the · 2019-07-19 · decoding the dna of the toyota production system let's look at how...

96 ARTWORK BY AMY NINC

TheToyota

story hasbeen intensively

researched andpainstakingly

documented,yet what reallyhappensinside thecompanyremains amystery.

Here's newinsight into

the unspokenrules that

give Toyotaits competitiveedge.

Decoding theDNA of theToyotaProductionSystemby Steven Spear and H. Kent Bowen

THE TOYOTA PRODUCTION SYSTEM HASlong heen hailed as the source of Toyota'soutstanding performance as a manufacturer.

The system's distinctive practices -its kanhan cardsand quality circles, for instance - have heen widelyintroduced elsewhere. Indeed, following their owninternal efforts to henchmark the world's hest man-ufacturing companies, GM, Ford, and Chrysler haveindependently created major initiatives to developToyota-like production systems. Companies thathave tried to adopt the system can be found in fleldsas diverse as aerospace, consumer products, metalsprocessing, and industrial products.

What's curious is that few manufacturers havemanaged to imitate Toyota successfully-eventhough the company has been extraordinarily openahout its practices. Hundreds of thousands of exec-utives from thousands of husinesses have touredToyota's plants in Japan and the United States.Frustrated hy their inability to replicate Toyota'sperformance, many visitors assume that the secretof Toyota's success must lie in its cultural roots.But that's just not the case. Other Japanese compa-nies, such as Nissan and Honda, have fallen shortof Toyota's standards, and Toyota has successfullyintroduced its production system all around theworld, including in North America, where the com-pany is this year huilding over a million cars, mini-vans, and light trucks.

So why has it heen so difficult to decode the Toy-ota Production System? The answer, we helieve, isthat observers confuse the tools and practices theysee on their plant visits with the system itself. Thatmakes it impossible for them to resolve an apparentparadox of the system - namely, that activities, con-nections, and production flows in a Toyota factoryare rigidly scripted, yet at the same time Toyota'soperations are enormously flexible and adaptable.Activities and processes are constantly being chal-lenged and pushed to a higher level of performance,enabling the company to continually innovate andimprove.

To understand Toyota's success, you have to un-ravel the paradox - you have to see that the rigidspecification is the very thing that makes the flexi-bility and creativity possible. That's what we cameto realize after an extensive, four-year study of theToyota Production System in which we examinedthe inner workings of more than 40 plants in theUnited States, Europe, and Japan, some operatingaccording to the system, some not. We studied hothprocess and discrete manufacturers whose productsranged from prefabricated housing, auto parts andfinal auto assembly, cell phones, and computerprinters to injection-molded plastics and aluminum

HARVARD BUSINESS REVIEW September-October 1999 97

DECODING THE DNA OF THE TOYOTA PRODUCTION SYSTEM

extrusions. We studied not only routine productionwork hut also service functions like equipmentmaintenance, workers' training and supervision,logistics and materials handling, and process designand redesign.

We found that, for outsiders, the key is to under-stand that the Toyota Production System creates acommunity of scientists. Whenever Toyota definesa specification, it is establishing sets of hypothesesthat can then he tested. In other words, it is follow-ing the scientific method. To make any changes,Toyota uses a rigorous problem-solving processthat requires a detailed assessment of the currentstate of affairs and a plan for improvement that is,in effect, an experimental test of the proposedchanges. With anything less than such scientificrigor, change at Toyota would amount to littlemore than random trial and error-a blindfoldedwalk through life.

The fact that the scientific method is so in-grained at Toyota explains why the high degree ofspeciflcation and structure at the company does notpromote the command and control environmentone might expect. Indeed, in watching people doingtheir jobs and in helping to design production pro-cesses, we learned that the system actually stimu-lates workers and managers to engage in the kind ofexperimentation that is widely recognized as thecornerstone of a learning organization. That iswhat distinguishes Toyota from all the other com-panies we studied.

The Toyota Production System and the scientificmethod that underpins it were not imposed onToyota-they were not even chosen consciously.The system grew naturally out of the workings ofthe company over five decades. As a result, it hasnever been written down, and Toyota's workersoften are not able to articulate it. That's why it's sohard for outsiders to grasp. In this article, we attemptto lay out how Toyota's system works. We try tomake explicit what is implicit. We describe fourprinciples - three rules of design, which show howToyota sets up all its operations as experiments,and one rule of improvement, which describes howToyota teaches the scientific method to workers atevery level of the organization. It is these rules-and

Steven Spear is an assistant professor of business admin-istration at Harvard Business School in Boston. H. KentBowen is the Bruce Rauner Professor of Business Admin-istration, also at Harvard Business School. ProfessorBowen is the coauthor of "Regaining the Lead in Manu-facturing" {HBR September-October 1994}.

