SloanSelectLeadership Collection

Enabling Bold Visions

The Leader’s New Work: Building Learning Organizations

Are Your Subordinates Setting You Up To Fail?

Bridging Faultlines In Diverse Times

Leading at the Enterprise Level

How to Lead a Self-Managing Team

Enabling Bold Visions

W I N T E R 2 0 0 8 V O L . 4 9 N O. 2

R E P R I N T N U M B E R 4 9 2 0 2

Douglas A. Ready and Jay A. Conger

Please note that gray areas reflect artwork that has been intentionally removed. The substantive content of the article appears as originally published.

It was a beautiful, sunny day in Miami. Inside a darkened auditorium, however,

3,500 senior executives of one of the world’s largest financial companies, flown

in for just this purpose, sat awaiting their new CEO. The scene resembled the

beginning of a pro basketball game: Loud rock music set the beat, and laser im-

ages peppered the crowd and the stage. As the audience members stomped their

feet and swayed with the music, the CEO appeared on stage. His task? To pitch his

newly minted enterprise vision to the already energized crowd.

His speech, as well choreographed as the buildup, was brief and upbeat.

The message was simple: “We will become ‘the breakout firm.’ And to achieve

breakout status, we will rely on the three I’s of innovation, integration, and

inspiration.” Given the tight schedule of events for this one-day meeting, the

CEO took no questions. But the speech had caught the executives’ attention,

and the new vision created a buzz of excitement.

Months later, the buzz had worn off. Reality had intruded, as it always

does. The company, successful for more than a hundred years, had always

proceeded conservatively. Its organizational culture was anything but “break-

out.” It wasn’t innovative; its businesses, products and customers were not

integrated; and its leaders were not known for being inspirational.

In short, following their initial enthusiasm, the company’s senior execu-

tives were having a hard time reconciling the new vision with day-to-day

realities as they met in planning sessions with the next levels of management

and front-line employees. Within the top executive team, there were few role

models for the three I’s, and there were no signs of rewards for behaving as

breakout leaders.

Within a year, the term “breakout” had slipped quietly into the back-

ground. Few senior executives could even remember the three I’s. About two

years after the speech in Miami, the CEO declared victory and went on to lead

another company. But the confusion sowed by the unrealized vision cost the

company both market share and the harder to quantify elements of trust and

momentum with employees and customers.

This gap between inspiration and implementation is a common one. We set

out to find out why, and what CEOs and their top teams could do to translate

bold visions into operational realities. (See “About the Research,” p. 72.) In this

article, we’ll first explain several common reasons behind the derailment of

Enabling Bold Visions

Douglas A. Ready is visiting professor of organizational behavior at London Business School and president of the International Consortium for Executive Development Re-search, based in Lexington, Massachusetts. Jay A. Conger is the Henry Kravis Research Professor of Leadership Studies at Claremont McKenna College in Claremont, California. Professors Ready and Conger have collaborated on several articles, including “Why Leadership-Development Efforts Fail,” MIT Sloan Management Review, spring 2003. Comment on this article or contact the authors through smrfeedback@mit.edu.

A CEO’s new vision often

blurs into an indistinct

image once the initial

blitz is over. To ensure

that the vision is more

than just a daydream,

companies should follow

a five-phase model that

some organizations have

used successfully to avoid

disaster or complacency.

Douglas A. Ready

and Jay A. Conger

70 MIT SLOAN MANAGEMENT REVIEW WINTER 2008

L E A D E R S H I P

SLOANREVIEW.MIT.EDU

WINTER 2008 MIT SLOAN MANAGEMENT REVIEW 71SLOANREVIEW.MIT.EDU

bold visions. We will then offer a framework that executives can

use to ensure that their visions become more than just pipe

dreams. Examples of companies that have successfully followed

this path will both illustrate the challenges and provide guidance

to those who want to actively address them.

Why Many Bold Visions DerailNew visions can fade away for many apparently unique reasons.

However, as a result of a search of the literature on organiza-

tional change and our discussions with executives from 40

companies, we were able to categorize the reasons for failure into

a set of themes.1

Failing to Focus How many initiatives can any company handle at

once? In our experience, an organization can be confronted with

a dozen or more major initiatives at a time. If the top team

attempts to wrestle with all of them at once, middle managers

and front-line employees will likely be confused. The bold vision

will get mired in a haze of other priorities, with various initiative

champions roaming through the company like prophets crying

in the wilderness.

Sitting Out This Dance It’s awkward dancing without a partner. But

that’s essentially what too many leaders of the companies we

researched tried to do: They started the dance by explaining the

vision but didn’t engage a partner to make it work. They didn’t

communicate how others in the company could learn the steps to

the new dance. Moreover, employees could walk away from the

dance floor without fear, since there was no short-term price to

pay for not engaging to make the new vision an operating reality.

Skipping the Skill Building By its nature, a bold vision can’t simply

be retrofitted onto a company. (If it could, it would be a tame or

gentle new vision.) But many of the companies we examined did

not invest to develop the new skills their people would need to

realize the vision. Some leaders didn’t even want to hear that their

organizations lacked the capability to execute it. For example, the

CEO of the “breakout” financial services company didn’t want to

talk about the need to develop collaborative managers and leaders

— an obvious prerequisite for the integration plank of his vision.

Mismatching Messages and Metrics Much is made of people’s ten-

dency to talk the talk without walking the walk — and that

general problem applies in this instance, too. Many executives

told us that a bold new vision often bumped up against subtle

reinforcements of business as usual. For example, it would

become clear through the stories floating through the organiza-

tion that yesterday’s leadership behaviors were still seen to be of

paramount importance in spite of the need for changes to fit the

vision. Further, we learned that performance management pro-

cesses and organizational performance measures often remain

unchanged, even though newly articulated visions require new

behaviors and mindsets.

Clashing Powers In many cases, an organization’s political

dynamics and culture can become a major barrier to the success

of the new vision. Powerful groups are often tied to existing core

activities that made the company thrive in the past. If the new

vision threatens the supremacy of the old guard, political in-

fighting may stall it out and defeat the change agents.

Neglecting the Talent Pipeline According to the managers in our

study, the challenge of bringing about a new vision is, in

essence, a change management challenge. What’s needed to make

it work? Layers of change agents, from top to bottom — a critical

mass of people who embody the behaviors and values of the new

call to action. But many companies lack such pipelines of talent,

filled with people who are capable of this critical task.

These six problem areas explain why many bold visions fail

to produce measurable change, far less ignite their companies to

reach new heights. But a model of change can help companies

avoid this trap.

The Change ModelWhile our research uncovered reasons for the failure of bold

visions, it also led us to discern a pattern followed by compa-

nies that successfully translated a vision into action. We

observed five critical activities, performed in sequence, that

together form a systems approach to enabling visions. This

approach can be described as a framework for leading large-

scale, enterprisewide change initiatives. (See “The Five-Phase

Model for Enabling Visions,” p. 73.)

Phase I: Framing the Agenda If an organization’s top executives are

spending much of their time attending to competing “top priorities,”

they’ll gradually put the CEO’s vision on the back burner, where it

will slowly simmer down to nothing. The leaders in our study who

successfully brought about change framed their organizations’ chal-

lenges as compelling stories that created an urgent agenda for action.

They used the stories to set priorities among the many challenges

facing their organizations, with a laser focus on the need to drive

performance and to develop the right organizational climate. They

often cast the need to drive performance in present and future terms

— as a matter of finding pathways to future growth while maintain-

ing the edge that enabled the company to deliver value to customers

in the short term. In addressing the need to create an organizational

climate that was consistent with the company’s performance objec-

tives, these change leaders took into consideration both external

(brand, corporate social responsibility) and internal (culture, values,

behaviors) concerns. (See “The Enterprise Agenda,” p. 74.)

72 MIT SLOAN MANAGEMENT REVIEW WINTER 2008 SLOANREVIEW.MIT.EDU

Consider the approach to change taken by the British Broad-

casting Corp. Like many established media companies, the BBC

has been under assault by the proliferation of media options, a

problem that has intensified with the advent of inexpensive and

easy-to-use digital technology. Customer choice has fragmented,

and the network’s market share has steadily eroded. Gone are the

days when a popular sitcom could attract more than 28 million

viewers, as one did in the 1960s.

The company needed to change to meet these challenges, but

it had to navigate risks. It had to chart a course that would not

destroy its proud heritage or instigate a flight of the talent that

had always been the bedrock of its core competence.

When Mark Thompson was hired as the BBC’s new director

general a few years ago, he inherited an organization deeply en-

trenched in the analog world. Innovators elsewhere in the

industry were already moving with dispatch into the digital

world. But Thompson didn’t move immediately to bring about

big changes. He took some time to make sure he understood the

core competencies of the organization and how they matched up

against the future direction of the industry. He got to know the

cutting-edge elements of the future of the business as well as the

best way to frame the BBC’s new vision for success.

After a great deal of listening to industry experts, customers

and employees, Thompson framed the BBC’s new enterprise

agenda: Creative Future. His message was simple. In order to

drive performance, the BBC would invest heavily in digital tech-

nology and become leaner and smaller (by becoming better at

aligning programming with an improved customer segmentation

capability). And in order to create the climate suitable to achieve

these performance objectives, the BBC’s leaders and managers

would need to become more agile and collaborative. This fram-

ing of the agenda was clear: As the competitive landscape shifted,

the BBC would not sacrifice quality, but it would make tough

decisions about programming and head count.

Another storied company facing the imperative to change

recently was Deutsche Bank AG, based in Frankfurt, Germany.

In this case, globalization, not technology, was the primary force

at work. By the early 1990s, the 150-year-old retail and commer-

cial bank still had its roots firmly planted in Germany. Its top

executives and managers were virtually all German, and nearly

70% of its core business was conducted in Germany. But a stra-

tegic analysis of the company’s future prospects revealed the

likelihood of trouble in the near future: The bank’s growth po-

tential in both retail and commercial banking in Germany

amounted to roughly 2% per year.

In contrast, growth prospects in other areas were much more

robust: 10% per year in capital markets and investment banking,

5% in asset and wealth management, and unquantifiably huge

growth prospects in the Asia Pacific region and throughout the

rest of Europe. But to exploit these opportunities, the bank would

have to undergo two major transformations — from domestic to

global player and from a narrow focus to a broad one, as a fully

integrated financial services institution with new core competen-

cies in capital markets and investment banking.

Facing possible extinction if it continued on its present path,

Deutsche Bank, led by Josef Ackermann, responded with a bold new

vision: the One Bank initiative. Ackermann, chairman of the Deutsche

Bank Group’s executive committee, knew that One Bank wouldn’t

work if it was just a slogan. “We had to make sure that we were mov-

ing from a one-culture bank to a one-bank culture.” And today the

bank’s executives are quick to point out that One Bank is only a part

of the story of the company’s impressive transformation.

In framing the bank’s enterprise agenda, Ackermann launched

the Business Realignment Program in conjunction with the One

Bank initiative. This program ensured that the Group’s executives

and managers would know precisely what the company’s priori-

ties were and where the Group was heading over the next decade.

The program also placed the capital markets and investment

banking divisions at the core of the bank’s growth strategy while

stressing the importance of integration. Again, driving perfor-

mance and creating climate — within a construct of ruthless

prioritization — were the twin goals.

As Ackermann put it, “We must view ourselves as one bank and

one team with one shared goal. We intend to implement the idea of

‘One Bank-One Team’ as swiftly as possible. Utmost priority will be

given to removing the silo mentality.”2 To reinforce this dual em-

phasis, Ackermann approved new leadership standards and cultural

values that were directly linked to the One Bank initiative. With

these changes, Ackermann made it clear that performance and cli-

mate would go hand in hand at Deutsche Bank.

Sometimes companies find themselves in dire situations but

still need a leader to spell out, in detail, just how serious the

L E A D E R S H I P

To gain a better understanding of the challenges compa-

nies face in enabling bold visions, we conducted research

on approximately 40 global companies. We conducted sur-

veys within the companies and then came up with an

initial research protocol for the dialogue component of the

project. We then met with representatives from those com-

panies to engage in discussions — mostly in groups, but

sometimes individually — about how to move bold visions

from the initial phase of inspiration to a sustainable aspect

of company operations. Our goal: to determine a best-

practice model. Finally, we gathered in-depth detail from

people in several companies that had successfully turned

visions into reality.

