Time-and-Motion Regained - Sustained Excellence through People

Time-and-MotionRegained

by Paul S. Adler

Reprint 93101

Harvard Business Review

HarvardBusinessReviewJANUARY-FEBRUARY 1993

Reprint NumberWALTER J. SALMON CRISIS PREVENTION: 93106

HOW TO GEAR UP YOUR BOARD

EXECUTIVE ACTION 93106THE FIGHT FOR GOOD GOVERNANCE

MICHAEL TREACY AND CUSTOMER INTIMACY AND OTHER VALUE DISCIPLINES 93107FRED WIERSEMA

PAUL S. ADLER TIME-AND-MOTION REGAINED 93101

RICHARD RAPAPORT TO BUILD A WINNING TEAM: 93108AN INTERVIEW WITH HEAD COACH BILL WALSH

RICHARD L. HUBER HOW CONTINENTAL BANK OUTSOURCED 93102ITS “CROWN JEWELS”

HBR CASE STUDYMARTHA NICHOLS THIRD-WORLD FAMILIES AT WORK: 93105

CHILD LABOR OR CHILD CARE?

IN QUESTIONALAN M. WEBBER WHAT’S SO NEW ABOUT THE NEW ECONOMY? 93109

FOUR CORNERSPAUL LAWRENCE AND JOINT VENTURES IN RUSSIA: 93104CHARALAMBOS VLACHOUTSICOS PUT THE LOCALS IN CHARGE

FIRST PERSONPETER T. JOHNSON HOW I TURNED A CRITICAL PUBLIC INTO 93103

USEFUL CONSULTANTS

DRAWINGS BY CHRISTOPHER BING Copyright © 1992 by the President and Fellows of Harvard College. All rights reserved.

Standardization is the death of creativity.Time-and-motion regimentation prevents con-

tinuous improvement. Hierarchy suffocates learning.U.S. manufacturing is in the throes of revolution,

and assumptions like these are becoming the newconventional wisdom about work. This new gospelsets up Frederick Winslow Taylor and his time-and-motion studies as the villain. It asserts that quality,productivity, and learning depend on managemen-t’s ability to free workers from the coercive con-straints of bureaucracy. It insists that detailed stan-dards, implemented with great discipline in a hierarchical organization, will inevitably alienateemployees, poison labor relations, stifle initiativeand innovation, and hobble an organization’s capac-ity to change and to learn.

But what if, as I believe, this new creed is wrong?What if bureaucracy can actually be designed to encourage innovation and commitment? What ifstandardization, properly understood and practiced,should prove itself a wellspring of continuouslearning and motivation?

In Fremont, California, a GM-Toyota joint ven-ture called New United Motor Manufacturing Inc.,NUMMI, for short, has succeeded in employing an

innovative form of Taylor’s time-and-motion regi-mentation on the factory floor not only to createworld-class productivity and quality but also to increase worker motivation and satisfaction.What’s more, NUMMI’s intensely Taylorist proce-dures appear to encourage rather than discourageorganizational learning and, therefore, continuousimprovement.

This outcome seems surprising because fordecades our attitudes toward work have beenshaped by a chain of reasoning that has led us to ex-pect (and guaranteed that we would get) a viciouscircle of escalating managerial coercion and em-ployee recalcitrance. The reasoning runs some-thing like this:M When tasks are routine and repetitive, efficiencyand quality require standardized work procedures.M High levels of standardization rob jobs of their in-trinsic interest, reducing motivation and creativity.

With workers defining their own job standards, quality and productivity at the Fremont plant went from worst to best.

Time-and-MotionRegained

Paul S. Adler is associate professor at the University ofSouthern California School of Business Administration.He has recently edited two collections of essays: Tech-nology and the Future of Work and Usability: TurningTechnologies into Tools. This article is based on a two-year study of the New United Motor Manufacturing Inc.plant in Fremont, California.

by Paul S. Adler

TIME-AND-MOTION

98 HARVARD BUSINESS REVIEW January-February 1993

M Demotivating work leads to dysfunctional em-ployee behavior such as absenteeism, highturnover, poor attention to quality, strikes, evensabotage.M Counterproductive behavior by the work forcerequires more authoritarian management, more hierarchical layers, and even higher levels of standardization.

In short, Taylorism leads inevitably to work-force discontent and union belligerence, which inturn lead inevitably to higher levels of bureaucraticexcess. The organization of work comes to build onthe dehumanizing logic of coercion and reluctantcompliance. Meanwhile, quality, profits, and jobsatisfaction all suffer.

NUMMI’s experience flies directly in the face of this thinking. That’s because the second step inthis chain of reasoning is false. Formal work stan-dards developed by industrial engineers and im-posed on workers are alienating. But proceduresthat are designed by the workers themselves in a continuous, successful effort to improve produc-tivity, quality, skills, and understanding can hu-manize even the most disciplined forms of bureau-cracy. Moreover, NUMMI shows that hierarchy canprovide support and expertise instead of a merecommand structure.

What the NUMMI experiment shows is that hi-erarchy and standardization, with all their knownadvantages for efficiency, need not build on the log-ic of coercion. They can build instead on the logic of learning, a logic that motivates workers and taps their potential contribution to continuous improvement.

In practice, NUMMI’s “learning bureaucracy”achieves three ends. First, it serves management byimproving overall quality and productivity. Second,it serves workers by involving them in the designand control of their own work, increasing their mo-tivation and job satisfaction, and altering the bal-ance of power between labor and management.Third, it serves the interests of the entire organiza-tion – management and the work force – by creatinga formal system to encourage learning, to captureand communicate innovation, and to institutional-ize continuous improvement.

