You are on page 1of 12


discussions, stats, and author profiles for this publication at:

Learning to Lead at Toyota

Article in Harvard business review June 2004

Source: PubMed


132 1,464

1 author:

Steven J Spear
Massachusetts Institute of Technology


All content following this page was uploaded by Steven J Spear on 25 February 2017.

The user has requested enhancement of the downloaded file.

Toyotas famous production

system makes great carsand
with them great managers.
Learning to Lead at
Heres how one American
hotshot learned to replicate
Toyotas DNA.
by Steven J. Spear

Reprint R0405E
Toyotas famous production system makes great carsand with them
great managers. Heres how one American hotshot learned to replicate
Toyotas DNA.

Learning to Lead at
by Steven J. Spear

Toyota is one of the worlds most storied com- otas tools and tacticskanban pull systems,
panies, drawing the attention of journalists, cords, production cells, and the likeand not
researchers, and executives seeking to bench- on its basic set of operating principles. In our

mark its famous production system. For good article, we identied four such principles, or
reason: Toyota has repeatedly outperformed rules, which together ensure that regular work
its competitors in quality, reliability, produc- is tightly coupled with learning how to do the
tivity, cost reduction, sales and market share work better. These principles lead to ongoing
growth, and market capitalization. By the end improvements in reliability, exibility, safety,
of last year it was on the verge of replacing and efciency, and, hence, market share and
DaimlerChrysler as the third-largest North protability.
American car company in terms of produc- As we explained in the article, Toyotas real
tion, not just sales. In terms of global market achievement is not merely the creation and
share, it has recently overtaken Ford to be- use of the tools themselves; it is in making all
come the second-largest carmaker. Its net in- its work a series of nested, ongoing experi-
come and market capitalization by the end of ments, be the work as routine as installing
2003 exceeded those of all its competitors. But seats in cars or as complex, idiosyncratic, and
those very achievements beg a question: If large scale as designing and launching a new
Toyota has been so widely studied and copied, model or factory. We argued that Toyotas
why have so few companies been able to much-noted commitment to standardization is
match its performance? not for the purpose of control or even for cap-
In our 1999 HBR article, Decoding the turing a best practice, per se. Rather, standard-
DNA of the Toyota Production System, H. izationor more precisely, the explicit speci-
Kent Bowen and I argued that part of the prob- cation of how work is going to be done before it
lem is that most outsiders have focused on Toy- is performedis coupled with testing work as it

