situation: This application is constructed from data obtained in interviews with a large
service company. Acme (not the actual company name) employs a pool of sixteen secretaries to
service its corporate offices. A secretarial staff supervisor administers the pool. Two years
before our interviews, a centralized secretarial pool had been reorganized. Each of the sixteen
secretaries is assigned to a department and supports up to six individuals. The secretaries earn
two weeks annual leave taken at the discretion of the secretary upon approval by the secretarial
staff supervisor. Additionally, secretaries occasionally have jury duty or take unscheduled sick
leave.
Currently, the secretaries are frustrated by their job assignments and work loads—feeling
that they are requested to complete more work than can be handled and that they are asked to
7
complete assignments for which they have not been trained or are not responsible. Conversely,
the managers are frustrated by the slowness with which assignments are completed and by the
errors contained in the resulting reports. This example will be used to illustrate the procedure for
constructing a CRT.
Step 1. List the Undesirable Effects.
In our interviews with Acme managers, secretaries, and the staff supervisor, we identified
the following UDEs:
1. Secretaries are frustrated.
2. Secretaries make mistakes.
3. The manager/department appears inept.
4. Work is completed slowly.
5. Secretaries constantly ask managers for clarification.
6. Managers do not trust the system.
7. Secretaries often change departments.
8. The secretary is not trained.
Step 2. Test each UDE for clarity using the Clarity Reservation.
Is the UDE a clear and concise statement? The original UDEs were edited to improve
clarity.
1. Secretaries become frustrated with their work assignment/load.
2. Secretaries make mistakes.
8
3. The manager/department appears inept.
4. Work gets postponed/completed more slowly than expected.
5. Work must iterate between the manager and secretary several times.
6. Management loses trust in the secretarial support system.
7. Secretaries are rotated among the managers frequently.
8. Assignments are made for which the secretary is not trained.
Step 3. Search for a causal relationship between two of the UDEs.
Both entities 2 and 3 and entities 8 and 2 seem to be related. [“Entity” describes any
statement in the CRT. Many entities are added during the CRT construction; however, only a
few, those on our original list, are UDEs.] We check 2 and 3 first. Numbering is used to quickly
locate the entity on the diagram, nothing more (see step 10, Table 2). An asterisk identifies each
of the original eight UDEs.
Step 4. Determine which UDE is the cause and which is the effect.
Read “IF cause THEN effect”. This test is called the Causality Reservation.
Occasionally the cause and effect might be reversed (Tautology Reservation); check for this error
using the following statement: “Effect BECAUSE cause.” If “secretaries make mistakes” then
“the manager/department appears inept”; or, “the manager/department appears inept” because
“secretaries make mistakes.” This relationship seems logical and causal except for the modifiers
(see step 9, Table 2). The relationship is better expressed as: if “secretaries make mistakes” then
“sometimes the manager/department appears inept.” As we build the tree we adjust the modifiers
9
based on further causality relationships. Other UDEs are causally related to UDE 3. This part of
the CRT is provided in Figure 1.
<Place Figure 1 here>
Step 5. Continue the process of connecting the UDEs until all are connected.
Entity 4 causes entity 3; entity 8 causes entity 2; entity 1 causes entity 2; entity 6 causes
entity 3; and entity 4 causes entity 5. These relationships are given in Figure 2.
<Place Figure 2 here>
Step 6. Apply the Clarity Reservation.
The remaining steps are not necessarily applied in the number sequence presented. They
clarify the causality by modifying the entity wording or structure. There appears to be a
relationship between entity 7 and cause 1. (See Figure 3a.) If “secretaries become frustrated with
their work assignment/load,” then “secretaries are rotated among managers frequently.” The
clarity reservation again requires the insertion an intervening entity. (See Figure 3b.) If
“secretaries become frustrated with their work assignment/load,” then “secretarial staff
supervisor reassigns work/secretary to satisfy the secretary.” If “secretarial staff supervisor
reassigns work/secretary to satisfy the secretary,” then “secretaries are rotated among managers
frequently.”
