The early literature that shaped the project management profession in the years before the Project Management Institute was born can be classified into five categories: management theory, the nature of projects, how to schedule and control projects, how to manage projects, and critical analyses.
Management Theory
In his recently published bibliography of project management [PMI®, 1999], David Cleland cites a 1949 book by Henri Fayol as one of the earliest writings recognizing the use of teams to accomplish specific tasks (i.e., projects). In 1961, Burns and Stalker recognized the differences between mechanistic and organic organizational forms in The Management of Innovation. Also that year, Joan Woodward identified alternative ways in which work could be organized. She recognized four categories as being appropriate for management as projects: “units to customer requirements; prototypes; large equipment or stages; small batches to customer order.” Cleland cites Feinberg's 1969 The Prometheus Project: Mankind's Search for Long-Range Goals for adding insights.
Learning begins to jell into knowledge when ideas and practices hit the printed page. In celebration of PMI's 30th anniversary, let's look at the scholars and practitioners who first put project management into words.
Recent—and still incomplete—library research has also identified a line of writings dealing with the efficiency and effectiveness of research and development, especially on large defense systems procurement. One name that appears in this literature is J. Sterling Livingston, known to have been instrumental in the development of PERT COST.
The Nature of Projects
In 1959, a Business Week article by Astrachan provided an overview of PERT/CPM techniques along with their historical development and application. This article, apparently the first evidence of a growing concern with projects, is still relevant today.
Paul Gaddis, a practitioner with Westinghouse Corporation, wrote about the project manager in the Harvard Business Review that same year. And in September 1959, there was an article similar to Astrachan's by R.L. Martino in Oil/Gas World.
James Kelley was trying to understand road-building projects as he reported on his research at a symposium on Computers and Management Decisions at Case Institute of Technology early in 1957. Later that year he wrote an internal (to Remington Rand UNIVAC) progress report on “The Construction Scheduling Problem.” In 1959, he and Morgan Walker (then with DuPont) presented the first article on Critical Path Project Scheduling (CPPS).
Francis M. Webster Jr., Ph.D., PMP, is a Fellow of the Project Management Institute and acts as PMI Historian. For many years he was PMI's editor-in-chief. Longtime members of the Institute know he was also the voice behind PM Network's “Olde Curmudgeon.”
Another early article exploring the nature of projects was “On the Anatomy of Development Projects,” published in 1960. The author, Peter Norden, conducting research for his Ph.D. dissertation while at IBM, was trying to understand how they could improve performance on developmental projects. Parts of his research were presented as early as May 1959. He cited the following notes on an American Management Association seminar on Budgeting and Controlling Engineering in the fall of 1958: ‘(1) It is possible to relate past or completed projects to new projects in arriving at a usable factor for estimating project costs. (2) There are regularities in all projects. (3) It is absolutely necessary to break down a project into small known components for estimating purposes.”
Reader Service Number 050
Norden also cited a survey to determine the nature and variety of practices in the management of R&D projects, along with a private communication, “Notes on Engineering Manpower Forecasting” with a P.R. Erdwinn, dated 1 April 1957. These references are cited to illustrate the high level of interest in R&D projects at this time.
An article in Armed Forces Management in February 1961 profiled Lt. Gen. Bernard A. Schriever, then commander of the Air Research and Development Command. It credited him with creating the “concurrency concept” as he applied it to the development of the Atlas intercontinental ballistic missile weapons system. To implement concurrency, he initiated a system, Program Evaluation Procedure (PEP), “that in many ways led to PERT, only without mechanization.” PEP was later computerized. Development of the concurrency concept began in 1954. A March 1971 Project Management Quarterly article by Richard Olsen notes that “The term ‘project management’ did not come into popularity until after 1954 when [then Colonel] Schriever put together a team of men to manage a crash project for the U.S. Air Force's missile program. It was not long after this (1958) that PERT was developed to aid in the management of the U.S. Navy's Polaris project.” This article mentions that Vice Adm. W.F.ľ Raborn, program manager for the Polaris project, visited Schriever to learn how the Atlas program was being managed sometime before PERT was developed. Thus, it appears that, more than anyone else, Schriever can be identified as the “Father of Modern Project Management.”
How to Schedule Projects
Prior to World War II the most common method of scheduling projects was the stillpopular Gantt chart, documented by Henry Gantt in 1919.
Adameski presented an even earlier application of graphics, the “Harmonogram,” in a paper in 1931. In an article in the March 1976 Project Management Quarterly, Edward Marsh indicates that this technique was being used in the late 1800s. Each task was represented by a strip of paper, the length of which signified its duration. These strips were moved on a chart marked in the same duration units. Thus, it is similar to a Gantt chart with the axes reversed.
About 1936, a report was published by Goodyear Tire and Rubber describing “Line of Balance,” a technique for scheduling small-lot production jobs. It became the basis for scheduling projects as evidenced by a later U.S. Navy report dated November 1966.