harvard business review may 2004 page 1

Learning to Lead at Toyota

is being done. The end result is that gaps be- would yield continuous learning and improve-
tween what is expected and what actually oc- ments, and to teach others to do the same.
curs become immediately evident. Not only
are problems contained, prevented from prop- The Program
agating and compromising someone elses Dallis arrived at Toyotas Kentucky headquar-
work, but the gaps between expectations and ters early one wintry morning in January
reality are investigated; a deeper understand- 2002. He was greeted by Mike Takahashi (not
ing of the product, process, and people is his real name), a senior manager of the Toyota
gained; and that understanding is incorporated Supplier Support Center (TSSC), a group re-
into a new specication, which becomes a tem- sponsible for developing Toyotas and supplier
porary best practice until a new problem is plants competency in TPS. As such, Takahashi
discovered. (See the sidebar The Power of was responsible for Dalliss orientation into
Principles.) the company. Once the introductory formali-
It is one thing to realize that the Toyota Pro- ties had been completed, Takahashi ushered
duction System (TPS) is a system of nested ex- Dallis to his car and proceeded to drive not to
periments through which operations are con- the plant where Dallis was to eventually work
stantly improved. It is another to have an but to another Toyota engine plant where Dal-
organization in which employees and manag- lis would begin his integration into the com-
ers at all levels in all functions are able to live pany. That integration was to involve 12 inten-
those principles and teach others to apply sive weeks in the U.S. engine plant and ten
them. Decoding the DNA of Toyota doesnt days working and making observations in Toy-
mean that you can replicate it. ota and Toyota supplier plants in Japan. The
So how exactly does a company replicate it? content of Dalliss trainingas with that of
In the following pages, I try to answer that any other Toyota managerwould depend on
question by describing how a talented young what, in Takahashis judgment, Dallis most
American, hired for an upper-level position at needed.
one of Toyotas U.S. plants, was initiated into Back to Basics. Bob Dalliss rst assignment
the TPS. His training was hardly what he at the U.S. engine plant was to help a small
might have expected given his achievements. group of 19 engine-assembly workers improve
With several degrees from top-tier universi- labor productivity, operational availability of
ties, he had already managed large plants for machines and equipment, and ergonomic
one of Toyotas North American competitors. safety.1 For the rst six weeks, Takahashi en-
But rather than undergo a brief period of cur- gaged Dallis in cycles of observing and chang-
sory walk-throughs, orientations, and introduc- ing individuals work processes, thereby focus-
tions that an incoming fast-track executive ing on productivity and safety. Working with
might expect, he learned TPS the long, hard the groups leaders, team leaders, and team
wayby practicing it, which is how Toyota members, Dallis would document, for in-
trains any new employee regardless of rank or stance, how different tasks were carried out,
function. It would take more than three who did what tasks under what circumstances,
months before he even arrived at the plant in and how information, material, and services
which he was to be a manager. were communicated. He would make changes
Our American hotshot, whom well call Bob to try to solve the problems he had observed
Dallis, arrived at the company thinking that he and then evaluate those changes.
already knew the basics of TPShaving bor- Dallis was not left to his own devices, de-
rowed ideas from Toyota to improve opera- spite his previous experience and accomplish-
tions in his previous joband would simply be ments. Meetings with Takahashi bracketed his
ne-tuning his knowledge to improve opera- workweek. On Mondays, Dallis would explain
Steven J. Spear ( is an tions at his new assignment. He came out of the following: how he thought the assembly
assistant professor at Harvard Business his training realizing that improving actual op- process worked, based on his previous weeks
School in Boston. He is the author, with erations was not his jobit was the job of the observations and experiences; what he thought
H. Kent Bowen, of Decoding the DNA workers themselves. His role was to help them the lines problems were; what changes he and
of the Toyota Production System, understand that responsibility and enable the others had implemented or had in mind to
which was published in the September them to carry it out. His training taught him solve those problems; and the expected impact
October 1999 issue of HBR. how to construct work as experiments, which of his recommendations. On Fridays, Taka-