10
<Place Figure 3 here>
The insertion of additional entities into the CRT is based on the creator’s knowledge of
the situation (interviews with Acme employees in this case) or from the creator’s general
knowledge. The creator should verify the existence of the entity. Validation is required
whenever entities are added to the CRT as required by steps 6, 7, and 8.
Step 7. Apply the Cause Insufficiency Reservation.
Entity 7 causes entity 8 (Figure 4a), but 7 alone is insufficient. If “secretaries are rotated
among managers frequently,” then “assignments are made for which the secretary is not trained.”
Another cause must be added to cement the relationship. (See Figure 4b.) If “secretaries are
rotated among managers frequently,” AND “different departments/managers require different
skill sets of their secretaries,” then “assignments are made for which the secretary is not trained.”
The ellipse (representing AND) across both arrows indicates that either cause alone is
insufficient to result in the proposed effect – all causes must exist for the effect to occur.
<Place Figure 4 here>
The original UDEs are now connected. (See Figure 5.) It is apparent that a loop exists
within the CRT: the secretaries are rotated because they make mistakes, and they make mistakes
because they are rotated. This vicious cycle must be broken to restore the effectiveness of the
organization.
11
<Place Figure 5 here>
Step 8. Apply the Additional Cause Reservation.
Although the original UDEs are now connected, the tree is still incomplete. Entity 7 has
an additional cause reservation. Reassignment by the secretarial staff supervisor due to
frustration is not the only cause. Each secretary takes two weeks annual leave at any time
approved by the staff supervisor or may unexpectedly stay home sick. Since the company
continues business when a secretary is absent, the secretaries are shifted among
managers/departments to cover for absent employees. When a secretary supporting a higher-
level manager is absent for a period, four or five secretaries might be reassigned to support
higher-level managers. The absence of one secretary creates a domino effect down several levels
of management, shown in Figure 6.
<Place Figure 6 here>
There must be some reason that entity 39 exists. When questioned, the managers and
staff supervisor responded that it is company policy that the highest skilled secretary available is
always assigned to the highest level of management. This relationship is shown in Figure 7.
<Place Figure 7 here>
12
One need not identify all possible causes of an effect; only enough causes to account for
80% or more of the occurrences of an effect. A simple test for whether to include an additional
cause would be to ask the question, “Does this additional cause alone account for many
occurrences of the effect?”
Completing the CRT
Using the clarity, cause insufficiency, and additional cause reservations as above and
restating several entities led to the more complete CRT shown in Figure 8. Some readers may
feel the CRT is simply the creator’s perception of reality; or as Greenleaf states: “As I look
through my window on the world I realize that I do not see all. Rather, I see only what the filter
of my biases and attitudes of the moment permits me to see (quoted in Cutting & Kouzmin,
2000, p. 479).” The creator’s biases and perceptions are present in the CRT, as in any mental
model. In this application, various managers using the secretarial support system and secretaries
within the secretarial support system identified the UDEs of the system. The TP tools are a
relatively simple application of Boolean logic to establish causality. Yes, errors in logic can
occur. However, the careful application of the CLR by the CRT developers and the subsequent
review of the CRT by an outside reviewer—in this case, several Acme managers and
secretaries—reduce systematic bias.
Although forty-two entities appear on the final CRT, two trees were constructed
originally—one representing the views of the managers and the other representing the views of
the secretaries. By approaching the problem space from the points of view of both the managers
and the secretaries, we were able to get an overview of the entire system rather than the view
13
from only one perspective. Consider the famous poem of John Godfrey Saxe about the six blind
men and the elephant. Each man was absolute in his conviction. Each of them was correct from
his perspective, but none understood the true nature of the elephant—hence, all were wrong. The
CRT conveys a system level understanding of the situation.
<Place Figure 8 here>
WHAT NEEDS TO BE CHANGED?
An entity that does not have an arrow entering it—the entity is not caused by some other
entity—is referred to as a root cause. A root cause can be classified into either of two categories.