The earliest evidence of a successful attempt to schedule projects using computers was ‘Application of a Technique for Research and Development Program Evaluation,” a 1959 article published in Journal of the Operations Research Society, in which Vice Adm. Raborn attributed a savings of a year in the Navy's Polaris program as a result of using PERT.
CPPS refers to the time-cost trade-off concept developed by UNIVAC and DuPont. It was first presented at the Eastern Joint Computer Conference in 1959 in a paper, “Critical-Path Planning and Scheduling” by James E. Kelley Jr. and Morgan R. Walker. CPPS had two significant advantages. First, the calculations could be repeated several times in much less time than with manual methods. Second, from these several iterations, the least cost schedule could be determined, considering other costs associated with the project. This was a significant advantage. A disadvantage at the time was that there were not many UNIVAC computers on which the calculations could be performed and each iteration took a significant amount of computer time. Another disadvantage was that the linear programming technique and the size of computers severely limited the number of activities that could be processed. Soon these disadvantages outweighed the advantages. Kelley and Walker then simplified the technique, eliminating the time-cost trade-off feature, and presented the Critical Path Method (CPM). For a more complete version of this story, see The Origins of CPM: A Personal History [PM Network, Feb. 1989].
Receiving much less publicity was a report published about 1960 at Wright-Patterson Air Force Base on their approach to PERT, called PEP. The significant difference was in the network diagram. In PEP, no two real activities were connected directly. A dummy activity was used in between. This resulted in substantially more dummy activities and to PEP's early demise. The early, manually prepared PEP charts presumably resembled what we now call a connected bar (Gantt) chart.
Following these developments came a flood of articles, some explaining the theory underlying the calculations in these techniques, others criticizing the theory, and some suggesting extensions. One paper argued that three time estimates in PERT were not adequate to describe the shape of the distribution. The presenter wanted at least five time estimates per activity. Another argued for the inclusion of other distributions such as a simple triangular distribution. The former elicited groans from practitioners while the latter has been adopted in current commercially available software programs.
Extensions included the consideration of resources required by time period versus by availability. An early article by Berman presenting the basic concepts for this was published in 1964. An article by VanSlyke in 1963 presented a method for overcoming the greatest deficiency of PERT, the critical path being dependent upon only the set of activities on that path. This new approach to PERT, the Monte Carlo approach, is widely used today to assess risk in the time dimension.
One of the most interesting extensions actually appeared in 1962 as a letter to the editor in the Journal of the Operations Research Society. Titled “Critical Path Planning: PERT Integration,” it combined the time-cost trade-off concept with the probability calculations of PERT to determine how much extra costs would be required to achieve specific odds of completing the project by a specified schedule date. It was used to schedule real estate construction projects at General Motors on which there was a penalty for late completion.
A major publication appeared in 1962. A Guide for the Implementation of PERT/COST, published by the U.S. Department of Defense, had an introductory memo signed by then-Secretary of Defense Robert S. McNamara and Associate Administrator of NASA Robert C. Seamans Jr., encouraging its use by defense/space contractors. It generated a flurry of activity, including meetings sponsored by the DOD explaining it to contractors and a “PERT/COST Coordinating Council” to resolve problems in its implementation as well as act as advocates in its members’ employing companies. A later directive by McNamara called for each DOD service to implement PERT/COST on two major defense system programs. This was done, but the results were less than dramatic and defense contractors were resistant. PERT/COST slipped into oblivion, eventually succeeded by Cost/Schedule Control Systems Criteria (C/SCSC).
During this time, John Fondahl was conducting research at Stanford University on scheduling for the construction industry He was responsible for developing the SPRED computer program. A report was published detailing the program and Fondahl published an article in the August 1968 Western Construction titled “Let's Scrap the Arrow Diagram.” Largely through his efforts, the precedence diagram (activity-on-the-node) supplanted the arrow diagram (activity-on-the-arrow) that had become popular due to PERT and CPM. His graphic method was identical to that used by industrial engineers and computer programmers to draw flow charts. But it was not until the advent of project scheduling software developed specifically for the micro- or personal computer that the switch was made by most users.
Two other techniques developed in the 1960s, DCPM and GERTS, added considerable additional capabilities beyond the classical PERT/CPM models, such as defining parallel paths for developing a single deliverable and looping back for rework if a component does not meet performance requirements. These are situations often encountered on R&D work.
The first book on PERT, apparently an applications manual prepared within Federal Electric Corporation, was published in 1963. Another book written by two academicians, Muth and Thompson, was probably for classroom use. The first comprehensive book on the theory of project scheduling, Project Management With CPM, PERT and Precedence Diagramming, by Moder, Phillips, and Davis, was published in 1964. It remains the most thorough treatment of the mathematics of scheduling, with the most recent edition dated 1983. CPM in Construction Management, authored by J.J. O'Brien in 1965, was more practitioner oriented.
