Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Managing Uncontrollable Growth John H. Bennett* The explosion in the acquisition and use of minicomputers ant! microcomputers is by now quite familiar. Over the past 10 years I have observed this growth from the point of view of a large, diver- sified, multinational corporation, United Technologies Corpora- tion (UTCl. United Technologies is a Fortune 100 corporation whose com- panies manufacture many kinds of high technology products. These include Pratt & Whitney Aircraft engines, Sikorsky heli- copters, Carrier air-conclitioning systems, Otis elevators, Mostek semiconductors, and a broac! range of electrical and electronic cle- vices and controls. All of these companies are experiencing the current microcomputer revolution. Since UTC allows its divisions a high degree of autonomy in management, each has reacted in its own way, both to the pressures of technological change and to the guidelines proviclec3 by the corporate office. Thus, UTC offers a unique opportunity to study a variety of companies, determine what the major common problems are, ant! see what managerial responses seem to be most successful. Most organizations using computer technology experience growth in the number of terminals used. While this growth rate is *John H. Bennett is corporate director of data processing for United Technologies Corporation, Hartford, Connecticut. 45
46 MANAGING MICROCOMPUTERS steady, it is by and large related to the rate of installation of sys- tems on a central computer. Thus, it is nominally under control. Workstation growth is a different story. Proceeding from a smaller base, word processors, personal com- puters, engineering design stations, and intelligent remote job- entry stations are growing at a much faster rate. Many of these applications require little or no immediate support from an orga- nization's main data center. The systems are inexpensive enough to get over or around the hurdles of cost justification that corpo- rate bureaucracies have placed in the path of computer acquisi- tion. Their growth is thus difficult to control. However, manage- ment of the process is obviously necessary for a couple of reasons. Many applications will eventually require data access ant! data in- tegrity; others clearly should be carried out on a central computer. The reasons for this rapid growth are threefold: service, eco- nomics, and control. Most computer service organizations or de- partments have a backlog of up to two years of cost-justified ap- plications. Customers see minicomputers and microcomputers as a shortcut around this bottleneck. Their cost is relatively easy to justify, and users have at least the appearance of being in control of their own destiny. The impact of these developments on computer service organi- zations has been devastating. The first symptoms of change took the systems departments by surprise. User departments pre- sented requests for special-purpose, minicomputer-based termi- nal systems for special applications. The first requests seemed innocuous enough: a computer for a laboratory; a dedicated test application; a stand-alone, computer-aided drafting application. Such projects were outside the normal range of competence or responsibility of most data processing departments, and they were clearly limiter! in scope. Or were they? At United Technologies over 500 of the Digital Equipment Cor- poration PDP Its have now been brought in for these "limited" applications. There are over 50 VAXs with their full network of supported terminals, virtually all of which are beyond the direct control of the systems department. The same drama is about to be played out in the area of manufacturing support. The first microcomputers were brought into my corporation by their owners, who were convinced of their utility but could not get approval to purchase them. In one division I found that an em- ployee had set up his own micro in a conference room and provided
MANAGING UNCONTROLLABLE GROWTH 47 a sign-up sheet so that other employees could schedule time to use it. Needless to say, I felt obligated to mention to management that an Apple computer is far too inexpensive to have expensive professional employees queueing up for it. Now, of course, the problem has changed to one of moderating the rate of acquisition. For the first time computer service organi- zations are being threatened by their own technologies. The warn- ings are clear: Improve service or lose control. A two-year backlog is no longer viable. The IBM solution is not the only acceptable one. And corporate ciata can no longer be accessible only to infor- mation specialists. The two-year backlog is, of course, the result of a rational at- tempt to establish priorities for a large workload. An unfortunate consequence is that small jobs, important only to one person or one department, never seem to get done. A number of options are available to computer service organizations (CSOs) to attack this backlog. System generators and application generators are one obvious choice. These tools can make a programmer up to five times more effective. Such tools also tend to be self-clocumenting, making later maintenance easier. Another tool-relatect problem has been the willingness of some departments to sacrifice service to the user in favor of operational convenience. Many CSO managers who have pointed to IBM or some other preferred venclor's solution to a problem as if it were the only solution are now waking up to find their customers in- stalling another venclor's equipment and running proprietary software. Ant} in the process their own responsibilities have been reduced to clownIoacling data files to the new system. More en- lightened departments have tried to provide modern tools for us- ers to get at their data. Some CSOs have even taken the offensive and initiated educa- tion programs to show users what systems can do for them, in- cluding what approaches are likely to be successful and what the limitations are. These programs may be adciressec! to a specific group, such as a manufacturing unit, or to management in gen- eral. Such efforts improve the ocicis that whatever approach a user decides on-mini, micro, or main computer will be reasonable ant} will have a high chance for success. Microcomputer growth has produced a number of problems for general management as well as for systems departments. The problem of computer security, both external ant! internal, is now a
