Technical and institutional factors in the emergence of project management

JPMA-01500; No of Pages 12 Available online at www.sciencedirect.com International Journal of Project Management xx (2013) xxx xxx www.elsevier.com/locate/ijproman Technical and institutional factors in the emergence of project management Stephen B. Johnson Center for Space Studies, National Institute for Science, Space, and Security Centers, University of Colorado at Colorado Springs, 1420 Austin Bluffs Parkway, Colorado Springs, CO 80918 USA Abstract This paper explores the fundamental question of why the practice and discipline of project management emerged during the 1940s through the 1960s in the United States. Although projects have been around for millennia, not until the middle of the 20th century in the U.S. military industrial academic complex did project management become formalized in institutional processes and as an academic discipline. The paper argues that technical complexity and novelty were the primary factors driving project management and its engineering counterpart systems engineering, as a new organizational form. Institutional factors such as the need for legal separation between government and industry created important secondary effects that drove the particular forms in which project management evolved. This paper uses examples from large scale, complex projects of the 1940s through 1960s in the aerospace and computing industries to tease out the fundamental technical and institutional factors that led to the emergence of project management in these two key American industries during this period. 2013 Elsevier Ltd. APM and IPMA. All rights reserved. Keywords: History; Systems analysis; Configuration management; Systems engineering; Systems integration; Weapon system; Intercontinental ballistic missile; Polaris; SAGE; Air defense; Von Braun; Schriever; Ramo-Wooldridge; MITRE; Phased planning; Lincoln Laboratory; Atomic bomb; V-2; Matrix management; Complexity 1. Introduction Recent renewed interest in history among project management (PM) researchers has led to calls to: move beyond the single case study method to address methods and ideas between and across projects (Packendorff, 1995), to understand the evolution of project methods over time and their changing contexts (Engwall, 2003), to understand the many kinds of projects and their behaviors, functions, and measures of success (Söderlund, 2004), call attention to early PM's flexibility and propensity to experiment (Lenfle and Loch, 2010). All of these issues, along with the seven schools of PM research identified in Söderlund (2011) require deeper historical understanding than can be gained by normative assessments of single projects. Viewed historically, PM is a major step in the evolution of how managers gained (or attempted to gain) control of organizations, 14197 Furrow Road, Larkspur, CO 80118, USA. Tel./fax: +1 719 487 9833. E-mail addresses: sjohns22@uccs.edu, kujaketuri@gmail.com. technologies, and workers. Management as a career path developed with the creation of railways in the 19th century in the United States (Chandler, 1977). To manage large-scale, distributed railroad organizations, managers borrowed Army methods in developing systematic management, which they used to control schedules, finances, and cargo. Upper management used systematic management to control mid-level managers and office workers (Yates, 1989). At the turn of the century, Frederick Winslow Taylor developed scientific management, which enabled managers, allied with engineers, to control factory operations and workers (Kanigel, 1998). These methods made the Ford assembly line possible (Nelson, 1992). Taylorist methods morphed into the Quality Control movement in Japan and then into Total Quality Management, which propagated around the world after World War II (Tsutsui, 1998). Project Management came into being in the 1940s 60s in the United States (U.S.) military industrial academic complex (Morris, 1994), in conjunction with operations research and systems engineering (Johnson, 1997). Within this context, PM became the primary managerial technique to develop complex 0263-7863/$36.00 2013 Elsevier Ltd. APM and IPMA. All rights reserved. http://dx.doi.org/10.1016/j.ijproman.2013.01.006

2 S.B. Johnson / International Journal of Project Management xx (2013) xxx xxx new products and technologies. We will see that project management formed as a response to difficult technical and organizational problems with complex military projects. This paper will demonstrate that project management is an evolving technique of organizational problem solving, which was created in a specific time, place, and culture to resolve specific problems. 2. Precursors: World War II World War II was a crucible in which scientifically sophisticated technologies were rushed from research to development to production to operations. Several particularly complex technologies and operations required new methods of analysis, coordination, and organization. These included radar, the German V-2 ballistic missile, the American B-29 bomber, and the atomic bomb. 2.1. Operations research to Project RAND Operations research was initially created in the run-up to World War II in Great Britain to help create an effective air defense system from a network of radar stations linked to fighter squadrons. Radar technology enabled detection of aircraft by reflected radio signals. German bombing raids on London during World War I made the problem particularly urgent. Scientists were crucial in determining how to link radar detection of bombers to the direction of fighters to intercept them. Operations research's reputation was made when the new radar-based air defense system proved itself as a key factor in British victory in the Battle of Britain (Buderi, 1996). Soon British operations research scientists were tackling problems of anti-aircraft gunnery, submarine detection, bomber navigation, and a host of other pressing issues. By 1943 American scientists were using operations research to study critical problems such as antisubmarine warfare, aerial operations to mine Japanese harbors, and bomber formations to ensure maximum protection against enemy fighters (Rau, 2000). In early 1946, the Commander of the Army Air Forces funded Douglas Aircraft Corporation's Project RAND to study intercontinental warfare. As its interactions with Air Force leaders to provide scientific expertise created potential conflicts of interest for Douglas in contract bids, Douglas executives spun it off into the non-profit RAND Corporation. RAND Corporation became an influential force in the development of systems approaches over the next two decades (Smith, 1966). As we will see, for project management, operations research is of significance because when applied to the study of technical feasibility, it became systems analysis, the first step in deciding whether to create a military project. This began to link proto-project management with military, technically advanced projects. 