To discuss this article, join HBR's authors and readers inthe HBR Forum at www.hbr.orglforum.

98

not the specific practices and tools that people ob-serve during their plant visits-that in our opinionform the essence of Toyota's system. That is why wethink of the rules as the DNA of the Toyota Produc-tion System. Let's take a closer look at those rules(for a summary, see the sidebar "The Four Rules").

Rule 1: How People WorkToyota's managers recognize that the devil is in thedetails; that's why they ensure that all work is highlyspecified as to content, sequence, timing, and out-come. When a car's seat is installed, for instance,the bolts are always tightened in the same order, thetime it takes to turn each bolt is specified, and sois the torque to which the holt should be tightened.Such exactness is applied not only to the repetitivemotions of production workers hut also to the activ-ities of all people regardless of their functional spe-cialty or hierarchical role. The requirement thatevery activity be specified is the first unstated ruleof the system. Put this baldly, the rule seems sim-ple, something you'd expect everyone to under-stand and he ahle to follow easily. But in reality,most managers outside Toyota and its partnersdon't take this approach to work design and execu-tiori -everi when they think they do.

The Four Rules

The tacit knowledge that underlies the ToyotaProduction System can be captured in four basicrules. These rules guide the design, operation, andimprovement of every activity, connection, andpathway for every product and service.The rules areas follows:

Rule 1: All work shall be highly specified as tocontent, sequence, timing, and outcome.

Rule 2: Every customer supplier connection must bedirect, and there must be an unambiguous yes-or-noway to send requests and receive responses.

Rule 3; The pathway for every product and servicemust be simple and direct.

Rule 4: Any improvement must be made inaccordance with the scientific method, under theguidance of a teacher, at the lowest possible level inthe organization.

All the rules require that activities, connections, andflow paths have built-in tests to signal problemsautomatically. It is the continual response to problemsthat makes this seemingly rigid system so flexible andadaptable to changing circumstances.

HARVARD BUSINESS REVIEW September-October 1999


Let's look at how operators at a typical U.S. autoplant install the front passenger seat into a car.They are supposed to take four bolts from a card-board box, carry them and a torque wrench to thecar, tighten the four bolts, and enter a code into acomputer to indicate that the work has been donewithout problems. Then they wait for the next carto arrive. New operators are usually trained by ex-perienced workers, who teach by demonstratingwhat to do. A seasoned colleague might be availableto help a new operator with any difficulties, such asfailing to tighten a bolt enough or forgetting to enterthe computer code.

This sounds straightforward, so what's wrongwith it? The problem is that those specifications ac-tually allow-and even assume-considerable varia-tion in the way employees do their work. Withoutanyone realizing it, there is plenty of scope for anew operator to put the seat into the vehicle differ-ently than an experienced employee would. Someoperators might put the front bolts in after the rearbolts; some might do it the other way around. Someoperators might put each bolt in and then tightenthem all; others might tighten as they go along. Allthis variation translates into poorer quality, lowerproductivity, and higher costs. More important, ithinders learning and improvement in the organiza-tion because the variations hide the link betweenhow the work is done and the results.

At Toyota's plants, because operators [new andold, junior and supervisory) follow a well-definedsequence of steps for a particular job, it is instantlyclear when they deviate from the specifications.Consider how workers at Toyota's Georgetown,Kentucky, plant install the right-front seat into aCamry. The work is designed as a sequence of seventasks, all of which are expected to be completedin 5 5 seconds as the car moves at a fixed speedthrough a worker's zone. If the production workerfinds himself doing task 6 (installing the rear seat-bolts) before task 4 [installing the front seat-bolts),then the job is actually being done differently thanit was designed to be done, indicating that some-thing must be wrong. Similarly, if after 40 secondsthe worker is still on task 4, which should havebeen completed after 31 seconds, then something,too, is amiss. To make problem detection even sim-pler, the length of the floor for each work area ismarked in tenths. So if the worker is passing thesixth of the ten floor marks [that is, if he is 33 sec-onds into the cyclej and is still on task 4, then heand his team leader know that he has fallen behind.Since the deviation is immediately apparent, workerand supervisor can move to correct the problemright away and then determine how to change the

specifications or retrain the worker to prevent a re-currence. [See the sidebar "How Toyota's WorkersLearn the Rules" for a short description of the pro-cess by which workers learn how to design work inthis way.)

