Management Discovers The Human Side Of Automation
Business Week - September 29, 1986 - By J. Hoerr, M.A. Pollack and D.E. Whiteside

 

Companies Are Finding That Workers Are The Key To Making Technology Pay Off

Like thousands of companies in the early 1980s, Shenandoah Life Insurance Co. marched eagerly into the world of high technology. It installed a $2 million system to computerize processing and claims operations at its Roanoke (Va.) headquarters. But the results were disappointing. It still took 27 working days and handling by 32 clerks in three departments--to process a typical application for a policy conversion.

Shenandoah's problem stemmed from its bureaucratic maze, not from defects in the technology. Only by radically reorganizing its work system could it reap the benefits of automation. The company grouped the clerks in "semiautonomous," teams of five to seven members. Each team now performs all the functions that once were spread over three depart meets. Team members learned new skills, bringing them greater job satisfaction and better pay. As a result, the typical case-handling time dropped to two days, and service complaints were practically eliminated. By 1986, Shenandoah was processing 50% more applications and queries with 10% fewer employees than it did in 1980.

The productivity gains at Shenandoah Life are part of a powerful synergism taking root in the U.S. - the pairing of people with automation. American managers are finally learning what the Japanese discovered years ago: The solution to fading competitive ability, sluggish productivity growth, and poor quality cannot be found in the mythical black box of a miraculous technology. To realize the full poter~tial of automation, leading-edge companies are integrating workers and technology in "sociotechnical" systems that revolutionize the way work is organized and managed.

This is an immensely important trend, one that is producing a new model of job design and work relations that will shape the workplace well into the 21st century. Nevertheless, the changeover isn't occurring fast enough. The great wave of automation that has swept through offices and factories since 1980 is losing momentum, largely because not enough companies are adopting the innovative work practices that get the most out of au~mation. Many managers are reluctant to "run the kind of social revolution at work that is needed to make technology pay for itself," says productivity expert George H. Kuper, who heads the Manufacturing Studies Board, a research arm of the National Academy of Sciences.

 

People Problems

With or without work reforms, computer-based technology is having an enormous impact on workers. In one way or another, it has changed the jobs of 40 million to 50 million people, almost half of the U. S. workforce. It has made some jobs more challenging and "deskilled" others. It has caused severe dislocations at specific work sites by eliminating jobs, raising a fundamental question of whether government and business are investing enough money and expertise in retraining displaced workers. For the entire nonfarm economy, however, technological change helped produce a 10.4 million increase in jobs between 1979 and this year.

But contrary to the engineers' vision of factories run by robots, the high-tech workplace depends more than ever on people. "There will be fewer of them, but the ones who are there will be critical," says Gerald I. Susman, an expert on work and technology at Pennsylvania State University. Mistakes by poorly trained, poorly motivated workers can cause enormous damage, as demonstrated by the nuclear accidents at Three Mi]e Island and Chernobyl. Says Lyman D. Ketchum, a pioneering consultant on teamwork: "We're moving increasingly into dangerous, unforgiving technologies that can't be operated safely with uncommitted people."

Most important, it is becoming evident that advanced computer technology calls for a radical change in traditional work practices. The old "scientific managementt' method of dividing work into discrete tasks that require little skill or training becomes obsolete in a computerized workplace where many functionsincluding materials handling, assembly, inventory control, and testing are inte grated by computer. "The integration no longer makes it possible to define jobs indindually or measure individual performance," says Richard E. Walton of Harvard University. ``It requires a collection of people to manage a segment of technology and perform as a team.''

 

Global Competition

For these reasons, more companies are installing work systems that emphasize broader-based jobs, teamwork, participative managers, and multiskilled workers. The innovations include a range of other labor policies aimed at developing iicommitted" workers, including enhanced job security, continuous training programs, and compensation schemes that reward group performance (table). Industries such as autos, steel, and communications have been moving slowly in this direction with the cooperation of their unions since the beginning of the "quality-of-work-life" movement in the 1970s.

But the new innovations go far beyond QWL reforms that involve workers in problem-solving groups or otherwise aim at making jobs more satisfying. Now the movement is being fueled by global competition and the need for a high rate of product innovation. The average life cycle of an electronics product, for example, is only three to five years. Experts say that while the U. S. may not be able to compete with countries that turn out standardized products and parts at low wages, it can create new market niches for customized products. But manufacturers must be able to switch quickly from one product line to another, and flexible work systems when combined with computer-based technology give them that ability.

The new "paradigm?" as organizational behavior specialists call it, will gradually replace the old system characterized by authoritarian management and an extreme division of labor epitomized by the assembly line. The new approach often entails sociotechnical planning that is, integrating the psychological and social needs of workers with technological requirements in designing a new plant or redesigning an old one. Haward's Walton conists the old "control" paradigm with the new model of "commitment".