A core driver is a condition that is beyond the control of the problem solver. Examples of a core
driver might include government regulation, environmental conditions, or human nature. The
second category of root cause, the core problem is the concern of the problem solver. A core
problem is a condition over which the problem solver can exert some control or influence. Upon
initial examination of the final CRT, all ten of the root causes, listed below, might appear to be
core drivers rather than core problems.
19. There are no formal initial training requirements.
21. Different departments/managers require different skill sets of their secretaries.
31. The work scheduling system is informal and ineffective.
37. The secretary's workload is not considered.
41. Managers know which secretary possesses the skills required.
14
44. Managers aren’t notified of the shift.
82. Secretaries try hard to please the manager(s).
101. Secretaries take time off for annual/sick leave.
102. The highest-skilled secretary available is always assigned to the highest level of
management.
109. The manager does not have time to do things that cannot/have not been delegated.
This phenomenon occurs because the core problem has existed for such a long time that all
concerned parties have accepted the core problem as fact, or as a given, unchangeable condition.
Notice that none of these root causes appeared in the initial list of UDEs.
The problem solver identifies those core problems that if reversed or eliminated destroy
the UDEs. What are the core problems of this CRT? Entity 102, “the highest skilled secretary
available is always assigned to the highest level of management,” entity 31 “the work scheduling
system is informal and ineffective,” and entity 19 “there are no formal initial training
requirements” are connected to most of the UDEs. A core problem should be connected to at
least 70% of the UDEs. In fact, every entity on the diagram except for the other seven root
causes can be traced to one of these three core problems. If these problems can be addressed then
almost all of the UDEs disappear.
The initial CRT (Figure 5) revealed a loop that in essence states that as secretaries are
rotated among managers, problems arise that cause the secretaries to be rotated among the
managers yet again. Approximately, forty distinct loops can be found upon examination of the
final CRT (Figure 8). These loops can be traced to three of the core drivers: 102 “the highest-
skilled secretary available is always assigned to the highest level of management”; 31 “the work
15
scheduling system is informal and ineffective”; and, 19 “there are no formal initial training
requirements”.
“The highest-skilled secretary available is always assigned to the highest level of
management” is an Acme policy and as such can be changed; therefore, this core driver is a core
problem. Reserving the highest-skilled secretary or a highly skilled secretary for use as a
replacement for any secretary who is absent might replace policy. In this way, no department
would be required to work with a secretary of lesser skill than is normally present. Additionally,
the work of a single department is disrupted by an absence, rather than disrupting the work of
many departments by re-assigning several secretaries.
However, elimination of the core problem 102 is not, by itself, sufficient to eliminate the
looping. The rotation of secretaries can also be traced to core driver 31 “the work scheduling
system is informal and ineffective.” The lack of a formal, effective work scheduling system can
also be changed. A formal, effective work scheduling system might be as simple as a job sign-up
board located at the secretary's workstation. All jobs in queue would be visible and could be
processed on a first-in, first-out basis or due-date basis established by the secretarial staff
supervisor. If a manager needed to have a job expedited or placed in queue ahead of other jobs,
then it would be his responsibility to clear such a request with the secretarial staff supervisor and
the managers whose jobs would be delayed.
Core driver 19 “there are no formal initial training requirements” also contributes to the
UDEs. Requiring initial skills training and requiring the various departments/managers to
standardize on one set of applications software is reasonably easy to implement. Core driver 21
“different departments/managers require different skill sets of their secretaries” actually
16
encompasses two thoughts: first, the different departments required different software
proficiencies such as word processing or spreadsheet or presentation graphics skills; and second,
the different departments used different software suites. As a move to standardize on a single
software suite was already underway, we chose to not address this as a core problem.
Prior to the creation of the CRT, the secretarial pool supervisor was fighting fires on a
daily basis. She was only addressing the symptoms of the underlying problems, and without
addressing the core problems the system was failing and would have continued to fail.