Even complex and infrequent activities, such astraining an inexperienced workforce at a new plant,launching a new model, changing over a productionline, or shifting equipment from one part of a plantto another, arc designed according to this rule. Atone of Toyota's suppliers in Japan, for example,equipment from one area of the plant was moved tocreate a new production line in response to changesin demand for certain products. Moving the machin-ery was broken into 14 separate activities. Each ac-tivity was then further subdivided and designed asa series of tasks. A specific person was assigned to doeach task in a specified sequence. As each of the ma-chines was moved, the way the tasks were actually

How Toyotas WorkersLearn the Rules

If the rules of the Toyota Production System aren'texplicit, how are they transmitted? Toyota's managersdon't tell workers and supervisors specifically how todo their work. Rather, they use a teaching and learningapproach that allows their workers to discover therules as a consequence of solving problems. Forexample, the supervisor teaching a person theprinciples of the first rule will come to the work siteand, while the person is doing his or her job, ask aseries of questions:

• How do you do this work?

• How do you know you are doingthis work correctly?

• How do you know that the outcomeis free of defects?

• What do you do if you havea problem?

This continuing process gives the person increasinglydeeper insights into his or her own specific work. Frommany experiences of this sort, the person graduallylearns to generalize how to design all activitiesaccording to the principles embodied in rule 1.

All the rules are taught in a similar Socratic fashionof iterative questioning and problem solving.Although this method is particularly effertive forteaching, it leads to knowledge that is implicit.Consequently, the Toyota Production System has sofar been transferred successfully only when managershave been able and willing to engage in a similarprocess of questioning to facilitate learning by doing.

•LHARVARD BUSINESS REVIEW September-October 1999 99


The Experiments of the Toyota Production System

When organizations are managed according to the four rules,individuals are repeatedly conducting experiments, testing inoperation the hypotheses built Into the designs of individual

work activities, customer-supplier connections, pathways, andimprovement efforts.The hypotheses, the way they are tested,and the response if they are refuted are summarized below.

Rule Hypotheses

1 The person or machine can do theactivity as specified.

If the activity is done as specified, thegood or service will be defect free.

Signs of a problem

The activity is not done asspecified.

The outcome is defective.

Customers' requests will be for goodsand services in a specific mix andvolume.

The supplier can respond tocustomers' requests.

Every supplier that is connectedto the flow path is required.

Any supplier not connected tothe flow path is not needed.

A specific change in an activity,connection, or flow path will improvecost, quality, lead time, batch size, orsafety by a specific amount.

Responses don't keep pacewith requests.

The supplier is idle, waitingfor requests.

Responses

Determine the true skill level of the personor the true capability of the machineand train or modify as appropriate.

Modify the design activity.

Determine the true mix and volume ofdemand and the true capability of thesupplier; retrain, modify activities, or reassigncustomer-supplier pairs as appropriate.

A person or machine isU not actually needed.

A nonspecified supplierprovides an intermediategood or service.

t

The actual result is differentfrom the expected result.

Determine why the supplier was unnecessary,̂and redesign the flow path.

Learn why the nonspecified supplier wasactually required, and redesign the flow path.

Learn how the activity was actually performedor the connection or flow path was actuallyoperated. Determine the true effects of thechange. Redesign the change.

done was compared with what was expected ac-cording to the original design, and discrepancieswere immediately signaled.

In calling for people to do their work as a highlyspecified sequence of steps, rule i forces them totest hypotheses through action. Performing theactivity tests the two hypotheses implicit in its de-sign: first, that the person doing the activity is capa-ble of performing it correctly and, second, that per-forming the activity actually creates the expectedoutcome. Rememher the seat installer? If he can'tinsert the seat in the specified way within the spec-ified amount of time, then he is clearly refuting atleast one of these two hypotheses, thereby indicat-ing that the activity needs to he redesigned or theworker needs to be trained.

Rule 2: How People Connectwhere the first rule explains how people performtheir individual work activities, the second ruleexplains how they connect with one another. We

express this rule as follows: every connection mustbe standardized and direct, unambiguously specify-ing tbe people involved, tbe form and quantity oftbe goods and services to be provided, the way re-quests are made by each customer, and tbe expectedtime in which the requests will be met. The rulecreates a supplier-customer relationship betweeneach person and tbe individual who is responsiblefor providing tbat person with eacb specific good orservice. As a result, there are no gray zones in decid-ing who provides what to whom and when. When aworker makes a request for parts, there is no confu-sion about tbe supplier, tbe number of units re-quired, or tbe timing of tbe delivery. Similarly,wben a person needs assistance, there is no confu-sion over who will provide it, how tbe help will betriggered, and what services will he delivered.

