Part III -- Some Independent Views
Responsibility For Quality In A Project
dardization in all aspects of construction practices is potentially damaging to the long-term viability of construction organizations through discouraging innovation and adaptability.
The commission of inquiry has performed a complex task in a generally thorough manner. They have proposed several recommendations for improving the operations of professional and construction firms in an effort to ensure that disasters like the Station Square collapse do not happen again. It is unfortunate that, in performing their tasks, the commission did not see the usefulness in consulting individuals who could have offered advice on the potential impact of their recommendations on project organizations. The commission needed to be made aware that the specific recommendations they have proposed have important implications for the organizations and their ability to adapt and operate in a dynamic and changing workplace. This section was intended to offer an organization theory perpective on the findings of the commission of inquiry, in effect, to make the commission aware that suggestions offered to improve specific aspects of construction firms' operations may have added undesirable side effects that the commission has not considered and project organizations are currently not in the position to address.
The Inquiry Commissioner quite correctly focused recommendations on improving the safety of the product of a project. The purpose of this portion of the examination of the incident and the subsequent recommendations is to clarify the opportunities for managing the overall quality, including safety, of the product of a project in light of recent developments in quality management in industry in general.
MODERN QUALITY MANAGEMENT
The management of quality has been. for the last decade, one of the most dynamic areas of concern in many industries. In volume production industries. the relevant measure of defects has changed from percent (parts per hundred) to ppm (parts per million). The “costs of quality” have been reduced from 15 cents to only 3 cents out of every sales dollar. These statistics can easily be misunderstood because the “costs of quality”, or really the costs associated with trying to achieve quality have changed dramatically during the decade. While we have traditionally accounted for only those costs associated with quality control, the “costs of quality” is a much broader concept today resulting in magnitudes which are startling at the least. The above numbers seem small when expressed in cents but when seen as $15 million out of a $100 million project, the magnitude of the Station Square development, it is easily seen as potentially greater than the profit from such a job. Thus, not only must the objective of the Inquiry be to ensure the safety of the public but to also permit the realization of the savings available from the approaches of modern quality management (MQM).
It is necessary to recognize exactly how such savings are realized in company after company in the volume production industries to understand how it can be achieved in project oriented endeavors. There are four basic costs of quality—internal failure, external failure, appraisal, and prevention.
The failure costs include costs of correcting defects during the project, as well as, after project completion. The costs of prevention and appraisal include quality control and inspection but go beyond conventional concepts of even these areas.
Under the classical approach to quality, failure costs have typically been 80% of total quality costs, about equal between internal and external. By this experience then, these could have been a total of about $12 million for Station Square. Appraisal and prevention have accounted for about 10% each, or perhaps $3 million for this development. After complete installation of MQM, experience has shown that the total amount spent for appraisal and prevention is actually reduced somewhat but the failure costs are reduced dramatically. Thus, it is conceivable that quality costs might have been as little as $3 million, a savings of $12 million on this project.
The definition of quality has changed from “fitness for use” to “conformance to requirements/specificstions.” The rationale for this is well documented in “A Case History” in Chapter 4 of Quality is Free [1, pp. 41-49]. The implications of this change are somewhat more subtle, however, and require more consideration than is appropriate in this article. Some essential concepts of MQM are that
- Quality cannot be inspected into a product, it must be designed in.
- To achieve quality output, attention must be focused on the process, i.e., process control, and generally statistical process control.
- The standard for quality is zero defects. To expect less is to ensure that less will be achieved.
To control the process then it is necessary to focus on the process using a systems view. The systems view is generally described in terms of “Input — Process — Output.” In the past, the focus has been on the input and the output with too little attention paid to the process. MQM focuses on the process, understands its variability, reduces the variability, and measures the output of the process to ensure that it stays in control. This is especially complex in projects due to the difficulty of establishing standards by which to manage change itself, the inherent characteristic of projects.
THE MANAGEMENT OF CHANGE
Indeed, it is this very characteristic which creates the dilemma in the Station Square project. While the mission of this effort was to change the nature of a piece of real estate to create a major commercial development, it was the changes in the requirements defined for this development which lead to its difficulties. This is a paradox of project management, i.e., the management of change, that changes of any kind in the project are most likely to lead to difficulties in the project.
The standard method for managing changes in the requirements in a project is “change control.” A properly designed change control system provides careful checks and balances on all changes whether they be requirements, methods, materials or assumptions. Change control systems tend to be very formal in large organizations. They are more likely to be informal the smaller the organization. This appears, based on the information in the Commissioner’s report, to have been a major problem at Station Square. Significant changes in requirements were followed by significant changes in assumptions with less than adequate checks and balances. Even two reviews of design calculations and assumptions failed to bring these errors to light. Thus, a possible recommendation might have been to require that a more formal change control and documentation (as built) system be used on all projects involving “major buildings” as defined by the Commissioner.