2.2. Systems integration: the B-29 and Mark 56 Gun Director One of the key systems approaches to organize highly technical projects that became essential in the early Cold War was systems integration. It formed in projects to develop complex aircraft and radar-directed anti-aircraft artillery. To understand the linkage between the technical problem-solving and organizational methods that became project management, one must first understand the pre-war process of aircraft design, testing, and production. Specialized organizations for aircraft development did not exist in the United States until the 1930s, leaving aviation initially dominated by individual aviators and their companies. The Army Air Corps procurement process started with the release of specifications for industry. Contractors built a prototype known as an X-Model, which the Air Corps tested by flying it. Test flights resulted in change recommendations, which were incorporated into a Y-Model prototype, whose design also addressed production considerations. After further flight tests, the contractor released production drawings. The Air Corps issued these for production bids, and the winning contractor then built the specified number of aircraft. The Air Corps then added weapons, radios, and other gear, and released the resulting aircraft into the field. To manage this process, the Air Corps typically assigned a single project officer with a small staff. This process changed significantly during the Second World War, as the Air Corps became the Army Air Forces. As the military hurried to put aircraft immediately into production, Congress allowed procurement officers write letters of intent to contractors to rush aircraft into production immediately, with cost negotiations deferred and costs reimbursed. Consequently, the Army Air Forces procurement staff expanded dramatically. As the Air Forces found design problems in testing and combat, aircraft were shipped from production lines to modification centers, which installed the latest changes. After this, the government separately installed weapons, navigation systems, and communications equipment. These procedures did not suffice for complex aircraft like the B-29 and the P-61. For these aircraft, weapons were directly integrated into the airframe, with (analog) computer-controlled gunnery and a pressurized interior. Project officers organized committees to develop and integrate the airframe, electronics, and armaments together into an entire weapon system (Johnson, 2002b). Systems integration also became an issue in the development of the US Navy's Mark 56 Gun Fire Control System, which connected radar to an analog computer that controlled the fire of its anti-aircraft gun. Institutions created problems due to the division of labor within the Navy's Bureau of Ordnance, and between the Massachusetts Institute of Technology Radiation Laboratory, the Navy, and other contractors. Ivan Getting, who headed the project for the Radiation Laboratory, believed that the Navy's difficulties with automated gun directors were due to Bureau of Ordnance's practice of dividing the work into small subcomponents, after a brief initial effort to define the system. This division of labor did not work because of the tightly coupled relationship between the radar, the gun itself, and the gun director's computers, which had to factor in the movement of the ship on which the gun was mounted, as well as the movement of enemy aircraft. In early 1945, Getting proposed changing the Radiation Laboratory's responsibilities from its historical role of designing

S.B. Johnson / International Journal of Project Management xx (2013) xxx xxx 3 equipment, building prototypes, preparing design schematics, and consulting the relevant organizations, to become a true project manager. Getting recommended adding new roles for the Radiation Laboratory, including sending representatives to meetings, testing, assisting with manufacturing, test, and training procedures, receiving copies of contractor correspondence and drawings, accessing contractor facilities, and critiquing designs, specifications, and equipment before their designs were frozen for production. The only authority not included was direct schedule and budget control. Getting acquired his requested authority, leading to successful development of the Mark 56. This model of systems integration became important as Radiation Laboratory scientists, engineers, and managers, propagated it into post-war projects and organizations (Mindell, 2002). 2.3. The V-2 Before and during World War II, Germany developed the world's first ballistic missile, known as the A-4 (Assembly-4) to its developers, but widely known as the V-2, or Vengeance Weapon 2. The project began in 1932 when German Army Ordnance decided to pursue rocketry as a way to deliver explosives farther than the largest artillery of World War I. Army Ordnance recruited the twenty-one year old Wernher von Braun, a member of the amateur rocketry group to direct the effort. He consequently developed liquid propulsion rockets for the Army while also a student at the University of Berlin. In 1932 3, it was little more than an expensive graduate student project, but the ascension of the Nazis in 1933 led to increased funding. Von Braun's small team successfully launched two A-2 rockets in December 1934 (Neufeld, 1995), encouraging Army Ordnance to develop much larger missiles. The desire for secrecy strongly influenced Army Ordnance's strategy to develop the V-2 in-house, or everything under one roof, as the Army's project leader Walter Dornberger later phrased it. This ensured that even though the Army issued contracts to industry and academia to develop rocket engines, guidance systems, high-speed wind tunnels, and launch facilities, the Army team firmly controlled the project performed the integration and testing of the missiles in its own facilities. Von Braun became a legendary project manager, with the technical skills to perform system integration himself, as well as the political and bureaucratic tasks of running a large organization. The major problem von Braun faced was the development and integration of radically new technologies, in which trial and error development, alternate technological paths, and scientific analysis all played prominent roles. Outsiders were amazed that von Braun himself could integrate the entire project (Neufeld, 2007). Particularly important in ballistic missile development was the role of successful failures, in which the team learned from failures, improved the design, and tried again. Despite von Braun's considerable skills, cost and schedule control were problematic, as attested by a 1938 commemorative booklet with the following satirical poem: Der Braun führt das Entwicklungsheer, Paperkrieg fällt ihm oftmals schwer, Das ganze Werk im Stillen Lacht, Wenn