Page images
PDF
EPUB

(The membership of the Task Force consisted mostly of people from the field centers.) With two-thirds of the agency's budget, the Office of Manned Space Flight (OMSF) was the most powerful program office in NASA. And conditions were uniquely favorable for Dr. George Mueller, who headed OMSF, to take the initiative. The first lunar landing was anticipated within the next month; the Air Force had just cancelled its Manned Orbiting Laboratory, the only remotely conceivable contender in manned space flight; and NASA had a new Administrator brought in from the outside who did not yet have his hands on the levers of influence. By then, there was a consensus among OMSF managers that the Shuttle was feasible, that it would draw on the expertise gained in Apollo, and that there was enough support behind it to sell it to the rest of the agency and possibly DOD. What applies to the Shuttle applies with equal force to other flight projects. What happened was that the field organizations - Marshall Space Flight Center (headed by Wernher von Braun) and Manned Spacecraft Center (headed by Robert Gilruth) took the lead in proposing new projects and convincing the Headquarters people to go along.

In project work, considerations of substance can (and should) determine considerations of procedure. The decision on whether work will be done in house or by contract depends on the nature of the project. Except in special cases, the option of doing everything in house exists mainly as a theoretical possibility; most laboratories lack design, test, and fabrication personnel and the necessary facilities. A more common approach is for the laboratory to do the conceptual studies and overall management, with the detailed design and most of the hardware contracted out. Depending on the laboratory's strengths, one of two approaches is used:

• The laboratory acts as project manager, designer, and systems integrator, with specific components, or "black boxes," farmed out to subcontractors.

The laboratory acts as project manager, with a prime contractor handling the complete hardware program from detail design through delivery.

It is important to understand that, in contracting out, an agency like NASA is not simply making a virtue of necessity. NASA stands in relation to its contractors somewhat as an industrial firm like General Motors stands in relation to its suppliers. Antitrust considerations aside, General Motors has been in a position to encourage competition among its suppliers, playing one off against another, as well as seeking (and getting) the most favorable combination of cost and design: turning to this firm for tires, to another for suspension system components, to a third

for engine valves and temperature controls.* Vertical integration could hardly promise as much, and that only at much greater cost and internal complexity. Similar considerations have been at work in NASA, the Air Force, and the Atomic Energy Commission from the beginning. For reasons discussed in Chapter III, a reversion to the arsenal system after the Second World War was not politically feasible. But especially since the mid-1950s (see the Polaris case study later in this chapter), the large Federal technology development agencies have deliberately encouraged potential contractors to bid for agency work, to sponsor their own self-initiated independent research and development reimbursable as a percentage of overhead, and to submit unsolicited proposals against the time when they would be incorporated in formal requests for proposals.

Whether the benefits of large-scale technology development carried out by contractors still outweigh the costs is open to question. The cost of entry into aerospace technology development in the late 1950s and early 1960s was lower than it is today but it was still large. Today it is prohibitively expensive for a firm to enter in any capacity other than that of subcontractor for a federally sponsored program. Some of NASA's largest contractors in the 1960s, like Boeing and Grumman, have largely withdrawn: to concentrate on commercial work in the case of Boeing, on defense work in the case of Grumman. All of this has greatly enhanced the Federal role in technology development and has made careful planning and thinking within the government and its advisory structure much more important than it has been in the past. It has also enhanced the importance of giving more freedom to contractors and to put as much of the responsibility for the project management and execution on the contractor. All of this must be done using procurement regulations that were written by people who were not always sensitive to these considerations.

Contracting out projects has important advantages over doing work in house, especially where agency personnel ceilings are fixed. In Chapter VIII we shall consider the use of contractors for support services

everything from trash removal to writing computer programs to

* Indeed, GM may consider it even more important to find suppliers for a variety of parts than to stimulate competition among suppliers of a single part — "Firm A for tires today, Firm B for tires tomorrow." The reason is that GM's overwhelming size vis-a-vis its suppliers gives it formidable bargaining power. Uniroyal, which manufactures the tires for many GM cars, needs GM far more than GM needs Uniroyal. Moreover, the recent agreement between GM and Toyota will place the original equipment manufacturers who supply GM in a position analogous to those defense contractors who are losing business. As more parts are manufactured overseas, these suppliers will be stuck with excess capacity, while coming under pressure from GM to reduce costs still further and improve productivity.

managing entire installations. But NASA's philosophy of using contractors has been based on certain principles which carry over into its technology development work: The rapid buildup of large projects has precluded reliance on government employees alone; it is Federal policy not to develop capabilities that are already available in the private sector; and it is better to let the up-and-down swings in manpower take place in the contractor, rather than the civil service, work force (ref. 108). Also, contract employees do not normally count against an agency's personnel ceiling. Within limits, agencies like NASA and DOD have great flexibility in their use of contractors, even extending to the right, affirmed by the Federal courts, to lay off their own employees before laying off contracted personnel (ref. 109).