The payoff can be significant. Many plants that were designed with sociotechnical methods and use the most radical innovation, semiautonomous teams, are 30% to 50% more productive than their conventional counterparts. In most plants, these teams manage themselves without first-line supervisors, determine their own work pace within parameters set by management, schedule their own vacations, and have a voice in hiring and firing team members and deciding when they qualify for raises. This is a relatively new creature on the U.S. industrial scene, and both managers and workers give it high praise.

Ten years ago, fewer than two dozen manufacturing plants in the U. S. organized work on a team basis. Today teamwork is used in several hundred offices and factories' especially new, highly automated plants with small work forces of 25 to 500 people. One example is a diesel engine plant jointly owned by Cummins Engine Co. and J. I. Case Co. in Whittakers, N. C. Teamwork, says plant manager John C. Read, brings out "an entrepreneurial cowboy spirit', in American workers. `'When this spirit gets wrapped into team efforts to figure out why a machine went down and if management gets out of the way itts a tremendously powerful tool."

Many workers like teamwork for its greater variety of tasks, compared with repetitive jobs on a conventional assembly line. That's true of Randy Gilbert, a 10-year veteran at General Motors Corp. who now is an elected team coordinator in Buick City, GM's showcase plant that combines high technology and Japanese management methods. '`Once in a while I get bored and switch jobs with someone just to relieve the tedium,', he says. That wasn't possible before.

 

Not Fast Enough?

But technology experts say teamwork and other innovative practices are not spreading fast enough. Although the sociotechnical revolution is here to stay, plants that use teamwork still constitute only a small minority of U. S. workplaces.

If teamwork produces such good results, why haven't more companies tried it? For one thing, it requires a drastic change in management style and methods. The old idea that a manager's main function is to control workers is replaced with the concept that a manager should encourage employees to use initiative. This goes against the grain of everything managers have been taught since the early years of the century, says Lyman Ketchum, who helped design one of the first sociotechnical plants in the U.S.' ~ Gaines Foods Inc. plant that opened in Topeka, Kan., in 1971. To accept the commitment model of work, he says, managers have to go through a 'ipersonal paradigm shift, which is a deep psychological process."

John B. Myers, vice-president for human resources at Shenandoah Life, adds that most managers are comfortable with old- style bureaucracies in which orders are passed from top to bottom. "Bureaucratic organizations become habit-forming, just like cigarettes," he says. `'That's why they don't change."

The slowness to change may have implications for productiYity growth. The rise in output per man-hour in the U. S. has lagged behind that of Japan and European nations for more than a decade. Bureau of Labor Statistics economists estimate that productivity will increase at an average annual rate of 1.7% through the mid-1990s, about double the rate of the past 10 years, largely because of new technology.

This projection, however, is based on a continuing high rate of technological innovation that may not happen. The 1980s started with glowing predictions of pushbutton factories linked to executive suites in vast computer networks. And the "paperless," electronic office was said to be just around the corner.

But the unmanned factory has not arrived. Computer-integrated manufacturing (CTM), in which shop-floor machines are operated by a central computer, is a reality in only a few plants. And offices are still struggling with primitive computer networks.

 

Pulling Back

Investment in new technology is not increasing nearly as fast as was predicted in the early 1980s. In early September, Dataquest Inc., San Jose (Calif.) market research firm, lowered its projection of industrial automation sales in 1990 by 13%, to $34 billion. Robot sales are also slowing down. '`I'm very discouraged,'' says Richard M. Cyert, president of Carnegie-Mellon University. `'For the future of manufacturing in this country, we have to find a way to move automation ahead at a faster pace."

Many companies are pulling back from overambitious automation projects. In its blueprint to convert seven plants to produce a new midsize car? GM had intended to install more than 1~000 robots and replace conventional car assembly lines with the automatic guided vehicle (AGV), a moving `'island" of car parts. Now? however, only three plants will be converted. In addition to financial and marketing reasons, GM has had problems integrating technology with its management systems, observers say.

The complexities involved in making CIM work are stymieing automation efforts at many companies. Few have managed to tie the major management functions of engineering, production, and marketing into a single, computerized information system. This kind of linking is necessary for companies to gain one of the larger benefits of computer technology -- eliminating layers of middle managers and technicians who now do this work. In addition, computerizing new techniques such as just-in-time inventory control may have as much impact in cutting production costs as eliminating direct labor and managers, says Penn State's Susman. "But reducing inventory tightens the couplings between parts of an organization," he adds, "and this requires workers who know what they're doing."