How to Manage Projects
Generally the early literature in this area addressed three issues: managing projects, organizational concepts, and using the matrix approach to organizing for projects.
Managing Projects. Many of the earliest articles on projects were in engineering forums. For example, Keith Davis’ 1962 article, “The Role of Project Management in Scientific Management,” was published in the IRE Transactions on Engineering Management. In 1964, Cleland, writing in Business Horizons, answered the question “Why Project Management?” In 1965, he published “Project Management: An Innovation in Management Thought and Theory” in the Air University Review. John M. Stewart's “Making Project Management Work” was an early article that gave an overview of the use of project management in contemporary organizations [Business Horizons, Fall 1965]. Roland W. Gutsch presented “Project Management by Means of a Network System of the Third Generation” in 1969 at the first meeting of INTERNET, the European counterpart to PMI.
Several books were published in quick succession by Miller, Martino, Archibald and Villoria, Cleland and King, Steiner and Ryan, and Lock. All of these at least implied that projects were large undertakings, for the government, and generally associated with defense or space efforts or with construction. Thus, the perception developed that project management was only applicable to such projects. This was probably appropriate at the time due to the capabilities of the computer programs for scheduling and to the computers themselves.
One of the most insightful but underrated books published in this time period was James E. Webb's Space Age Management: The Large-Scale Approach [1969]. Webb, Administrator of the National Aeronautics and Space Administration in this time period, describes policies and techniques he used to manage NASA during the space race and the realization of landing a man on the moon.
Organizational Concepts. In 1961, Kast and Rosenzweig in “Weapons System Management and Organizational Relationships” [Academy of Management Journal] described the use of project teams in the management of military weapons systems. Morrison critically analyzed how the U.S. Air Force used project teams in the design, development, and production of military systems using “The 375 Series” of manuals as of 1967. Marquis and Straight presented an analysis of “Organizational Factors in Project Performance” published by M.I.T. in 1965. They showed the relative importance of criteria used in evaluation of project performance in R&D. They found that technical performance was overwhelmingly the most important criteria. They also found that the importance of other criteria varied depending on the role of the observer. Cleland presented the organizational issues involved in project management in 1966. Lawrence and Lorsch first wrote about the role of the “Integrator” in 1967 in the Harvard Business Review, and expanded on this concept in their book Organizations and Environment: Managing Differentiation and Integration in 1969.
Interesting contributions during this period were typologies of project organizational approaches. In 1962, Keith Davis described the expediter, coordinator, confederation manager, and general manager as roles in scientific management. Cleland referred to an assistant to and matrix approach in a 1967 Business Horizons article. Steiner and Ryan mentioned influence, matrix, and pure project approach in their 1968 book Industrial Management.
The Matrix Approach. The matrix approach to managing projects was a substantial departure from bureaucratic theory. Thus, it quickly became a fertile subject for academic research and writing. John F. Mee's “Matrix Organization” in the July 1964 Business Horizons is believed to be the first use of the term “matrix” to describe an organizational concept. Shull published the book Matrix Structure and Project Authority for Optimizing Organizational Capacity in 1965. Cleland wrote about “Understanding Project Authority” in 1967, describing the use of both de jure and de facto authority in the matrix organization. Later that year he joined Munsey to write “Who Works with Whom” for the Harvard Business Review, explaining how to use a Linear Responsibility Chart to assign authority and responsibility in the matrix organization. Cleland went on to explicate the conflict over authority and responsibility between the project manager and the functional manager in a matrix organization in “The Deliberate Conflict” [Business Horizons, 1968]. Thus the concept of a matrix organization was well established by 1968.
Critical Analysis
In 1962, Pocock reviewed the efficacy of PERT in program planning and predicted a bright future for the development of techniques and for a general theory of managing projects. In 1966 Poust and Rubin were concerned with the problems of laboratory managers in managing research and development. Wearne and Cunningham published a book on Problems and Efficiency in the Management of Engineering Projects in 1966. Rubin wrote about “Experience as a Factor in the Selection and Performance of Project Managers” for IEEE in 1967. Farris described the motivational factors for scientists working on projects in “Some Antecedents and Consequences of Scientific Performance,” also published in IEEE Transactions in 1969.
THUS, THE CONCEPTS of modern project management were largely defined in the late 1950s and the 1960s in literature scattered among a variety of journals and proceedings. One of the services of PMI has been to provide, since 1969, a single set of publications and proceedings in which the majority of project management literature has been concentrated.
The range of concepts and techniques presented in the early years, compared to those currently in use may provide a picture of what is to come. There is much in the early literature that is not in use today. Many of the techniques died prematurely due to limited computing capabilities. The ability of project managers to learn and accept so many new ideas and techniques at one time was also an important limitation. As the tools now available become second nature to project managers, they will seek more capabilities. Some of those capabilities may be reborn from the early literature, for, as Ecclesiastes once wrote, “There is nothing new under the sun.”