harvard business review may 2004 page 2

Learning to Lead at Toyota

hashi reviewed what Dallis had done, compar- proved productivity and ergonomics made the
ing actual outcomes with the plans and expec- machines malfunction more often. Rather, be-
tations they had discussed on Monday. fore the changes were made, there was enough
In the rst six weeks, 25 changes were im- slack in the work so that if a machine faulted,
plemented to individual tasks. For instance, a there was often no consequence or inconve-
number of parts racks were recongured to nience to anyone. But with Dalliss changes,
present materials to the operators more com- the group was able to use 15 people instead of
fortably, and a handle on a machine was repo- 19 to accomplish the same amount of work. It
sitioned to reduce wrist strain and improve er- was also able to reduce the time required for
gonomic safety. Dallis and the rest of the group each task and improve workload balance. With
also made 75 recommendations for redistribut- a much tighter system, previously inconse-
ing their work. These were more substantial quential machine problems now had signi-
changes that required a reconguration of the cant effects.
work area. For instance, changing the place After Dallis had improved the human tasks
where a particular part was installed required in the assembly line, Takahashi had him switch
relocating material stores and moving the light to studying how the machines worked. This
curtains, along with their attendant wiring and took another six weeks, with Takahashi and
computer coding. These changes were made Dallis again meeting on Mondays and Fridays.
with the help of technical specialists from the Takahashi had Dallis, holder of two masters
maintenance and engineering departments degrees in engineering, watch individual ma-
while the plant was closed over the weekend, chines until they faulted so that he could inves-
after Dalliss fth week. tigate causes directly. This took some time. Al-
Dallis and Takahashi spent Dalliss sixth though work-method failures occurred nearly
week studying the groups assembly line to see twice a minute, machine failures were far less
if the 75 changes actually had the desired ef- frequent and were often hidden inside the ma-
fects. They discovered that worker productivity chine.
and ergonomic safety had improved signi- But as Dallis observed the machines and the
cantly, as shown in the exhibit The U.S. En- people working around them, he began to see
gine Plant Assembly LineBefore and After. that a number of failures seemed to be caused
Unfortunately, the changes had also reduced by peoples interactions with the machines. For
the operational availability of the machines. instance, Dallis noticed that as one worker
This is not to say that the changes that im- loaded gears in a jig that he then put into the
machine, he would often inadvertently trip the
trigger switch before the jig was fully aligned,
causing the apparatus to fault. To solve that
problem, Dallis had the maintenance depart-
The Power of Principles ment relocate the switch. Dallis also observed
The insight that Toyota applies underly- discrete-parts fabrication and assembly another operator push a pallet into a machine.
ing principles rather than specic tools operations, has based its Alcoa Business After investigating several mechanical failures,
and processes explains why the com- System (ABS) on the TPS rules. Alcoa he realized that the pallet sometimes rode up
pany continues to outperform its com- claims that ABS saved the company $1.1 onto a bumper in the machine. By replacing
petitors. Many companies have tried to billion from 1998 to 2000, while improv- the machines bumper with one that had a dif-
imitate Toyotas tools as opposed to its ing safety, productivity, and quality. ferent cross-section prole, he was able to elim-
principles; as a result, many have ended In another example, pilot projects ap- inate this particular cause of failure. Direct ob-
up with rigid, inexible production sys- plying the rules at the University of servation of the devices, root-cause analysis of
tems that worked well in the short term Pittsburgh Medical Center and other each fault, and immediate reconguration to
but didnt stand the test of time. health care organizations have led to remove suspected causes raised operational
Recognizing that TPS is about apply- huge improvements in medication ad- availability to 90%, a substantial improvement
ing principles rather than tools enables ministration, nursing, and other critical though still below the 95% target that Taka-
companies that in no way resemble processes, delivering better quality care hashi had set for Dallis.
Toyota to tap into its sources of success. to patients, relieving workers of nonpro- The Master Class. After 12 weeks at the U.S.
Alcoa, a company whose large-scale pro- ductive burdens, as well as providing engine plant, Takahashi judged that Dallis had
cessesrening, smelting, and so on costs savings and operating efciencies. made progress in observing people and ma-
bear little resemblance to Toyotas chines and in structuring countermeasures as