About the Research

WINTER 2008 MIT SLOAN MANAGEMENT REVIEW 73SLOANREVIEW.MIT.EDU

problems are. This bad-news-first approach can be critical to the

framing of a bold vision that will get a company moving away

from the status quo as quickly as possible. In 1999, when Carlos

Ghosn took the helm at Nissan Motor Co. Ltd., he stressed sev-

eral dismal facts. Nissan had just lost its position as Japan’s

second-largest carmaker. Market share had, in fact, been declin-

ing for two and a half decades. Fewer than one Nissan model in

10 was profitable. Against this background, Ghosn told the com-

pany that the Nissan Revival plan was the company’s “last

chance.” Employees got the message.

Phase II: Engaging the Organization Once change leaders have

framed their agendas, they must do whatever they can to distrib-

ute “ownership” of that vision as broadly as possible. But this has

to be done the right way. We found that while managers and even

executives yearn for more leadership in their organizations, they

don’t want to be cast simply as followers. Instead, they want a

different style of leadership, one characterized by authentic col-

laboration and broad-based engagement.

This finding gets to the heart of why we emphasize that bold

visions must be enabled rather than merely executed. Visions that

are truly bold will create tension and anxiety within an organiza-

tion’s business-as-usual stakeholders. Great change leaders

understand this, and they don’t make those who have succeeded

under yesterday’s business model enemies of the new vi-

sion. They invite differing views and perspectives and

listen carefully to those who contribute legitimate argu-

ments concerning the tensions created by the newly

framed enterprise agenda.

Carlos Ghosn was particularly keen to solicit new

ideas and to get new voices to speak up. At the beginning

of his tenure at Nissan, he created nine cross-functional

teams and staffed them with middle managers in their

30s and 40s from different functions, divisions and coun-

tries. Each team had only 10 members. To keep discussions

fluid, a facilitator was assigned to each; to ensure that the

teams had authority and were not influenced by a single

function, each had two board-level sponsors.

Only three guidelines were set out for the teams’ work.

They had one goal: to make proposals to reduce costs and

develop the business. They had one rule: no sacred cows,

no constraints. And they had one deadline: three months

for final recommendations. In the end, the cross-func-

tional teams were a powerful vehicle for returning Nissan

to profitability. Critically, Ghosn had not simply issued

orders from the top about how to realize the new vision.

Instead, he had engaged the company’s managers to find

their own routes to success.

Mark Thompson’s strategy to engage the BBC’s stake-

holder communities with his Creative Future vision

started at the top. The director general engaged the BBC’s board

of directors in a series of dialogues that eventually produced a set

of recommendations for restructuring the BBC along four crite-

ria: creative, digital, simple and open. The board owned the

overarching direction of the restructuring. It left to Thompson

the task of engaging his managers and executives so that the re-

structuring would come to life in day-to-day operations.

Thompson followed through by bringing teams together from

different BBC groups — Vision, Journalism, Audio & Music and

Future Media and Technology. The task was to work through the

details of how the BBC would transform itself from a collection

of separate units to a unified organization that relied heavily on

cross-unit collaboration. Eventually this engagement process also

paved the way for Thompson to restructure the company into

three multimedia output groups and to merge the Future Media

and Technology team with the new Operations team. Collabora-

tion moved from being an articulated but unrealized value to an

ingrained way of working at the BBC.

Similarly, Deutsche Bank’s Ackermann launched a variety of

engagement initiatives to enable his vision. For senior-level lead-

ers, he led a series of dialogues called the Spokesman’s Challenge.

These were open discussions led by Ackermann and his top ex-

ecutive team that focused on listening to the concerns of his

enterprise leadership team. The outcome? A better understand-

Companies that successfully put new visions of the enterprise into

practice follow a process that takes the organization through five

phases. In today’s competitive landscape, this is an iterative pro-

cess: While companies may be able to stick with a single vision for

a few years, most will need to return to the process regularly.

The Five-Phase Model for Enabling Visions

Energizing theOrganization

Through the Powerof People

Framing theEnterprise

Agenda

EngagingMultiple Layers ofthe Organization

BuildingMission-Critical

Capabilities

Connecting theDots by Creating

Alignment

74 MIT SLOAN MANAGEMENT REVIEW WINTER 2008 SLOANREVIEW.MIT.EDU

ing of the tensions that might get in the way of making the One

Bank initiative come to life.

Ackermann also created the One Bank Task Force, a group

focused on engaging successive layers of management so that

they would also own the vision and become part of the solution.

In addition, he skillfully used Deutsche Bank’s Learning and

Leadership Development Operation to initiate programs that

were directly linked to implementing the One Bank vision.

While Deutsche Bank and the BBC faced compelling reasons to

change, their situations were not as dire as that of Mattel Inc. in

2000. In May of that year, Bob Eckert became CEO of an organiza-

tion that was in deep trouble. The company had fired its previous

CEO over the disastrous acquisition of an electronic games com-

pany. Morale was low and finger-pointing was high. Eckert set out

to engage the organization with a vision of greater collaboration

— a vision that would eventually be called One Mattel. In that

spirit, a team of line managers approached Eckert and asked if

they could craft a set of new values for the company conveying

collaboration. The team came back with four values that all built

on the word “play” — Play Fair, Play Together, Play with Passion

and Play to Grow. Since these values accurately captured the aspi-

rations of many employees, they were quickly and enthusiastically

adopted by the organization. Just as important to the success of

One Mattel: the engagement of the company’s line managers.

Phase III: Building Mission-Critical Capabilities While engagement

with the new vision is critical, it’s not enough. The development

of new capabilities is also necessary, as companies usually have

gaps in what they want to do and what they can do.

The process of building mission-critical capabilities is no

small challenge, however. We identified two significant problems

in the course of our research. First, there are the practical con-

cerns: How do you go about identifying which capabilities will

provide differentiated competitive advantage? Second, there are

political concerns: How do you identify and address capability

gaps without turning the process into an enterprisewide blame

game? Either or both of these concerns can become cancerous if

matched with a CEO with a short attention span who “just wants

to get on with it.” This underscores the importance of following

the phases sequentially. After all, it is difficult for a CEO to have

a serious conversation about building new capabilities without

first framing the enterprise agenda and engaging the organiza-

tion in an active dialogue.

Without proper leadership, the BBC’s Creative Future initia-

tive could have foundered during this phase. In order to realize

the full benefits of Thompson’s bold vision, the BBC required

new skills and capabilities and cross-unit collaboration. But the

organization was constructed of durable silos with a “craftsman-

ship culture” in each silo that guided career development decisions

of its managers and employees.

Addressing these challenges was not the work of a moment

— or of a single program. Thompson attacked the problems in a

variety of ways. He brought the BBC’s technologists together

from across the organization to create what he referred to as a

“powerhouse resource.” He brought in or developed the expertise

required to launch new multiplatform initiatives — that is, initia-

tives that transcended a single medium or product. He oversaw the

introduction of Stepping Stones, a program to create a better un-

derstanding of different parts of the Vision Studios and to enable

BBC producers to work more flexibly. He launched Hot Shoes, a

developmental program that encouraged hundreds of staff to try

out their expertise in other disciplines throughout the BBC, allow-

ing them to see what new skills they might need to span

boundaries or develop capabilities for new growth areas in the in-

dustry. Finally, he initiated the Journalism College Web site,

enabling more than 8,500 journalists to sharpen their skills in edi-

torial policy and in emerging fields of the craft. The sum of these

programs was a newly skilled base of employees who could work

across platforms and had a deep understanding of digital media.

Despite the differences in industry, Deutsche Bank’s capabil-

ity-building challenges were strikingly similar to those of the

BBC. Its organizational structure was also formed by silos, and

An enterprise agenda must take into account, on one axis,

both present and future performance — both the compa-

ny’s future growth strategy and its current customer value

proposition. This reflects the company’s focus on “doing.”

On the other axis, the agenda must reflect an organiza-

tion’s state of “being” — the way employees and other

insiders are affected by the climate of change, as well as

the way customers and other external stakeholders per-

ceive the company’s new vision.

The Enterprise Agenda

Creating Climate

“Being”

DrivingPerformance

“Doing”

Growth Strategy

Value Proposition

InternalExternal

L E A D E R S H I P

WINTER 2008 MIT SLOAN MANAGEMENT REVIEW 75SLOANREVIEW.MIT.EDU

its future depended upon how well Ackermann could break

them down.

But in Deutsche Bank’s case, more than just skill building was

needed. The bank also had to cut costs deeply in order to become

more operationally efficient. Ackermann shed some of the bank’s

portfolio of businesses and acquired and reorganized new busi-

nesses that would serve as growth platforms. He targeted markets

in other parts of Europe, the Americas and particularly Asia. He

made English the official language of Deutsche Bank and care-

fully moved executives and managers across business and

geographic boundaries to build a global mind-set and a collab-

orative culture. And he used the leadership and learning programs

not only as phase II engagement mechanisms but also as critical

capability-building tools for his managers.

Phase IV: Connecting the Dots By Creating Alignment Bold visions

rarely succumb to frontal attacks. The companies that collabo-

rated with us in our study mentioned repeatedly that bold visions

get derailed by more subtle means: by mind-sets, systems and

processes that are out of sync or, ironically, by the unintentional

messages sent out by the very executives who are trying to enable

the visions to take root.

For example, one of us was engaged by one of Korea’s leading

“chaebols” (conglomerates) to facilitate a leadership initiative.

The goal was to build a global mind-set among the group’s senior

executive team. In the audience were 55 Korean business-unit

presidents, all men. The discussion was proceeding well until the

topic of reconstructing the top team came up. The suggestion was

made to establish a guideline metric that would call for one or

two of the group’s presidents to be either female or non-Korean.

The suggestion was greeted with amused disbelief. But that

response illuminates a broad problem that afflicts people in any

country: Companies and their employees are hard-wired to resist

change. We have observed similar roadblocks when companies

attempt to align processes — such as those for performance man-

agement and rewards — with new visions.

For example, many companies today are trying to break down

silos to deliver integrated solutions to their customers. They are

trying to enact “one company” visions. And yet they frequently

continue to measure and reward their managers solely on their

ability to meet unit objectives. Such outdated processes do noth-

ing to encourage cross-unit collaboration in the service of

enhanced customer-value propositions.

What can be done to align processes with a new vision? When

Josef Ackermann first launched Deutsche Bank’s One Bank ini-

tiative, there was a good deal of discussion but little actual

business transacted as a result of cross-unit collaboration. That

began to change when two of the bank’s executives, Anshu Jain,

head of Global Markets, and Pierre de Weck, head of Private

Wealth Management, signed a Global Partnership Agreement.

The agreement set the stage for a

robust collaboration between the

two divisions. The two business

heads figured out how to work

together to satisfy clients’ needs,

as well as how to share revenues

and profits from those clients.

This wasn’t the only example

of parts of the bank learning to

collaborate. Other approaches to

aligning the company’s systems,

processes and mind-sets with the

One Bank vision include:

■ executive development activi-

ties undertaken as a team, with

leaders from different divisions

taking part;

■ the establishment of a unit-by-

unit performance measurement

process that tracked business gen-

erated by boundary spanning;

■ changes to the executive succes-

sion process that took into

consideration the extent to which

emerging leaders exhibited cross-

boundary behaviors;

■ and the institution of the One

Bank award, a formal recognition

by the Group Executive Commit-

tee of exceptional behavior in

service of the new vision.

The alignment of vision and

process can also be reinforced by changing organizational struc-

ture and its support mechanisms. For example, following the

restructuring of the BBC’s departments, Mark Thompson cre-

ated a new Operations Group. The group was set up to bring

leadership, direction and consistency to the BBC’s big infra-

structure projects, a critical element of the Creative Future

vision. Thompson also linked cost-cutting measures in func-

tional departments to the enablement of the vision, since cost

savings could be funneled directly into new investments. Thus,

members of the finance department could see how their efforts

to develop new accounting and finance systems were contribut-

ing $100 million toward the Creative Future initiative. By such

simple yet creative means are the hearts and minds of large or-

ganizations won.

Phase V: Energizing the Organization Through the Power of People No

vision of the enterprise can come to life without the enthusiastic

support and follow-through of literally thousands of managers

Most leadership

studies focus

on individuals.

But all too

often, the

problem is not

that leaders

are failing their

companies, but

rather that a

company’s talent

policies are

failing its leaders.