The real question that concerns us here is whetherpeople interact differently at Toyota than they do atother companies. Let's return to our seat installer.When he needs a new container of plastic bolt cov-ers, be gives a request to a materials bandler, wbo is

100 HARVARD BUSINESS REVIEW September-October 1999


tbe designated bolt-cover supplier. Commonly,such a request is made witb a kanban, a laminatedcard tbat specifies the part's identification number,the quantity of parts in the container, and the loca-tions of the part supplier and of tbe worker (the cus-tomer} who will install it. At Toyota, kanban cardsand other devices like andon cords set up directlinks between the suppliers and the customers. Theconnections are as smooth as the passing of the batonin tbe best Olympic relay teams because tbey arejust as carefully tbought out and executed. For ex-ample, tbe number of parts in a container and tbenumber of containers in circulation for any givenpart are determined by tbe physical realities of tbeproduction system-tbe distances, tbe changeovertimes, and so on. Likewise, the number of workersper team is determined by tbe types of problems ex-pected to occur, tbe level of assistance the teammembers need, and tbe skills and capabilities of theteam's leader.

Other companies devote substantial resources tocoordinating people, but tbeir connections generallyaren't so direct and unambiguous. In most plants, re-quests for materials or assistance often take a con-voluted route from the line worker to tbe suppliervia an intermediary. Any supervisor can answer anycall for help because a specific person bas not beenassigned. Tbe disadvantage of that approach, as Toy-ota recognizes, is that when something is everyone'sproblem it becomes no one's problem.

Tbe requirement that people respond to supply re-quests within a specific time frame further reducesthe possibility of variance. That is especially true inservice requests. A worker encountering a problemis expected to ask for assistance at once. The desig-nated assistant is then expected to respond immedi-ately and resolve the problem within the worker'scycle time. If the worker is installing a front seatevery 55 seconds, say, then a request for help mustbe answered and dealt with in less than tbe 55 sec-onds. If tbe problem cannot be resolved in less tban55 seconds, tbat failure immediately cballengesthe hypotheses in this customer-supplier connec-tion for assistance. Perhaps the request signal isambiguous. Perhaps the designated assistant has toomany other requests for help and is busy or is nota capable problem solver. Constantly testing tbe hy-potheses in this way keeps the system fiexiblc, mak-ing it possible to adjust the system continually andconstructively.

The striking thing about the requirement to askfor help at once is that it is often counterintuitiveto managers who are accustomed to encouragingworkers to try to resolve problems on their own be-fore calling for bclp. But tben problems remain hid-

den and are neither shared nor resolved company-wide. The situation is made worse if workers beginto solve problems tbemselves and then arbitrarilydecide when tbe problem is big enougb to warrant acall for belp. Problems mount up and only getsolved much later, hy which time valuable infor-mation about the real causes of the problem mayhave heen lost.

Rule 3: How the Production LineIs ConstructedAll production lines at Toyota have to be set up sotbat every product and service flows along a simple,specified patb. Tbat patb sbould not cbange unlessthe production line is expressly redesigned. In prin-ciple, tben, there are no forks or loops to convolutethe flow in any of Toyota's supply chains. That's thethird rule.

To get a concrete idea of what that means, let'sreturn to our seat installer. If he needs more plasticbolt covers, he orders them from the specific mater-ial bandler responsible for providing him with holtcovers. That designated supplier makes requests tohis own designated supplier at the off-line store inthe factory who, in turn, makes requests directly tohis designated supplier at the holt cover factory'sshipping dock. In this way, the production linelinks each person who contributes to tbe produc-tion and delivery of tbe product, from the Toyotafactory, tbrougb the molding company, to even theplastic pellet manufacturer.

The point is that when production lines are de-signed in accordance with rule 3, goods and servicesdo not flow to the next available person or macbinebut to a specific person or macbine. If for some rea-son tbat person or machine is not available, Toyotawill see it as a problem tbat might require tbe lineto be redesigned.

Tbe stipulation that every product follow a sim-ple, prespecified path doesn't mean that each pathis dedicated to only one particular product, how-ever. Quite tbe contrary: eacb production line at aToyota plant typically accommodates many moretypes of products than its counterparts do at othercompanies.

The third rule doesn't apply only to products-itapplies to services, like help requests, as well. If ourseat installer, for example, needs help, that tooeomes from a single, specified supplier. And if thatsupplier can't provide the necessary assistance, she,in turn, has a designated helper. In some of Toyota'splants, this pathway for assistance is three, four, orfive links long, connecting the shop fioor workerto tbe plant manager.



