Successfully introducing project management techniques into an organization
C. W. Dane, C. F. Gray, and B. Woodworth
School of Business, Oregon State University
Corvallis, Oregon 97331
Project management techniques used as scheduling systems for large projects are applied less frequently than the advocates of such techniques would like to admit. In eighteen virtually identical government units which appeared to be ideally suited for network analysis, only four successfully utilized the technique on a continuing basis. Nine factors which were thought to explain the success or failure of such techniques were identified. Of the nine, the two factors with statistical significance were (1) hierarchical level of the person making the introduction and (2) the stated purpose, or use, of the techniques.
Network techniques, such as CPM and PERT, have the potential for assisting a project manager in scheduling individual activities to meet long range organizational objectives but to do so in such a way as to control scarce project resources. The techniques can be shown as providing significant assistance in the planning and organizing functions of project management. The techniques have the potential for significant benefits but we find that there is considerable resistance to their acceptance in most project oriented organizations. There is a growing concern among students of network techniques about the problem of application and retention.
The motivating reason for this study was a desire to identify those elements or characteristics which have a significant influence on the successful retention of a network scheduling technique. The circumstance which enabled the study was the fact that the authors were able to observe an organization which has many subunits similar to each other in organizational structure. Each subunit faced the identical problem which, in the opinion of the authors, was ideally suited to network analysis. The personnel involved were, for the most part, college educated and open to suggested improvements in their decision making processes. Whereas some form of network analysis was introduced in all but one of the subunits, the number of retentions was very small. There were differences among the subunits in the manner in which the network techniques were introduced. Some of these differences may allow one to predict whether or not there will be continued utilization of network analysis.
The Forest Service, an agency of the U.S. Department of Agriculture, has national forests located throughout the United States. In the states of Oregon and Washington, there are 19 national forests which administer 23 million acres of land and provide over 25 percent of the saw timber sold from national forest land in the continental United States.
Usually each forest is administered by a forest supervisor. His staff is composed of technical specialists in personnel, engineering, and resources. The latter have historically included specialists in timber growing and selling, range, water and recreation, fish, and wildlife; recently, in the face of growing environmental concern, specialists in landscape architecture, anthropology, and political science have been added. Underneath the forest supervisor and his staff are three to five districts, each headed by a district manager who must supervise 20-100 people. The district ranger reports to the forest supervisor and often uses the services of the supervisor’s staff.
The organizational structure is fairly uniform throughout the 19 national forests covered in this study. Minor variations existed in the way the district level engineering skills were provided. In many cases, engineers were grouped in work units at the same organizational level as the district but provided engineering skills to more than one district. In these 19 forests, the organizational purpose and personnel specialties tended to be similar. Of the many tasks facing the forest supervisors and district rangers in the western United States, one of the most complex is the scheduling and coordination of the people and activities required to prepare a portion of a forest for a timber sale. A given ranger district may be handling as many as 50 timber sales concurrently, and each timber sale may require as many as 30 skill categories, which must come from many levels and departments within the forest.
The number of skills and of people within a skill are relatively fixed in the short run. Since different timber sales often demand the same resources at the same time, and since the number of personnel is relatively fixed, each forest faces a complex job of scheduling the activities and personnel necessary to prepare sales for the scheduled dates. Thus a national forest seems ideally suited for the application of network analysis systems.
The Introduction of Network Analysis
In light of the massive scheduling task facing Forest Service personnel, it was logical to expect individual managers to begin seeking ways to assist them in this area. The beginnings of this search occurred in the early 1960s. During the middle 1960s formal training programs were established to further promote the use of these techniques. By 1970, 18 of the 19 forests in the region had experimented with some form of network analysis in their timber-sale activities. The interest was so wide spread that the Forest Service research organization developed a computer package called “Critical Path Man Scheduling” or CPMS. CPMS was a significant breakthrough in network techniques at the time being capable of handling the multi-project, multi-resource nature of a forest’s timber sale problem. CPMS can analyze 99 concurrent timber sales (or other projects) containing up to 4,500 activities requiring 99 different resource skills.
In short, the need was there and the assistance to deal with this need was available. Virtually all of the individual organizations took advantage of this assistance but very few of the units retained the technique on a continuing basis. It thus provided a unique opportunity to examine the methods of introduction and utilization of the technique and to identify the reasons underlying the retention or rejection of project management techniques by managers who seemingly could obtain significant benefits from them.
Factors Affecting Successful Retention
Prior to beginning this study it was necessary to define what constituted a successful introduction of a project management technique. A successful introduction was defined in terms of the length of time the technique remained in use and this retention time was determined to be continued usage for at least three years following the date of introduction. The attempt to identify the true causal factors affecting and explaining a successful retention was carried out by personal interview. The person interviewed was either the person who actually made the introduction or was closely associated with the introduction and history of technique usage in the particular forest.
As mentioned earlier, there were numerous similarities among the units involved in this study. Foremost of these similarities were organizational structure, task requirements, and type of personnel. With respect to the personnel, they all appeared to be highly motivated to do an effective job, were college educated, and skilled in their job performance. These elements combined to make the various units essentially homogeneous thus eliminating them as causal factors affecting retention of the technique.