76 MIT SLOAN MANAGEMENT REVIEW WINTER 2008 SLOANREVIEW.MIT.EDU

and employees. Individuals in every business, function and re-

gion play a critical role in enabling a bold vision to move from

inspiration to implementation.

But do companies have the talent they need to see a vision

through to completion? The respondents in our study expressed

deep concern about this issue. Specifically, they worried that their

leaders were placing too little emphasis on identifying and nur-

turing the talent pools that would be needed. The problem was

not a lack of talent management and succession systems — 97%

of those participating in the research indicated that their compa-

nies had such systems in place. However, an almost equally large

percentage indicated that their companies lacked a free and flow-

ing pipeline of talent sufficient to execute their organizations’

competitive strategies. Pressed further, the respondents indicated

that most companies’ talent management systems were not in

sync with their future capability requirements.

This finding leads us to conclude that leadership failures are

not so much a matter of struggling individuals, the focus of most

leadership studies. Instead, such failures stem from executive in-

attention to the connections between talent requirements and

competitive capability requirements. All too often, the problem is

not that leaders are failing their companies, but rather that a

company’s talent policies are failing its leaders.

Great change leaders are deeply committed to, engaged in and

accountable for building robust talent processes. They do not

delegate this job to the human resources department. The BBC’s

Mark Thompson brought in a new head of BBC People, as it is

called, with the express purpose and mandate to create a human

capital strategy that is directly linked to the corporation’s Creative

Future vision. All division heads engaged in and signed off on the

strategy and committed to implement it with the same rigor they

would apply to their marketing or operations strategies.

Deutsche Bank’s Ackermann demanded that the company’s

entire leadership curriculum be re-engineered in accordance

with the Business Realignment Program and the One Bank vi-

sion. The bank’s programs now are focused directly on global

leadership, managing networks and organizational complexity

and leading across business and geographic boundaries — all

essential ingredients in making the One Bank vision a sustain-

able success.

Even companies that do not face immediate threats need to

consider how to keep their people energized. Starbucks Corp., for

example, has yet to face major barriers to its growth. In 2006,

however, the company launched the Grow and Stay Small initia-

tive. Its focus has been to determine how Starbucks ensures that

its partners (employees) and customers have the same positive

experiences that they had when the company was much smaller

as well as identify the forces affecting the culture. Through the

engagement of partners and customers, the company is identify-

ing the common ground needed to “stay small” culturally while

building support for enterprisewide initiatives that build on this

common ground.

Continuous RevisionJust a few decades ago, organizations could stay the course with a

strategy for a period of years. The idea that a new vision would

be needed, perhaps with some frequency, would have been treated

with mild amusement, if not outright derision. But competitive

realities have forced executives to rethink what their companies

are doing, and how they are doing it, over and over again. Auto

manufacturers like Nissan see profitability and market share

evaporating. Media companies like the BBC face the hostile

world of disruptive technology. Financial institutions like

Deutsche Bank discover that a one-country focus is a path to

extinction. In such conditions, bold visions are called for — and

will be called for again, probably sooner than any executive

would like.

Let’s stipulate: A bold vision that fundamentally misreads the

competitive landscape has no chance of success, regardless of the

process used to make it reality. The vision has to be on target, so-

phisticated, inspiring and far-seeing. Once those difficult criteria

have been met, however, a process is needed to take the vision from

its birth — as words thundered from a podium by the CEO as rock

star, for example — to a new way of doing business. By following

the model outlined here, executives can sustain the momentum,

and see through the changes, of enabling bold new visions.

REFERENCES

1. See, for example, J.P. Kotter, “Leading Change: Why Transformation Efforts Fail,” Harvard Business Review (March-April 1995): 59-67; L. Bossidy, R. Charan and C. Burck, “Execution: The Discipline of Getting Things Done” (New York: Crown Business, 2002); D.A. Ready, “Leading at the Enterprise Level,” MIT Sloan Management Review 45, no. 3 (spring 2004): 87-91; D.A. Ready, “How Storytelling Builds Next-Gener-ation Leaders,” MIT Sloan Management Review 43, no. 4 (summer 2002): 63-69; J.C. Collins and J.I. Porras, “Building Your Company’s Vi-sion,” Harvard Business Review (September 1996): 65-77; M.A. Roberto and L.C. Levesque, “The Art of Making Change Initiatives Stick,” MIT Sloan Management Review 46, no. 4 (summer 2005): 53-60; C.K. Prahalad and G. Hamel, “The Core Competence of the Corpora-tion,” Harvard Business Review 68, no. 3 (May-June 1990): 79-91; M. Beer and N. Nohria, “Cracking the Code of Change,” Harvard Business Review (May-June 2000): 133-141; D.A. Ready and J. Conger, “Make Your Company a Talent Factory,” Harvard Business Review (June 2007): 68-77; E.G. Chambers, M. Foulon, H. Handfield-Jones, S.M. Hankin and E.G. Michaels III, “The War for Talent,” McKinsey Quarterly 3 (August 1998): 44-57; D.C. Hambrick, “The Top Management Team: Key to Strategic Success,” California Management Review 30, no. 1 (fall 1987): 88-108; and G. Probst and K. Marmenout, “Deutsche Bank: Be-coming a Global Leader With European Tradition,” European Case Clearing House no. 306-529-1 (Geneva: University of Geneva, 2007).

2. Probst, “Deutsche Bank.”

Reprint 49202. Copyright © Massachusetts Institute of Technology, 2008. All rights reserved.

L E A D E R S H I P

The Leader’s New Work: Building LearningOrganizations

Peter M. Senge

Reprint 3211

MassachusettsInstitute of Technology

Fall 1990

Volume 32Number 1

MIT

SloanManagement

Review

7Fall 1990

HUMAN BEINGS are designed for learning.No one has to teach an infant to walk,or talk, or master the spatial relation-

ships needed to stack eight building blocks thatdon’t topple. Children come fully equipped withan insatiable drive to explore and experiment.Unfortunately, the primary institutions of oursociety are oriented predominantly toward con-trolling rather than learning, rewarding individu-als for performing for others rather than for cul-tivating their natural curiosity and impulse tolearn. The young child entering school discoversquickly that the name of the game is getting theright answer and avoiding mistakes — a man-date no less compelling to the aspiring manager.

“Our prevailing system of management hasdestroyed our people,” writes W. EdwardsDeming, leader in the quality movement.1

“People are born with intrinsic motivation, self-esteem, dignity, curiosity to learn, joy in learn-ing. The forces of destruction begin with tod-dlers — a prize for the best Halloween costume,grades in school, gold stars, and on up throughthe university. On the job, people, teams, divi-sions are ranked — reward for the one at the top,punishment at the bottom. MBO quotas, incen-tive pay, business plans, put together separately,division by division, cause further loss, unknownand unknowable.”

Ironically, by focusing on performing forsomeone else’s approval, corporations create thevery conditions that predestine them to mediocreperformance. Over the long run, superior perfor-mance depends on superior learning. A Shellstudy showed that, according to former planningdirector Arie de Geus, “a full one-third of theFortune ‘500’ industrials listed in 1970 had van-ished by 1983.”2 Today, the average lifetime ofthe largest industrial enterprises is probably lessthan half the average lifetime of a person in anindustrial society. On the other hand, de Geusand his colleagues at Shell also found a smallnumber of companies that survived for seventy-five years or longer. Interestingly, the key to theirsurvival was the ability to run “experiments inthe margin,” to continually explore new businessand organizational opportunities that create po-tential new sources of growth.

If anything, the need for understanding howorganizations learn and accelerating that learn-ing is greater today than ever before. The olddays when a Henry Ford, Alfred Sloan, or TomWatson learned for the organization are gone. Inan increasingly dynamic, interdependent, andunpredictable world, it is simply no longer pos-sible for anyone to “figure it all out at the top.”The old model, “the top thinks and the localacts,” must now give way to integrating thinkingand acting at all levels. While the challenge isgreat, so is the potential payoff. “The person

OVER THE PAST two years, business academics and senior managers have begun talking aboutthe notion of the learning organization. Ray Stata of Analog Devices put the idea succinctly inthese pages last spring: “The rate at which organizations learn may become the only substan-tial source of competitive advantage.” And in late May of this year, at an MIT-sponsored con-ference entitled “Transforming Organizations,” two questions arose again and again: How canwe build organizations in which continuous learning occurs? and, What kind of person can best leadthe learning organization? This article, based on Senge’s recently published book, The FifthDiscipline: The Art and Practice of the Learning Organization, begins to chart this new territory, de-scribing new roles, skills, and tools for leaders who wish to develop learning organizations.

The Leader’s New Work: Building Learning Organizations

tnemeganaM fo loohcS naolS TIMegneS .M reteP

Peter M. Senge is Director of the Systems Thinking and Organiza-tional Learning program at the MIT Sloan School of Management.

who figures out how to harness the collective ge-nius of the people in his or her organization,” ac-cording to former Citibank CEO Walter Wriston,“is going to blow the competition away.”

Adaptive Learning and Generative LearningThe prevailing view of learning organizationsemphasizes increased adaptability. Given the ac-celerating pace of change, or so the standardview goes, “the most successful corporation ofthe 1990s,” according to Fortune magazine,“will be something called a learning organiza-tion, a consummately adaptive enterprise.”3 Asthe Shell study shows, examples of traditionalauthoritarian bureaucracies that responded tooslowly to survive in changing business environ-ments are legion.

But increasing adaptiveness is only the firststage in moving toward learning organizations.The impulse to learn in children goes deeperthan desires to respond and adapt more effective-ly to environmental change. The impulse tolearn, at its heart, is an impulse to be generative,to expand our capability. This is why leadingcorporations are focusing on generative learning,which is about creating, as well as adaptive learn-ing, which is about coping.4

The total quality movement in Japan illus-trates the evolution from adaptive to generativelearning. With its emphasis on continuous ex-penmentation and feedback, the total qualitymovement has been the first wave in buildinglearning organizations. But Japanese firms’ viewof serving the customer has evolved. In the earlyyears of total quality, the focus was on “fitness tostandard,” making a product reliably so that itwould do what its designers intended it to doand what the firm told its customers it woulddo. Then came a focus on “fitness to need,” un-derstanding better what the customer wantedand then providing products that reliably metthose needs. Today, leading edge firms seek tounderstand and meet the “latent need” of thecustomer—what customers might truly valuebut have never experienced or would never thinkto ask for. As one Detroit executive commentedrecently, “You could never produce the MazdaMiata solely from market research. It required aleap of imagination to see what the customermight want.”5

Generative learning, unlike adaptive learning,

requires new ways of looking at the world,whether in understanding customers or in un-derstanding how to better manage a business.For years, U.S. manufacturers sought competi-tive advantage in aggressive controls on invento-ries, incentives against overproduction, and rigidadherence to production forecasts. Despite theseincentives, their performance was eventuallyeclipsed by Japanese firms who saw the chal-lenges of manufacturing differently. They real-ized that eliminating delays in the productionprocess was the key to reducing instability andimproving cost, productivity, and service. Theyworked to build networks of relationships withtrusted suppliers and to redesign physical pro-duction processes so as to reduce delays in mate-rials procurement, production set up, and in-process inventory — a much higher-leverageapproach to improving both cost and customerloyalty.

As Boston Consulting Group’s George Stalkhas observed, the Japanese saw the significanceof delays because they saw the process of orderentry, production scheduling, materials pro-curement, production, and distribution as anintegrated system. “What distorts the system sobadly is time,” observed Stalk — the multipledelays between events and responses. “Thesedistortions reverberate throughout the system,producing disruptions, waste, andinefficiency.”6 Generative learning requires see-ing the systems that control events. When wefail to grasp the systemic source of problems,we are left to “push on” symptoms rather thaneliminate underlying causes. The best we canever do is adaptive learning.

The Leader’s New Work“I talk with people all over the country aboutlearning organizations, and the response is al-ways very positive,” says William O’Brien, CEOof the Hanover Insurance companies. “If thistype of organization is so widely preferred, whydon’t people create such organizations? I thinkthe answer is leadership. People have no realcomprehension of the type of commitment it re-quires to build such an organization.”7

Our traditional view of leaders — as specialpeople who set the direction, make the key de-cisions, and energize the troops — is deeplyrooted in an individualistic and nonsystemicworldview. Especially in the West, leaders areheroes — great men (and occasionally women)

LearningOrganizations

8Senge

who rise to the fore in times of crisis. So long assuch myths prevail, they reinforce a focus onshort-term events and charismatic heroes ratherthan on systemic forces and collective learning.