GM's going ahead with other projects, including a highly automated front-axle plant in Saginaw, Mich. When this plant reaches full production in late 1987, it will be run entirely by robots part of each working day. This plant illustrates the new U. S. emphasis on human skills. It will be operated by 38 hourly employees, all members of the United Auto Workers, who survived a stringent selection process. They are now being schooled in electronic? mechanical, and problem-solving skills and will have more than a year of training before the plant begins making axles.

This amount of training is new in the U. S., although the Japanese have routinely engaged in such comprehensive programs for years. It is one of the ways Japan invests in human resources so that automation will `'make people more productive" says Thomas J. Gallogly, a metalworking expert in the Commerce Dept.,s International Trade Administration. The Japanese also enlist workers in "quality circles" to solve production and quality problems. Lifetime job security is emphasized in Japan as well, although for limited numbers of workers in the larger companies.

 

Wider Scope

When American managers began touring Japanese plants in the mid-1970s, the Japanese stressed the importance of these human factors' recalls Kuper of the Manufacturing Studies Board, who led some of those tours. ``Our managers kept looking for the technological solution to the growing Japanese success," he says. `'The Japanese were trying to be honest with us, but we were too stupid to listen."

Now, U. S. employers are belatedly turning to the human side of technology, partly by borrowing the Japanese techniques but also by using other methods. For example, Japanese companies do not emphasize a fundamental redesign of jobs to make them more appealing to workers. Furthermore, when the Japanese use production teams, they usually keep them under the control of first-line foremen.

Indeed, the semiautonomous team idea originated in experiments at British coal mines in the late 1940s. Behanoral scientists at London's Tavistock Institute of Human Relations, led by Enc Trist now professor emeritus at the Wharton School concluded that industry needed a new paradigm of work organization. By stressing autonomous work groups, jobs of wider scope, and worker ~nvolvement in decision-making, Trist and his colleagues said, companies could adjust much more easily to fast changing market and political conditions.

Trist and others developed the `'sociotechnical systems"' concept of work design. STS calls for involving workers whenever possible in planning a new or redesigned plant, as auto workers have been involved in GM'S Saturn project. Usually, the technical design came first, and work flow and the placement of work stations followed. In designing an auto plant, for example, engineers would specify a conventional assembly line that allows only one social system: Workers must stay at fixed stations along the line, performing the same task every 30 seconds or so.

"Traditionally, jobs were designed with no capacity for people to initiate anything," says Harvey F. Kolodoy, a professor at the University of Toronto. "If things went wrong, you'd get an inflexible response. We should design jobs so that workers can be more than a pair of hands behaving in a mechanical way."

To give workers a greater variety of duties, a sociotechnical auto plant design would call for teams to assemble entire subunits of a car from parts moved through the plant on AGVS. Team members would be free to move around, rotate jobs, pace themselves within a much longer work cycle of perhaps five minutes or more, and have more control over product quality. Studies show that group assembly not only makes workers feel better but also produces higher quality.

The STS concept moved from Britain to Norway and Sweden, where Volvo used it in designing its plant at Kalmar, Sweden, which opened in 1974. Kalmar's work force is divided into about 20 production teams; each assembles a major unit of a car in an average of 20 minutes to 40 minutes. Production costs at Kalmar are 25% lower than at Volvo's conventional plants, and the company is building a new plant at Uddevalla based on the Kalmar experience.

Teamwork also began to appear in the U. S. in the 1960s. But for years it was confined to a handful of pioneering companies, including Procter & Gamble, Cummins Engine, Gaines Foods, Sherwin-Williams, the Packard Electric Div. of GM, Hewlett-Packard, TRW, and Best Food, a unit of CPG International.

In the past few years, scores of companies that traditionally set the patterns - industrial relations have adopted the concept. Among them are General Electric, Ford, most GM divisions and Westinghouse, as well as Xerox, Honeywell, Digital Equipment, and other high-tech companies. Shell Canada Ltd. runs four chemical and refinery plants with sociotechnical principles. Even the financial services industry is picking up the concept. In addition to Shenandoah Life, Lincoln National Life and Amenican Transtech have reorganized their paper processing operations into teams, and one giant insurance company, Aetna, is on the verge of doing so.

For the most part, the teamwork movement has been a quiet revolution. Many of the leading companies have not trumpeted their findings, partly because they believed their innovations provided a competitive edge. Now some of the pioneers are opening up a bit, and their evidence of superior performance in teamwork is impressive.