The Worst Plant in the World

NUMMI is housed in what was once the GeneralMotors assembly plant in Fremont, California, 35miles southeast of San Francisco, which opened in1963 and manufactured GM trucks and the Chevy

Malibu and Century. At the old GM-Fremont plant,work was organized along traditional Tayloristlines, with more than 80 industrial engineers estab-lishing assembly-line norms that managementthen did its best to impose on the work force, withthe predictable results.

Over the years, GM-Fremont came to be whatone manager called “the worst plant in the world.”

GM-Fremont had lowproductivity, abysmal quality,drug and alcohol abuse, and absenteeism over 20%.

Productivity was among the lowest of any GMplant, quality was abysmal, and drug and alcoholabuse were rampant both on and off the job. Absen-teeism was so high that the plant employed 20%more workers than it needed just to ensure an adequate labor force on any given day. The UnitedAuto Workers local earned a national reputation formilitancy; from 1963 to 1982, wildcat strikes andsickouts closed the plant four times. The backlog ofunresolved grievances often exceeded 5,000.

GM-Fremont reached its peak employment of6,800 hourly workers in 1978. Numbers then de-clined steadily to a little over 3,000 when GM final-ly closed the plant in February 1982.

Discussions between GM and Toyota about apossible joint venture began that same year. InFebruary 1983, the two companies reached anagreement in principle to produce a version of theToyota Corolla, renamed the Nova, at the Fremontplant, using Toyota’s production system. GMwould be responsible for marketing and sales; Toy-ota would take on product design, engineering, anddaily operations. The new entity, NUMMI, wouldmanufacture and assemble the car. Beginning in1986, the plant also made Corolla FXs. In 1988,both the Nova and the FX were phased out, and Fre-mont began building Corollas, Geo Prizms, and, asof late 1991, Toyota trucks.

The two companies’ objectives were comple-mentary. GM wanted to learn about Toyota’s pro-duction system. It also obtained a high-quality sub-compact for its Chevrolet division at a time whenGM’s market share was rapidly eroding. Toyotawanted to help defuse the trade issue by buildingcars in the United States. To do this, it needed tolearn about U.S. suppliers.

Toyota later claimed it had also wanted “to gainexperience with American union labor,” but at firstToyota wanted nothing to do with the UAW. As it

HARVARD BUSINESS REVIEW January-February 1993 99

happened, there was no alternative. GM offeredthem no other facility, and the UAW had de factocontrol of Fremont. Moreover, GM was afraid of a union backlash at other plants if it tried to set up the joint venture as a nonunion shop.

In September 1983, NUMMI and the unionsigned a letter of intent recognizing the UAW assole bargaining agent for the NUMMI labor force,specifying prevailing auto-industry wages and ben-efits, and stipulating that a majority of the workforce would be hired from among the workers laidoff from GM-Fremont. In return, the UAW agreedto support the implementation of a new productionsystem and to negotiate a new contract.

NUMMI was formally organized in February1984. Toyota contributed $100 million in capital,and GM supplied the Fremont plant. Hiring beganin May. Every applicant went through three days of production simulations, written examinations,discussions, and interviews. Managers and unionofficials jointly evaluated applicants for the hourlyjobs: team leader and team member. The union alsoplayed a role in selecting managers, except for the16 who came directly from GM and a group of about30 Toyota managers and production coordinatorswho came from Japan. The CEO, Tatsuo Toyoda,brought with him the prestige of the company’sfounding family.

Over the following 20 months, NUMMI hired2,200 hourly workers – 85% from the old GM-Fre-mont plant, among them the old union hierarchy.(Almost none of GM-Fremont’s salaried employees

The NUMMI productionsystem not only madepeople work harder, it madethem work smarter as well.

was rehired. In any case, many had long sincemoved to other GM plants.) Since GM-Fremont haddone little hiring for several years before it closed,the average age of the new work force was 41. Mosthad high school educations. About 26% were His-panic, 20% black, and 15% female.

The first group of 450 team leaders and the en-tire NUMMI management team attended a three-week training program at the Toyota plant in Ja-pan – Takaoka – on which NUMMI was modeled.These people then helped to set up the new plantand train workers.

The NUMMI production system required peopleto work harder than they had at GM-Fremont. Jobsat the old plant occupied an experienced worker

about 45 seconds out of 60. NUMMI’s norm is clos-er to 57 seconds out of 60. And because workershave to meet much higher quality and efficiencystandards, they have to work not only harder butsmarter as well.

By the end of 1986, NUMMI’s productivity washigher than that of any other GM facility and morethan twice that of its predecessor, GM-Fremont. Infact, NUMMI’s productivity was nearly as high asTakaoka’s, even though its workers were, on aver-age, ten years older and much less experienced withthe Toyota production system. Quality, as rated byinternal GM audits, customer surveys, and Con-sumer Reports was much higher than at any otherGM plant and, again, almost as high as Takaoka’s.

Equally important, absenteeism has droppedfrom between 20% and 25% at the old GM-Fre-mont plant to a steady 3% to 4% at NUMMI; sub-stance abuse is a minimal problem; and participa-tion in the suggestion program has risen steadilyfrom 26% in 1986 to 92% in 1991. When GM-Fre-mont closed its doors, it had more than 2,000grievances outstanding. As of the end of 1991, some700 grievances had been filed at NUMMI altogetherover the course of eight years. The overall propor-tion of employees describing themselves as “satis-fied” or “very satisfied” has risen progressively tomore than 90%.

In 1990, Toyota announced that it would invest$350 million in an additional assembly line to builda Toyota truck for the U.S. market. So NUMMIhired 650 hourly workers on top of the 3,100 – plus400 salaried personnel – already employed. The firsttrucks rolled off the line in August 1991.

