CHAPTER V The Structure of Technology Obstacles to Technical Innovation The laboratory is the linchpin of technology development. To understand where and how (or if) innovation occurs, one must begin here. With certain significant exceptions, the organizational structures tend to be quite similar for all research and development organizations, independent of the organization's functions. This is so because all of these institutions live with the built-in conflict between flexibility and the organization's formal mission. Professional people need the flexibility to start new projects or terminate existing ones; change the distribution of effort between in-house staff and contractors; transfer funds between tasks or projects without the need for prior approval by the laboratory director; or encourage people involved in fundamental research to communicate with those doing applied research, and even to transfer from one group to the other. Indeed, at the project level the most important function of the manager may be to motivate the project team, rather than to make decisions which are both unilateral and final (ref. 74). Yet there is a limit easier to sense than to define precisely to the flexibility a laboratory director will allow his professional staff. The laboratory is constrained (or driven) by its mission, by its budget, by the particular skills it needs, and, paradoxically, by that need to innovate which tends to destroy its stability. In a laboratory without a strong sense of mission, flexibility may degenerate into a situation where professionals all "do their own thing." But in a laboratory where every research task is yoked to an overriding agenda, there may be no room for that relatively modest amount of basic research needed to keep the organization abreast of the state of the art, to prepare for new goals and missions. Thus the burdens of research management are imposed by the nature of the organization. There is, first, the problem that once any organization attains a certain size, coordinating the work of the various divisions. consumes much of senior management's time. Different departments are sealed off from each other; the paperwork needed to process (say) a procurement action increases; and routine tends to drive out innovation. But our second point innovation may be regarded (and quite accurately) as a threat to the status quo.* Where innovation is perceived as a threat, management can deal with it in a number of ways: allowing the effort to continue, but isolating it from the rest of the organization; reducing the level of effort, so that innovation never attains critical mass; compartmentalizing innovation, so that it occurs in one part of the organization, but not in the others; laying down development criteria so stringent that no research effort can ever satisfy them; or converting radical innovation into routine, incremental improvements in existing systems (ref. 75). In large organizations, whether corporations or government agencies, a major program, one with long lead times and a large budget, acquires enormous momentum. While corporations are more likely than a government agency to cancel an unsuccessful project, the contrast is usually overdrawn. True, even on the most optimistic projections, the hydrogen fusion research sponsored by the Department of Energy will not lead to commercial production until well into the next century. On a smaller time scale, it took the Boeing B-47 bomber 7.8 years and the Boeing B-52 bomber 9.4 years to attain operational capability in the late 1940s and the early 1950s. Other Federal development projects not only met schedules, but surpassed them. The nuclear submarine is only the best-known example; it took the Thor and Atlas ballistic missiles only 3.5 and 5.2 years, respectively, to go from program approval to first operational squadron, instead of the 6.8 years first projected (ref. 76). There are also commercial projects which require extremely long lead times. Consider one current example: Exxon's research and pilot-testing of a surfactant (detergent) method to coax more oil out of the ground. Research on an enhanced surfactant process began in the mid-1960s and led to a pilot project in 1969 to 1971. In turn, the results of the project led to research to develop a process to reduce the salinity content in the test reservoirs. A pilot test of the process began in 1980 and was considered a technical success in 1981. It will take another ten years before field testing for commercial viability will make a decision to proceed with full-scale development possible (ref. 77). * In this connection, Jacques Gansler's observations about the defense industry are worth citing. "The relative inelasticity of . . . demand is a particularly interesting factor in defense R&D; it implies that if you come up with a new idea for, say, a better airplane, you will simply be replacing the old design (which may well have been your own), with very little likelihood of being able to create increased demand, since the number of airplanes to be procured is a function of the force structure, not of cost or performance. Thus, technological advances that originate in the civilian sector are likely to be immediately applied in that sector, with very little thought given by the firm to military application. Only much later is it likely that a defense-oriented firm might pick up the idea and perhaps begin to apply it." Jacques Gansler, The Defense Industry (Cambridge, Mass.: MIT Press, 1980), pp. 304-305. Whether government or commercial, projects such as these tend to force innovation into narrow channels. In a large research organization some, and in a smaller one, most professional staff, may be working on small, carefully-defined areas into which the research effort is parceled. Of course, as work continues, the researchers will be further down the learning curve. But the danger in these larger projects is that once an all-out commitment is made, it becomes difficult, if not impossible, to admit that the organization is on the wrong track. As Donald Schon has noted, "Large-scale developments of the kind undertaken by supercorporations or the military may proceed for months or years beyond the point where they should have stopped; they continue because of massive commitments to errors too frightening to reveal . . . In these cases, the personal commitment of the people involved in the development, the apparent logic of investment, and the fear of admitting failure, all combine to keep the project in motion until it fails of its own weight." (ref. 78.) It is not that laboratory directors and the officials to whom they report are unaware of these problems. Some Federal agencies have tried