The third rule runs contrary to conventional wis-dom about production lines and pooling resources-even contrary to how most people think the ToyotaProduction System works. According to receivedwisdom, as a product or service is passed down theline, it should go to the next machine or person avail-able to process it furtber. Similarly, most people as-sume that help should come from the first availableperson ratber tban from a specific person. At oneauto parts supplier we studied, for example, most oftbe parts could be stamped on more tban one pressmacbine and welded at more than one welding sta-tion. Before the company adopted the Toyota sys-tem, its practice was to pass each part on to tbe firstavailable press machine and to the first availahlewelder. When the plant switched over, under Toy-ota's guidance, each type of part followed only oneproduction path through the plant.

By requiring that every pathway be specified, tberule ensures that an experiment will occur eachtime the path is used. Here the hypotheses emhed-ded in a pathway designed according to rule 3 arethat every supplier connected to the pathway is nec-essary, and any supplier not connected is not neces-sary. If workers at the auto parts supplier foundthemselves wanting to divert production to another

On-Demand Production at the Aisin Mattress Factory

Aisin Seiki produces 850 varieties of mattresses, distinguished by size, firmness,covering fabric, quilting pattern, and edge trim. Customers can order any one ofthese in a retail store and have it delivered to their homes in three days, yet Aisinmaintains an inventory at the plant equal to just 1.5 days of demand.To be able todo so, Aisin has made thousands of changes in individual work activities, in theconnections linking customers and suppliers of intermediate goods and services,and to the overall production lines.This table captures how dramatic the results ofthose changes have been.

1986 1988 1992 1996 1997

Styles

Units per day

Units per person

Productivity index

Finished-goods inventory (days)

Number of assembly lines

200

160

100

30

325

230

11

138

670

360

13

175

2.5 1.8

machine or welding station, or if they began turningfor belp to someone other than their designatedhelpers, they'd conclude that their actual demand orcapacity didn't match their expectations. And there

would also be no ambiguity about whicb pressor welder was involved. Again, tbe workers wouldrevisit the design of their production line. Thusrule 3, like rules i and 2, enables Toyota to conductexperiments and remain fiexible and responsive.

Rule 4: How to ImproveIdentifying problems is just the first step. For peo-ple to consistently make effective cbanges, theymust know how to change and who is responsiblefor making tbe cbanges. Toyota explicitly teaebespeople how to improve, not expecting them to learnstrictly from personal experience. That's where therule for improvement comes in. Specifically, rule 4stipulates that any improvement to productionactivities, to connections hetween workers or ma-chines, or to pathways must be made in accordancewith the scientific method, under the guidance ofa teacher, and at the lowest possible organizationallevel. Let's look first at bow Toyota's people learnthe scientific method.

How People Learn to Improve. In 1986, Aisin Seiki,a Toyota Group company that made complex prod-ucts such as power trains for the auto industry, cre-ated a line to manufacture mattresses to absorb ex-

cess capacity in one of its plants.' Since 1986, its range bas grown

from 200 to 850 types of mat-tresses, its volume bas grownfrom 160 mattresses per day to550, and its productivity basdoubled. Here's an example ofhow they did it.

On one of our visits to thisplant, we studied a team of mat-tress assembly workers whowere being taugbt to improvetbeir problem-solving skills hyredesigning tbeir own work.Initially, the workers had beenresponsible for doing only tbeirown standardized work; theybad not been responsible forsolving problems. Then theworkers were assigned a leaderwho trained them to frame prob-

3 2 lems better and to formulateand test hypotbeses- in otber

I words, be taugbt tbem bow touse tbe scientific metbod to

design their team's work in accordance witb thefirst three rules. The results were impressive. Oneof the team's accomplishments, for instance, was toredesign the way edging tape was attached to tbe

750 850

530

20

550

26

197 208

1.5 1.5



Toyota's Commitment to Learning

All the organizations we studied that are managedaccording to the Toyota Production System share an ^overarching belief that people are the most significant Mcorporate asset and that investments in their knowledgeand skills are necessary to build competitiveness. That'swhy at these organizations all managers are expected tobe able to do the jobs of everyone they supervise and alsoto teach their workers how to solve problems according tothe scientific method. The leadership model applies asmuch to the first-level "team leader" supervisors as it doesto those at the top of the organization. In that way,everybody at Toyota shares in the development of humanresources. In effect, there is a cascading pathway forteaching, which starts with the plant manager, thatdelivers training to each employee.

To reinforce the learning and improvement process,each plant and major business unit in the Toyota Groupemploys a number of Toyota Production Systemconsultants whose primary responsibility is to help senior •managers move their organizations toward the ideal. •These "learner-leader-teachers" do so by identifying evermore subtle and difficult problems and by teachingpeople how to solve problems scientifically.