Fear, Selection, Socialization

NUMMI’s remarkable turnaround poses an obvi-ous question: How is it possible to convert a plantfrom worst to best quality and from dismal to su-perlative productivity over the course of a fewmonths? The most obvious answers are not entire-ly satisfying.

For example, fear. The GM-Fremont plant closedin 1982, and the people rehired by NUMMI didn’tgo back to work until 1984. Two years of unemploy-ment can produce a great deal of cooperation. Infact, some NUMMI workers believe managementmakes deliberate use of the specter of another plantclosure as a veiled threat to keep people in line. Butthe chairman of the union bargaining committeepoints out that while the old plant’s closure obvi-ously made workers more receptive to NUMMI’s

TIME-AND-MOTION


new approach, a return to old coercive managementmethods would have produced a rapid return to oldantagonistic work-force behavior patterns.

A second possibility is that management weededout troublemakers in the rehiring process. But infact NUMMI rehired the entire union hierarchyand many well-known militants. In general, veryfew applicants were screened out. The union even

won a second chance for some who failed drug teststhe first time around.

A third answer is that NUMMI made use of acomprehensive socialization process during hiringto instill a new set of values in the new work force.Certainly, NUMMI did its best to shape and alterthe attitudes of both workers and managers. For ex-ample, the company tried to undercut the custom-

Voices from the Factory Floor: Excerpts from Interviews with Team Leader

I’ll never forget when I was first hired by GM manyyears ago. The personnel manager who hired us gotthe...workers who were starting that day into a roomand explained: “You new employees have been hiredin the same way we requisition sandpaper. We’ll putyou back on the street whenever you aren’t needed anymore.” How in the hell can you expect to foster a loyaland productive work force when you start out hearingstuff like that? At NUMMI, the message when wecame aboard was “Welcome to the family.”

Team Leader Once you start working as a real team, you’re not

just work acquaintances anymore. When you reallyhave confidence in your co-workers, you trust them,you’re proud of what you can do together, then you be-come loyal to them. That’s what keeps the absen-teeism rate so low here. When I wake up in the morn-ing, I know there’s no one out there to replace me ifI’m feeling sick or hung over or whatever.…At NUM-MI, I know my team needs me.

Team Leader The average worker is definitely busier at NUMMI

than he was at Fremont. That’s the point of the NUMMI production system and the way it ties togeth-er standardized work, no inventories, and no qualitydefects. The work teams at NUMMI aren’t like the au-tonomous teams you read about in other plants. Herewe’re not autonomous, because we’re all tied togetherreally tightly. But it’s not like we’re just gettingsqueezed to work harder, because it’s the workers whoare making the whole thing work – we’re the ones thatmake the standardized work and the kaizen sugges-tions. We run the plant – and if it’s not running right,we stop it. At GM-Fremont, we ran only our own littlejobs. We’d work really fast to build up a stock cushionso we could take a break for a few minutes to smoke acigarette or chat with a buddy. That kind of “hurry upand wait” game made work really tiring. There wasmaterial and finished parts all over the place, and halfof it was defective anyway. Being consistently busywithout being hassled and without being overworkedtakes a lot of the pain out of the job. You work harder

at NUMMI, but I swear it, you go home at the end ofthe day feeling less tired – and feeling a hell of a lot bet-ter about yourself!

Team MemberIn our standardized work training, our teachers told

us we should approach our fellow team members andsuggest ways to improve their jobs. Hell, do you seeme trying that with a team member who’s six-foot-four and weighs 250 pounds? You’d be picking me upoff the floor if I tried that.…Standardized work is a joke as far as I can see. We’re supposed to go to man-agement and tell them when we have extra seconds tospare. Why would I do that when all that will happenis that they’ll take my spare seconds away and workme even harder than before? I’d rather just do the jobthe way I’m already comfortable with. I’m no fool.

Department ManagerOur assumption at NUMMI is that people come to

work to do a fair day’s work. There are exceptions, andyou would be foolish to ignore them. But 90% of peo-ple, if you give them a chance to work smarter and im-prove their jobs, and if they find that by doing thatthey have created free time for themselves, will spon-taneously look for new things to do. I’ve got hundredsof examples. I don’t think that people work harder atNUMMI than in other plants. Not physically anyway.But the mental challenge is much greater.

Team Leader I don’t think industrial engineers are dumb. They’re

just ignorant. Anyone can watch someone else doing ajob and come up with improvement suggestions thatsound good....And it’s even easier to come up with theideal procedure if you don’t even bother to watch theworker at work, but just do it from your office, on pa-per. Almost anything can look good that way. Evenwhen we do our own analysis in our teams, some ofthe silliest ideas can slip through before we actuallytry them out.

There’s a lot of things that enter into a good job de-sign....The person actually doing the job is the onlyone who can see all factors. And in the United States,engineers have never had to work on the floor – not


ary we-they divisions between workers and man-agement by eliminating special parking and eatingfacilities for managers and by introducing an identi-cal dress code – uniforms – for everyone. Manage-ment also devoted a great deal of attention to eachindividual hire and welcomed each personally tothe company that was going to build “the finestvehicles in America.”

However much these three factors – fear of un-employment, selection, and socialization – mayhave contributed to the final outcome, they do notadequately explain NUMMI’s continuing successor its ability to let workers draw improved motiva-tion and greater satisfaction from a system thatplaces them in a more regimented and bureaucraticenvironment and makes them work harder and

Managers, Workers, and Union Officialslike in Japan. So they don’t know what they don’tknow.…Today we drive the process, and if we needhelp, the engineer is there the next day to work on itwith us.