Wernher die Termine macht. (Neufeld, 2007) (Braun leads the development army, red tape often drives him crazy. The whole works silently laughs when Wernher makes his deadline). From a team of nearly 350 workers in May 1937, by the time of its first successful test flight in October 1942 it had thousands of workers. With Hitler's endorsement, Von Braun's team turned its attention to production. The need for labor led to a 1943 decision to use foreign concentration camp workers, the lever by which the SS (Schutzstaffel) gained ever-more control over the program. After the British bombing of V-2 facilities at Peenemünde in August 1943, the SS used slave labor to build underground facilities and produce the rockets. The first V-2s were fired at London and Paris in September 1944, and V-2 launches continued from Holland until March 1945. Over its lifetime, the approximate cost of the project was half a billion 1940s dollars, the largest German project of the war (Neufeld, 1995). As Germany collapsed and the Russians approached, the SS ordered the team to western Austria, where von Braun surrendered to the Americans on 2 May. The United States, the Soviet Union, the United Kingdom, and France all sought to acquire German rocket scientists and the remaining rocket components and facilities, with the spoils going to the victors in roughly the above-noted order. The United States Army offered jobs to von Braun and much of his team, and moved them to Fort Bliss, near El Paso, Texas, to construct and test V-2s (Neufeld, 2007). The Soviet Union acquired many of the technicians and a few high-level engineers, and established its own facility in eastern Germany to acquire V-2 technology, under Sergei Korolev (Siddiqi, 2000). 2.4. The Manhattan Project The Manhattan Project, the American effort to build an atomic bomb, was the single largest technical project of the war, costing roughly $2 billion and managed by Leslie Groves of the Army Corps of Engineers. Physicist Robert Oppenheimer led the bomb development project, based at Los Alamos, New Mexico, where physicists ultimately evolved from an academic-style organization to a product-based organization, and developed two versions of the atomic bomb, which were dropped on Japan in August 1945 and shocking the world with its power (Rhodes, 1986). Lenfle (2008) provides an overview of the technical problems of organizational innovations involved in the project. The Manhattan Project was important for project management primarily because it provided a powerful example of project management of a scientifically-based product. When conferences on project management began in the 1960s, the Manhattan Project was inevitably one of the major examples described. It associated project management with scientific prestige and power, a crucial factor in its adoption and recognition.

4 S.B. Johnson / International Journal of Project Management xx (2013) xxx xxx 3. The Early Cold War and strategic weapons The development of the atomic bomb at the end of World War II, and the subsequent development of the much more powerful hydrogen (fusion) bomb in the early 1950s seemed to change the nature of warfare. Nations with nuclear weapons, which by the end of the 1950s included the United States, Soviet Union, Great Britain, and France, put tremendous resources into means of delivering them. A lower priority was the effort to detect and shoot down bombers, and later to detect ballistic missiles quickly enough to launch nuclear missile counterstrikes. While the United States and Soviet Union both initially used long-range bombers, the prospects of uniting nuclear weapons with unstoppable ballistic missiles beckoned as a better alternative. Ballistic missiles reduced the delivery of nuclear bombs from hours to 30 min or less, making rapid detection and response essential. The pressure to develop nuclear delivery systems and the means to detect and defend against them was the spur to the organizational innovation known as project management. 3.1. Aircraft acquisition: the weapon system approach In 1947, the Army Air Forces became the independent service, the United States Air Force. Scientific expertise was harnessed by the creation of the Research and Development Board, which formalized relationships established during the war. Of significance for project management was the passing of the Procurement Act of 1947, which gave the military permanent use of its wartime authority contracting methods. Most important was the continued use of cost-plus contracts, the primary vehicle to contract academia and industry for high-risk research. The cost-plus contract required a significant military bureaucracy to monitor contracts and costs. The existence of the cost-plus contract and its technical bureaucracy provided the resources for the military to pursue technical innovation and to manage it (Kaufman, 1996). In 1949, several Air Force officers argued successfully for greater attention to research and development, leading to the establishment of a new separate command, the Air Research and Development Command, with a new Air Staff position, the Deputy Chief of Staff, Development (DCS/D). The first DCS/D was General Gordon Saville, formerly the commander of Air Defense Command. On his staff was Dr. Ivan Getting, of systems integration fame, and his deputy Colonel Bernard Schriever. (Johnson, 2002b). Air Research and Development Command developed new weapon designs, which it then turned over to Air Materiel Command, which controlled manufacturing and operations. Both Air Materiel Command and Air Research and Development Command officers worked with small project offices as had been done before World War II, with separate procurements for the airframe, engines, and armaments, but now buttressed by the larger bureaucracy available for cost-plus contracting. The problems with this approach became apparent upon the outbreak of the Korean War in June 1950, when the Air Force found that many of its existing and new aircraft were not ready for combat. This problem led to a study called Combat Ready Aircraft, published in April 1951. One of the major issues it highlighted was insufficient coordination and direction of the complete weapon that included the airframe, engines, components, armaments, and logistics support. Dean Wooldridge of the new Ramo-Wooldridge Corporation later described how many prime contractors in the aircraft industry took a black box approach to the airborne instrument field. The job of the instrumentation was considered to be of secondary importance. Their responsibility was to be sure that their components fitted properly into the space allotted and that they operated well. With such instruments now integral to the weapon's purpose, this no longer sufficed (Stambler, 1955). The study team, which included Schriever, recommended that the Air Force create an organization that controlled the complete weapon, by adding planning, budgeting, programming, and