Many of these individuals have received intensivetraining at Toyota's Operations Management ConsultingDivision. OMCD was established in Japan as an outgrowthof efforts by Taiichi Ohno- one of the original architects ofthe Toyota Production System - to develop and diffuse thesystem throughout Toyota and its suppliers. Many ofToyota's top officers- including Toyota Motor's newpresident, Fujio Cho- have honed their skills within OMCD.During their OMCD tenure, which can extend for a periodof years, Toyota's employees are relieved of all lineresponsibilities and instead are charged with leadingimprovement and training activities in the plants of Toyotaand its suppliers. By supporting all of Toyota's plant andlogistical operations in this way, OMCD serves as a trainingcenter, building its consultants' expertise by giving themopportunities to solve many difficult problems and teachothers to do the same.

In 1992,Toyota founded the Toyota Supplier SupportCenter (TSSC) in the United States to provide NorthAmerican companies with training in the ToyotaProduction System. Modeled on OMCD,TSSC has givenworkshops to more than 140 companies and directassistance to 80. Although most of these companies areauto suppliers, few are exclusively Toyota suppliers;participants come from other industries and fromuniversities, government organizations, and industryassociations. Indeed, much of the research for this paperwas derived from the experience of one of the authors,who was a member of a TSSC team for five months,promoting the Toyota Production System at a plant thatsupplies Toyota and two other auto assembly plants.

mattresses, thereby reducing the defect rate by90%. (See the exhibit "On-Demand Production atthe Aisin Mattress Factory.")

To make changes, people are expected to pre-sent the explicit logic of the hypotheses. Let'slook at what that can involve. Hajime Ohba, gen-eral manager of the Toyota Supplier Support Cen-ter, was visiting a factory in which one of TSSC'sconsultants was leading a training and improve-ment activity (for a description of the role of theToyota Production System promotion centers,see the sidebar "Toyota's Commitment to Learn-ing"). The consultant was helping factory em-ployees and their supervisor reduce the manufac-turing lead time of a particular line, and Ohhawas there to evaluate the group's progress.

Group members began their presentation bydescribing the steps by which their product wascreated -delineating all the problems they identi-fied when they had first studied the process forchanging over a machine from making one partto making another, and explaining the specificchanges they had made in response to each of thoseprohlems. They concluded hy saying, "When westarted, the changeover required 15 minutes. Wewere hoping to reduce that by two-thirds- toachieve a five-minute changeover - so that wecould reduce batch sizes by two-thirds. Becauseof the modifications we made, we achieveda changeover time of seven and a half minutes- areduction of one-half."

After their presentation, Ohba asked why thegroup members had not achieved the five-minutegoal they had originally established. They were abit taken aback. After all, they had reduced thechangeover time by 50%, yet Ohba's questionsuggested he had seen opportunities for evengreater improvement that they had missed. Theyoffered explanations having to do with machinecomplexity, technical difficulty, and equipmentupgrade costs. Ohba responded to these replieswith yet more questions, each one meant to pushthe consultant and the factory people to articu-late and challenge their most basic assumptionsabout what could and could not he changed-as-sumptions that both guided and constrained theway they had solved their problems. Were theysure four bolts were necessary? Might thechangeover be accomplished with two? Werethey certain that all the steps they included in thechangeover were needed? Might some be com-bined or eliminated? In asking why they had notachieved the five-minute goal, Ohha was not sug-gesting that the team had failed. Rather, he wastrying to get them to realize that they had not



fully explored all their improvement opportunitiesbecause they had not questioned their assumptionsdeeply enough.

There was a second reason for Ohha's persis-tence. He was trying to show the group memhersthat their improvement activity had not heen car-ried out as a bona fide experiment. They had estah-hshed a goal of five minutes hased on the premisethat faster changeovers and smaller batches are bet-ter than slower changeovers and larger batches. Buthere they were confusing goals with predictionsbased on hypotheses. The goal was not a predictionof what they helieved they would achieve throughthe specific improvement steps they planned totake. As a result, they had not designed the im-provement effort as an experiment with an explieit,clearly articulated, verifiahle hypothesis of theform, "If we make the following specific changes,we expect to aehieve this specific outcome." Al-though they had reduced the changeover time con-siderably, they had not tested the hypotheses im-plicit in their effort. For Ohba, it was critical thatthe workers and their supervisor realize that howthey made changes was as important as whatchanges they made.

Who Does the Improvement. Frontline workersmake the improvements to their own jobs, andtheir supervisors provide direetion and assistanceas teachers. If something is wrong with the way aworker eonnects with a particular supplier withinthe immediate assembly area, the two of themmake improvements, with the assistance of theircommon supervisor. The Aisin team we describedearlier, for example, consisted of the assembly lineworkers and the supervisor, who was also their in-structor. When changes are made on a larger scale,Toyota ensures that improvement teams are createdconsisting of the people who are directly affectedand the person responsible for supervising the path-ways involved.