Acme realized that the system was in disarray and had intended to re-centralize its
secretarial pool before examining the CRT. Such a move would have cost a considerable sum of
money as the mini-pools are on various floors of three different buildings. Acme's solution
would also have diminished the one-to-one contact of secretary and manager/department and
increased the makespan of each job. The three possible solutions suggested by the CRT are less
intrusive, require lower initial and ongoing investment, and, most importantly, eliminate the
UDEs in the CRT.
WHAT HAPPENED AND WHERE DO WE GO FROM HERE?
The Secretarial Pool CRT
Spence & Brucks (op. cit.) found that experts perform much better than novices on
structurable problems though the experts’ confidence decreases as the problem structure
decreases and that decision makers resort to simplifying heuristics when solving less structured
problems. They concluded that “experts outperform novices because they can impose meaningful
17
structure onto judgment-intensive, nontrivial problems…this effect is most apparent when
solving problems without access to a decision aid” (Spence & Brucks, op. cit., p. 244).
The Acme employees—the experts in this case—had decided to re-centralize the
secretarial staff. A move that they acknowledged would be expensive and would likely cause
other problems. We—the novices—had access to a decision aid, the CRT. The secretarial staff
supervisor, selected secretaries and managers, and the vice-president, felt that the CRT was
correct in identifying policies that were creating serious problems in secretarial and managerial
productivity and morale. Without the use of the CRT, the VP would have centralized the
secretarial staff into a single pool, again. The novices outperformed the experts.
The CRT and the System View
The CRT graphically represents the inter-related conditions existing in the system. The
CRT logically links the UDEs to the core drivers. Too often in business, managerial problem-
solving attacks the symptoms of some underlying problem rather than treating the problem
directly. These actions are analogous to treating the symptoms of a biological infection—fevers,
chills, etc.—instead of attacking the cause of the infection. Long ago, the medical profession
recognized the futility of such efforts and routinely performs diagnostic tests to identify the cause
of the ailment so as to properly treat it. The CRT allows businesses to perform a diagnostic test
similar to what is practiced in the medical profession.
Our task in this research is not to solve the secretarial pool problem but rather it is to
teach the procedure for constructing a CRT. We feel the CRT provides a mechanism for
identifying the impact of policies, procedures, and actions in one functional area on other
18
organization areas. Both the traditional and three cloud approaches are useful. These same rules
apply to constructing the most complex CRTs using either approach – no more rules are required.
Validation of the causality and existence of entities and their relationship is always necessary.
Triangulation is often used when the biggest problems in the system are identified from different
perspectives—in this example, the managers, secretaries, and secretarial staff supervisor
provided three different perspectives of the problem environment. In this manner we ensure that
we are objectively viewing the system. To this end, the three cloud approach to CRT
construction offers the advantage of looking at an environment from three different perspectives.
Addressing Structurable & Unstructured problems
Smith’s (op. cit.) definitions of structured, structurable and unstructured problems imply
that structurable problems are a superset of structured problems and that there is no intersection
between these sets of problems and the set of unstructured problems. Our view of Smith’s
problem classification overlaid on a business and its environment is provided in Figure 9. This
view combines Smith’s definitions with Deming’s (1994) concern for the business as a system.
We believe that the TP are the advances in the state-of-the-art problem solving for which Smith
(op. cit.) calls to convert unstructured problems into structurable problems and structurable
problems into structured problems.
<Place Figure 9 here>
19
The structured problem with its simplistic solution seldom provides the correct solution
to a business situation. It ignores the true goal of the larger system—instead attempting to
optimize a local or functional objective. It seldom recognizes the interaction of decisions and
policies in this and other areas on the overall business, and, therefore, seldom recognizes the core
problem. A similar argument can be made with respect to structurable problems when
considering Smith’s contention that a structurable problem becomes structured with additional
information or by breaking the problem into more manageable sub-problems. Managers and
secretaries created their own solutions—bypassing the work scheduling system and taking annual
or sick leave, respectively—to the problems they saw in the secretarial support system, but these
solutions caused other problems and created a vicious cycle.