harvard business review may 2004 page 3

Learning to Lead at Toyota

experiments to be tested. However, Takahashi peded the worker and lengthened cycle times.
was concerned that Dallis still took too much of Dalliss workmate could not speak English, and
the burden on himself for making changes and no translator was provided, so the two had to
that the rate at which he was able to test and learn to communicate through the physical en-
rene improvements was too slow. He decided vironment and through models, drawings, and
it was time to show Dallis how Toyota practiced role-playing. Afterward, Dallis speculated that
improvements on its home turf. He and Dallis the logic of starting with overburden was to
ew to Japan, and Dalliss rst three days there get buy-in from the worker who was being
were spent working at Toyotas famous Kamigo asked to do his regular job while being inter-
engine plantwhere Taiichi Ohno, one of the rupted by a non-Japanese-speaking stranger.
main architects of TPS, had developed many of There is also semantic signicance in the
his major innovations. On the morning of their phrasing: Focusing on overburden empha-
arrival, Takahashi unleashed the rst of several sizes the impact of the work design on the per-
surprises: Dallis was to work alongside an em- son. By contrast, focusing on waste suggests
ployee in a production cell and was to make 50 that the person is the problem.
improvementsactual changes in how work Dallis applied the approach he had learned
was doneduring his time there. This worked at the U.S. engine plant. On day one, he spent
out to be one change every 22 minutes, not the the rst three hours observing his new work-
one per day he had been averaging in his rst mate, and by the shifts end proudly reported
ve weeks of training. that he had seven ideas, four of which he and
The initial objective set for Dallis was to re- his workmate had implemented. Then Taka-
duce the overburden on the workerwalk- hashi unleashed his next surprise: He told Dal-
ing, reaching, and other efforts that didnt add lis that two Japanese team leaders who were
value to the product and tired or otherwise im- going through the same trainingpeople with
jobs far less senior than the one for which Dal-
lis was being preparedhad generated 28 and
31 change ideas, respectively, within the same
The U.S. Engine Plant Assembly Line amount of time. Somewhat humbled, Dallis
picked up the pace, looking for more opportu-
Before and After nities to make improvements and trying even
The following table describes the im- safety (eliminating four processes and more quick and dirty methods of testing
pact of the changes Dallis made to the improving the rest). But machine avail- ideas: bolting rather than welding things, tap-
U.S. engine plant assembly line during ability actually decreased during the ing rather than bolting, and holding rather
his rst six weeks there. He made sub- period from 90% to 80%. In Dalliss sec- than tapinganything to speed up the rate of
stantial improvements in productiv- ond six weeks, he and his team were feedback. By 11 am on the second day, he and
ityreducing the number of workers able to restore availability back to 90%, his coworker had built the list to 25 ideas. Taka-
and cycle times. He and the group also but this was still below the 95% target. hashi would visit the machine shop while they
made signicant improvements in were working, ask what Dallis was concentrat-
ing on, and then follow up with very specic
Before After queries about the change idea. Before I could
Productivity give a speculative answer, recalled Dallis, he
Number of operators 19 15
sent me to look or try for myself.
Dallis found that his ability to identify and
Copyright 2004 Harvard Business School
Publishing Corporation. All rights reserved.

Cycle time 34 seconds 33 seconds

resolve problems grew with practice, and by
Total work time/engine 661 seconds 495 seconds
the morning of the third day, he had moved
Ergonomics* from examining the details of individual work
Red processes 7 1 routines to looking at problems with how the
Yellow processes 2 2 production cell as a whole was laid out and
the effects on workers physical movements:
Green processes 10 12
There were two machines, with gauges and
Operational availability 90% 80% parts racks. A tool change took eight steps on
one and 24 on the other. Was there a better
* Processes were rated from worst (red) to best (green) on the basis of their layout that would reduce the number of steps
ergonomicsa formula that took into account weight lifted, reaching, twist-
ing, and other risk factors. and time? We gured out how to simulate the