Leadership in learning organizations centerson subtler and ultimately more important work.In a learning organization, leaders’ roles differdramatically from that of the charismatic deci-sion maker. Leaders are designers, teachers, andstewards. These roles require new skills: the abili-ty to build shared vision, to bring to the surfaceand challenge prevailing mental models, and tofoster more systemic patterns of thinking. Inshort, leaders in learning organizations are re-sponsible for building organizations where peopleare continually expanding their capabilities toshape their future — that is, leaders are responsi-ble for learning.

Creative Tension: TheIntegrating PrincipleLeadership in a learning organization starts withthe principle of creative tension.8 Creative tensioncomes from seeing clearly where we want to be,our “vision,” and telling the truth about where weare, our “current reality.” The gap between thetwo generates a natural tension (see Figure 1).

Creative tension can be resolved in two basic

ways: by raising current reality toward the vision,or by lowering the vision toward current reality.Individuals, groups, and organizations who learnhow to work with creative tension learn how touse the energy it generates to move reality morereliably toward their visions.

The principle of creative tension has longbeen recognized by leaders. Martin LutherKing, Jr., once said, “Just as Socrates felt thatit was necessary to create a tension in themind, so that individuals could rise from thebondage of myths and half truths . . . so mustwe . . . create the kind of tension in societythat will help men rise from the dark depthsof prejudice and racism.”9

Without vision there is no creative tension.Creative tension cannot be generated from cur-rent reality alone. All the analysis in the worldwill never generate a vision. Many who are oth-erwise qualified to lead fail to do so because theytry to substitute analysis for vision. They believethat, if only people understood current reality,they would surely feel the motivation to change.They are then disappointed to discover that peo-ple “resist” the personal and organizationalchanges that must be made to alter reality. Whatthey never grasp is that the natural energy forchanging reality comes from holding a picture ofwhat might be that is more important to peoplethan what is.

But creative tension cannot be generated fromvision alone; it demands an accurate picture ofcurrent reality as well. Just as King had a dream,so too did he continually strive to “dramatize theshameful conditions” of racism and prejudice sothat they could no longer be ignored. Visionwithout an understanding of current reality willmore likely foster cynicism than creativity. Theprinciple of creative tension teaches that an accu-rate picture of current reality is just as important asa compelling picture of a desired future.

Leading through creative tension is differentthan solving problems. In problem solving, theenergy for change comes from attempting toget away from an aspect of current reality thatis undesirable. With creative tension, the ener-gy for change comes from the vision, fromwhat we want to create, juxtaposed with cur-rent reality. While the distinction may seemsmall, the consequences are not. Many peopleand organizations find themselves motivated tochange only when their problems are badenough to cause them to change. This worksfor a while, but the change process runs out ofsteam as soon as the problems driving the

SloanManagement

Review

9Fall 1990

change become less pressing. With problemsolving, the motivation for change is extrinsic.With creative tension, the motivation is intrin-sic. This distinction mirrors the distinction be-tween adaptive and generative learning.

New RolesThe traditional authoritarian image of the leaderas “the boss calling the shots” has been recognizedas oversimplified and inadequate for some time.According to Edgar Schein, “Leadership is inter-twined with culture formation” Building an orga-nization’s culture and shaping its evolution is the“unique and essential function” of leadership.10 Ina learning organization, the critical roles of lead-ership — designer, teacher, and steward — haveantecedents in the ways leaders have contributedto building organizations in the past. But eachrole takes on new meaning in the learning orga-nization and, as will be seen in the following sec-tions, demands new skills and tools.

Leader as DesignerImagine that your organization is an ocean linerand that you are “the leader.” What is your role?

I have asked this question of groups of man-agers many times. The most common answer,not surprisingly, is “the captain.” Others say,“The navigator, setting the direction.” Still oth-ers say, “The helmsman, actually controlling thedirection,” or, “The engineer down there stokingthe fire, providing energy,” or, “The social direc-tor, making sure everybody’s enrolled, involved,and communicating.” While these are legitimateleadership roles, there is another which, in manyways, eclipses them all in importance. Yet rarelydoes anyone mention it.

The neglected leadership role is the designer ofthe ship. No one has a more sweeping influencethan the designer. What good does it do for thecaptain to say, “Turn starboard 30 degrees,”when the designer has built a rudder that willonly turn to port, or which takes six hours toturn to starboard? It’s fruitless to be the leader inan organization that is poorly designed.

The functions of design, or what some havecalled “social architecture,” are rarely visible; theytake place behind the scenes. The consequencesthat appear today are the result of work donelong in the past, and work today will show its

benefits far in the future. Those who aspire tolead out of a desire to control, or gain fame, orsimply to be at the center of the action, will findlittle to attract them to the quiet design work ofleadership.

But what, specifically, is involved in organiza-tional design? “Organization design is widely mis-construed as moving around boxes and lines,” saysHanover’s O’Brien. “The first task of organizationdesign concerns designing the governing ideas ofpurpose, vision, and core values by which peoplewill live.” Few acts of leadership have a more en-during impact on an organization than building afoundation of purpose and core values.

In 1982, Johnson & Johnson found itself fac-ing a corporate nightmare when bottles of its best-selling Tylenol were tampered with, resulting inseveral deaths. The corporation’s immediate re-sponse was to pull all Tylenol off the shelves of re-tail outlets. Thirty-one million capsules were de-stroyed, even though they were tested and foundsafe. Although the immediate cost was significant,no other action was possible given the firm’scredo. Authored almost forty years earlier by presi-dent Robert Wood Johnson, Johnson & Johnson’scredo states that permanent success is possibleonly when modern industry realizes that:• service to its customers comes first;• service to its employees and managementcomes second;• service to the community comes third; and• service to its stockholders, last.

Such statements might seem like motherhoodand apple pie to those who have not seen theway a clear sense of purpose and values can affectkey business decisions. Johnson & Johnson’s cri-sis management in this case was based on thatcredo. It was simple, it was right, and it worked.

If governing ideas constitute the first designtask of leadership, the second design task in-volves the policies, strategies, and structuresthat translate guiding ideas into business deci-sions. Leadership theorist Philip Selznick callspolicy and structure the “institutional embodi-ment of purpose.”11 “Policy making (the rulesthat guide decisions) ought to be separatedfrom decision making,” says Jay Forrester.12

“Otherwise, short-term pressures will usurptime from policy creation.”

Traditionally, writers like Selznick andForrester have tended to see policy making andimplementation as the work of a small numberof senior managers. But that view is changing.Both the dynamic business environment and themandate of the learning organization to engage

LearningOrganizations

10Senge

people at all levels now make it clear that thissecond design task is more subtle. HenryMintzberg has argued that strategy is less a ratio-nal plan arrived at in the abstract and imple-mented throughout the organization than an“emergent phenomenon.” Successful organiza-tions “craft strategy” according to Mintzberg, asthey continually learn about shifting businessconditions and balance what is desired and whatis possible.13 The key is not getting the rightstrategy but fostering strategic thinking. “Thechoice of individual action is only part of . . . thepolicymaker’s need,” according to Mason andMitroff.14 “More important is the need toachieve insight into the nature of the complexityand to formulate concepts and world views forcoping with it.”

Behind appropriate policies, strategies, andstructures are effective learning processes; theircreation is the third key design responsibility inlearning organizations. This does not absolve se-nior managers of their strategic responsibilities.Actually, it deepens and extends those responsi-bilities. Now, they are not only responsible forensuring that an organization have well-devel-oped strategies and policies, but also for ensuringthat processes exist whereby these are continuallyimproved.

In the early 1970s, Shell was the weakest ofthe big seven oil companies. Today, Shell andExxon are arguably the strongest, both in sizeand financial health. Shell’s ascendance beganwith frustration. Around 1971, members ofShell’s “Group Planning” in London began toforesee dramatic change and unpredictability inworld oil markets. However, it proved impossi-ble to persuade managers that the stable world ofsteady growth in oil demand and supply theyhad known for twenty years was about tochange. Despite brilliant analysis and artful pre-sentation, Shell’s planners realized, in the wordsof Pierre Wack, that they “had failed to changebehavior in much of the Shell organization.”15

Progress would probably have ended there, hadthe frustration not given way to a radically newview of corporate planning.

As they pondered this failure, the planners’view of their basic task shifted: “We no longersaw our task as producing a documented view ofthe future business environment five or ten yearsahead. Our real target was the microcosm (the‘mental model’) of our decision makers.” Onlywhen the planners reconceptualized their basictask as fostering learning rather than devisingplans did their insights begin to have an impact.

The initial tool used was “scenario analysis,”through which planners encouraged operatingmanagers to think through how they wouldmanage in the future under different possiblescenarios. It mattered not that the managers be-lieved the planners’ scenarios absolutely, onlythat they became engaged in ferreting out theimplications. In this way, Shell’s planners condi-tioned managers to be mentally prepared for ashift from low prices to high prices and from sta-bility to instability. The results were significant.When OPEC became a reality, Shell quickly re-sponded by increasing local operating companycontrol (to enhance maneuverability in the newpolitical environment), building buffer stocks,and accelerating development of non-OPECsources — actions that its competitors tookmuch more slowly or not at all.

Somewhat inadvertently, Shell planners haddiscovered the leverage of designing institution-al learning processes, whereby, in the words offormer planning director de Geus,“Management teams change their shared men-tal models of their company, their markets, andtheir competitors.”16 Since then, “planning aslearning” has become a byword at Shell, andGroup Planning has continually sought outnew learning tools that can be integrated intothe planning process. Some of these are de-scribed below.

Leader as Teacher“The first responsibility of a leader,” writes re-tired Herman Miller CEO Max de Pree, “is todefine reality.”17 Much of the leverage leaders canactually exert lies in helping people achieve moreaccurate, more insightful, and more empoweringviews of reality.

Leader as teacher does not mean leader as au-thoritarian expert whose job it is to teach peo-ple the “correct” view of reality. Rather, it isabout helping everyone in the organization,oneself included, to gain more insightful viewsof current reality. This is in line with a popularemerging view of leaders as coaches, guides, orfacilitators.18 In learning organizations, thisteaching role is developed further by virtue ofexplicit attention to people’s mental models andby the influence of the systems perspective.

The role of leader as teacher starts with bring-ing to the surface people’s mental models of im-portant issues. No one carries an organization, amarket, or a state of technology in his or her

SloanManagement

Review

11Fall 1990

head. What we carry in our heads are assump-tions. These mental pictures of how the worldworks have a significant influence on how weperceive problems and opportunities, identifycourses of action, and make choices.

One reason that mental models are so deeplyentrenched is that they are largely tacit. IanMitroff, in his study of General Motors, arguesthat an assumption that prevailed for years wasthat, in the United States, “Cars are status sym-bols. Styling is therefore more important thanquality.”19 The Detroit automakers didn’t say, “Wehave a mental model that all people care about isstyling.” Few actual managers would even saypublicly that all people care about is styling. Solong as the view remained unexpressed, there waslittle possibility of challenging its validity or form-ing more accurate assumptions.

But working with mental models goes beyondrevealing hidden assumptions. “Reality,” as per-ceived by most people in most organizations,means pressures that must be borne, crises thatmust be reacted to, and limitations that must beaccepted. Leaders as teachers help people restruc-ture their views of reality to see beyond the super-ficial conditions and events into the underlyingcauses of problems — and therefore to see newpossibilities for shaping the future.

Specifically, leaders can influence people toview reality at three distinct levels: events, pat-terns of behavior, and systemic structure.

Systemic Structure(Generative)

Patterns of Behavior(Responsive)

Events(Reactive)

The key question becomes where do leaders pre-dominantly focus their own and their organization’sattention?

Contemporary society focuses predominantlyon events. The media reinforces this perspective,with almost exclusive attention to short-term,dramatic events. This focus leads naturally to ex-plaining what happens in terms of those events:“The Dow Jones average went up sixteen pointsbecause high fourth-quarter profits were an-nounced yesterday.”

Pattern-of-behavior explanations are rarer, incontemporary culture, than event explanations,

but they do occur. “Trend analysis” is an exam-ple of seeing patterns of behavior. A good edito-rial that interprets a set of current events in thecontext of long-term historical changes is anoth-er example. Systemic, structural explanations goeven further by addressing the question, “Whatcauses the patterns of behavior?”