The CRT provides a realistic examination of a surface problem, its relationships to other
surface problems within and across functions, and its relationships to the core drivers and core
problems. The scope of the problem and solution can be extended depending on the distinction
between core drivers and core problems. The objective of the CRT is to identify the core
problems driving many of the seemingly unrelated and independent problems.
Summary
Goldratt’s TP offer an alternate approach to studying ill-defined problems. Structured
problems are commonly solved with simple but independent solutions. These solutions fail to
consider the appropriate system and its goal. Structurable problems are addressed by using a
CRT to better understand the causal relationships within and across functions and from core
problems to UDEs and to the system goal. Many unstructured problems can now be reclassified
20
as structurable problems with proper identification of the system, its goals and its constraints.
Significant data need not be collected to study every aspect of the system. Only eight to ten
UDEs – drawn from different functions and different levels to gain a systems perspective – are
required to develop a good understanding of an organization when viewed as the system and of
its core problems and drivers.
In this manner, the TP provides structure for many structurable problems and some
unstructured problems. If, as Smith (op. cit., p. 1497) comments, “…it is the behavior of
problem solvers that we advert to in making ascriptions of structure to problems”, then the TP
significantly increase the number of structured and structurable problems; serve as the decision
aid to decrease the expert-novice performance differential; and increase the confidence the
problem solver has in his recommendation.
Many problems are not well defined because we have not identified the reference system
to identify the core problems and core drivers. We should study the whole system to identify the
core problem and how to solve it. We need to identify the causality to assure we are addressing
the core problems and not just symptoms.
Directions for the Future
While the CRT in Figure 8 appears complex (it takes several hours to construct a
logically sound CRT using the CLR), it is an effective problem identification and communication
tool (recall this secretarial pool problem existed several years without identifying the correct
problem). Goldratt (1992) identified physical constraints and traditional measurements as major
constraints to improving organizational performance. However, in many cases, the constraint
21
blocking organizational improvement is a policy or procedure that has been in place for several
years. In most cases, organizations address only the symptoms of the core problem. This is
where the endless list of new buzzwords comes into play. In some cases these tools may have
been effective, but in many they were not (Dumond, 1995; Tatikonda & Tatikonda, 1996). The
TP help identify the core problem, identify and construct win-win solutions, construct effective
implementation plans, and communicate effectively in an organization. Before a company
spends money on big improvement projects, managers should logically think through their UDEs
to ensure they are attacking a core problem that will result in true organizational improvement.
While we have focused only on the CRT, the six TOC TP taken together provide managers and
business students skills to lead the organization in the new century.
REFERENCES
Altier, W.J. (1999), The Thinking Manager's Toolbox: Effective Processes For Problem Solving
And Decision Making, Oxford University Press, London.
Anderson, J. R. (1985), Cognitive Psychology and Its Implications, Freeman, New York.
Balconi, M. (2002), “Tacitness, Codification of Technical Knowledge, And the Organisation of
Industry”, Research Policy, vol. 31 no. 3, pp. 357-379.
Bransford, J. D., Sherwood, R. D., and Sturdevant, T. (1987), “Teaching Thinking And Problem
Solving”, in Boykoff Baron, J., and Sternberg, R. J. (eds.), Teaching Thinking Skills:
Theory And Practice, Freeman, New York.
Campbell, D. (1960), “Blind Variation and Selective Retention in Creative Thought As In Other
Knowledge Processes”, Psychological Review, vol. 67, pp. 380-400.
22
Chen, C. (1999), “A Protocol Analysis Model for Investigating Computer Supported Problem-
Solving Activities”, Information Technology, Learning and Performance Journal, vol. 17
no. 2, pp. 35-43.
Cox, J. F., III, Blackstone, J. H., Jr. and Schleier, J. G., Jr. (2003), Managing Operations: A
Focus on Excellence, North River Press, Great Barrington, MA.
Cutting, B., and Kouzmin, A. (2002), “The Emerging Patterns Of Power In Corporate
Governance – Back To The Future In Improving Corporate Decision Making”, JMR
Journal of Marketing Research, vol. 15 no. 5, pp. 477-511.