Thus tbe process remains the same even at thehighest levels. At Aisin's mattress faetory, wefound that the plant manager took responsibilityfor leading the change from three production linesback to two (the number had risen to three to copewith an increase in product types). He was involvednot just because it was a big change but also be-cause he had operational responsibility for oversee-ing the way work fiowed from the feeder lines tothe final assembly lines. In this way, Toyota en-

Countermeasures in the Toyo

Toyota does not consider any of the tools or practices -such as kanbans or andon cords, which so many outsidershave observed and copied-as fundamental to the ToyotaProduction System.Toyota uses them merely as temporaryresponses to specific problems that will serve until abetter approach is found or conditions change.They'rereferred to as "countermeasures," rather than "solutions,"because that would imply a permanent resolution to aproblem. Over the years, the company has developeda robust set of tools and practices that it uses ascountermeasures, but many have changed or even beeneliminated as improvements are made.

So whether a company does or does not use anyparticular tool or practice is no Indication that It is trulyapplying Toyota's rules of design and improvement. Inparticular, contrary to the Impression that the concept ofzero inventory is at the heart of the Toyota system, we'veobserved many cases in which Toyota actually built up itsInventory of materials as a countermeasure.The idealsystem would in fact have no need for inventory. But, inpractice, certain circumstances may require it:• Unpredictable downtime or yields. Sometimes aperson or a machine is unable to respond on demandwhen a request is made because of an unexpected

mechanical breakdown. For this reason, safety stockis held to protect the customer against randomoccurrences.The person responsible for ensuring thereliability of a machine or process owns that inventoryand strives to reduce the frequency and length ofdowntimes so that the amount of the safety stock canbe reduced.• Time-consuming setups. Difficulties in switching amachine from processing one kind of product to anothercan prevent a supplier from responding immediately.Therefore, suppliers will produce the product in batchsizes greater than one and hold the excess as inventory soit can respond immediately to the customer. Of course,suppliers will continually try to reduce the changeovertime to keep batch sizes and stores of inventory as smallas possible. Here, the owners of both the problem and thecountermeasure are the machine operator and the teamleader, who are responsible for reducing changeover timesand batch sizes.

• Volatility in the mix and volume of customer demand.In some cases, variations in customers' needs are so largeand unpredictable that it is impossible for a plant toadjust its production to them quickly enough. In thoseinstances, buffer stock is kept at or near the shipping point

104 HABTARD BUSINESS REVIEW September-October 1999


sures that problem solving and learning take placeat all levels of the company. Of course, as we havealready seen, Toyota will bring in external expertsas necessary to ensure the quality of the learningprocess.

In the long term, the organizational structures ofcompanies that follow the Toyota Production Sys-tem will shift to adapt to the nature and frequencyof the problems they encounter. Since the organiza-tional changes are usually being made at a very lowlevel, however, they can be hard for outsiders to de-tect. That's hecause it is the nature of the problemsthat determines who should solve them and how theorganization is designed. One consequence is thatdifferent organizational structures coexist quitehappily even in the same plant.

Consider Toyota's engine-machining plant inKamigo, Japan. The plant has two machine divi-sions, each of which has three independent produc-tion shops. When we visited in summer 1998, theproduction people in the first machine division an-swered to shop heads, and the process engineersanswered directly to the head of the division. How-ever, in the second machine division, the engineerswere distrihuted among the three shops and, like

as a countermeasure.The buffer stock also serves as asignal to production and sales managers that the personwho works most directly with the customer must helpthat customer eliminate the underlying causes of anypreventable swings in demand.

In many cases, the same type of product is held indifferent types of inventory.Toyota does not pool itsvarious kinds of inventory, even though doing so wouldreduce its inventory needs in the short term.That mightsound paradoxical for a management system so popularlyknown to abhor waste. But the paradox can be resolvedwhen we recognize that Toyota's managers and workersare trying to match each countermeasure to eachproblem.

There's no link between the reason for keeping safetystock-process unreliability-and the reason for keepingbuffer stock-fluctuations in customer demand.To poolthe two would make it hard to distinguish between theseparate activities and customer-supplier connectionsinvolved.The inventory would have many owners, andthe reasons for its use would become ambiguous.Pooling the inventory thus muddles both the ownershipand cause of the problems, making it difficult to introduceimprovements.