control (programming here means allocation of the funds). A single prime contractor should be hired whenever possible. The associate contractor method was also acceptable, in which the Air Force integrated components purchased from contractors. A single Air Force project office was to manage the project, so that this office, and not the component engineers, decided what design changes would be made. Under this new weapon system concept, the Air Force purchased management of new weapon system development. Under the old regime, the Air Force requested and acquired data from a subcontractor and passed that to the integrating contractor. When the integrating contractor needed subcontractor data, it went through the Air Force to acquire it. Under the new concept, the Air Force no longer acted as an intermediary, thus saving time and money (Reed, 1953). Given the Air Force's separation of development under Air Research and Development Command from manufacturing and operations under Air Materiel Command, the new Weapon System Project Office would be run by one Air Research and Development Command officer and one Air Material Command officer, with the Air Research and Development Command officer in charge through development, and the Air Materiel Command officer through production and operations. This weapon system approach spread through the Air Force over the next couple years and were institutionalized in the Cooke Craigie Procedures of March 1954 (Johnson, 2002b). 3.2. Air Force ballistic missiles In the immediate aftermath of World War II, ballistic missiles were an intriguing but low priority technology in the United States. However, each service put some funding into ballistic missiles to preclude the possibility of another service gaining control of the technology. The Army snagged the von Braun team, and was also developing the Corporal ballistic missile at the California Institute of Technology's Jet Propulsion Laboratory. The Navy began development of the Viking rocket with Martin Company. The USAF funded Convair's MX-774 long-range missile program, and a number of other small-scale efforts. After the first Soviet atomic bomb test in 1949, ballistic missile programs were accelerated. The Army accelerated Corporal development and moved von Braun's team to

S.B. Johnson / International Journal of Project Management xx (2013) xxx xxx 5 Huntsville, Alabama to develop the Redstone ballistic missile. The Air Force restarted funding for Convair to develop the MX-774, renamed the Atlas (Heppenheimer, 1997). Further impetus for American ballistic missile projects came after American (1952) and Soviet (1955) testing of the hydrogen bomb. In his position as head of Air Research and Development Command's Development Planning Office, Colonel Schriever learned of the successful American test in 1953. Mathematician John von Neumann informed him that the weight of hydrogen bombs would soon be significantly reduced. This made intercontinental ballistic missiles (ICBM) much smaller and now feasible. Schriever assumed that the Soviet Union would also recognize this possibility, and he successfully argued the case for all-out ICBM development all the way to President Eisenhower (Neufeld, 1990; Sheehan, 2009). Armed first with the USAF's top priority, and later with the nation's highest priority, Schriever established the Western Development Division in Inglewood, California in July 1954. Here, Schriever established a discipline-based organization modeled on academia, and added a new element: the Ramo- Wooldridge Corporation as systems engineering and integration contractor, working with the Air Force to direct contractors. Believing that he needed a far more highly educated, scientific workforce than was available with Convair and other aircraft manufacturers, Schriever directed Ramo-Wooldridge to hire them. Ramo-Wooldridge personnel worked side-by-side with Air Force officers, providing technical direction to other contractors, with Air Force officers providing government authority. Ramo-Wooldridge performed research studies and technical evaluations, prepared and analyzed specifications, prepared work statements, maintained the project control room, performed systems engineering, and tracked relevant science for the ICBM project. Schriever used the associate contractor option with Ramo-Wooldridge acting as the Atlas integrator. Convair bitterly fought this arrangement but lost, in part due to Ramo- Wooldridge's technical contributions, such as determining that a new nose cone design would reduce the overall weight of the ICBM from 460,000 to 240,000 lb, and the number of engines from 5 to 3. Though Schriever had the authority to make technical decisions and control budgets, Air Force procedures still required that he gain concurrence on five different budget appropriations, and 38 approvals or concurrences from the Air Force or Department of Defense. This slowed progress and forced Schriever and his staff to spend inordinate amounts of time simply getting signatures. Schriever made the case for simplification, and in November 1955 he won the authority to bypass these many reporting lines and control one budget through a new Ballistic Missile Committee. His own reporting and budget was separate from the Air Force's regular procedures. Other organizations had to support Schriever, but could no longer hinder his progress or decisions. Schriever's management methods featured parallel approaches, simultaneously developing the missile, its components, and ground launch sites. Concurrency extended to include an entire separate missile design, which became the two-stage Titan ICBM, built by Martin Company. To keep track progress and coordinate tasks, he instituted so-called Black Saturday weekly meetings. Based on the model of the Navy's shipboard combat information centers, he created a management control center to track schedules (Johnson, 2002a, 2002b). The early ballistic missile tests uncovered many problems, with roughly 50% of launches failing. This led to new processes to improve testing, quality control, and new equipment such as thermal vacuum chambers and vibration tables. One of the major difficulties, first uncovered in testing of the Thor missile (a shorter range intermediate range ballistic missile) was that many failures were due to mismatches between the design schematics and the flown vehicle, due to changes made on the vehicle as it progressed through testing to launch. To address this, one of Schriever's lieutenants created configuration control, which ensured that no changes were made to the vehicle unless cleared by a configuration control committee. This ensured coordination of changes with all relevant groups, and thus that the vehicle flew with known and approved equipment. The idea of configuration control developed independently on Jet Propulsion Laboratory's Sergeant missile (a follow-on to Corporal), and perhaps also at Boeing on the