Thus the final shape of a project will depend on a good deal more than the availability of funds to get the work done. There are many interdependent elements: the ability of the lead center or laboratory to define a mission, the availability of contract support, the particular capabilities of the lead center or laboratory, and the speed with which the center or laboratory can move from preliminary analysis to design definition to a definition in detail of the project approach. Clearly, center and laboratory managers must be able to "cost out" projects. Project costs are normally estimated by one of two methods. The first is through detailed comparison of previous similar projects of known costs; the other is to generate costs by a complete "from-the-ground-up" work breakdown, sometimes with the assistance of computer models. This latter approach, a kind of zero-based budgeting, examines the efforts involved in every element by the required manhours. Finally, all the elements are added up to develop the overall cost. The first method is easier and is as reliable if, and only if, the comparison program really is similar.

Just how difficult cost estimating of large projects really is, can be shown by table 4, which reviews NASA's early projects.

Where the agency is buying production-line items, where (for example) spacecraft and experiment design were established before the start of the project, accurate estimates of project costs are possible. But cost estimating is and will remain very much an art, until completely standardized project hardware with experiments that can be "plugged in" has been developed. That time probably will never arrive for advanced technology development projects of the kind considered in this book.

There is considerable variation in project structure, most of it involving the degree to which the organization is "projectized" versus reliance on "functional organizations."* Suppose a project requiring 200

* An intermediate approach is known as matrix management. Here, employees are temporarily assigned to a project, while remaining on the rolls of the parent organization.

Table 4. Cost Growth in Selected R & D Projects, 1958 to 1966, in Millions of Dollars

[blocks in formation]

These are the dates at which the cost calculations start.

2 Cost ratio based on total cost.

Source: Arnold Levine, Managing NASA in the Apollo Era, Washington, DC: NASA Scientific & Technical Information Branch, 1982), p.155, after memorandum of DeMarquis Wyatt, Assistant Administrator for Program Planning and Analysis, April 10, 1969.

people at its peak. In a fully projectized operation, all 200 people would work directly in the project office and would “get their paychecks" from the project manager. For example, the office that managed the Pioneer 10 and 11 missions to Jupiter and Saturn was fully projectized (fig. 39).

There were three major groups - spacecraft, experiments, and operations each with its manager. Each person had specific assignments: to follow the spacecraft subsystem as it was developed by TRW, the prime contractor, to follow from one to three experiments, or to follow functions like launch, tracking, and data acquisition.

With the functional approach, a small project office of (say) ten people would be established. They would control the project, but the bulk of the work would be farmed out by task order to functional organizations such as the Mechanical Design Section, the Test Laboratory, the Electronic Design Section, and the like.

Each of these approaches has certain advantages and disadvantages. A projectized organization can maintain tighter control of the project, both technically and fiscally. Personnel can devote all of their attention to the project; the project office is responsible for a given assignment from beginning to end; and the organization is tailor-made to fit the job (ref. 110). The principal drawback is the inefficient use of manpower. In the functional approach, when someone is needed on the project only one day a week, he can do other work during the remainder of the week. In a project organization, he tends to sit on his hands the other four days. The problem with the functional approach is threefold: It conflicts with the desire of functional managers to build a technical expertise in their sections; responsibility for a given job is diffused; and only part-time attention can be given to any one project (ref. 111).

Obviously, no idealized description of a project approach can do full justice to the range of projects within even one agency. A given project will be affected by many variables other than those discussed. It may be affected by a midcourse change of goals by the sponsoring agency; by an increase in the length of the project approval process relative to the length of the project; and by the jeopardy to the careers of project managers in committing themselves to long-term projects running to five or more years. To examine how goals of a sponsoring organization tend to shape the kinds of projects sponsored, we have selected three projects for analysis. One was a large space-flight project sponsored by NASA; the second was a successful weapons development project; and the third involved work on several fronts by a leading industrial laboratory. Case Studies

The Orbiting Astronomical Observatories (ref. 112). One of the interesting byproducts of the Second World War was that it made space astronomy possible. With the delivery of captured V-2 rockets, American

« PreviousContinue »