Deming, W. E. (1994), The New Economics for Industry, Government, Education, MIT Center
for Advanced Engineering Study, Cambridge, MA.
Dumond, E. J. (1995), “Learning from the Quality Improvement Process: Experience from U. S.
Manufacturing Firms”, Production and Inventory Management Journal, vol. 36 no. 4, pp.
7-13.
Dettmer, H. W. (1995), “Quality and The Theory of Constraints”, Quality Progress, vol. 28 no.
4, pp. 77-83.
Dettmer, H. W. (1997), Goldratt’s Theory of Constraints, ASQ Quality Press, Milwaukee, WI.
Fazlollahi, B., and Vahidov, R. (2001), “A Method For Generating Alternatives By Decision
Support Systems”, Journal of Management Information Systems, vol. 18 no. 2, pp. 229-
250.
Goldratt, E. M. (1992), The Goal: A Process of Ongoing Improvement, 2nd Revised Edition,
North River Press, Great Barrington, MA.
Goldratt, E. M. (1994), It’s Not Luck, North River Press, Great Barrington, MA.
23
Gattiker, T. F., and Boyd, L. H. (1999), “A Cause-And-Effect Approach To Analyzing
Continuous Improvement At An Electronics Company”, Production and Inventory
Management Journal, vol. 40 no. 2, pp. 26-31.
Getzels, J., and M. Csikszentmihalyi (1976), The Creative Vision: A Longitudinal Study Of
Problem-Finding In Art, Wiley-Interscience, New York.
Kwolek, M., and Cox, J. F., III (1996), “The Use of the Current Reality Tree to Identify Core
Problems in a Military Logistics Organization”, Journal of Systems Improvement, vol. 1
no. 2, pp. 23-36.
Israni, A. (1995), “Application of Theory of Constraints in the Semiconductor Industry”, in
McMullen, T. (Chair), Make Common Sense a Common Practice. Symposium conducted
at the meeting of the 1995 APICS Constraints Management Symposium and Technology
Exhibit, Phoenix, Arizona.
Lamont, L. M., and Friedman. K. (1997), “Meeting The Challenges To Undergraduate Marketing
Education”, Journal of Marketing Education, vol. 19 no. 3, pp. 17-30.
Moon, S. A., (1996), “Untitled”, in McMullen, T. (Chair), Make Common Sense a Common
Practice. Symposium conducted at the meeting of the 1996 APICS Constraints
Management Symposium and Technology Exhibit, Detroit, Michigan.
Mordoch, A., (1996), “TOC and the Inherent Problem of Supply Base Management”, in
McMullen, T. (Chair), Make Common Sense a Common Practice. Symposium conducted
at the meeting of the 1996 APICS Constraints Management Symposium and Technology
Exhibit, Detroit, Michigan.
24
Murphy, R. E. Jr., (1996), “Breakthrough Performance in the Semiconductor Industry”, in
McMullen, T. (Chair), Make Common Sense a Common Practice. Symposium conducted
at the meeting of the 1996 APICS Constraints Management Symposium and Technology
Exhibit, Detroit, Michigan.
Nickerson, R. S. (1994), “The Teaching of Thinking and Problem Solving”, in Sternberg, R. J.
(ed.), Thinking and Problem Solving, Academic Press, New York.
Palady, P., and Snabb, T. (2000), TAPS: A Total Approach To Problem Solving, PAL
Publications: Detroit, MI.
Rasiel, E. M., and Friga, P. N. (2001), The McKinsey Mind: Understanding and Implementing
the Problem Solving Tools and Management Techniques of the World's Top Strategic
Consulting Firm, McGraw Hill College Division, New York.
Roadman, C. H. (1996), “Air Force Medicine: Differential Diagnosis and Therapy or How the
Thinking Process Has Helped Us Rethink Our Healthcare System”, in McMullen, T.
(Chair), Make Common Sense a Common Practice. Symposium conducted at the meeting
of the 1996 APICS Constraints Management Symposium and Technology Exhibit,
Detroit, Michigan.