harvard business review may 2004 page 4

Learning to Lead at Toyota

change before getting involved with heavy ma- each with three to seven membersmanaged
chinery to move the equipment for real, Dallis and presented their improvement projects. In
said. By the time the three days were up, he one case, a group leader was exploring ways
had identied 50 problems with quality of reducing machine changeover times and
checks, tool changes, and other work in his ma- establishing a more even production pace for
chine shop35 of which had been xed on the an injection-molding process. In another, a
spot. (The effects of these changes are summa- group leader was looking for ways to reduce
rized in the exhibit The Kamigo Report downtime in a machining operation. In all
Card.) the presentations, the group leaders ex-
Takahashi had Dallis conclude his shop- plained the problems they were addressing,
oor training by presenting his work to the the processes they used to develop counter-
plant manager, the machine shop manager, measures, and the effect these countermea-
and the shops group leaders. Along the way, sures had on performance. Dallis quickly real-
Dallis had been keeping a careful log of the ized that people at all levels, even those
changes and their effects. The log listed oper- subordinate to the one for which he was
ations in the shop, the individual problems being developed, were expected to structure
Dallis had observed, the countermeasure for work and improvements as experiments.
each problem, the effect of the change, and
the rst- and second-shift workers reactions Lessons Learned
to the countermeasure. (For a snapshot of the Although Takahashi at no point told Dallis
log, see the exhibit Excerpts from Dalliss exactly what he was supposed to learn from
Log.) Photographs and diagrams comple- his experience, the methodology of the train-
mented the descriptions. During the presen- ing just described is so consistent and specic
tations, Dallis reported, the plants general that it reveals at least four fundamental prin-
manager, the machine shops manager, and ciples underlying the system. Together with
its group leaders were engaged in what [I and the rules we described in our 1999 article, the
the other] lowly team leaders said. Two- following lessons may help explain why Toy-
thirds [of the audience] actively took notes ota has remained the worlds preeminent
during the team leaders presentations, asking manufacturer.
pointed questions throughout.
After Dallis made his presentation, Taka- Lesson 1
hashi spent the remaining week showing him Theres no substitute for direct observation.
how Toyota group leaderspeople responsi- Throughout Dalliss training, he was required
ble for a few assembly or machining teams, to watch employees work and machines oper-

The Kamigo Report Card

During his three days at Kamigos machining shop, Dallis documented the effects of the 50 changes he made to work motion (the physical
movements of assembly-line workers) and cell layout. The changes are categorized according to the nature of the activitywalking, reaching,
or other movements. They cut about half a mile of walking per shift per operator in addition to reducing ergonomic and safety hazards.

Quality checks* Tool changes* Other work

Copyright 2004 Harvard Business School
Publishing Corporation. All rights reserved.

Walking Reaching Other Walking Reaching Other

Number 8 8 13 7 4 5 5
of changes

Effect 20-meter 2-meter Elimination of 50-meter 180-cm Improvement Elimination of

of changes reduction reduction tripping risk, reduction reduction of ergonomics, tripping risk,
(50%) in reaching organization of per tool in organization simplication
per check tools to reduce change reaching to reduce risk of oil change
risk of confusion of confusion

* Quality checks were performed two to three times an hour, and tool changes were made once an hour.

harvard business review may 2004 page 5

Learning to Lead at Toyota

ate. He was asked not to gure out why a such precision. Dalliss rst six weeks at the
machine had failed, as if he were a detective U.S. engine plant meant that he had up to
solving a crime already committed, but to sit 23,824 opportunities to observe complete
and wait until he could directly observe its fail- work cycles. Because his work was limited to
ureto wait for it to tell him what he needed a 19-person line, he could view more than a
to know. thousand work cycles per person. That gave
One of the group leader presentations at Ka- him deep insight into the lines productivity
migo described this principle in action. In a and safety.
project to improve machine maintenance, it
became clear to the group that machine prob- Lesson 2
lems were evident only when failures occurred. Proposed changes should always be struc-
In response, the shops group leaders had re- tured as experiments.
moved opaque covers from several machines In the scientic method, experiments are
so that operators and team leaders could hear used to test a hypothesis, and the results are
and see the inner workings of the devices, thus used to rene or reject the hypothesis. Dal-
improving their ability to assess and anticipate liss problem solving was structured so that
problems with the machines. This is a very dif- he embedded explicit and testable assump-
ferent approach from the indirect observation tions in his analysis of the work. Throughout
on which most companies relyreports, inter- his training, therefore, he had to explain
views, surveys, narratives, aggregate data, and gaps between predicted and actual results.
statistics. Not that these indirect approaches In his meetings with Takahashi at the U.S.
are wrong or useless. They have their own engine plant, for example, he was required
value, and there may be a loss of perspective to propose hypotheses on Monday and the
(the big picture) when one relies solely on di- results of his experiments on Friday. In Ja-
rect observation. But direct observation is es- pan, he had to present his changes as tests of
sential, and no combination of indirect meth- causal relationships, stating the problem he
ods, however clever, can possibly take its place. saw, the root cause he suspected, the change
Dalliss previous experience managing he had made, and the countermeasures ac-
plants might have prepared him to look at op- tual effect on performance.
erations of greater scale and scope, but had Ta- Of course, many people trying to improve a
kahashi given him a project with greater scope, process have some idea of what the problems
Dallas might not have learned to observe with are and how to x them. The difference with