UAW Official One thing I really like about the Toyota style is that

they’ll put in a machine to save you from bendingdown. The Toyota philosophy is that the workershould use the machine and not vice versa.…It wouldbe fine if the robots worked perfectly – and the engi-neers always seem to imagine that they will. But theydon’t, so the worker ends up being used by the ma-chine. At NUMMI, we just put in a robot for installingthe spare tire – that really helps the worker, because itwas always a hell of a tiring job. It took awhile, and wehad to raise it in the safety meetings and argue aboutit. And they came through. That would never happenat GM-Fremont – you never saw automation simply tohelp the worker.

UAW Official In the future we’re going to need union leaders with

more technical and management knowledge. We’remuch more involved now in deciding how the plantoperates. That stretches our capabilities. Managementis coming to us asking for our input....The old ap-proach was much simpler – “You make the damneddecision, and I’ll grieve it if I want.” Now we need tounderstand how the production system works, to takethe time to analyze things, to formulate much moredetailed proposals. This system really allows us totake as much power as we know what to do with.

UAW Official Now when I try to explain [NUMMI] to old UAW

buddies from other plants…they figure that I’m forcedto say all this stuff because they shut our plant downand I had no choice. They figure going along with theteam concept and all the rest was just the price we hadto pay to get our jobs back. I explain to them that theplant is cleaner, it’s safer, we’ve got more say on im-portant issues, and we have a real opportunity to buildour strength as a union. I explain to them that ourmembers can broaden their understanding of the man-

ufacturing system and build their self-esteem, andthat the training we’ve gotten in manufacturing, prob-lem solving, quality, and so on can help them reachtheir full potential and get more out of their lives. I ex-plain to them that in a system like this, workers havegot a chance to make a real contribution to society –we don’t have to let managers do all the thinking. Butthese guys just don’t see it. Maybe it’s because theyhaven’t personally experienced the way NUMMIworks. Whatever the reason, they just see it all asweakening the union. Someone like Irving Bluestoneprobably understands what we’re doing. He had theidea a long time ago: if the worker has the right to votefor the president of the United States, he ought to havethe right to participate in decisions on the shop floor.

Team Member In the old days, we had to worry about management

playing its games, and the union was there to defendus. But now, with the union taking on its new role, it’snot as simple as before, and we have to worry aboutboth the management games and the union games. Idon’t want the type of union muscle we used to have.You could get away with almost anything in the oldplant, because the union would get you off the hook. Itwas really crazy. But it wasn’t productive.

Team Leader There are people here who will tell you they hate

this place. All I say is: actions speak louder thanwords. If people were disgruntled, there’s no way thatwe’d be building the highest quality vehicle. Youwouldn’t have a plant that’s this clean. You would stillhave the drug problems we had before. You would stillhave all the yelling and screaming. You can’t force allthat. And try this: go into any of the bathrooms, andyou’ll see there’s no graffiti. If people have a problemwith their manager, they don’t have to tell him on thebathroom wall. They can tell him to his face. And theboss’s first words will be: “Why?” Something’s hap-pened here at NUMMI. When I was at GM, I remem-ber a few years ago I got an award from my foreman forcoming to work for a full 40 hours in one week. A cer-tificate! At NUMMI, I’ve had perfect attendance fortwo years.

TIME-AND-MOTION


faster. The most critical piece of that explanationlies in the production system itself and in the poli-cies and practices that buttress it.

The NUMMI Production System

The idea of a production system is itself some-thing of a novelty in many U.S. manufacturingplants. All factories have production techniques,procedures, and policies, but these usually com-prise not so much a system as an ad hoc accumu-lation of responses to changing and often contra-dictory business and design demands. NUMMI’sproduction system is a finely tuned, superbly inte-grated whole, refined by Toyota over decades ofmanufacturing experience.

The basic techniques are familiar at least inname. The assembly line is a just-in-time operationthat does away with work-in-progress and makesquality assurance the responsibility of each workstation. The application of kaizen, or continuousimprovement, includes an extraordinarily activesuggestion program, constant refinement of proce-dures, and the designation of special kaizen teamsto study individual suggestions or carry out specificimprovement projects. Every machine and processis designed to detect malfunctions, missing parts,and improper assemblies automatically. Every jobis carefully analyzed to achieve maximum efficien-cy and quality. Job rotation is standard; workers arecross-trained in all team assignments and then al-lowed to shift from one task to another. Plannedproduction leveling eliminates variation in dailyand weekly schedules.

This system is essentially the same one Toyotauses in Japan, the same one many American manu-facturers are now beginning to adopt. But NUM-MI’s approach is distinctive in two respects: first,

The NUMMI approach hastwo distinctive features: a commitment to the socialcontext of work and a focus on standardization.

its strong commitment to the social context inwhich work is performed, and, second, its intensefocus on standardized work.

In terms of social context, NUMMI seeks tobuild an atmosphere of trust and common purpose.

NUMMI maintains exceptional consistency in itsstrategies and principles, it carefully builds consen-sus around important decisions, and it has pro-grams ensuring adequate communication of resultsand other essential information.

The basic structural unit is the production team,of which NUMMI has approximately 350, eachconsisting of five to seven people and a leader. Theidea is that small teams encourage participative de-cision making and team bonding. Four teams com-prise a group, led by a group leader who representsthe first layer of management.

Above and beyond the production teams, the big-ger team is everyone – all the workers, team lead-ers, managers, engineers, and staff in the plant aswell as NUMMI’s suppliers. Toyota leadershipwants workers to understand that the company isnot the property of management but of everyone to-gether. In NUMMI’s view, the primary purpose andresponsibility of the management hierarchy is tosupport the production teams with problem-solv-ing expertise.