48 MANAGING MICROCO~UTERS concern at the highest levels of corporate management. With more managers and executives using terminals and personal com- puters, the problems of data access and corresponding data secu- rity can no longer be avoidecI. UTC is only one of many companies considering new guidelines for computer security. Cost is another management problem. In more and more corpo- rate negotiations over system support, service wins out over com- patibility. Terminal-basec! systems are more expensive than their batch report counterparts, whether they are mini, micro, or main computer based. At UTC, for example, the cost of hardware is growing both in absolute dollars and as a percentage of the ciata processing budget, despite the advancing technology. United Technologies Corporation has used a number of ap- proaches to manage the latest stage of the computer revolution. The most obvious is the use of report writers and other high-level languages. These allow ciata processing professionals to respond to requests for service that require occasional reports drawn from existing data bases or the calculation of specific results on an in- frequent basis. Some of these tools are simple enough to be taught to end users as well. Along with the personal computer, these new tools are the ma- jor focus of the information centers that are currently springing up throughout United Technologies. The information center at the Essex Group Headquarters, for example, has achieved na- tional acclaim. It uses terminals, microcomputers, and user- friendly software to help its customers perform a variety of tasks that might otherwise have to be done by data processing. Another response to service problems is to tackle the backlog clirectly by increasing programmer productivity. There is in- creased interest today in the use of automates! programming aids such as program generators. In the area of program or system generators UTC has been somewhat conservative. The Raytheon product, Reacli-Cocle, is used at one company; the Burroughs product, ICING, is used at another; a third unit using a home- grown applications generator is probably the best of the three. I cannot overemphasize the power of a well-clesignec} application generator. One of UTC's smaller divisions installed a broad range of financial and administrative systems over a two-and-one-half- year period. During that time the systems development group never incluciec3 more than two-and-one-half employees. All appli- cations were fully customized, not packages. The latest was a
MANAGING UNCONTROLLABLE GROWTH 49 complex purchasing system (10,000 lines of COBOL code) begun in May and turned over to the users for testing in August. UTC is perhaps unique in the area of management education. In 1982 the company deciclec! to invest in a major effort to educate its executives in the capabilities of computers. The method cho- sen was to develop a training program for executives. The result- ing program, implemented in late 1982, is called the Executive Personal Computing Workshop. This workshop, a three-day ses- sion concluctec! at the Uniter! Technologies Research Center, is restricted! to senior managers ant! executives. It offers a hands-on approach using the IBM personal computer and the Context MBA software. Following the program, participants receive a personal computer for their "business use." Executives in this program are taught to use an array of stan- dard tools: automated spreadsheets, word processing, business graphics, and communications. No job-specific applications or programming techniques are taught. Although the three-clay time period does not permit inclusion of database applications, a self-study database moclule has been developed for later study. Considerable effort was spent developing this program, which in- cludes custom training materials rather than manuals. Indiviclual workstations are connected to the instructor's station through a switching network that allows participants to view the instruc- tor's screen on their own screens by flipping a switch. The pro- gram has been well accepted and is one of the most popular the corporation has ever run. It suggests the type of imagination that ciata processing organizations will need in the future if they are to regain control of their enterprise. The current upheaval has some predictable longer-term effects as well. First, it is safe to assume that by the end of the decade virtually every white-collar worker, ant! many others, will have or have access to some form of workstation. The intelligence built onto that station will vary with user need. Similarly, whether it is connected to a minicomputer or a mainframe will depend on the application. Second, this rapid growth in workstations will im- pose an enormous training requirement on organizations. Al- ready companies are springing up to meet this neecl, and vendors are unbundling their training to reap additional profit. Even com- puter service organizations will not be immune to this training requirement due to the acquisition of new productivity tools. What will happen to computer service departments in the fu
50 MANAGING MICROCOMPUTERS sure? Clearly there will be far less need for mundane programming skills given the use of productivity tools both by the systems de- partment and by customers. Although computers will be much more widely distributed than they are tociay, systems depart- ments will neither wither away nor will their leaclers become super information executives. On the contrary, these departments will return to their original function of providing professional services consisting of computer support, file management, systems analy- sis, and program generation. They will be smaller, highly profes- sional organizations. Finally, if our organizations have an apparent destiny, we must decicle how we can best lead them toward it. There are two sides to this question, which can be illustrated by two stories. The first story is about the traveling salesman who, after searching several hours for his destination, finally encounters a farmer. When asked how to reach the destination, the farmer sacITy replies, "You can't get there from here." Luckily, for most of us, the situation is not that baci. For some companies, however, that statement is all too true. Faced with fragmented systems responsibilities, internal bickering over resources, ant} no central control or management commitment, the only way they can clear with the coming changes is to restructure, to become different · . Organizations. The seconc! story comes from Lewis Carroll's classic, Alice's Adventures irz Wonderlar~cI. Alice, who is lost, asks one of the con- suitants of her day, the Cheshire Cat, for directions: "Would you tell me, please, which way I ought to go from here?" "That depends a good deal on where you want to get to," said the Cat. "I don't much care where-" said Alice. "Then it doesn't matter which way you go," said the Cat. "-so long as I get somewhere," added Alice, as an explanation. "Oh, you're sure to do that," said the Cat, "if you only walk long enough." This exchange reflects the more normal situation in many com- panies now trying to make decisions about which way to go with computer technology. The possible paths are management ap- proaches. The goals where we want to go-are strategic clirec- tions. We must choose: · Do we want to be on the leaching edge of the technology and possibly achieve a strategic advantage over our competitors? If so, the cost is high, ant} so is the risk of costly blunders. We have
MANAGING UNCONTROLLABLE GROWTH 51 to develop many of our own tools. Cost-benefit analysis as a man- agement toot is virtually useless. · Are we willing to stay a step behind the leacling edge? If so, we are less likely to achieve a competitive advantage. However, the cost is less and so is the risk. Most applications can be econom- ically justified. But some decisions will still have to be justified on strategic grounds. · Do we want to stay well back from the leading ecige, imple- menting only proven and therefore cost-effective systems changes? If so, we will have minimum cost and minimum risk of failure. We are, however, at maximum risk of finding ourselves at a competitive disacivantage. I cannot recommend which path to take. Each company must make its own choice, and it should choose a path consistent with its management approach in other areas. But most importantly, it should not pick a path by clefault and it should not let its techni- cians do the choosing.