Minuteman program (Johnson, 2002b; Sheehan, 2009). On the next-generation solid-rocket ICBM, the Minuteman, configuration control was transformed into configuration management, which became one of the most important tools of the project manager. It was likely created by Minuteman prime contractor, Boeing, or perhaps by its Air Force manager Colonel Samuel Phillips. The configuration control board was expanded to include financial and legal representatives, and anyone that requested a design change was required to come before the board. Managers required that all affected parties provide cost and schedule estimates before approving any change. This small but profound process modification for the first time linked technical changes to costs and schedules. In turn, this provided the basis for managers to control the scientists and engineers. It is much like Peter Drucker's Management by Objectives, where knowledge workers provided the information needed for managers to control their tasks. It linked the hierarchy of managers to the working groups of engineers and scientists through the Configuration Control Board. Implementation of configuration management, along with the improved testing processes and designs (particularly the solid rocket motor instead of more complex liquid-propellant engines) improved the reliability of Minuteman ballistic missiles from the 50% range to over 90% (Johnson, 2002a). 3.3. Polaris On the same day in November 1955 that he authorized Schriever's separate budget and control authority, the Secretary of Defense authorized a joint Army Navy ballistic missile program. The Navy quickly created its Special Projects Office, which followed Schriever's example by being outside of the Navy's regular budget allocations. The Special Projects Office was created ostensibly because none of the Navy's regular design bureaus had the skills to create the novel new system. The Army Navy program reported to the same Department of

6 S.B. Johnson / International Journal of Project Management xx (2013) xxx xxx Defense Ballistic Missile Committee as Schriever, and was explicitly formed as a project organization. After switching from liquid to solid propellants, in December 1956 the Navy severed its connection to von Braun's Army rocket team, and the Polaris program was born (Sapolsky, 1972). Like the Air Force, the Navy used associate contractors on Polaris, and did much of the technical integration itself because no single contractor had the relevant knowledge. However, to perform this integration, it contracted for parallel studies with its existing contractors and had Special Project Office naval staff and technical bureaus decide among the competing ideas. The competition of ideas also led directly to competition for contracts. Core technical capabilities resided in the contractors, which included the Massachusetts Institute of Technology for its inertial guidance systems. Polaris consciously fostered new managerial methods, highlighted by the development of the Program Evaluation and Review Task (PERT) network planning tool to portray and manage schedules. As Sapolsky (1972) describes, PERT was important, but hardly crucial. The Navy used a management center and weekly staff meetings, though it is unclear if the Navy adopted these ideas from the Air Force. Polaris utilized parallel development for some key technologies such as navigation. Thus it adopted or re-created some of the Air Force's methods, but also created some of its own, most significantly internal competition among its contractors, and PERT. 3.4. The rocket team The von Braun rocket team, now the Army Ballistic Missile Agency in Huntsville, Alabama, was from 1950 developing an improved Redstone missile. This was followed by a larger Jupiter rocket, which in January 1958 launched the first American satellite, Explorer I, into space. It also began the development of the much larger Saturn rocket. When the United States Army lost the long-range missile competition to the Air Force, the Army Ballistic Missile Agency's large rocket program became expendable, and the newly-formed National Aeronautics and Space Administration (NASA) inherited it. In January 1960 the Army Ballistic Missile Agency became the NASA Marshall Space Flight Center, with its primary project being the development of Saturn for the soon-to-be-approved manned lunar program, Apollo. Von Braun's approach to project management and systems engineering in some ways differed significantly from those of the USAF and Navy. At the Army Ballistic Missile Agency and in Marshall's early years, the team used the same in-house, test-and-learn development methods it had used for the V-2. Contractors manufactured some components under strict government control. Once the bugs had been worked out of the design generally by investigating failures, the Army Ballistic Missile Agency then contracted manufacturing of the production missile or space launch vehicle (Committee on Government Operations, 1959). One significant change from the German model was the April 1952 directive for the Army Ballistic Missile Agency to utilize prime contractors for production. It contracted with automaker Chrysler to produce Redstone and Jupiter missiles (Neufeld, 2007). Von Braun remained team leader and system integrator, with the team organized in laboratories that developed specific rocket components (Dunar and Waring, 1999). Von Braun (1956) described his approach to PM as being like that of a gardener. He stated that technical projects like rockets are best performed by a team that grew slowly and organically like a tree or a flower allowing the team's organizations and personnel work out relationships to each other. They needed the freedom to learn, and the maximum delegation of authority and responsibility for their products, loosely coordinated by an experienced technical leader (Sato, 2005). Von Braun required all team members to have automatic responsibility to either fix or communicate problems they found, whether or not it was their official responsibility. He utilized a system of Weekly Notes by which all managers two levels below him supplied a single page of progress reports and technical issues. Von Braun responded to these notes by writing comments into the margins; this marked-up set of notes was then copied and distributed to the entire organization, becoming popular reading. This clever method ensured technical issues were rapidly and effectively communicated across the entire organization (Tompkins, 1993). He disapproved of formal controls, whether for PM or systems engineering (Apollo Executives, 1964). This in-house, organic approach worked because von Braun's team had been together developing rockets since the late 1930s. Formal coordination methods such as systems engineering were unnecessary because the team members knew each other and their technology intimately. On Apollo, NASA forced von Braun's team to greater use of contractors, partly