The most substantive expression of this big-teamstrategy is the no-layoff policy spelled out in NUM-MI’s collective-bargaining agreement with theunion. Recognizing that “job security is essential toan employee’s well being,” NUMMI agrees “that it will not lay off employees unless compelled to do so by severe economic conditions that threatenthe long-term viability of the Company.” NUMMIagrees to take such drastic measures as reducingmanagement salaries and assigning previously sub-contracted work to bargaining unit employees be-fore resorting to layoffs.

Management sees the no-layoff policy as a criti-cal support for its overall production strategy notonly because it reinforces the team culture, but alsobecause it eliminates workers’ fear that they arejeopardizing jobs every time they come up with anidea to improve efficiency.

Workers came to trust this no-layoff commit-ment when in 1988 poor sales of the Nova broughtcapacity utilization down to around 60%. Workersno longer needed on the assembly line were not laidoff but instead assigned to kaizen teams and sent totraining classes.

Another important support for NUMMI’s teamconcept is its radically simplified job classifica-tion system. Where GM-Fremont had 18 skilledtrades classifications, NUMMI has two. WhereGM-Fremont had 80 hourly pay rates, at NUMMIall production workers get the same hourly rate –currently $17.85 – regardless of their jobs, exceptthat team leaders get an extra 60 cents. There are noseniority-, performance-, or merit-based bonuses.


Important as money is, equity is more importantstill in reducing tensions and resentments.

The second distinctive feature of NUMMI’s sys-tem is standardization. Typically, American com-panies approach team empowerment by allowingteams considerable autonomy in how they accom-plish tasks. NUMMI, in contrast, is obsessive aboutstandardized work procedures. It sees what oneNUMMI manager has called “the intelligent inter-pretation and application of Taylor’s time-and-mo-tion studies” as the principal key to its success. Thereference to Taylor may be jarring, but it fits.

Standardized Work...

At GM-Fremont, industrial engineers did alltime-and-motion analysis and formal job design,and workers tended to view them with resentmentor contempt. The problem, as one union official de-scribed it, was that management assumed a “divineright” to design jobs however it saw fit. Industrialengineers with no direct experience of the work be-yond capsule observation would shut themselves ina room, ponder various potentials of the humanbody, time the result, and promulgate a task design.Or so it seemed to workers, whom no one ever con-sulted despite their intimate familiarity with thespecific difficulties of the work in question.

Normally, when an industrial engineer presentedone of these pedantically designed jobs to a supervi-sor, the supervisor would politely accept it, thenpromptly discard it in favor of the more traditionalkick-ass-and-take-names technique. The worker,in turn, usually ignored both engineer and foremanand did the job however he or she was able – except,of course, when one of them was looking. If an in-dustrial engineer was actually “observing” – stop-watch and clipboard in hand – standard practicewas to slow down and make the work look harder.The entire charade was part of an ongoing game ofcoercion and avoidance. Multiply this scenario bytwo shifts and thousands of workers, and the resultis anything but the rational production of a high-quality car.

At NUMMI, in radical contrast to GM-Fremont,team members themselves hold the stopwatch.They learn the techniques of work analysis, descrip-tion, and improvement. This change in the designand implementation of standardized work has far-reaching implications for worker motivation andself-esteem, for the balance of power between work-ers and management, and for the capacity of thecompany to innovate, learn, and remember.

The job design process itself is relatively simple.Team members begin by timing one another withstopwatches, looking for the safest, most efficientway to do each task at a sustainable pace. They pickthe best performance, break it down into its fun-damental parts, then explore ways of improvingeach element. The team then takes the resultinganalyses, compares them with those of the othershift at the same work station, and writes the de-tailed specifications that become the standard workdefinition for everyone on both teams.

Taking part in the group’s analytical and descrip-tive work involves every team member in a com-mitment to perform each task identically. In onesense, therefore, standardized work is simply ameans of reducing variability in task performance,which may seem a relatively trivial achievement.In fact, however, reduced variability leads to a whole series of interconnected improvements: M Safety improves and injuries decline becauseworkers get a chance to examine all the possiblesources of strain and danger systematically. M Quality standards rise because workers haveidentified the most effective procedure for each job. M Inventory control grows easier, and inventorycarrying costs go down because the process flowsmore smoothly.M Job rotation becomes much more efficient andequitable, which makes absences less troublesome. M Flexibility improves because all workers are nowindustrial engineers and can work in parallel to re-spond rapidly to changing demands. For example,NUMMI can convert to a new line speed in four tosix weeks, a process that might easily have takensix months to a year at GM-Fremont, with its engi-neers frantically recalculating thousands of tasks

If orders decline, NUMMI canslow the production line toproduce fewer cars. In thesame situation, GM-Fremonthad to lay off an entire shift.

and trying to force the new standards on workers. Infact, GM-Fremont never even attempted anythingas demanding as a line-speed change. If orders de-clined, GM-Fremont had to lay off an entire shift.NUMMI’s new capacity to alter line speed means,among other things, that the plant can accommo-date a drop in orders by slowing production. M Standardized work also has the overall benefit ofgiving control of each job to the people who know it

TIME-AND-MOTION


best. It empowers the work force. Not surprisingly,NUMMI discovered that workers bought into theprocess quite readily. As one manager put it, “Theyunderstood the technique because it had been doneto them for years, and they liked the idea becausenow they had a chance to do it for themselves.”

...and Continuous Improvement

Yet by far the most striking advantage of stan-dardized work is that it gives continuous improve-ment a specific base to build on. As one managerput it, “You can’t improve a process you don’t un-derstand.” In this sense, standardization is the es-sential precondition for learning.

Indeed, standardization is not only a vehicle anda precondition for improvement but also a directstimulus. Once workers have studied and refinedtheir work procedures, problems with materialsand equipment quickly rise to the surface. More-over, since each worker is now an expert, each workstation is now an inspection station – and a centerof innovation.