to control the development process, whether by stimulating competition between laboratories or by introducing management-decision points at important stages of a project. In the 1960s, NASA, for example, instituted "phased project planning," whereby management could intervene at four stages in the life of a project. The project, so the theory went, would begin with a study of alternative approaches (Preliminary Analysis), followed by the selection of one of them (Definition). This would lead to a Design stage, culminating in the development of a mock-up or "breadboard" of project hardware. In the final stage (Development/Operations), the contractor, in cooperation with NASA scientists and engineers, would prepare the final hardware design, leading to development, fabrication, testing, and operation. While phased project planning was a successful management method for controlling a well-defined project, it did not help at all in determining whether or not a project should be terminated. Projects have occasionally been terminated in NASA but these terminations have little relation to the "phase" in which the project was in at the time of cancellation. In any case, for so complex a project as the Space Shuttle, it is often impossible to state accurately at what “phase" the Shuttle — as opposed to its subsystems — has arrived. In sum, the pressures on research laboratories are generated externally and from within. The external dangers stem from having to hew strictly to one narrowly-defined mission. The pressures generated within the laboratory are toward routine and conservatism. Where a laboratory is one of several within an agency, it is imperative that one of them, or divisions within all of them, do some basic research beyond their current mission. Technical innovation is a somewhat mysterious process, one easily swamped by the conservatism of a large organization. In The Sources of Invention, perhaps the most thorough investigation of the subject, the authors conclude that "the forces which make for innovation are so numerous that they are not fully understood." (ref. 79.) It is disconcerting, but true, that important discoveries are as likely to be made for aesthetic reasons "craftsmanship for craftsmanship's sake" — as for economic or military reasons. This was the case in aeronautics, in the development of wing flaps, streamlining, and stressed skin metal construction (ref. 80). It is even more disconcerting that some important inventions taken up by government agencies originated outside government laboratories or the industries from which the invention might have been expected to come. The development of the jet engine is a classic example: The pioneering work was done in two countries (Britain and Germany) by men who were either unconnected with the aircraft industry or were specialists in airframe design; no significant development originated with the aircraft engine manufacturers; despite the engine's military value, governments were reluctant to support it; and not until the engine manufacturers awoke to the significance of the jet engine was its development assured (ref. 81). Innovation in the laboratory is never a foregone conclusion. Yet some laboratories have been remarkably productive, and it is important to understand why. We shall briefly consider the formal organization of technology development centers, although the only justification for any organizational scheme may be fairly arbitrary and tailored to the particular individuals who work in the organization. We shall then examine three cases of technology development in search of clues perhaps the nature of certain test facilities, the freedom of researchers to bypass the organizational structure, the coupling of the researcher with the ultimate user as to what generates new ideas, new hardware, new technologies. Finally, we shall consider at length how one agency, NASA, manages its research and technology program and, in particular, how it reconciles centralized control with discretionary research at the field centers. Organizational Structure of Research and Development Institutions Most, if not all, technology development centers are organized at four levels: the branch or group level, the division, the directorate or department, and the office of the director and his staff (fig. 34). A group working in basic research or the early stages of technology development is generally quite small - somewhere between five and twenty people, or ten to forty if support personnel are included. What distinguishes a branch or group or section is that it generally has one and only one research objective. Director Deputy Director Directorate or Division Level Branch or Group Level FIGURE 34.-Typical schematic organization of a large research institution. The actual technical work is done at the branch or group level. The division tends to be the most important organizational element because it is still close to the technical work yet has enough resources to manage that it can claim management attention. The next organizational level is the research or technical division, consisting of two to five research groups working on related topics. Thus division-level organizations tend to have 50 to 200 people and form the first organizational unit whose leader has formal financial relations with the headquarters financial office. It is normally the lowest administrative level at which business with the sponsoring agency is transacted. A division-level organization may also be responsible for one or more of the laboratory's research and test facilities: a large particle accelerator at one of the former Atomic Energy Commission laboratories, a major wind tunnel at one of the NASA research centers, or a contractor-operated computer facility supporting all of a laboratory's operations. At most research installations with 1 000 or more people on the staff, the next level is the directorate or department. A department will normally have 200 to 500 people and will be organized according to function, to discipline, or — where the laboratory is responsible for a large project to project. This, for instance, was the case with the Viking project, which led to the successful landing of two probes on Mars in 1976. Viking was run out of the NASA-Langley Research Center, and the project manager reported directly to the center director. A directorate or department may also be organized according to discipline, as in the case of the Chemistry Department at the Lawrence Livermore National Laboratory. This department coordinates the laboratory's work in various areas of chemistry as well as providing the support for chemical diagnostics of nuclear explosions. The individuals heading departments |