because Apollo was so large that Marshall could not develop everything in-house, and partly because U.S. politics demanded it. George Mueller, who came to NASA from Ramo-Wooldridge and headed the human spaceflight program, required Marshall to use Air Force-inspired systems engineering and project management methods. Von Braun initially responded by creating the relatively small Industrial Operations organization, which housed the project managers that managed contractors, project schedules and budgets. These groups began to use PERT and other formal methods such as configuration management. However, the technical muscle remained with Marshall's laboratories which reported directly to von Braun. More formalized project management and systems engineering finally gained momentum at Marshall when Apollo program funding began to decrease at Marshall in the late 1960s. Marshall was forced to diversify beyond rocketry, and its informal, decentralized methods no longer sufficed. As the Germans retired and as they began to develop rovers, space stations and science satellites, Marshall personnel no longer knew all the details of these very different applications. Nor did the new American personnel have decades of experience together. In these circumstances, von Braun and Marshall finally began to adopt systems engineering and more formal project management (Johnson, 2002a; Sato, 2005; Tompkins, 1993). Von Braun's management methods were an interesting outlier to the Air Force-dominated trend. It required long term relationships and projects, which in the mobile and fast-moving culture of the

S.B. Johnson / International Journal of Project Management xx (2013) xxx xxx 7 United States was not typical. They did not spread beyond Marshall and did not last far beyond the retirement of the German V-2 engineers. 3.5. Automating strategic defense Defending against incoming aircraft and missiles required early detection of incoming aircraft or missiles with radar or other sensors, assembling the information determine their trajectories and targets, and then deploying fighters and ground-to-air missiles to shoot down aircraft, and to command counterstrikes before the missiles and bombs hit. For the U.S. military, from the late 1940s to the late-1950s the primary concern was Soviet nuclear-armed bombers coming over the Arctic. This air defense problem drove the creation of the world's first computer-based air defense system, known as SAGE (Semi-Automatic Ground Equipment). It contained the International Business Machines (IBM) AN/FSQ-7, the first large scale real-time computer system, with by far the most sophisticated software in existence at that time. The development of SAGE, the AN/FSQ-7, and associated software was critical for project management because many of the leaders of the future computer and software industries learned their craft on SAGE. SAGE originated from Whirlwind, a Navy-funded Massachusetts Institute of Technology (MIT) project to develop an analog computer to analyze aircraft stability and control. The project leader, Jay Forrester, shifted Whirlwind to a general purpose real-time digital computer to automate functions typically performed by humans in combat direction centers. Direction centers were the locations that received information about incoming aircraft and other threats, in which commanders decided how to respond, and then sent commands to implement the countermeasures. When in 1951 the Air Force decided to automate air defense because of rapidly increasing aircraft speeds, it selected Whirlwind. MIT created Project Lincoln in 1952, which soon became the Lincoln Laboratory. IBM won the contract to manufacture the production versions of the computer (Hughes, 1998; Johnson, 2002b). Lincoln was granted significant autonomy and flexibility to bend procurement regulations, with guidance from Air Force Development Planning Office chief Ivan Getting. Lincoln's organization was very centralized for a university project, requiring extensive documentation and testing of every component and its integration into the computer. Forrester had different series of memoranda for administration, conferences, schedules, engineering notes, and official reports. He strongly emphasized systems engineering, being one of the first to use this term. A block diagrams group developed and controlled the design through what would later be called a computer's architecture. When IBM joined the project, these formal mechanisms helped, but a series of meetings between the MIT and IBM engineers and managers was found necessary. Lincoln's engineers found themselves directing the project, but without having any formal authority. It did so through Technical Information Releases, which without directing anyone stated what needed to be accomplished. Other contractors knew to interpret these as directives. One major difficulty was the exponential growth of the computer's software, a problem which would plague generations of organizations to come. SAGE software was particularly difficult, because there was no easy way to determine exactly what it should do. To figure this out, the SAGE protoprogrammers worked directly with Air Force operations personnel to determine what they did in a command center. From this interaction Lincoln personnel determined what could be automated. In addition, the software had to understand the functions of the radar and other sensory inputs, and of the fighter aircraft and guided missiles. It became the system's brain, coordinating all of the parts. The software soon overwhelmed Lincoln, and in 1955 RAND Corporation was brought in to address the problem, since it had been simulating direction centers for several years. Lincoln and RAND subdivided the programming and specifications into 35 subprograms, and assigned teams to each. They greatly simplified the software specifications, but even so the software totaled over 100,000 instructions, far larger than any other program of the time. Desiring to focus on research instead of development, MIT decided to withdraw from the project after initial development was complete, and RAND expressed similar misgivings about the huge programming effort, which distracted from its research. In 1956 7 RAND spun off SAGE programming into a new non-profit corporation, the System Development Corporation. Many software programmers trained at Lincoln and System Development Corporation were hired by the budding computer industry, along with their colleagues at IBM, which itself became the titan of the computer industry for the next twenty years. In 1958 MIT separated systems engineering work on air defense systems from Lincoln into the new non-profit MITRE Corporation (Johnson, 2002b). 4. Standardizing and propagating project management The major military projects of the 1950s featured high national priority, large budgets, short schedules, and high levels of novelty and complexity. From these common features PM emerged as a practice in the U.S. military industrial academic complex. From common patterns among these projects emerged project management standards, which in turn became the basis for project management as a discipline. These became standard practices in many industries, as they learned from the military, NASA, and the aerospace and computing industries. 4.1. From operations research to systems analysis After World War II, operations research applications expanded dramatically in the United States and Great Britain. While most practitioners worked on military problems, the prestige of physicists and mathematicians helped propagate operations research into other fields (Fortun and Schweber, 1993). The first major operations research textbook by Morse and Kimball, 1950 highlighted World War II applications, but in McCloskey and