At GM-Fremont, worker suggestions were apt tomeet a brick wall of indifference. At NUMMI, engi-neers and managers are meant to function as a sup-port system rather than an authority system. Whena team can’t solve a problem on its own, it can seekand get help. When a worker proposes complex in-novation, engineers are available to help assess thesuggestion and design its implementation.

The difference between traditional Taylorismand the learning-oriented NUMMI version resem-bles the difference between computer software de-signed to be “idiot-proof” and the kinds of comput-er systems that are meant to leverage and enhancetheir users’ capabilities. The first “de-skills” theoperator’s task to an extent that virtually elimi-nates the possibility of error, but it also eliminatesthe operator’s ability to respond to unpredictableevents, to use the system in novel ways or adapt itto new applications. The idiot-proof system may beeasy to use, but it is also static and boring. Leverag-ing systems make demands on the operator. Theytake time to learn and require thought and skill touse, but they are immensely flexible, responsive,and satisfying once mastered.

The difference goes deeper yet. At GM-Fremont –where work procedures were designed to be idiot-proof – the relationship between production systemand worker was adversarial. Standards and hierar-chy were there to coerce effort from reluctant work-ers. If the system functioned as expected and the

operator was sufficiently tractable and unimagina-tive, the two together could turn out a fair product.There was little the operator could improve on,however, and the role of the system was utterlyrigid until it broke down, whereupon everythingstopped until a specialist arrived.

At NUMMI, the relationship of workers to theproduction system is cooperative and dynamic. In-stead of circumventing user intelligence and initia-tive, the production system is designed to realize asmuch as possible of the latent collaborative poten-tial between the workers and the system.

Suggestion programs illustrate the two approach-es to organizational technology design. At manycompanies, suggestion programs are idiot-proof andopaque. They are designed primarily to screen outdumb ideas, and the basic review criteria, the iden-tity of the judges, the status of proposals, and thereasons for rejection are all a black box as far as theworkers are concerned. Predictably, a lot of theseprograms sputter along or die out altogether.

At NUMMI, the program is designed to encour-age a growing flow of suggestions and to help work-ers see and understand criteria, evaluators, process,status, and results. Like a computer system de-signed to leverage rather than de-skill, the programhelps employees form a mental model of the pro-gram’s inner workings. Not surprisingly, workersmade more than 10,000 suggestions in 1991, ofwhich more than 80% were implemented.

In systems that de-skill and idiot-proof, technolo-gy controls, indeed dominates, workers. In systemsdesigned for what experts call usability, the opera-tor both learns from and “teaches” the technology.Using learned analytical tools, their own experi-ence, and the expertise of leaders and engineers,workers create a consensual standard that theyteach to the system by writing job descriptions. Thesystem then teaches these standards back to work-ers, who, then, by further analysis, consultation,and consensus, make additional improvements.Continual reiteration of this disciplined process ofanalysis, standardization, re-analysis, refinement,and restandardization creates an intensely struc-tured system of continuous improvement. And thesalient characteristic of this bureaucracy islearning, not coercion.

This learning orientation captures the imagina-tion. People no one had ever asked to solve a prob-lem, workers who never finished high school, menand women who had spent 20 years or more in theauto industry without a single day of off-the-jobtraining found themselves suddenly caught up inthe statistical analysis of equipment downtime,putting together Pareto charts. One worker report-


ed that he did literally a hundred graphs before hegot one right.

A woman on the safety committee in the bodyshop described how she applied kaizen techniquesto her kitchen at home after a fire on her stove. Sheanalyzed the kitchen layout, installed a fire extin-guisher, and relocated her pot tops so she could usethem to smother flames. In short, she subjectedherself and her home work space to the formalproblem-solving procedures she had learned at theNUMMI plant.

The paradoxical feature such stories have in com-mon is their enthusiasm for a form of disciplinedbehavior that both theory and past practice seem torule out. This paradox grows from our failure to dis-tinguish between what Taylorist, bureaucratic pro-duction systems can be and what, regrettably, theyhave usually been.

The Psychology of Work

The chain of reasoning by which disciplinedstandardization leads inescapably to coercion, re-sentment, resistance, and further coercion seems toturn Taylorism and bureaucracy into what sociolo-gist Max Weber called an iron cage. Taylorism andbureaucracy may have a devastating effect on inno-vation and motivation, the reasoning goes, buttheir technical efficiency and their power to en-force compliance seem to be the perfect tools fordealing with employees assumed to be recalcitrant.Taylor himself at least occasionally endorsed thiscoercive view of work. Italics bristling, he once

wrote, “It is only through the enforced standardiza-tion of methods, enforced adoption of the best im-plements and working conditions, and enforced co-operation that this faster work can be assured. Andthe duty of enforcing the adoption of standards andof enforcing this cooperation rests with the man-agement alone.”

Against this background, it is hardly surprisingthat most managers and academics, at least in theWest, have come to believe that Taylorism and bu-reaucracy will inevitably alienate workers andsquander their human potential. But the psycholog-ical assumption underlying this expectation is thatworkers are incapable of delayed gratification.Managers seem to believe that performance willimprove only as work comes more and more to re-semble free play – our model of an intrinsically mo-tivating activity. Indeed, it is an an elementary ax-iom of economics that work is something thatworkers will always avoid.

NUMMI demonstrates the error of imputing in-fantile psychology to workers. Interviews withNUMMI team members suggest, in fact, that thiswhole historical accumulation of assumptions ob-scures three sources of adult motivation that theNUMMI production system successfully taps into:

First, the desire for excellence. Second, a mature sense of realism. Third, the positive response to respect and trust. The first of these – the desire to do a good job, the

instinct for workmanship – comes up again andagain in conversations with workers. The NUMMIproduction system and the training that went withit increased both the real competence of workersand their feelings of competence. Workers talk a lot

Workers once ashamed of their products are now inclined to let car owners know that they “helped build this one.”