The failure of the design reviews suggests that more rigor was needed in this process. This must be approached cautiously lest the reviews and process control increase the costs and time required for the design process in a prohibitive manner. A deceptively simple solution to this is to develop manuals of recommended practices, standard documents and practices, and predetermined methods. While these have the obvious advantage of ease of administration and avoidance of risk, they have some inherent disadvantages.
As pointed out in a companion article in this set, such an approach tends toward bureaucratization and rigidity. While such standards are certainly appropriate, care must be exercised in their prescription. To rigidly define responsibilities at the macro level is sure to reduce flexibility, innovativeness, and creativity in adapting to the changes in technology, requirements, and opportunities in the future.
The costs and time for applying such standards can be substantial and different standards are often conflicting. What can be overlooked, however, are the chances that excessive dependency on such standards can lead to a “mental laziness” which results in failure to check for effects of interactions in a specific design. An approach is needed which avoids excessive costs and time and provides the highest possible assurance of identifying the really serious consequences of errors. Such an approach is well documented and practiced in many organizations today. It generally goes by the title of “Quality Assurance.”
THE INDEPENDENT QUALITY ASSURANCE MANAGER
An independent quality assurance manager must ensure that work performed on a project conforms to an array of standards while not incurring exorbitant cost. Some of these standards are client requirements, guiding physical principles, legal requirements, specifications, and drawings. To ensure that quality standards are met, a quality assurance manager must both design a process in an appropriate level of detail and use that process to control quality. The following four step approach can be used to both design and control the appropriate process for meeting quality standards on projects. The steps are:
- Identify the relevant quality issues for the particular project.
- Categorize the issues according to their levels of importance.
- Audit the process to determine whether the standards/specificstions are met.
- Use information regarding any deviations from standards for the twofold purpose of correcting the immediate problems and preventing a reoccurrence from happening.
IDENTIFY POSSIBLE QUALITY PROBLEMS
The potential quality problems can he considered according to the probability that they will occur and the consequences if they do occur as shown in Figure 1, [2, p,194 ].
CATEGORIZE THE PROBLEMS
The second step in managing a quality process is to categorize the quality problems so that conformance with quality standards can be maintained at a low cost. If all quality problems were treated the same, the cost of 100% inspection would be prohibitive. Yet if certain types of quality failures occurred, the results could be catastrophic, Therefore, potential quality problems should be categorized in such ways that facilitate the economic attainment of quality standards. These problems can first be categorized according to the level of consequences should they occur. The potential quality problems can be categorized as to consequences using the following four seriousness classes as follows.
“Class I. Negligible. Will not result in personnel injury or product damage.
Class II. Marginal. Can be counteracted or controlled without injury to personnel or major product damage.
Class III. Critical. Will cause personnel injury or major product damage, or will require immediate corrective action for personnel or product survival.
Class IV. Catastrophic. Will cause death, severe injury to personnel, or product loss.” [2, p. 197]
Next, the causes of the potential quality problems should be identified using techniques such as Failure Mode and Effect Analysis and Fault Tree Analysis [2, p. 183 & p. 197]. Generally, as shown above, the potential problems will have different probabilities of occurrence and, therefore, must be sought out with differing intensities.
A project manager is often faced with making tradeoffs between cost, schedule, and performance. The logic behind Figure 1 can be used to help a project manager decide whether to make performance tradeoffs. Negligible technical performance issues (Class I) might deal with issues such as aesthetics or minor rework, The consequences of these could often be accepted even though the probability may be high. Marginal quality issues such as level of achieving non-critical specified performance can sometimes be accepted. As quality defects move toward the catastrophic consequence end of the spectrum (Class IV), they should never be compromised.
The consequences of quality problems can be viewed along several different dimensions. One dimension is the clarity of problem definition, A poorly understood problem has a greater chance of having unexpected consequences and, as such, should be considered quite seriously. To ensure proper recognition of consequences any tradeoff affecting quality should be accomplished by rewriting the specifications rather than failing to meet them.
Another dimension is the function of a particular part, assembly, or system. Any part of a building that is overhead (such as the parking deck) is more prone to causing a catastrophic problem than an area that is underfoot. Likewise, other specialties on a project (such as electrical or high pressure gases) have potentially serious problems.
A third dimension to consider in looking for quality problems is the process of designing and producing the project. Certain design activities are more critical than others. For example, the calculations and assumptions used to determine the sizes and strengths of the structural elements were highly critical.