M

the production workers, answered to the variousshop heads. Neither organizational structure is in-herently superior. Rather, the people we inter-viewed explained, prohlems in the first divisionhappened to create a situation that required theengineers to learn from one another and to pool en-gineering resources. By contrast, the prohlems thatarose in the second division required the produc-tion and engineering people to cooperate at the levelof the individual shops. Thus the organizationaldifferences reflect the fact that the two divisionsencountered different prohlems.

Toyota's Notion of the IdealBy inculcating the scientific method at all levels ofthe workforce, Toyota ensures that people willclearly state the expectations they will be testingwhen they implement the changes they haveplarmed. But beyond this, we found that people incomipanics following the Toyota Production Sys-tem share a common goal. They have a commonsense of what the ideal production system wouldbe, and that shared vision motivates them to makeimprovements beyond what would be necessarymerely to meet the current needs of their customers.This notion of the ideal is very pervasive, and webelieve it is essential to understanding the ToyotaProduction System.

When they speak of the ideal, workers at Toyotado not mean something philosophically abstract.They have a concrete definition in mind, one that isremarkahly consistent throughout the company.Very specifically, for Toyota's workers, the outputof an ideal person, group of people, or machine:

• is defect free (that it, it has the features and perfor-mance the customer expects);

• can he delivered one request at a time (a batch sizeof one);

• can be supplied on demand in the version requested;• can be delivered immediately;• can be produced without wasting any materials,

labor, energy, or other resources (such as costs as-sociated with inventory); and

• can he produced in a work environment that issafe physically, emotionally, and professionallyfor every employee.

We consistently found people at plants that usedthe Toyota Production System making changes thatpushed operations toward this ideal. At one com-pany that produced electromechanical products, forexample, we found that workers had come up witha number of ingenious error-detecting gauges that


DECODING THE DNA OF THE TOYOTA PRODUCTION SYSTI M

generated a simple, unambiguous yes-or-no signalto indicate whether their output was free of defects -as specified in the ideal. At yet another plant, whichmanufactures injection-molded parts, we found thatworkers had reduced the time it took to change alarge molding die from an already speedy five min-utes to three minutes. This allowed the company toreduce the hatch sizes of each part it produced by40%, bringing it closer to the ideal batch size of one.As Toyota moves toward the ideal, it may temporar-ily hold one of its dimensions to be more importantthan another. Sometimes this can result in practicesthat go against the popular view of Toyota's opera-tions. We have seen cases where Toyota keeps higherlevels of inventory or produces in batch sizes largerthan observers generally expect of a just-in-time op-eration, as we describe in the sidebar "Countermea-sures in the Toyota Production System."

Toyota's ideal state shares many features of thepopular notion of mass customization-the ahilityto create virtually infinite variations of a productas efficiently as possible and at the lowest possihlecost. In the final analysis, Toyota's ideal plantwould indeed be one where a Toyota customer

'...and now, please stay tuned for America'sScariest Business Mergers."

could drive up to a shipping dock, ask for a cus-tomized product or service, and get it at onceat the lowest possible price and with no defects. Tothe extent that a Toyota plant -or a Toyota worker'sactivity-falls short of this ideal, that shortcomingis a source of creative tension for further improve-ment efforts.

The Organizational Impact of the RulesIf the rules make companies using the Toyota Pro-duction System a community of scientists perform-ing continual experiments, then why aren't theseorganizations in a state of chaos? Why can one per-son make a change without adversely affecting thework of other people on the production line? Howcan Toyota constantly introduce changes to its op-erations while keeping them running at full tilt? Inother words, how does Toyota improve and remainstable at the same time?

Once again, the answer is in the rules. By makingpeople capable of and responsible for doing and im-proving their own work, by standardizing connec-tions between individual customers and suppliers,

and by pushing the resolution of con-nection and flow problems to the lowestpossible level, the rules create an orga-nization with a nested modular struc-ture, rather like traditional Russiandolls that come one inside the other.The great henefit of nested, modularorganizations is that people can imple-ment design changes in one part with-out unduly affecting other parts. That'swhy managers at Toyota can delegate somuch rcsponsihility without creatingchaos. Other companies that follow therules will also find it possible to changewithout experiencing undue disruption.

Of course, the structures of othercompanies have features in commonwith those that follow the Toyota Pro-duction System, but in our research wefound no company that had them allthat did not follow the system. It mayturn out in the end that you can buildthe structure only by investing the timeToyota has. But we believe that if a com-pany dedicates itself to mastering therules, it has a better chance of replicat-ing Toyota's DNA- and with that, itsperformance. ^

Reprint 99509

To order reprints, see the last page of this issue.


the · 2019-07-19 · decoding the dna of the toyota production system let's look at how...

Documents