TIME-AND-MOTION


system.” Even though this work lacks the kind ofintrinsic interest that would bring a worker in on a free Sunday, for example, the difference betweenthe levels of motivation at NUMMI and at GM-Fre-mont spells the difference between world-class andworst-in-class.

The third explanation of increased motivation isthe respect and trust that management showsworkers in NUMMI’s ongoing operations. For ex-ample, when the plant first began operations, thenew NUMMI managers responded quickly to re-quests from workers and union representatives foritems like new gloves and floor mats, which sur-prised workers used to seeing requests like theseturn into battles over management prerogative.

After a few months of getting everything theyasked for, workers and union representatives start-ed trying to think of ways to reciprocate. Eventual-ly, they decided that chrome water fountains wereunnecessary and told management they’d foundsome plastic ones for half the price. A few weekslater, management upped the ante one more timeby giving work teams their own accounts so theycould order supplies for team members without pri-or approval from management. This kind of behav-ior led workers to conclude that they did indeedshare common goals with management.

Power and Empowerment

The NUMMI production system confronts uswith a set of formalized procedures that seem de-signed not primarily as instruments of dominationbut as elements of productive technique that allparticipants recognize as tools in their own collec-tive interest. Management and labor support theNUMMI system. In fact, the first and overwhelm-ing fact to emerge from interviews is that no one atNUMMI wants to go back to the old GM-Fremontdays. Whatever their criticisms and whatever theirpositions, everyone feels that NUMMI is a far supe-rior work environment.

NUMMI’s no-layoff policy, management effortsto build an atmosphere of trust and respect, theNUMMI production system – especially the stimu-lus of its learning orientation – all help to explainthis attitude. Beyond these formal policies, how-ever, there are two more factors that help explainNUMMI’s success with workers. The first of these,as we’ve seen, is the psychology of work. The finalpiece of the puzzle has to do with power.

There are two kinds of power to consider: hierar-chical power within the organization and the power

about expertise, pride, and self-esteem. One UAWofficial named “building a quality product” as oneof the strategic goals that the union found mostcompelling at NUMMI. Perhaps the most strikingstory about pride in all the interviews came from a team leader:

Before, when I saw a Chevy truck, I’d chuckle to my-self and think, “You deserve that piece of crap if youwere stupid enough to buy one.” I was ashamed tosay that I worked at the Fremont plant. But when Iwas down at the Monterey Aquarium a few week-ends ago, I left my business card – the grunts evenhave business cards – on the windshield of a parkedNova with a note that said, “I helped build this one.”I never felt pride in my job before.

The second element of motivation is a maturesense of realism – in this case, the understandingthat unless NUMMI constantly improves its per-formance, competitors will take its market and itsworkers’ jobs. A useful psychological theory cannotassume that workers are so captive to the pleasureprinciple that their only source of motivation is theimmediate pleasure of intrinsically meaningfulwork. The evidence suggests that at least some ofthe workers at NUMMI are powerfully motivatedby the simple recognition that international com-petition now forces them to “earn their money theold-fashioned way.”

Other things being equal, work that is intrinsi-cally motivating – as opposed to mundane and rou-tine – is better than work that isn’t. But workers atNUMMI recognize that other things are not equal,and they are realistic in their recognition of havinghad an unlucky draw in terms of education and op-portunity. They see automobile assembly as work

Some workers take powerfulmotivation from theknowledge that they have to “earn their money theold-fashioned way.”

that can never have much instrinsic value, but theyunderstand that their own motivation levels cannevertheless vary from strongly negative, at GM-Fremont, to strongly positive, at NUMMI.

“What we have here is not some workers’utopia,” said one NUMMI worker. “Working on anassembly line in an automobile factory is still a lousy job….We want to continue to minimize the negative parts of the job by utilizing the new


balance between labor and management. NUMMItakes a distinctive approach to both.

In terms of hierarchical layers, NUMMI is a fairlytypical U.S. manufacturing plant, and in this sense,as well as in work-flow procedures, it is a very bu-reaucratic organization. NUMMI’s structure is notflat. It has several well-populated layers of middlemanagement. But consistent with the idea of turn-ing the technologies of coercion into tools for learn-ing, the function of hierarchy at NUMMI is notcontrol but support.

Decisions at NUMMI are made by broad verticaland horizontal consensus. At first glance, decisionmaking appears to be somewhat more centralizedthan at most U.S. factories, but this is because con-sensus-based decision making draws higher andlower layers into a dialogue, not because higher lev-els wield greater unilateral control. Both ends of thehierarchical spectrum are drawn into more deci-sion-making discussions than either would experi-ence in a conventional organization.

The contrast with the popular approaches to em-powerment is striking. At one U.S. telecommuni-cations company, the model organization today is a plant of 90 workers in self-managed teams, allreporting to a single plant manager. The company’sold model included a heavy layer of middle manage-ment whose key function was to command andcontrol, so it is easy to understand the inspiring ef-fect of the new approach. But at NUMMI, middlemanagement layers are layers of expertise, not ofrights to command, and if middle managers haveauthority, it is the authority of experience, mastery,and the capacity to coach.

As for the second aspect of power, many ob-servers have assumed that the intense discipline ofToyota-style operations requires complete man-agement control over workers and elimination ofindependent work-force and union power. But atNUMMI, the power of workers and the union localis still considerable. In some ways, their power hasactually increased. In fact, it may be that the NUMMI model has succeeded only because of thishigh level of worker and union power.