8 S.B. Johnson / International Journal of Project Management xx (2013) xxx xxx Trefethen, 1954 published a text that highlighted operations research's application to commercial and social problems, including supermarkets, sales, the printing industry, and agriculture. Logistics and inventory management became a common commercial operations research application. By the early 1960s, operations research became an important element of management science, addressing scheduling, inventory management and queuing, and statistical analyses, as described in Enrick's (1965) text. PERT and CPM (Critical Path Management) were direct applications of operations research methods to management scheduling problems. While CPM was developed in a production operations environment, PERT evolved in large-scale development. PERT's use became so prevalent that some considered it and PM nearly synonymous, as shown in Martino's (1964) CPM primer and (1968) project management texts. Operations research's evolution into systems analysis is equally significant. Systems analysis is simply operations research applied to determine technology feasibility in a project's early stages. Project RAND and its successor the RAND Corporation provided several early examples of systems analysis, such as the path-breaking 1946 study of the feasibility of Earth-orbiting satellites. RAND economists Charles Hitch and Alain Enthoven combined technical feasibility studies with economic criteria to assess project benefits and costs. When Robert McNamara became Secretary of Defense in 1961, he hired Hitch to become the Comptroller for the Department of Defense, along with Enthoven (Hughes, 1998). McNamara had used statistics to support Army logistics in World War II and after the war rose through the Ford Motor Company to become president by reforming the company's finances (Shapley, 1993). Upon hearing Hitch's presentation on financial reforms to the strategic nuclear forces, McNamara exclaimed, That's exactly what I want [but with] one change. Do it for the entire defense program. And in less than a year. This was so fast and unexpected that the military services were caught completely off-guard. The heart of the new Programming, Planning, and Budgeting System was allocation of funding by the product or use, instead of by service or organization. It enshrined systems analysis tied to economic analysis as the starting point for all proposed U.S. military projects. 4.2. Standardizing systems management By the late 1950s, institutional turmoil rocked the Air Force's ICBM and air defense programs. In air defense, the withdrawal of MIT and the creation of MITRE and System Development Corporation were symptoms of weak military management. Air defense involved the Army, Navy, and Canadian military as well as the Air Force these organizations controlled many of the sensors, facilities, ground-to-air missiles, and interceptors. The Sputnik crisis led to USAF studies of how to better organize research and development, and to the elevation of Schriever to head Air Research and Development Command in April 1959. Schriever's appointment ensured the use of his management methods, except for its use of contractors like Ramo-Wooldridge. Ramo-Wooldridge, which became TRW, faced strong opposition from aircraft contractors, who argued that it was acquiring inside information to build its own business. Their objections gained force when TRW won satellite contracts. The resulting Congressional investigations led in 1960 to the separation of Air Force technical direction work from TRW to a new non-profit, The Aerospace Corporation. That same year, Schriever's team began to develop new management procedures based on ICBM precedents. These new 375-series of Systems Management regulations governed major USAF development programs. One significant battle in the creation of the 375 procedures pitted Schriever's ICBM process, which assumed that program requirements could be defined at program startup, versus the air defense conception, in which requirements were learned and defined during a project's course. Schriever's procedures won. Ironically, Schriever's independence and authority for Air Force research and development was undermined by his success. When McNamara became Secretary of Defense, he was committed to strong program management. Schriever's methods fit the bill, and in 1965 the Department of Defense made them the basis for its new research and development regulations (Johnson, 2002b). McNamara also instituted a preliminary design phase after feasibility studies for all Department of Defense projects to ensure that decision-makers had good cost and schedule estimates before committing to a new major project. If these estimates showed insufficient performance compared to the cost, the proposed project was cancelled (Johnson, 2000). 4.3. Project management as a practice and discipline How did researchers and managers view these technical and institutional issues and developments? A common pattern that aircraft company managers first perceived was project proliferation, and the need to strengthen the role of managers in each of them. Within the aircraft industry of the 1940s and early 1950s, the typical project loosely coordinated strong technical departments such as the airframe, instrumentation, controls, etc. The implementation of the Air Force weapon system approach mandated complete and absolute authority for the project manager over his project. However, technical specialists had the right of appeal channeled through the authority of his technical department head. (Uhl, 1954) With the proliferation of Cold War projects, a single company often managed several projects. This led to dual reporting lines, which were eventually called matrix management, a standard throughout the aerospace industry (Bergen, 1954; Lanier, 1956). Project managers were frustrated by scientists and engineers, who had such a radically different way of thinking This was often completely foreign to the manager; his values are too different, and he finds it difficult to understand what makes professionals in technical fields behave so queerly. (Orth, 1957) Managers assumed that they had to plan, supervise, and control company activities, and that there were definite ways in which these activities should be carried out. Another typical assumption of the period was that an individual's success is primarily a product of his position on the management ladder. These assumptions derived from Taylorist strategies to expropriate