Yet another way to consider potential problems is their systemic effect on a project. One type of systemic problem is how different subsystems interact with each other. Another type of systemic effect is how the build-up of multiple small problems results in a large problem. Both of these systemic effects occurred on the Station Square project as changes in the parking system interacted with the structural system. Further, systemic problems can be in the customary way of performing tasks. The sidewalk contractor, accustomed to being penalized for underbuilding, erred in favor of over building. The variability of the process was too great and failure resulted.
AUDIT THE PROCESS
The next step is to audit the process to determine whether quality standards are being achieved. There are several methods of auditing the process including peer review, preaward surveys, postaward conferences, self inspection, etc. Regardless of the auditing method that is used, a process must be devised to determine which quality standards to inspect and how frequently. As stated above neither 100% inspection nor simple random sampling is cost effective. What should be used instead is stratified random sampling similar to that shown in Figure 2.
Figure 2. Sampling Frequency vs Potential Quality Problems
This is actually a variation of stratified random sampling since it prescribes 100% inspection for catastrophic problems and progressively less inspection as the consequences of failure decrease. Since the probability of failure is highest for Class I failures, the smallest sample sizes are required to detect these problems. As the seriousness of the consequences increase the probabilities of them occurring decrease. Therefore, greater precision is required in sampling to identify them. Class IV failures, although potentially catastrophic, seldom have high probabilities. Therefore, 100% inspection is probably warranted.
The viability of the concept of random sampling has been recognized as a practical tool in accounting for nearly three decades as indicated by the discussion of it in a popular cost accounting text [3, pp. 746-61]. The advantages of stratified random sampling are also well documented in texts on sampling techniques [4, pp. 65-110]
FIX THE PROBLEM AND THE CAUSE
The final step is to use the knowledge of quality problems to correct mistakes and to prevent the same problems from happening again. One aspect of prevention is to maintain discipline and ethical standards within the profession. Competitive pressures can erode these standards, however, so it is essential to follow one of the central precepts of quality assurance, i.e., separation of responsibility.
A major requirement of the Nuclear Regulatory Commission imposed on all projects under its supervision is that the Quality Management function cannot report to the position which is responsible for performing the work. While the work on Station Square is not as critical as that on a nuclear power station, neither is the complexity of the quality assurance function. Thus, an individual not otherwise associated with performing the work of the project could provide this function in a manner consistent with this requirement for separation of responsibility.
Many organizations are turning to modern quality management methods to deliver quality products at reasonable cost. As an example, the Department of Defense (DOD) has recently adopted a total quality management approach in dealing with contractors. The emphasis has shifted from inspection of final output to analysis of the production process in order to identify the causes of quality problems. There are high hopes that this shift from old to new quality approaches will lead to better products at lower cost. If a large bureaucratic organization like the DOD can effectively use modern quality approaches, small contractors should easily be able to adjust to these ideas and benefit handsomely.
1. Crosby, Phillip B.. 1979. Quality is Free. New York: McGraw-Hill.
2. Juran, J. M., & Gryna, Frank M. Jr.. 1980. Quality Planning and Analysis, 2nd Ed., New York: McGraw-Hill.
3. Horngren, Charles T.. 1962. Cost Accounting, A Managerial Emphasis. Englewood Cliffs, N. J.: Prentice-Hall, Inc.
4. Cochran, William G.. 1953. Sampling Techniques. New York: John Wiley & Sons, Inc.,
In the case of the Station Square roof collapse, it is not necessary to range far afield in a search for underlying causes. So far as I can determine, these words are the only words in boldface type in the Final Report of the Inquiry Commissioner:
The three practices -- fast-tracking, bidding for professional services and fragmentation... cause me a great deal of concern. (p. 42)
In this brief commentary on the “Socio/Economic/Political Implications” of the tragedy, I shall build upon the Commissioner's conclusions, indicating why I think his analysis, while on the mark, did not go far enough in exploring the underlying problems—problems I take to be increasingly universal, indeed inevitable, when competitive bidding and contracts are combined.
PRESSURES OF COMPETITIVE BIDDING
From my perspective, the stage was set for disaster very early on. When eight bids were received on the structural engineering work, the third lowest bidder ($20,500 for a 5.4 million building) was chosen, perhaps because the owners considered this firm more qualified than the others (the Commissioner gave no reason for the choice). More to the point, the winning bid was “negotiated down to $17,000, perhaps because the owners assumed that the winning bidder had overstated costs, the lowest bid was the correct price (even though the bidder was rejected), or both. This initial and relatively petty form of corner-cutting flew in the face of what is widely known among those associated with competitive bidding practices.