Excerpts from Dalliss Log

Throughout his training, Dallis kept a precise log of identied problems, proposed solutions, expected results, and actual outcomes. Records like
the one below are essential to the Toyota Production System, as they help encourage the precision that is necessary for true experimentation.
The following excerpt shows two of the problems Dallis identied. Note that he obtained approval of his changes from the people actually doing
the work. Thats because at the end of the day, the people doing the work must own the solution. This kind of hierarchical inversion is a common
feature of Toyota operations.

Problem # Location Description Countermeasure Result Date Shift 1 Shift 2

approval approval
Copyright 2004 Harvard Business School
Publishing Corporation. All rights reserved.

4 Station 6R Team member walks Move rst-piece 4-meter May 8 Yes Yes
4 meters to get and check gauge from reduction
then return rst-piece table to shelf in walk/tool
check gauge during between stations change
tool changes 5 and 6

58 Part Team member walks Remove light pole Reduce Not done Yes Yes
gauging 5 steps to return cams (obstruction) and walk
area to return chute, walking move part gauge 2 steps (Pending help from
around light pole 45 maintenance department)

harvard business review may 2004 page 6

Learning to Lead at Toyota

TPSand this is keyis that it seeks to fully ments increased gradually. When Dallis started
understand both the problem and the solution. at the U.S. engine plant, he conducted single
For example, any manager might say, Maybe factor experiments, changing small, individ-
the parts rack should be closer to the assem- ual work elements rather than taking a system
blers hand. If we move it here, Ill bet itll perspective. Whats more, his efforts there
shave a few seconds off the cycle. Were he to started with individual work methods, pro-
try this and nd that it saved six seconds, he gressing to more complex and subtle machine
would probably be quite pleased and consider problems only when he had developed his ob-
the problem solved. servation and problem-solving skills over the
But in the eyes of a Toyota manager like Ta- six weeks. Thus, he moved from problems that
kahashi, such a result would indicate that the were easier to observe to those that were
manager didnt fully understand the work that harder. If each learning cycle is kept small and
he was trying to improve. Why hadnt he been bounded, then the learner can make mistakes
more specic about how far he was going to and the consequences will not be severe. This
move the rack? And how many seconds did he approach increases the learners willingness to
expect to save? Four? If the actual savings is six take risks and learn by doing. Dalliss training
seconds, thats cause for celebrationbut also at Kamigo mirrored this progression: He be-
for additional inquiry. Why was there a two- gan, once again, with work-method issues of
second difference? With the explicit precision overburden before moving on to machines.2
encouraged by Takahashi, the discrepancy
would prompt a deeper investigation into how Lesson 4
a process worked and, perhaps more impor- Managers should coach, not x.
tant, how a particular person studied and im- Dalliss training not only gave him insight into
Dallis was asked not to proved the process. how Toyota delivers continuous improvement
but also helped him understand the unique re-
figure out why a machine Lesson 3 lationships between Toyotas managers and
Workers and managers should experiment as workers. Dallis himself had been rewarded by
had failed but to sit and frequently as possible. his previous employer for being a problem
wait until he could At Toyota, the focus is on many quick, simple ex- solver, albeit one with a more participative
periments rather than on a few lengthy, complex and inclusive approach than most. What he
directly observe its ones. This became particularly evident when saw at Toyota, by contrast, was workers and
failureto wait for it to Dallis went to Japan. Whereas in the United low-level managers constantly solving prob-
States he made 25 changes in six weeks (before lems. Indeed, the more senior the manager,
tell him what he needed the weekend blitz during which 75 were com- the less likely he was to be solving problems
pleted), in Japan he had to make 50 changes in 2 himself.
to know.
shifts, which meant an average of one change Toyota managers act as enablers. Through-
every 22 minutes. This encouraged Dallis to out Dalliss training, Takahashione of Toy-
learn from making small incremental changes otas most senior operational managersposi-
rather than large system-design changes. He tioned himself as a teacher and coach, not as a
would observe work actually being done, technological specialist. He put Dallis through
quickly see where struggles were occurring, then experiences without explicitly stating what or
rapidly test his understanding by implementing how he was to learn. Even when specic skills
a countermeasure, thereby accelerating the rate were imparted, these were purely to assist Dal-
at which he discovered contingencies or inter- liss observation and experimentation. For in-
ferences in the process. This is precisely the way stance, Takahashi showed Dallis how to ob-
Toyota workers practice process improvement. serve an individual worker in order to spot
They cannot practice making a change, be- instances of stress, wasted effort, and so on,
cause a change can be made only once. But they and he explicitly advised Dallis on how to de-
can practice the process of observing and testing velop prototypes. But at no point did he sug-
many times. gest actual process improvements. Rather, he
To ensure that Dallis received the practice directed Dallis on how to nd opportunities
he needed and that he internalized his under- for those improvements (as in, study this per-
standing of it, Takahashi structured Dalliss son or that machine, looking for various types
training so that the complexity of his experi- of stress, strain, or faults) and on how to de-