In some sense, all three levels of explanationare equally true. But their usefulness is quite dif-ferent. Event explanations — who did what towhom — doom their holders to a reactive stancetoward change. Pattern-of-behavior explanationsfocus on identifying long-term trends and assess-ing their implications. They at least suggest how,over time, we can respond to shifting conditions.Structural explanations are the most powerful.Only they address the underlying causes of be-havior at a level such that patterns of behaviorcan be changed.

By and large, leaders of our current institu-tions focus their attention on events and patternsof behavior, and, under their influence, their or-ganizations do likewise. That is why contempo-rary organizations are predominantly reactive, orat best responsive — rarely generative. On theother hand, leaders in learning organizations payattention to all three levels, but focus especiallyon systemic structure; largely by example, theyteach people throughout the organization to dolikewise.

Leader as StewardThis is the subtlest role of leadership. Unlike theroles of designer and teacher, it is almost solely amatter of attitude. It is an attitude critical tolearning organizations.

While stewardship has long been recognized asan aspect of leadership, its source is still not wide-ly understood. I believe Robert Greenleaf cameclosest to explaining real stewardship, in his semi-nal book Servant Leadership.20 There, Greenleafargues that “The servant leader is servant first. . . .It begins with the natural feeling that one wantsto serve, to serve first. This conscious choicebrings one to aspire to lead. That person issharply different from one who is leader first, per-haps because of the need to assuage an unusualpower drive or to acquire material possessions.”

Leaders’ sense of stewardship operates on twolevels: stewardship for the people they lead andstewardship for the larger purpose or mission thatunderlies the enterprise. The first type arises froma keen appreciation of the impact one’s leadership

LearningOrganizations

12Senge

can have on others. People can suffer economi-cally, emotionally, and spiritually under ineptleadership. If anything, people in a learning or-ganization are more vulnerable because of theircommitment and sense of shared ownership.Appreciating this naturally instills a sense of re-sponsibility in leaders. The second type of stew-ardship arises from a leader’s sense of personalpurpose and commitment to the organization’slarger mission. People’s natural impulse to learnis unleashed when they are engaged in an en-deavor they consider worthy of their fullest com-mitment. Or, as Lawrence Miller puts it,“Achieving return on equity does not, as a goal,mobilize the most noble forces of our soul.”21

Leaders engaged in building learning organiza-tions naturally feel part of a larger purpose that goesbeyond their organization. They are part of changingthe way businesses operate, not from a vague philan-thropic urge, but from a conviction that their effortswill produce more productive organizations, capableof achieving higher levels of organizational successand personal satisfaction than more traditional orga-nizations. Their sense of stewardship was succinctlycaptured by George Bernard Shaw when he said,

This is the true joy in life, the being used for apurpose you consider a mighty one, the beinga force of nature rather than a feverish, selfishclod of ailments and grievances complainingthat the world will not devote itself to makingyou happy.

New SkillsNew leadership roles require new leadershipskills. These skills can only be developed, in myjudgment, through a lifelong commitment. It isnot enough for one or two individuals to devel-op these skills. They must be distributed widelythroughout the organization. This is one reasonthat understanding the disciplines of a learningorganization is so important. These disciplinesembody the principles and practices that canwidely foster leadership development.

Three critical areas of skills (disciplines) arebuilding shared vision, surfacing and challengingmental models, and engaging in systems thinking.22

Building Shared VisionHow do individual visions come together to cre-ate shared visions? A useful metaphor is the

hologram, the three-dimensional image createdby interacting light sources.

If you cut a photograph in half, each halfshows only part of the whole image. But if youdivide a hologram, each part, no matter howsmall, shows the whole image intact. Likewise,when a group of people come to share a visionfor an organization, each person sees an individ-ual picture of the organization at its best. Eachshares responsibility for the whole, not just forone piece. But the component pieces of thehologram are not identical. Each represents thewhole image from a different point of view. It’ssomething like poking holes in a window shade;each hole offers a unique angle for viewing thewhole image. So, too, is each individual’s visionunique.

When you add up the pieces of a hologram,something interesting happens. The image be-comes more intense, more lifelike. When morepeople come to share a vision, the vision be-comes more real in the sense of a mental realitythat people can truly imagine achieving. Theynow have partners, co-creators; the vision nolonger rests on their shoulders alone. Early on,when they are nurturing an individual vision,people may say it is “my vision.” But, as theshared vision develops, it becomes both “my vi-sion” and “our vision.”

The skills involved in building shared visioninclude the following:• Encouraging Personal Vision. Shared vi-sions emerge from personal visions. It is not thatpeople only care about their own self-interest —in fact, people’s values usually include dimen-sions that concern family, organization, commu-nity, and even the world. Rather, it is thatpeople’s capacity for caring is personal.• Communicating and Asking for Support.Leaders must be willing to continually sharetheir own vision, rather than being the officialrepresentative of the corporate vision. They alsomust be prepared to ask, “Is this vision worthy ofyour commitment?” This can be difficult for aperson used to setting goals and presuming com-pliance.• Visioning as an Ongoing Process. Buildingshared vision is a never-ending process. At anyone point there will be a particular image of thefuture that is predominant, but that image willevolve. Today, too many managers want to dis-pense with the “vision business” by going off andwriting the Official Vision Statement. Suchstatements almost always lack the vitality, fresh-ness, and excitement of a genuine vision that

SloanManagement

Review

13Fall 1990

comes from people asking, “What do we reallywant to achieve?”• Blending Extrinsic and Intrinsic Visions.Many energizing visions are extrinsic — that is,they focus on achieving something relative to anoutsider, such as a competitor. But a goal that islimited to defeating an opponent can, once thevision is achieved, easily become a defensive pos-ture. In contrast, intrinsic goals like creating anew type of product, taking an established prod-uct to a new level, or setting a new standard forcustomer satisfaction can call forth a new level ofcreativity and innovation. Intrinsic and extrinsicvisions need to coexist; a vision solely predicatedon defeating an adversary will eventually weakenan organization.• Distinguishing Positive from NegativeVisions. Many organizations only truly pull to-gether when their survival is threatened.Similarly, most social movements aim at elimi-nating what people don’t want: for example, anti-drugs, anti-smoking, or anti-nuclear arms move-ments. Negative visions carry a subtle message ofpowerlessness: people will only pull togetherwhen there is sufficient threat. Negative visionsalso tend to be short term. Two fundamentalsources of energy can motivate organizations: fearand aspiration. Fear, the energy source behindnegative visions, can produce extraordinarychanges in short periods, but aspiration enduresas a continuing source of learning and growth.

Surfacing and Testing Mental ModelsMany of the best ideas in organizations never getput into practice. One reason is that new insightsand initiatives often conflict with established men-tal models. The leadership task of challenging as-sumptions without invoking defensiveness requiresreflection and inquiry skills possessed by few lead-ers in traditional controlling organizations.23

• Seeing Leaps of Abstraction. Our minds lit-erally move at lightning speed. Ironically, thisoften slows our learning, because we leap to gen-eralizations so quickly that we never think to testthem. We then confuse our generalizations withthe observable data upon which they are based,treating the generalizations as if they were data.The frustrated sales rep reports to the home of-fice that “customers don’t really care about quali-ty, price is what matters,” when what actuallyhappened was that three consecutive large cus-tomers refused to place an order unless a largerdiscount was offered. The sales rep treats her

generalization, “customers care only aboutprice,” as if it were absolute fact rather than anassumption (very likely an assumption reflectingher own views of customers and the market).This thwarts future learning because she starts tofocus on how to offer attractive discounts ratherthan probing behind the customers’ statements.For example, the customers may have been sodisgruntled with the firm’s delivery or customerservice that they are unwilling to purchase againwithout larger discounts.• Balancing Inquiry and Advocacy. Most man-agers are skilled at articulating their views andpresenting them persuasively. While important,advocacy skills can become counterproductive asmanagers rise in responsibility and confront in-creasingly complex issues that require collabora-tive learning among different, equally knowl-edgeable people. Leaders in learningorganizations need to have both inquiry and ad-vocacy skills.24

Specifically, when advocating a view, theyneed to be able to:— explain the reasoning and data that led totheir view;— encourage others to test their view (e.g., Doyou see gaps in my reasoning? Do you disagreewith the data upon which my view is based?);and— encourage others to provide different views(e.g., Do you have either different data, differentconclusions, or both?).

When inquiring into another’s views, theyneed to:— actively seek to understand the other’s view,rather than simply restating their own view andhow it differs from the other’s view; and— make their attributions about the other and theother’s view explicit (e.g., Based on your statementthat . . . ; I am assuming that you believe . . . ;Am I representing your views fairly?).

If they reach an impasse (others no longer ap-pear open to inquiry), they need to:— ask what data or logic might unfreeze the im-passe, or if an experiment (or some other inquiry)might be designed to provide new information.• Distinguishing Espoused Theory fromTheory in Use. We all like to think that we holdcertain views, but often our actions reveal deeperviews. For example, I may proclaim that peopleare trustworthy, but never lend friends moneyand jealously guard my possessions. Obviously,my deeper mental model (my theory in use), dif-fers from my espoused theory. Recognizing gapsbetween espoused views and theories in use

LearningOrganizations

14Senge

(which often requires the help of others) can bepivotal to deeper learning.• Recognizing and Defusing DefensiveRoutines. As one CEO in our research programputs it, “Nobody ever talks about an issue at the8:00 business meeting exactly the same way theytalk about it at home that evening or over drinksat the end of the day.” The reason is what ChrisArgyris calls “defensive routines,” entrenchedhabits used to protect ourselves from the embar-rassment and threat that come with exposing ourthinking. For most of us, such defenses began tobuild early in life in response to pressures to havethe right answers in school or at home.Organizations add new levels of performanceanxiety and thereby amplify and exacerbate thisdefensiveness. Ironically, this makes it even moredifficult to expose hidden mental models, andthereby lessens learning.

The first challenge is to recognize defensiveroutines, then to inquire into their operation.Those who are best at revealing and defusing de-fensive routines operate with a high degree ofself-disclosure regarding their own defensiveness(e.g., I notice that I am feeling uneasy about howthis conversation is going. Perhaps I don’t under-stand it or it is threatening to me in ways I don’tyet see. Can you help me see this better?)

Systems ThinkingWe all know that leaders should help people seethe big picture. But the actual skills whereby lead-ers are supposed to achieve this are not well under-stood. In my experience, successful leaders oftenare “systems thinkers” to a considerable extent.They focus less on day-to-day events and more onunderlying trends and forces of change. But theydo this almost completely intuitively. The conse-quence is that they are often unable to explaintheir intuitions to others and feel frustrated thatothers cannot see the world the way they do.

One of the most significant developments inmanagement science today is the gradual coales-cence of managerial systems thinking as a field ofstudy and practice. This field suggests some keyskills for future leaders:• Seeing Interrelationships, Not Things, andProcesses, Not Snapshots. Most of us havebeen conditioned throughout our lives to focuson things and to see the world in static images.This leads us to linear explanations of systemicphenomenon. For instance, in an arms race eachparty is convinced that the other is the cause of

problems. They react to each new move as anisolated event, not as part of a process. So long asthey fail to see the interrelationships of these ac-tions, they are trapped.• Moving beyond Blame. We tend to blameeach other or outside circumstances for ourproblems. But it is poorly designed systems, notincompetent or unmotivated individuals, thatcause most organizational problems. Systemsthinking shows us that there is no outside —that you and the cause of your problems are partof a single system.• Distinguishing Detail Complexity fromDynamic Complexity. Some types of complexi-ty are more important strategically than others.Detail complexity arises when there are manyvariables. Dynamic complexity arises when causeand effect are distant in time and space, andwhen the consequences over time of interven-tions are subtle and not obvious to many partici-pants in the system. The leverage in most man-agement situations lies in understandingdynamic complexity, not detail complexity.• Focusing on Areas of High Leverage. Somehave called systems thinking the “new dismal sci-ence” because it teaches that most obvious solu-tions don’t work — at best, they improve mat-ters in the short run, only to make things worsein the long run. But there is another side to thestory. Systems thinking also shows that small,well-focused actions can produce significant, en-during improvements, if they are in the rightplace. Systems thinkers refer to this idea as theprinciple of “leverage.” Tackling a difficult prob-lem is often a matter of seeing where the highleverage lies, where a change — with a mini-mum of effort — would lead to lasting, signifi-cant improvement.• Avoiding Symptomatic Solutions. The pres-sures to intervene in management systems thatare going awry can be overwhelming.Unfortunately, given the linear thinking thatpredominates in most organizations, interven-tions usually focus on symptomatic fixes, notunderlying causes. This results in only tempo-rary relief, and it tends to create still more pres-sures later on for further, low-leverage interven-tion. If leaders acquiesce to these pressures, theycan be sucked into an endless spiral of increasingintervention. Sometimes the most difficult lead-ership acts are to refrain from interveningthrough popular quick fixes and to keep thepressure on everyone to identify more enduringsolutions.