Sacerdoti, E. D. (1977), A Structure For Plans And Behavior, Elsevier North-Holland, New
York.
Scheinkopf, L.J. (1996), Thinking for a Change: Putting the TOC Thinking Processes to Work,
St. Lucie Press, Boca Raton, FL.
Smith, G. (1988), “Towards A Heuristic Theory Of Problem Structuring”, Management Science,
vol. 34 no. 12, pp. 1489-1506.
25
Spence, M. T., and Brucks, M. (1997), “The Moderating Effects of Problem Characteristics On
Experts’ And Novices’ Judgments”, JMR, Journal of Marketing Research, vol. 34 no. 2,
pp. 233-247.
Sproull, B. (2001), Process Problem Solving: Guide for Maintenance and Operations Teams,
Productivity Press, Portland, OR.
Tatikonda, L. U., and Tatikonda, R. J. (1996), “Top Ten Reasons Your TQM Effort is failing to
Improve Profit”, Production and Inventory Management Journal, vol. 37 no. 3, pp. 5-9.
26
Table 1: Traditional method to construct a Current Reality Tree.
Step 1
List between 5 and 10 problems (called undesirable effects -- UDE’s) related to the
situation.
Step 2
Test each UDE for clarity -- is the UDE a clear and concise statement. This test is
called the clarity reservation.
Step 3
Search for a causal relationship between two of the UDEs.
Step 4
Determine which UDE is the cause and which is the effect. Read as “IF cause THEN
effect.” This test is called the causality reservation.
Step 5
Continue the process of connecting the UDEs using the If-Then logic.
Step 6
Many times the causality is strong to the person feeling the problem but doesn’t seem
to exist to others. In these instances, “clarity” is the problem. Use the clarity
reservation. Generally, entities between the cause and the effect are missing.
Step 7
Sometimes the cause by itself is not enough to create the effect. These cases are tested
with the cause insufficiency reservation and are improved by reading “IF cause AND
__________ THEN effect.”
Step 8
Sometimes the effect is caused by many independent causes. The causal relationships
are strengthened by the additional cause reservation. These cause-effect relationships
are called a "magnitudinal and
" for each cause increases the magnitude of the effect.
Each of the causes must be addressed individually to eliminate most of the effect.
Step 9
Sometimes an if-then relationship seems logical but the causality is not appropriate in
its wording. In these instances words like “some”, “few”, “many”, “frequently”,
“sometimes” and other adjectives can make the causality stronger.
Step 10
Numbering of UDEs on the CRT is for ease of locating entities only. An asterisk by a
UDE indicates that UDE was provided in the original list of UDEs.
27
3* The manager/
department appears
inept.
2* Secretaries
make mistakes.
Figure 1: The search for causal relationships.
28
2* Secretaries make
mistakes.
8* Assignments are
made for which the
secretary is not trained.
3* The manager/
department appears inept.
4* Work get postponed/
completed more slowly
than expected.
1* Secretaries
become frustrated
with their work
2* Secretaries make
mistakes.
3* The manager/
department appears
inept.
6* Management
loses trust in the
secretarial support
4* Work get postponed/
completed more slowly
than expected.
5* Work must iterate
between the manager and
secretary several times.
Figure 2: The search for more causal relationships.
29
7* Secretaries are rotated
among managers frequently.
1* Secretaries become
frustrated with their work
assignment/load.
7* Secretaries are rotated
among managers frequently.
1* Secretaries become
frustrated with their work
assignment/load.
a
b
52 Secretarial staff
supervisor reassigns work/
secretary to satisfy the
secretary.
Figure 3: Another example of the clarity reservation.
30
7* Secretaries are
rotated among
managers frequently.
8* Assignments are
made for which the
secretary is not trained.
7* Secretaries are
rotated among
managers frequently.
8* Assignments are
made for which the
secretary is not trained.
21 Different
departments/manager
s require different
skill sets of their
secretaries.
a.
b.
Figure 4: An example of cause insufficiency.