harvard business review may 2004 page 7

Learning to Lead at Toyota

velop and test possible countermeasures. With Takahashis help, Dallis worked with
Takahashi also gave Dallis the resources he the lines group leader and assistant manager
needed to act quickly. For example, at Kamigo, in order to develop the problem-solving skills
Dallis had the help of a maintenance worker to of the lines team members and team leaders.
move equipment, create xtures, relocate The point was for the team to learn to solve lit-
wires and pipes, and provide other skilled trade tle problems simultaneously so that the line
work so that he could test as many ideas as could recover quickly when problems oc-
possible. Takahashi and the shop manager also curred. For instance, the team realized that it
came to the cell of the machining operation to had difculties in keeping track of what work
review Dalliss ideas; they gave him tips on pi- needed to be done and in identifying problems
loting his changes before asking support work- as they occurred. It therefore had to improve
ers to make parts or relocate equipment. When its visual management of the workwhat
Dallis wanted to rotate some gauges that was going well, what was going wrong, and
tested parts, the shop manager showed him what needed to be done. Dallis sat down with
how to quickly and inexpensively make card- the group leader and assistant manager and set
board prototypes to test location, orientation, out a schedule for identifying specic prob-
size, and so on. lems and allocating responsibility for them
The result of this unusual managerworker across the team. As the team members ob-
relationship is a high degree of sophisticated served and developed countermeasures, Dallis
problem solving at all levels of the organiza- would drop by much as Takahashi had done,
tion. Dallis noted, As a former engine-plant asking them specic questions that would
person, I saw a line [at Kamigo] that was 15 oblige them to observe their allotted problems
years old but that had the capacity to build 90 more closely as they happened. To its delight,
I saw a line [at Kamigo] different engine types. It was amazing that the group hit its mark ahead of schedule and
they solved so many problems with such sim- raised operational availability to 99%.
that was 15 years old but ple equipment. Behind the changes was some Dallis had returned to America with an al-
pretty deep thinking. The basic company phi- tered focus. He had realized from the way Ta-
that had the capacity to losophy is that any operating system can be im- kahashi had managed his training, and from
build 90 different engine proved if enough people at every level are what hed seen of others training, that the ef-
looking and experimenting closely enough. forts of a senior manager like himself should
types. (After all, if only the big shots were expected to be aimed not at making direct improvements
make changes, all that little stuff would get but at producing a cadre of excellent group
overlooked.) The fact that Dallis, after just leaders who learn through continuous experi-
three months at the U.S. engine plant, was able mentation. The target of 95% operational avail-
to empower others to implement 50 improve- ability at the U.S. engine plant was the same,
ments at Kamigo, one of Toyotas top plants, of- but he now knew whose target it really was,
fers insight into why Toyota stays ahead of its and it wasnt his. At this point, Takahashi -
competitors.3 nally released Dallis from his training to take
on his full-time managerial responsibilities.
Back to America
To see if Dallis had learned the right lessons For anyone trying to understand how the
from his training, Takahashi sent him back to Toyota Production System really works, there
the U.S. engine plant where his instruction is probably no substitute for the kind of total
had begun. As we have seen, Dallis had al- immersion that Dallis received. TPS is a sys-
ready helped make substantial improvements tem you have to live to fully understand, let
in the assembly lines labor productivity and alone improve. Besides, anyone like Dallis
ergonomic safety before going to Japan. But coming into Toyota from the outside, regard-
he hadnt been able to raise operational avail- less of his or her experience, is coming into an
ability to 95%. Now, upon Dalliss return to organization with a long history of making
that plant, Takahashi had him attempt this improvements and modications at a pace
goal again. However, there was a marked de- few organizations have ever approached. No
parture from Dalliss earlier approach, in one can expect to assimilatelet alone recre-
which he primarily saw himself as a problem atesuch a strong and distinct culture in just
solver. a few weeks or even a few months. Neverthe-