While leaders who can articulate systemic ex-

SloanManagement

Review

15Fall 1990

planations are rare, those who can will leave theirstamp on an organization. One person who hadthis gift was Bill Gore, the founder and long-time CEO of W.L. Gore and Associates (makersof GoreTex and other synthetic fiber products).Bill Gore was adept at telling stories that showedhow the organization’s core values of freedomand individual responsibility required particularoperating policies. He was proud of his egalitari-an organization, in which there were (and stillare) no “employees,” only “associates,” all ofwhom own shares in the company and partici-pate in its management. At one talk, he ex-plained the company’s policy of controlledgrowth: “Our limitation is not financial re-sources. Our limitation is the rate at which wecan bring in new associates. Our experience hasbeen that if we try to bring in more than a 25percent per year increase, we begin to bog down.Twenty-five percent per year growth is a real lim-itation; you can do much better than that withan authoritarian organization.” As Gore tells thestory, one of the associates, Esther Baum, wenthome after this talk and reported the limitationto her husband. As it happened, he was an as-tronomer and mathematician at LowellObservatory. He said, “That’s a very interestingfigure.” He took out a pencil and paper and cal-culated and said, “Do you realize that in onlyfifty-seven and a half years, everyone in theworld will be working for Gore?”

Through this story, Gore explains the systemicrationale behind a key policy, limited growth rate— a policy that undoubtedly caused a lot of stressin the organization. He suggests that, at largerrates of growth, the adverse effects of attemptingto integrate too many new people too rapidlywould begin to dominate. (This is the “limits togrowth” systems archetype explained below.) Thestory also reaffirms the organization’s commit-ment to creating a unique environment for its as-sociates and illustrates the types of sacrifices thatthe firm is prepared to make in order to remaintrue to its vision. The last part of the story showsthat, despite the self-imposed limit, the companyis still very much a growth company.

The consequences of leaders who lack systemsthinking skills can be devastating. Many charis-matic leaders manage almost exclusively at thelevel of events. They deal in visions and in crises,and little in between. Under their leadership, anorganization hurtles from crisis to crisis.Eventually, the worldview of people in the organi-zation becomes dominated by events and reactive-ness. Many, especially those who are deeply com-

mited, become burned out. Eventually, cynicismcomes to pervade the organization. People haveno control over their time, let alone their destiny.

Similar problems arise with the “visionarystrategist,” the leader with vision who sees bothpatterns of change and events. This leader is bet-ter prepared to manage change. He or she canexplain strategies in terms of emerging trends,and thereby foster a climate that is less reactive.But such leaders still impart a responsive orienta-tion rather than a generative one.

Many talented leaders have rich, highly sys-temic intuitions but cannot explain those intu-itions to others. Ironically, they often end upbeing authoritarian leaders, even if they don’twant to, because only they see the decisions thatneed to be made. They are unable to conceptual-ize their strategic insights so that these can be-come public knowledge, open to challenge andfurther improvement.

New ToolsDeveloping the skills described above requiresnew tools — tools that will enhance leaders’conceptual abilities and foster communicationand collaborative inquiry. What follows is asampling of tools starting to find use in learningorganizations.

Systems Archetypes One of the insights of the budding, managerialsystems-thinking field is that certain types of sys-temic structures recur again and again. Countlesssystems grow for a period, then encounter prob-lems and cease to grow (or even collapse) well be-fore they have reached intrinsic limits to growth.Many other systems get locked in runaway viciousspirals where every actor has to run faster andfaster to stay in the same place. Still others lure in-dividual actors into doing what seems right local-ly, yet which eventually causes suffering for all.25

Some of the system archetypes that have thebroadest relevance include:• Balancing Process with Delay. In thisarchetype, decision makers fail to appreciate thetime delays involved as they move toward agoal. As a result, they overshoot the goal andmay even produce recurring cycles. Classic ex-ample: Real estate developers who keep startingnew projects until the market has gone soft, by

LearningOrganizations

16Senge

which time an eventual glut is guaranteed bythe properties still under construction.• Limits to Growth. A reinforcing cycle ofgrowth grinds to a halt, and may even reverse it-self, as limits are approached. The limits can beresource constraints, or external or internal re-sponses to growth. Classic examples: Product lifecycles that peak prematurely due to poor qualityor service, the growth and decline of communi-cation in a management team, and the spread ofa new movement.• Shifting the Burden. A short-term “solution”is used to correct a problem, with seeminglyhappy immediate results. As this correction isused more and more, fundamental long-termcorrective measures are used less. Over time, themechanisms of the fundamental solution mayatrophy or become disabled, leading to evengreater reliance on the symptomatic solution.Classic example: Using corporate human re-source staff to solve local personnel problems,thereby keeping managers from developing theirown interpersonal skills.• Eroding Goals. When all else fails, lower yourstandards. This is like “shifting the burden,” ex-cept that the short-term solution involves lettinga fundamental goal, such as quality standards oremployee morale standards, atrophy. Classic ex-ample: A company that responds to deliveryproblems by continually upping its quoted deliv-ery times.• Escalation. Two people or two organizations,who each see their welfare as depending on a rela-tive advantage over the other, continually react tothe other’s advances. Whenever one side gets ahead,the other is threatened, leading it to act more ag-gressively to reestablish its advantage, which threat-ens the first, and so on. Classic examples: Armsrace, gang warfare, price wars.• Tragedy of the Commons.26 Individuals keepintensifying their use of a commonly availablebut limited resource until all individuals start toexperience severely diminishing returns. Classicexamples: Sheepherders who keep increasingtheir flocks until they overgraze the commonpasture; divisions in a firm that share a commonsalesforce and compete for the use of sales repsby upping their sales targets, until the salesforceburns out from overextension.• Growth and Underinvestment. Rapidgrowth approaches a limit that could be elimi-nated or pushed into the future, but only by ag-gressive investment in physical and human ca-pacity. Eroding goals or standards causeinvestment that is too weak, or too slow, and

customers get increasingly unhappy, slowing de-mand growth and thereby making the neededinvestment (apparently) unnecessary or impossi-ble. Classic example: Countless once-successfulgrowth firms that allowed product or servicequality to erode, and were unable to generateenough revenues to invest in remedies.

The Archetype template is a specific tool thatis helping managers identify archetypes operat-ing in their own strategic areas (see Figure 2).27

The template shows the basic structural form ofthe archetype but lets managers fill in the vari-ables of their own situation. For example, theshifting the burden template involves two bal-ancing processes (“B”) that compete for controlof a problem symptom. The upper, symptomaticsolution provides a short-term fix that will makethe problem symptom go away for a while. Thelower, fundamental solution provides a more en-during solution. The side effect feedback (“R”)around the outside of the diagram identifies un-intended exacerbating effects of the symptomaticsolution, which, over time, make it more andmore difficult to invoke the fundamental solu-tion.

Several years ago, a team of managers from aleading consumer goods producer used the shift-ing the burden archetype in a revealing way. Theproblem they focused on was financial stress,

SloanManagement

Review

17Fall 1990

which could be dealt with in two different ways:by running marketing promotions (the symp-tomatic solution) or by product innovation (thefundamental solution). Marketing promotionswere fast. The company was expert in their de-sign and implementation. The results were high-ly predictable. Product innovation was slow andmuch less predictable, and the company had ahistory over the past ten years of product-inno-vation mismanagement. Yet only through inno-vation could they retain a leadership position intheir industry, which had slid over the past ten totwenty years. What the managers saw clearly wasthat the more skillful they became at promo-tions, the more they shifted the burden awayfrom product innovation. But what really struckhome was when one member identified the un-intended side effect: the last three CEOs had allcome from advertising function, which had be-come the politically dominant function in thecorporation, thereby institutionalizing the symp-tomatic solution. Unless the political valuesshifted back toward product and process innova-tion, the managers realized, the firm’s declinewould accelerate — which is just the shift thathas happened over the past several years.

Charting Strategic DilemmasManagement teams typically come ungluedwhen confronted with core dilemmas. A classicexample was the way U.S. manufacturers facedthe low cost-high quality choice. For years, mostassumed that it was necessary to choose betweenthe two. Not surprisingly, given the short-termpressures perceived by most managements, theprevailing choice was low cost. Firms that chosehigh quality usually perceived themselves as aim-ing exclusively for a high quality, high price mar-ket niche. The consequences of this perceived ei-ther-or choice have been disastrous, even fatal, asU.S. manufacturers have encountered increasinginternational competition from firms that havechosen to consistently improve quality and cost.

In a recent book, Charles Hampden-Turnerpresented a variety of tools for helping manage-ment teams confront strategic dilemmas creative-ly.28 He summarizes the process in seven steps:• Eliciting the Dilemmas. Identifying the op-posed values that form the “horns” of the dilem-ma, for example, cost as opposed to quality, orlocal initiative as opposed to central coordina-tion and control. Hampden-Turner suggests thathumor can be a distinct asset in this process

since “the admission that dilemmas even existtends to be difficult for some companies.”• Mapping. Locating the opposing values as twoaxes and helping managers identify where they seethemselves, or their organization, along the axes.• Processing. Getting rid of nouns to describethe axes of the dilemma. Present participles formedby adding “ing” convert rigid nouns into processesthat imply movement. For example, central con-trol versus local control becomes “strengtheningnational office” and “growing local initiatives.”This loosens the bond of implied opposition be-tween the two values. For example, it becomespossible to think of “strengthening national ser-vices from which local branches can benefit.”• Framing/Contextualizing. Further softeningthe adversarial structure among different valuesby letting “each side in turn be the frame or con-text for the other.” This shifting of the “figure-ground” relationship undermines any implicitattempts to hold one value as intrinsically superi-or to the other, and thereby to become mentallyclosed to creative strategies for continuous im-provement of both.• Sequencing. Breaking the hold of staticthinking. Very often, values like low cost andhigh quality appear to be in opposition becausewe think in terms of a point in time, not interms of an ongoing process. For example, astrategy of investing in new process technologyand developing a new production-floor cultureof worker responsibility may take time andmoney in the near term, yet reap significantlong-term financial rewards.• Waving/Cycling. Sometimes the strategicpath toward improving both values involvescycles where both values will get “worse” for atime. Yet, at a deeper level, learning is occurringthat will cause the next cycle to be at a higherplateau for both values.• Synergizing. Achieving synergy where signifi-cant improvement is occurring along all axes ofall relevant dilemmas. (This is the ultimate goal,of course.) Synergy, as Hampden-Turner pointsout, is a uniquely systemic notion, coming fromthe Greek syn-ergo or “work together.”

“The Left-Hand Column”:Surfacing Mental ModelsThe idea that mental models can dominate busi-ness decisions and that these models are oftentacit and even contradictory to what people es-

LearningOrganizations

18Senge

pouse can be very threatening to managers whopride themselves on rationality and judicious de-cision making. It is important to have tools tohelp managers discover for themselves how theirmental models operate to undermine their ownintentions.

One tool that has worked consistently to helpmanagers see their own mental models in actionis the “left-hand column” exercise developed byChris Argyris and his colleagues. This tool is es-pecially helpful in showing how we leap fromdata to generalization without testing the validityof our generalizations.

When working with managers, I start this ex-ercise by selecting a specific situation in which Iam interacting with other people in a way that isnot working, that is not producing the learningthat is needed. I write out a sample of the ex-change, with the script on the right-hand side ofthe page. On the left-hand side, I write what Iam thinking but not saying at each stage in theexchange (see sidebar).

The left-hand column exercise not onlybrings hidden assumptions to the surface, itshows how they influence behavior. In the exam-ple, I make two key assumptions about Bill: helacks confidence and he lacks initiative. Neithermay be literally true, but both are evident in myinternal dialogue, and both influence the way Ihandle the situation. Believing that he lacks con-fidence, I skirt the fact that I’ve heard the presen-tation was a bomb. I’m afraid that if I say it di-rectly, he will lose what little confidence he has,or he will see me as unsupportive. So I bring upthe subject of the presentation obliquely. When Iask Bill what we should do next, he gives no spe-cific course of action. Believing he lacks initia-tive, I take this as evidence of his laziness; he iscontent to do nothing when action is definitelyrequired. I conclude that I will have to manufac-ture some form of pressure to motivate him, orelse I will simply have to take matters into myown hands.