31
7* Secretaries are
rotated among
managers
frequently.
8* Assignments are
made for which the
secretary is not
trained.
21 Different
departments/managers
require different skill
sets of their
2* Secretaries make
mistakes.
1* Secretaries
become frustrated
with their work
23 Secretaries are
unable to learn what a
particular department/
manager likes.
5* Work must iterate
between the manager and
secretary several times.
4* Work get postponed/
completed more slowly than
expected.
3* The manager/
department appears inept.
52 Secretarial staff supervisor
reassigns work/secretary to
satisfy the secretary.
6* Management loses
trust in the secretarial
support system.
7 (loop)
52 (loop)
Figure 5: The initial Current Reality Tree.
32
7* Secretaries are
rotated among
managers
100 In a pool of 16
secretaries, at least
one secretary is
absent one day per
39 Secretaries are
shifted to cover
for absences.
101 Secretaries take time off
work for annual/sick leave.
52 Secretarial staff supervisor
reassigns work/secretary to
satisfy the secretary.
Figure 6: An example of the additional cause reservation.
101 Secretaries take time off
work for annual/sick leave.
102 The highest skilled
secretary is always
assigned to the highest
level of management
Figure 7: Another example of the additional cause reservation.
39 Secretaries are
shifted to cover
for absences.
33
7* Secretaries are rotated
among managers
frequently.
8* Assignments are made for
which the secretary is not
trained.
2* Secretaries make mistakes.
1* Secretaries
become frustrated
with their work
23 Secretaries are unable to
learn what a particular
department/manager likes.
3* The manager/
department
appears inept
52 Secretarial staff
supervisor reassigns work /
secretary to satisfy the
manager(s) / secretary.
7 (loop)
52 (loop)
100 In a pool of 16 secretaries, at
least one secretary is absent one
day per week.
39 Secretaries are shifted to
cover for absences.
101 Secretaries take time off work
for annual/sick leave.
102 The highest-skilled secretary is
always assigned to the highest level of
management.
103 Lower level managers
are unable to chose which
secretary is assigned to them.
21 Different departments/managers
require different skill sets of their
secretaries.
8
11 Different secretaries
have different skill sets.
51 Secretarial skill don’t
necessarily match the skill sets
required by different managers.
44 Managers aren’t
notified of the shift.
37 The secretary’s
current workload
is not considered.
21
19 There are no
formal initial
training
requirements.
6* Management loses trust in
the secretarial support system.
41 Manager know which
secretary possesses the
skills required.
61 Managers seek
out the secretary
whose skills are
required even if the
secretary is in
another department.
5* Work must iterate
between the manager
and secretary several
times.
71 The secretary
assignment
system is by-
passed.
104 The current
work of the moved
secretary is delayed.
105 The moved
secretary is
unfamiliar with the
new department/
manager and is
therefore less
efficient.
106 The work of
the absent
secretary is
27 Managers must oversee
more work than he/she should.
4* Work get
postponed/complete
d more slowly than
expected.
99 Managers
interrupt the work
currently being
processed by the
secretary to
expedite a task.
107 The manager
is required to do
some tasks
his/her self.
109 The manager
does not have time to
do things that
cannot/have not been
110 The manager
complains to the
secretarial staff
supervisor.
82
Secretaries
try hard to
please the
manager(s).
81 Some secretaries
are busier than the
system would
91 Some secretaries
are less busy than the
system would
31 The work scheduling
system is informal and
ineffective.
80 The
secretarial
staff
supervisor
counsels the
secretary
about
his/her work
46 The work of the
receiving
department is also
made late.
49
24
49 Managers
are less
efficient
27
48 The company
is less efficient.
24 Managers can’t
become too
dependent upon a
particular secretary.
7
50 Secretaries
are less efficient.
34 Managers don’t
train their secretary to
handle as many tasks
as might be possible.
1
81
48
4
Figure 8: The final Current Reality Tree.
34
Structured problem
and local solution
Structurable
problem
Unstructured problem
and solution space
Figure 9: Problem and solution domains.
35