The interviews were guided by ten objective type questions related to success factors involved in instituting organizational change. The questions were based on various journal articles dealing with change and on the personal experience of the authors. In addition, two open-ended subjective questions were used to gain insight into other factors which may have been overlooked when structuring the objective portion of the questionnaire. The complete questionnaire is shown in Appendix 1.
The interview results were used to identify nine factors which showed variation from unit to unit among the forests involved. These factors are as follows:
Date of introduction. Timing of the introduction was considered important because of the multi-project nature of the problem. The large number of activities involved in concurrent projects tended to discourage the use of network techniques prior to the advent of CPMS.
Rank of person making the introduction. Top level management support is generally regarded as an essential element for successfully introducing management science into an organization. The level was determined by identifying who had actually made the introduction in each forest and where they were on the organization chart.
Organizational scope of application. Timber sale preparation required the use of many functional specialists in the forest organization. Due to departmentalization, however, some applications were limited to a single function. Limited scope might tend to diminish effectiveness hence retention of the technique.
Planning Horizon. Managers had a wide range of time horizons which could be used for planning resale activities. This was due to a five year lead time normally used in timber sale preparation.
Announced purpose for introducing PERT/CPM. Discussions showed a rather broad range of reasons for introducing the technique. Some were very general, such as long range planning, whereas others were very specific such as scheduling the daily activities of the people involved in presale work.
Software awareness. The computational requirements of the technique might possibly influence the success since they are substantial. The existence of computer programs, and knowing of their existence, could eliminate this problem.
Use of standardized format. The computational requirements could have merely appeared formidable as a manager attempted to deal with as many as 50 apparently unique projects. Standardization of network, mostly by providing common definition, could reduce the complexity of dealing with the multi-project nature of the task.
Recognizing limited resources. Basic PERT/CPM ignores the fact that resources are limited. There are means of incorporating the concept of limited resources which could have enhanced the usefulness hence retention of the technique.
Updating. Updating is the practice of periodically reviewing the status of scheduled versus actual activities and rescheduling if necessary. Due to unforeseen interruptions, the ability to update was considered important. None of the available software packages had this capability although it could be done manually.
The most significant finding produced by this study is that all but one, 18 out of 19, of the units involved had introduced some form of network technique to analyze some aspect of timber presale planning. The managers involved obviously were aware of the potential match-up between PERT/CPM and presale planning. Of the 18 introductions, however, only four met the predetermined criterion for a successful introduction.
Multiple regression identified two factors which appear to explain a successful introduction of PERT/CPM analysis in this organization. These factors were the level of the person making the introduction and the stated purpose of the introduction. These factors were both found to be statistically significant at the five percent level of confidence. The discriminant model for these two factors is shown in Appendix 2.
Interpretations and Conclusions
The stimulus which prompted this study was the authors’ growing concern for the lack of retention of scientific techniques of managing large projects. Network techniques receive a lot of lip service and, in fact, are frequently applied but they seem to fall into disuse very quickly. The concern is why this happens and how it can be prevented.
The statistical analysis of the survey results did provide some insight into the problem in that the two factors listed above accounted for 53 percent of the variance in predicting a successful application of network techniques. Other factors, as yet undetermined, would explain the remaining variance. Thus the results cannot be considered all that useful. That top level management support is necessary is widely recognized as a prime factor in achieving successful utilization of any new technique. The results of this study merely confirm this conclusion. On the other hand, the stated purpose for introducing the technique might be more significant. The conclusion that a broadly stated purpose is the most effective mode of introduction seems to suggest that a major problem is gaining acceptance by those whose work responsibility will be affected by the technique. In other words, if a technique is introduced specifically to schedule, and thereby controls an individual, the technique will be resisted. With network techniques since they depend on data input from the individual involved, sabotage would be very easy to accomplish. With respect to the stated purposes, it is interesting to note the similarity among information requirements. In order to achieve the stated purpose “schedule coordination between individuals and/or units,” the system had to be designed so that it could “schedule individuals.” Actually then the system designed to “budget” which was the broadest and most successful stated purpose was identical in information requirements and output to the stated purpose of “schedule individuals.” This emphasizes the extreme importance of clearly conveying to the people involved the system’s intended use.
It is obvious that network techniques are not achieving the results desired by the managers who introduce them even though the techniques appear to be well suited to the type of problem confronting the organization. A follow-up question concerning this point revealed that although there were some benefits, these benefits were not sufficiently valuable considering the amount of effort required to obtain them. Clearly, then, it is necessary to determine what constitutes valuable information from, the standpoint of the manager who will be using the technique. It would also seem wise to examine the data collection process in order to streamline it. In other words, taking something “off the shelf’ is not satisfactory. This suggests that a system should be designed in consultation with the prospective user paying particular attention to desired output and to data collection procedures. The result will be a user oriented technique tailored to deal with a specific type of problem.
Discriminant analysis yields the following model:
This model was applied to the two categories of forests, i.e., success or failure which resulted in a discriminate value of 2.56. Any similar forest could be tested with this model. If the Y value falls below 2.56, the model predicts success. It is capable of predicting 77.8 percent of the known cases correctly. The F test showed the model to be significant at the .002 level.
Project Management Questionnaire