The exercise reveals the elaborate webs of as-sumptions we weave, within which we becomeour own victims. Rather than dealing directlywith my assumptions about Bill and the situa-tion, we talk around the subject. The reasons formy avoidance are self-evident: I assume that if Iraised my doubts, I would provoke a defensivereaction that would only make matters worse.But the price of avoiding the issue is high.Instead of determining how to move forward toresolve our problems, we end our exchange withno clear course of action. My assumptions about

Bill’s limitations have been reinforced. I resort toa manipulative strategy to move things forward.

The exercise not only reveals the need forskills in surfacing assumptions, but that we arethe ones most in need of help. There is no oneright way to handle difficult situations like myexchange with Bill, but any productive strategyrevolves around a high level of self-disclosure andwillingness to have my views challenged. I needto recognize my own leaps of abstraction regard-ing Bill, share the events and reasoning that are

The Left-Hand Column: An ExerciseImagine my exchange with a colleague, Bill, after he made a bigpresentation to our boss on a project we are doing together. I hadto miss the presentation, but I’ve heard that it was poorly received.

Me: How did the presentation go?Bill: Well, I don’t know. It’s really too early to say. Besides, we’rebreaking new ground here.Me: Well, what do you think we should do? I believe that the is-sues you were raising are important.Bill: I’m not so sure. Let’s just wait and see what happens. Me: You may be right, but I think we may need to do more thanjust wait.

Now, here is what the exchange looks like with my “left-handcolumn”:

What I’m Thinking What Is Said

Everyone says the presentation Me: How did the presenta-was a bomb. tion go?

Does he really not know how Bill: Well, I don’t know. It’sbad it was? Or is he not will- too early to say. Besides, we’reing to face up to it? breaking new ground here.

Me: Well, what do you thinkwe should do? I believe that theissues you were raising are im-portant.

He really is afraid to see the Bill: I’m not so sure. Let’s justtruth. If he only had more wait and see what happens.confidence, he could probablylearn from a situation like this.

I can’t believe he doesn’t realize Me: You may be right, but Ihow disastrous that presenta- think we may need to do moretion was to our moving ahead. than just wait.

I’ve got to find some way tolight a fire under the guy.

LearningOrganizations

20Senge

Hanover Insurance has gone from the bot-tom of the property and liability industryto a position among the top 25 percent ofU.S. insurance companies over the pasttwenty years, largely through the efforts ofCEO William O’Brien and his predeces-sor, Jack Adam. The following commentsare excerpted from a series of interviewsSenge conducted with O’Brien as back-ground for his book.Senge: Why do you think there is so muchchange occurring in management and organi-zations today? Is it primarily because of in-creased competitive pressures?O’Brien: That’s a factor, but not the most sig-nificant factor. The ferment in managementwill continue until we find models that aremore congruent with human nature.

One of the great insights of modern psy-chology is the hierarchy of human needs.As Maslow expressed this idea, the mostbasic needs are food and shelter. Thencomes belonging. Once these three basicneeds are satisfied, people begin to aspiretoward self-respect and esteem, and towardself-actualization — the fourth- and fifth-order needs.

Our traditional hierarchical organizationsare designed to provide for the first three levels,but not the fourth and fifth. These first threelevels are now widely available to members ofindustrial society but our organizations do notoffer people sufficient opportunities forgrowth.Senge: How would you assess Hanover’sprogress to date?O’Brien: We have been on a long journeyaway from a traditional hierarchical cul-ture. The journey began with everyone un-derstanding some guiding ideas about pur-pose, vis ion, and values as a basis forparticipative management. This is a betterway to begin building a participative cul-ture than by simply letting people in on “de-cision making.” Before there can be mean-ingful participation, people must sharecertain values and pictures about where weare trying to go. We discovered that peoplehave a real need to feel that they’re part of

an ennobling mission. But developing sharedvisions and values is not the end, only thebeginning.

Next we had to get beyond mechanical,linear thinking. The essence of our jobs asmanagers is to deal with “divergent” prob-lems — problems that have no simple an-swer. “Convergent” problems — problemsthat have a “right” answer — should besolved locally. Yet we are deeply condi-tioned to see the world in terms of conver-gent problems. Most managers t r y toforce-fit simplistic solutions and under-mine the potential for learning when di-vergent problems arise. Since everyonehandles the linear issues fairly well, com-panies that learn how to handle divergentissues will have a great advantage.

The next basic stage in our progressionwas coming to understand inquiry and ad-vocacy. We learned that real openness isrooted in people’s ability to continually in-quire into their own thinking. This requiresexposing yourself to being wrong — notsomething that most managers are rewardedfor. But learning is very difficult if you can-not look for errors or incompleteness inyour own ideas.

What all this builds to is the capabilitythroughout an organization to manage men-tal models. In a locally controlled organiza-tion, you have the fundamental challenge oflearning how to help people make good de-cisions without coercing them into makingparticular decisions. By managing mentalmodels, we create “self-concluding” deci-sions — decisions that people come tothemselves — which will result in deeperconviction, better implementation, and theability to make better adjustments when thesituation changes.Senge: What concrete steps can top managerstake to begin moving toward learning organiza-tions?O’Brien: Look at the signals you sendthrough the organization. For example, onecritical signal is how you spend your time. It’shard to build a learning organization if peo-ple are unable to take the time to thinkthrough important matters. I rarely set up an

Learning at Hanover Insurance

leading to my concern over the project, and beopen to Bill’s views on both. The skills to carryon such conversations without invoking defen-siveness take time to develop. But if both partiesin a learning impasse start by doing their ownleft-hand column exercise and sharing them witheach other, it is remarkable how quickly every-one recognizes their contribution to the impasseand progress starts to be made.

Learning Laboratories: Practice Fields for Management TeamsOne of the most promising new tools is thelearning laboratory or “microworld”: constructedmicrocosms of real-life settings in which man-agement teams can learn how to learn together.

The rationale behind learning laboratories canbest be explained by analogy. Although mostmanagement teams have great difficulty learning(enhancing their collective intelligence and ca-pacity to create), in other domains team learningis the norm rather than the exception — teamsports and the performing arts, for example.Great basketball teams do not start off great.They learn. But the process by which theseteams learn is, by and large, absent from modernorganizations. The process is a continual move-ment between practice and performance.

The vision guiding current research in man-agement learning laboratories is to design andconstruct effective practice fields for manage-ment teams. Much remains to be done, but thebroad outlines are emerging.

First, since team learning in organizations isan individual-to-individual and individual-to-system phenomenon, learning laboratories mustcombine meaningful business issues with mean-ingful interpersonal dynamics. Either alone isincomplete.

Second, the factors that thwart learning aboutcomplex business issues must be eliminated inthe learning lab. Chief among these is the inabil-ity to experience the long-term, systemic conse-quences of key strategic decisions. We all learnbest from experience, but we are unable to expe-rience the consequences of many important or-ganizational decisions. Learning laboratories re-move this constraint through system dynamicssimulation games that compress time and space.

Third, new learning skills must be developed.One constraint on learning is the inability of man- agers to reflect insightfully on their assumptions andto inquire effectively into each other’s assump-tions. Both skills can be enhanced in a learninglaboratory, where people can practice surfacingassumptions in a low-risk setting. A note of cau-tion: It is far easier to design an entertaining learn-ing laboratory than it is to have an impact on real management practices and firm traditions outside the learning lab. Research on management simu-lations has shown that they often have greater en-tertainment value than educational value. One of the reasons appears to be that many simulations do not offer deep insights into systemic structures causing business problems. Another reason is that they do not foster new learning skills. Also, there is no connection between experiments in the learn- ing lab and real life experiments. These are signifi- cant problems that research on learning laboratory design is now addressing.

Developing Leaders and LearningOrganizationsIn a recently published retrospective on or-ganizat ion deve lopment in the 1980s ,Marshall Sashkin and N. Warner Burke ob-serve the return of an emphasis on develop-ing leaders who can develop organizations.29

They also note Schein’s critique that most

SloanManagement

Review

21Fall 1990

appointment for less than one hour. If thesubject is not worth an hour, it shouldn’t beon my calendar.Senge: Why is this so hard for so manymanagers?O’Brien: It comes back to what you believeabout the nature of your work. The authoritar-ian manager has a “chain gang” mental model:“The speed of the boss is the speed of the gang.

I’ve got to keep things moving fast, because I’vegot to keep people working.” In a learning or-ganization, the manager shoulders an almostsacred responsibility: to create conditions thatenable people to have happy and productivelives. If you understand the effects the ideas weare discussing can have on the lives of people inyour organization, you will take the time.

top executives are not qualified for the task ofdeveloping culture.30 Learning organizations rep-resent a potentially significant evolution of orga-nizational culture. So it should come as no sur-prise that such organizations will remain adistant vision until the leadership capabilitiesthey demand are developed. “The 1990s may bethe period,” suggest Sashkin and Burke, “duringwhich organization development and (a new sortof) management development are reconnected.”

I believe that this new sort of management devel-opment will focus on the roles, skills, and tools forleadership in learning organizations. Undoubtedly,the ideas offered above are only a rough approxima-tion of this new territory. The sooner we begin seri-ously exploring the territory, the sooner the initialmap can be improved — and the sooner we will re-alize an age-old vision of leadership:

The wicked leader is he who the people despise.The good leader is he who the people revere.The great leader is he who the people say, “We

did it ourselves.”—Lao Tsu

References1P. Senge, The Fifth Discipline: The Art and Practice of the LearningOrganization (New York: Doubleday/Currency, 1990).

2A.P. de Geus, “Planning as Learning,” Harvard BusinessReview, March–April 1988, pp. 70–74.

3B. Domain, Fortune, 3 July 1989, pp. 48–62.

4The distinction between adaptive and generative learning has itsroots in the distinction between what Argyris and Schon have calledtheir “single-loop” learning, in which individuals or groups adjusttheir behavior relative to fixed goals, norms, and assumptions, and“double-loop” learning, in which goals, norms, and assumptions, aswell as behavior, are open to change, e.g., see C. Argyris and D.Schon, Organizational Learning: A Theory-in-Action Perspective(Reading, Massachusetts: Addison-Wesley, 1978).

5All unattributed quotes are from personal communicationswith the author.

6G. Stalk, Jr., “Time: The Next Source of Competitive Advantage,”Harvard Business Review, July–August 1988, pp. 41–51.

7Senge (1990).

8The principle of creative tension comes from Robert Fritz’ work oncreativity. See R. Fritz, The Path of Least Resistance (New York:Ballantine, 1989) and Creating (New York: Ballantine, 1990).

9M.L. King, Jr., “Letter from Birmingham Jail,” AmericanVisions, January–February 1986, pp. 52–59.

10E. Schein, Organizational Culture and Leadership (San Francisco:Jossey-Bass, 1985). Similar views have been expressed by manyleadership theorists. For example, see: P. Selznick, Leadership inAdministration (New York: Harper & Row, 1957); W. Bennisand B. Nanus, Leaders (New York: Harper & Row, 1985); andN.M. Tichy and M.A. Devanna, The Transformational Leader(New York: John Wiley & Sons, 1986).

11Selznick (1957).

12J.W. Forrester, “A New Corporate Design,” SloanManagement Review (formerly Industrial ManagementReview), Fall 1965, pp. 5–17.

13See, for example, H. Mintzberg, “Crafting Strategy,”Harvard Business Review, July–August 1987, pp. 66–75.

14R. Mason and I. Mitroff, Challenging Strategic PlanningAssumptions (New York: John Wiley & Sons, 1981), p. 16.

15P. Wack, “Scenarios: Uncharted Waters Ahead,” HarvardBusiness Review, September–October 1985, pp. 73–89.

16de Geus (1988).

17M. de Pree, Leadership Is an Art (New York: Doubleday, 1989), p. 9.

18For example, see T. Peters and N. Austin, A Passion for Excellence(New York: Random House, 1985) and J.M. Kouzes and B.Z.Posner,The Leadership Challenge (San Francisco: Jossey-Bass, 1987).

191. Mitroff, Break-Away Thinking (New York: John Wiley &Sons, 1988), pp. 66–67.

20R.K. Greenleaf, Servant Leadership: A Journey into the Nature ofLegitimate Power and Greatness (New York: Paulist Press, 1977).