harvard business review may 2004 page 8

Learning to Lead at Toyota

less, any company that develops and imple- who used it to teach Dallis. He directly observed Dalliss
work by creating short learning cycles with rapid feedback
ments a training program such as the one
so that he could continually reassess Dalliss knowledge and
Dallis participated in is sure to reap enor- skills, both to provide feedback in order to help him learn
mous dividends. The organization that ap- and to design the next learning increment.
plies the rules in designing its operations and
3. According to Takahashi, the expectation was that group
that trains its managers to apply those rules leaders at Kamigomanagers who supervised several oper-
will have made a good start at replicating the ating shops or cellswould spend 70% of their time doing
process improvement work. This time would often be
DNA of the Toyota Production System.
shared among three to four teams, implying that team lead-
erspeople managing one shop or cellwere expected to
spend a minimum of 20% of their time on improvement
1. Operational availability equals machine run time/ma- work.
chine use time. For instance, if a machine requires eight
minutes of process time to grind a surface, but, because of
jams and other interruptions, ten minutes are actually Reprint R0405E
spent from start to nish, then operational availability
To order, see the next page
would be 80%. Ideally, operationally availability would be
100%that is, the machine always runs when it is needed. or call 800-988-0886 or 617-783-7500
or go to
2. The incremental approach was also helpful to Takahashi,

harvard business review may 2004 page 9

Further Reading
The Harvard Business Review
Paperback Series

Harvard Business Review OnPoint Here are the landmark ideasboth

articles enhance the full-text article contemporary and classicthat have
with a summary of its key points and established Harvard Business Review as required
a selection of its company examples reading for businesspeople around the globe.
to help you quickly absorb and apply Each paperback includes eight of the leading
the concepts. Harvard Business articles on a particular business topic. The
Review OnPoint collections include series includes over thirty titles, including the
three OnPoint articles and an following best-sellers:
overview comparing the various
perspectives on a specific topic. Harvard Business Review on Brand
Product no. 1445

Harvard Business Review on Change

Product no. 8842

Harvard Business Review on Leadership

Product no. 8834

Harvard Business Review on Managing

Product no. 9075

Harvard Business Review on Measuring

Corporate Performance
Product no. 8826

For a complete list of the Harvard Business

Review paperback series, go to

To Order

For reprints, Harvard Business Review

OnPoint orders, and subscriptions
to Harvard Business Review:
Call 800-988-0886 or 617-783-7500.
Go to

For customized and quantity orders

of reprints and Harvard Business
Review OnPoint products:
Call Frank Tamoshunas at
or e-mail him at

page 10

View publication stats