S.B. Johnson / International Journal of Project Management xx (2013) xxx xxx 9 worker knowledge so as to gain control of the organization's productive process. Technical professionals and tasks contradicted these assumptions, which did not surprise Peter Drucker (1950), who had already diagnosed knowledge work and workers as the management challenge of the era (Drucker, 1950). How to manage knowledge workers was not obvious. According to Air Research and Development Command Commander Thomas Power in 1955, We are convinced that it is possible and necessary to plan, schedule, and manage research and development (Power, 1955). Another manager stated that attempts to apply standard control and evaluation techniques to the research program have proved very disappointing in many companies. He also noted that if the output to be evaluated by management is the ultimate economic result, it might seem logical to make this result the basis for controlling the progress of the R&D program. This is not generally feasible, however, because control requires indications of current progress, so immediate changes can be initiated if necessary. (Rubenstein, 1957) This was the problem. Managers had little idea how to measure research and development progress. For large-scale development, the solution was configuration management, in which managers required scientists and engineers to provide cost and schedule estimates for every change to a baselined design, which was then frozen.thisprovidedcost and schedule data, which when portrayed through tools like PERT and CPM enabled reasonable predictions of product cost and delivery. While PERT and CPM garnered accolades, the real breakthrough was configuration management, which provided the cost and schedule data without which these tools were useless. PM was a hot topic in the 1950s aircraft and missile industry, with many trade publications on the topic (though not usually with the project management label). The publication of Gaddis's (1959) seminal paper on project management in the Harvard Business Review in 1959 was important mainly because it broadcast this internal discussion to academics and business leaders in other industries. He put forth the fundamental characteristics of projects as organized by task (vertical structure) instead of by function (horizontal organization), the linkage to complex technologies, and to the management of a higher proportion of professionals. Gaddis noted that project managers are often flying blind because they had few means to measure performance so as to control the project. He had not yet learned of configuration management, which was just then being developed on Minuteman. Academic research on project management (PM) began as early as 1962 when Arizona State University management faculty member Keith Davis (1962) published a paper of empirical observations of the tasks that preoccupied project managers, mainly planning and controlling. Davis elaborated on Gaddis, arguing that the primary reason for PM organization is to achieve some measure of managerial unity, in the same way that physical unity is achieved with the product. The first books on PM were both published in 1968: Martino's (1968) Project Management, and Cleland and King, 1968 Systems Analysis and Project Management. Cleland was an associate professor of management at the Air Force Institute of Technology at Wright-Patterson Air Force Base, teaching officers how to manage weapon acquisition. Martino's book emphasized scheduling techniques such as PERT and CPM, while Cleland and King, 1968 book also included topics such as the systems concept, McNamara's reforms, and classical topics such as organization, authority, and control. The next year, the Project Management Institute was established in Philadelphia. Only one of the PMI's founding members was from the aerospace community, showing PM's spread. McNamara's Department of Defense reforms, which instituted PM and systems analysis across the military, received academic attention then and to the present. Even more attention-getting was the American space program, and in particular the Apollo moon landing in 1969. Congress held hearings to learn and propagate the secrets of NASA's program management (Apollo, 1969). NASA did not advertise that Apollo management was largely derived from the United States Army and the Air Force (Johnson, 2002a). American space and computer successes fed European fears of a technological gap, which some believed to be a gap in methods of organization (Servan-Schreiber, 1969), or a management gap (Spencer, 1970). These concerns motivated Europeans to create their own space programs and to collaborate with the United States to learn NASA's methods (Johnson 20002a). PM had become a significant factor in international industrial policy, as well as an object of academic interest and practical action. 5. Conclusion PM and its technical counterparts systems analysis and systems engineering were primarily created to resolve the combined issues of technical complexity and novelty. If a technology is complex, but not new, then organizations will have already adapted to it. Conversely, if a technology is novel but simple, then individuals and organizations will simply perform the tasks. In either of these cases individually, there is little reason to create new organizations or processes. However, in combination the technologies are too complex to easily understand, and being novel, there is little historical experience. These technical product uncertainties translated into cost and schedule uncertainties. Faced with cost and schedule overruns, managers had to find a way to manage technology development, which could not be done unless they found a way to address the large knowledge uncertainties. Why did the problem of new complex systems lead to project management? Revisiting the twin issues of novelty and complexity, if either occurs or neither occur, there is no reason to make changes to address technologies. However, when both occur together, classical management methods of systematic and scientific management do not work because these assume that managers can master workers' knowledge. For complex new knowledge, techniques of knowledge appropriation did not work because the new knowledge did not yet exist. Once scientists and engineers standardized that knowledge into production lines or organizational procedures, classical management methods could and did resume. In the interim, new methods to organize knowledge creation were needed; this was project management.