Procter & Gamble Co., which estabished its first team-based plants in the 1960s and now has 18 such sites, has always refused to comment publicly on the matter. However, it confirms remarks made in late 1984 by Sernor VicePresident David Swanson in a closed meeting at Harvard. Swanson said P&G'S teamwork plants were "30% to 40% more producive than their traditional counterparts and significantly more able to adapt quickly to the changing needs of the business.''

 

Person To Person

Cummins Engine has three teamwork plants, including the North Carolina site, and has used elements of the team approach at its older plants in Columbus, Ind. Cummins also has been reticent about the new-style plants. But in a recent interview with BUSINESS WEEK, Vice-President Ted L. Marsten said that Cummins is convinced that "this is the most costeffective way to run plants. In traditional plants, work was broken down to the lowest common denominator, and there was not a lot of flexibility. We created teams to get the work to flow in the most produclive way The people felt a lot better about the work they did, and we got a much higher quality product."

In Oregon, Tektronix Inc. converted a few years ago from assembly-line manufacturing in its metals group to teams. Each "cell" of 6 to 12 workers turns out a product that can be manufactured in relatively few steps. One particular cell now turns out as many defect-free products in three days as an entire assembly line did in 14 days with twice as many people. Xerox Corp. began using teams in some of its operations a few years ago and has found them to be `'at least 30% more productive', than conventionally organized operations, says Dominick R. Argona, manager of employee involvement.

The Gaines Foods plant in Topeka, which received heavy publicity in the early 1970s for its new style of management, has proved that teamwork there was not a passing fad. Fifteen years after it started production, the plant still uses teamwork- both in the office and on the shop floor. Plant manager Herman R. Simon says Topeka produces the same pet foods as a sister plant in Kankakee, Ill., at 7% lower labor costs. Once a unit of General Foods Corp., Gaines is now a wholly owned subsidiary of Anderson, Clayton & Co.

Shenandoah Life's decision to embrace teamwork is a classic illustration of why technology can't solve all problems. Even after installing an automated system, the company found that processing clerks were still, in effect, "passing papers from person to person electronically,", says Myers. "It made no sense to have a new technology and yet operate the old social system."

Since experimenting with one clerical team in 1983, Shenandoah has formed nine teams of employees who before worked in separate departments. Former first-line supervisors belong to a team that "advises" the processing groups. Shenandoah has never laid off employees, but the team system has enabled it to reduce the work force by 14%, down to 229, over the past year and a half.

 

Turf Issues

The team approach also can help shape the kinds of goods and services a company produces. Shenandoah~s disability income team which includes an actuary, an underwriter? and a marketing specialist took only six months to develop and market a policy amendment designed to attract new business. If all the skills needed to design a new product are contained in one team, Myers says, product development doesn't get hung up on the `'turf issues'' that arise when several departments are involved in planning a new product.

Because of the difficulty in changing the culture and management style in existing plants, most teamwork plants are ``greenfield'' sites, and most are nonunion. Indeed, the sociotechnical trend can present a problem for unions Where employees have been allowed a strong voice in decision-making and largely manage themselves in teams, organizers have had a tough time presenting a case for unionization.

A number of unions, however, have worked jointly with management to convert existing plants to the team concept. These include the Auto Workers, Electronic Workers, Clothing & Textile Workers' and Steelworkers. One successful example involves the Aluminum Workers and the Rohm & Haas Co. plexiglass plant in Knoxville, Tenn. Within four years after the plant began changing to team organization, productivity measured as square feet of plexiglass produced per worker-hour had risen some 60%.

But resistance to the concept remains fairly strong in these and other unions because it requires changes in many traditional union-management relations. Instead of multiple job classifications, for example, a teamwork plant usually has only one or two. Production and maintenance work, traditionally separated under the scientific management organization of work, tend to merge into one fluid work system.

 

High Anxiety

For all its productiveness, teamwork is very difficult to implement and keep working successfully. Changes in plant and corporate management' from participative to old-style managers, have doomed many a promising teamwork experiment. Personality conflicts in teams also cause problems. Indeed, says Cummins' John Read, tension levels in sociotechnical plants tend to be higher than in conventional workplaces. `'It's wrong to think of teamwork plants as merely happy places," he says. "But the tensions tend to be constructive, and they produce high performance.''

High performance is what U. S. industry needs if it is to make the most of computer-based automation, say technology experts. While the U. S. is still behind Japan in matching workers and new manufacturing techniques, it is trying to catch up. "We need work environments that produce continuous innovation in a highly competitive global economy," says Eric Trist. The new model of work relations is not yet the dominant one, he adds, but `'I'd be sad if we weren't getting close to that point by the end of the century."

By John Hoerr in New York and Michael A. Pollock in Washington, with David E. Whiteside in Detroit and bureau reports