What makes the NUMMI production system soenormously effective is its ability to make produc-tion problems immediately visible and to mobilizethe power of teamwork. Implemented with trustand respect, both these features of the system cre-ate real empowerment. Wielded autocratically,they would have the opposite effect. Visible controlcould easily turn into ubiquitous surveillance.Teamwork could become a means of mobilizingpeer pressure. A healthy level of challenge could de-generate into stress and anxiety.

The NUMMI production system thus gives man-agers enormous potential control over workers.With this potential power ready at hand, and underpressure to improve business performance, there isa real danger that the relationship will sooner or lat-er slide back into the old coercive pattern.

But such a slide would have an immediate andsubstantial negative impact on business perfor-mance, because labor would respond in kind. Analienated work force wipes out the very foundation

The new system gives workersgreat positive power to improve production andgreat negative power to disrupt it.

of continuous improvement and dries up the flowof worker suggestions that fuel it. And the lack ofinventory buffers means that disaffected workerscould easily bring the whole just-in-time produc-tion system to a grinding halt. Alongside workers’positive power to improve quality and efficiency,the system also gives workers an enormous nega-tive power to disrupt production.

In other words, NUMMI’s production system in-creases the power both of management over work-ers and of workers over management.

A system this highly charged needs a robust gov-ernance process in which the voices of manage-ment and labor can be clearly heard and effectivelyharmonized on high-level policy issues as well ason work-team operating issues. The union givesworkers this voice.

When, for example, workers felt frustrated bywhat they saw as favoritism in management’s se-lection of team leaders, the union largely eliminat-ed the problem by negotiating a joint union-man-agement selection process based on objective testsand performance criteria.

As one UAW official put it, “The key to NUM-MI’s success is that management gave up some ofits power, some of its traditional prerogatives. Ifmanagers want to motivate workers to contributeand to learn, they have to give up some of theirpower. If managers want workers to trust them, weneed to be 50-50 in making the decision. Don’t justmake the decision and say, ‘Trust me.’”

Union leaders and top management confer regu-larly on- and off-site to consider a broad range ofpolicy issues that go far beyond the traditionalscope of collective bargaining. The union local has

TIME-AND-MOTION


embraced the NUMMI concept and its goals. But itsability and willingness to act as a vehicle for workerconcerns adds greatly to the long-term effective-ness of the organization.

NUMMI’s ability to sustain its productivity,quality, and improvement record now depends onworkers’ motivation, which rests, in turn, on theperception and reality of influence, control, and eq-uitable treatment. It is in management’s own inter-est that any abuse of management prerogativesshould meet with swift and certain penalties. Thecontribution of labor’s positive power depends onthe reality of its negative power.

In this way, the union not only serves workers’special interests, it also serves the larger strategicgoals of the business by effectively depriving man-agement of absolute domain and helping to main-tain management discipline.

Empowerment is a powerful and increasinglypopular approach to reinvigorating moribund orga-nizations. The NUMMI case points up two of em-powerment’s potential pitfalls and suggests ways ofovercoming them.

First, worker empowerment degenerates into exploitation if changes at the first level of manage-ment are not continuously reinforced by changesthroughout the management hierarchy. Strong em-ployee voice is needed to ensure that shop-floorconcerns are heard at all levels of management.Without it, workers’ new power is little more thanthe power to make more money for management.

Second, worker empowerment degenerates intoabandonment if work teams fail to get the righttools, training in their use, and support in their im-plementation. Standardized work, extensive train-ing in problem solving, a responsive managementhierarchy, and supportive specialist functions arekey success factors for empowerment strategies.

Taylorist time-and-motion discipline and formalbureaucratic structures are essential for efficiencyand quality in routine operations. But these princi-ples of organizational design need not lead to rigidi-ty and alienation. NUMMI points the way beyondTaylor-as-villain to the design of a truly learning-oriented bureaucracy.Reprint 93101

Harvard Business School PublishingCustomer Service - Box 230-560 Harvard Way Boston, MA 02163Telephone: U.S. and Canada (800) 545-7685Outside U.S. and Canada: (617) 495-6117 or 495-6192Fax: (617) 495-6985Internet address: [email protected]

Harvard Business School Management Productions videos are produced by award winning documentary filmmakers. You’ll findthem lively, engaging, and informative.

Please inquire about HBR’s custom service and quantity discounts.We will print your company’s logo on the cover of reprints, collec-tions, or books in black and white, two color, or four color. The pro-cess is easy, cost effective, and quick.

Telephone: (617) 495-6198 or Fax: (617) 496-8866

For permission to quote or reprint on a one-time basis:Telephone: (800) 545-7685 or Fax: (617) 495-6985

For permission to re-publish please write or call:

Permissions EditorHarvard Business School PublishingBox 230-560 Harvard Way Boston, MA 02163Telephone: (617) 495-6849

HBR Subscriptions

HBR Article ReprintsHBR Index and Other CatalogsHBS CasesHBS Press Books

HBS Management ProductionsVideos

HBR Custom Reprints

Permissions

Harvard Business ReviewHarvard Business ReviewU.S. and CanadaSubscription ServiceP.O. Box 52623Boulder, CO 80322-2623

Telephone: (800) 274-3214Fax: (617) 496-8145

Outside U.S. and CanadaTower HouseSovereign ParkLathkill StreetMarket HarboroughLeicestershire LE16 9EFTelephone: 44-85-846-8888Fax: 44-85-843-4958

American Express, MasterCard, VISA accepted. Billing available.

Time-and-Motion Regained - Sustained Excellence through People

Documents

Transcript of Time-and-Motion Regained - Sustained Excellence through People