21L. Miller, American Spirit: Visions of a New Corporate Culture(New York: William Morrow, 1984), p 15.

LearningOrganizations

22Senge

Technology, 1990. All rights reserved.Copyright © Massachusetts Institute of

Reprint 3211

SloanManagement

Review

23Fall 1990

22These points are condensed from the practices of the fivedisciplines examined in Senge (1990).

23The ideas below are based to a considerable extent on thework of Chris Argyris, Donald Schon, and their Action Science colleagues:C. Argyris and D. Schon, Organizational Learning: a Theory-in-Action Perspective (Reading, Massachusetts: Addison-Wesley, 1978);C. Argyris, R. Putnam, and D. Smith, Action Science (San Francisco: Jossey-Bass, 1985);C. Argyris, Strategy, Change, and Defensive Routines (Boston: Pitman, 1985); andC. Argyris, Overcoming Organizational Defenses (Englewood Cliffs, New Jersey: Prentice-Hall, 1990).

24I am indebted to Diana Smith for the summary points below.

25The system archetypes are one of several systems diagramming and communication tools. See D.H. Kim, “Toward LearningOrganizations: Integrating Total Quality Control and SystemsThinking” (Cambridge, Massachusetts: MIT Sloan School ofManagement, Working Paper No. 3037-89-BPS, June 1989).

26This archetype is closely associated with the work of ecolo-gist Garrett Hardin, who coined its label: “The Tragedy of the Commons,” Science, 13 December 1968.

27These templates were originally developed by Jennifer Kemeny, Charles Kiefer, and Michael Goodman of Inno-vation Associates, Inc., Framingham, Massachusetts.

28C. Hampden-Turner, Charting the Corporate Mind (New York: The Free Press, 1990).

29M. Sashkin and W.W. Burke, “Organization Development in the 1980s” and “An End-of-the-Eighties Retrospective,” in Advances in Organization Development, ed. F. Masarik (Norwood, New Jersey: Ablex, 1990).

30E. Schein (1985).

Reprint 3211. Copyright © Massachusetts Institute of Technology, 1990. All rights reserved.

S U M M E R 2 0 0 9 V O L . 5 0 N O. 4

R E P R I N T N U M B E R 5 0 4 3 4

Jean-François Manzoni and Jean-Louis Barsoux

Setting You Up to Fail?

Please note that gray areas reflect artwork that has been intentionally removed. The substantive content of the ar-ticle appears as originally published.

SLOANREVIEW.MIT.EDU SUMMER 2009 MIT SLOAN MANAGEMENT REVIEW 43

Even reasonable subordinates sometimes play a leading role in triggering and perpetuating unproductive and often painful relationships with their bosses.

THE LAST 15 YEARS have seen an explo-

sion in the number of popular books focusing

on “bad bosses.” These publications have not

only described the “leader from hell” but also

advised subordinates on how to handle toxic

or incompetent managers.

More scholarly writers on management

have preferred to adopt the perspective of

bosses, but much of their writing shares the

same underlying emphasis on “what man-

agers are doing wrong,” highlighting the

many ways that bosses fail to engage their

employees, through lack of communica-

tion, authenticity, imagination or emotional

intelligence. No matter who is doing the

writing, the employees are usually portrayed

as well-intentioned, competent individuals

who, if half-decently managed, will perform

well. There is, of course, some talk about

the small proportion of “dead wood” —

employees who cannot meet the company’s

minimum performance threshold — but

the rest are assumed to have what it takes

to succeed under the “right kind” of lead-

ership. Implicitly or explicitly, subordinates

are treated as receptive individuals waiting

only for the boss to offer a productive

channel to their intrinsic energies. Indeed,

much of our own writing has highlighted

L E A D E R S H I P

Are Your Subordinates Setting You Up to Fail?Subordinates sometimes make it extremely difficult for their bosses to be good leaders. Executives who fail to understand the forces at play may find their careers in jeopardy.BY JEAN-FRANÇOIS MANZONI AND JEAN-LOUIS BARSOUX

THE LEADING QUESTIONPlenty is written about the mis-takes leaders make. But sometimes they fail mainly because their subordinates undermine them. How can good managers prevent that?

FINDINGS Employees — sometimes delib-erately, sometimes subconsciously — take steps to sabotage their boss’s success.

Leaders make the situation worse by not understanding why an employee might act this way.

Concerted efforts to engage employees in particular ways can forestall the problem and help you keep your job.

44 MIT SLOAN MANAGEMENT REVIEW SUMMER 2009 SLOANREVIEW.MIT.EDU

L E A D E R S H I P

the boss’s responsibility for creating unsatisfactory

relationships — a phenomenon we labeled the set-

up-to-fail syndrome.1 Bosses trigger this dynamic

inadvertently through a combination of premature

labeling (“I know I’ve been working with him for

only three weeks, but it looks like he won’t be very

effective”); over-monitoring; and cognitive biases

(expecting the employee to fail, the boss looks for

places where the employee is underperforming and

attributes the employee’s successes to external ele-

ments). These three factors distort the way bosses

notice, interpret and remember events.

Although we initially identified the subordi-

nate’s contribution to this dynamic, we saw it

mostly as retaliation for the boss’s actions. We did

not suspect that reasonable subordinates would in

some cases play a leading role in triggering and per-

petuating unproductive and often painful

relationships with their bosses.

Yet, working on boss-subordinate relationships

for more than a decade, we have encountered many

situations where subordinates, individually or col-

lectively, placed their bosses in no-win situations.

(See “About the Research.”) Bosses need to under-

stand why subordinates would engage in such

self-defeating behavior in order to avert or inter-

cept these dysfunctional dynamics — not only to

make sure their companies perform well but also,

sometimes, to save their own jobs.

Fast ImpressionsThe appointment of a new boss creates anxiety for

subordinates. They feel a strong urge to size up the

person to determine what they will have to do to

thrive under the new regime. They closely scan the

new boss’s comments and behavior for signs of what

type of boss this is. Is she here for a while or just pass-

ing through? Is she approachable? Does she listen? Is

she consistent? Subordinates rapidly form impres-

sions of their new boss’s competence, judgment,

fairness, receptiveness and interest in developing oth-

ers; these labels guide their interactions with the boss.

How fast do subordinates start to make up their

minds? Researchers asked employees to rate their

first impressions of their boss, based on questions

such as whether they and their boss shared their out-

look, perspective and values, or whether they

expected the boss to do well in the organization. The

findings showed that subordinates’ expectations of

the boss measured in the first five days of their rela-

tionship were strong predictors of subordinate

assessments of the quality of their working relation-

ship measured two weeks, six weeks and six months

later.2 The problem with these first impressions,

based as they are on the slender evidence of a few

early interactions, is that they serve as a prism for

processing ensuing information. Once subordinates

suspect that they have a “difficult boss,” they will find

the necessary proof — and through their behavior

(passivity, defensiveness, aggressiveness) can eventu-

ally turn the manager into a difficult boss — at which

point the circle closes and the prophecy is fulfilled.

Here’s how the process unfolds.

How Good Bosses Get MislabeledIt is surprisingly easy for an individual or even a

group of subordinates to mistake a good boss for a

bad one, thus initiating a downward spiral. The pro-

cess may be set in motion because of pre-existing

conditions or developments after the boss takes

charge. For example:

1. The Boss Walks in Prelabeled. This can occur

for two reasons:

Oversensitive Employees: The way individuals

view their new boss is likely to be influenced by how

they’ve been (mis)managed previously. Certain indi-

viduals may feel they have never received due credit, or

blame their failure to advance on a previous manager

(perceptions that may or may not be totally un-

founded). This can breed a strong sense of betrayal and

cynicism.3 From day one, these subordinates may be

on the lookout for the same kind of disrespect or mis-

treatment they received (or believe they received) from

the predecessor. For different reasons, former “favor-

ites” may also prove thin-skinned. Never having been

challenged on their shortcomings by an overindulgent

predecessor, they may be susceptible to negative feed-

back and may find the incoming boss hypercritical or

unfair. New bosses also need to beware of subordinates

who were actual or undeclared contenders for the

leadership role, or allies of a manager who didn’t get

the promotion. They too may enter the relationship

primed to focus on the new boss’s shortcomings rather

than strengths. Consciously or not, these individuals

may expect — or indeed want — the new boss to fail.

SUMMER 2009 MIT SLOAN MANAGEMENT REVIEW 45SLOANREVIEW.MIT.EDU

Team Preconceptions: The new boss will some-

times be adversely affected by the reputation of his

predecessor(s). In one company we studied, the new

incumbent was the fourth general manager in three

years. Team members had lost faith in bosses and were

reluctant to reengage because they didn’t want to be

disappointed yet again. The new boss came in with a

“dead man walking” label already hanging over him.

At the other extreme, incoming leaders may suffer

from the contrast with a much admired predecessor.

When the previous boss has created a strong collective

view of what the ideal leader looks like, it’s difficult for

a new leader to match that image, regardless of the

traits and abilities the newcomer brings to the role.

Research shows that incompatibility between the ac-

tual and ideal images of a boss strongly correlates with

dissatisfaction with the boss and lower-quality rela-

tionships.4 The circumstances surrounding the

predecessor’s departure also matter. The more unex-

pected the event, the more intense the employee

emotional reaction. The new boss may have to con-

tend with feelings of anger, grief, betrayal or anxiety,

especially among those who saw the outgoing leader

as a mentor, friend or coach. Research suggests that

negative emotions are associated with detailed gather-

ing of evidence that typically leads to unfavorable

evaluations of the target.5 Finally, even before the new

boss arrives, team members may tap their internal and

external social networks for information. Incoming

bosses can find themselves preceded by their own rep-

utations — or in some cases harsh nicknames (often

awarded for ripping out costs). Carlos Ghosn’s repu-

tation as “Le Cost Killer” preceded him at Nissan

Motor Co., as did that of John Mack “the Knife” when

he moved to Credit Suisse Group.

2. How the Boss Makes Things Worse. Subordi-

nates can also misconstrue the actions of their bosses.

This can happen to new and established bosses alike.

New Bosses: Subordinates pay close attention to

how their boss behaves, not just toward the team col-

lectively, but especially toward them as individuals.

They are on the lookout for indications of their rela-

tive standing within the group and how well they’re

doing. They watch their boss interact with their

colleagues. They notice who the boss spends time

with, what the boss says or does not say. This ex-

treme vigilance, particularly early on, can encourage

subordinates to make too much of casual comments,

rushed feedback or perceived slights. Routine over-

sights — such as the boss’s failing to follow up on a

suggestion or remarking on a success without men-

tioning everyone who contributed — can prompt

speculation that the boss is a “phony,” a “hard-ass” or

“plays favorites.” In their rush to protect themselves

from threats, subordinates may discern false or exag-

gerated patterns in the new boss’s behavior. Such snap

judgments reflect the “fundamental attribution

error,” where people tend to latch onto and overesti-

mate dispositional or personality-based factors when

explaining the behavior of others, while underesti-

mating situational factors.6 The leader’s early

decisions will also be scrutinized for meaning. Given

the new boss’s incomplete understanding of the con-

text, some of these decisions are bound to have

unforeseen consequences. Subordinates often under-

estimate the time and attention constraints on the

new boss, and the cost of delay or of acquiring more

ample information. Consequently, if some individu-

als suffer as a result of an early decision, the boss may

ABOUT THE RESEARCHOur work in the area of boss-subordinate relations started with a field study of 50 boss-subordinate dyads working in four manufacturing operations of Fortune 100 companies.i The 50 subordinates and their bosses were interviewed at least twice and completed a questionnaire between the two interviews, in which the bosses evaluated the performance of their subordinates and the subordinates described their boss’s behavior toward them.

Several meetings and interactions were observed directly, and many individ-uals working close to these dyads were also interviewed. In total, the study involved close to 400 hours of interviews and observations, 75% of which were taped and later transcribed, providing a rich base of information.

The strength of this approach lies, of course, in its potential to illuminate a gestalt or syndrome that is a set of mechanisms — subordinate expectations, emotions and behaviors, boss cognitions and responses — that combine and develop into a self-perpetuating process. An in-depth and longitudinal case study is able to capture causalities and potential feedback loops between leader behaviors and subordinate responses that are unlikely to show up in large-scale investigations, particularly those using cross-sectional designs.

To test the generalizability of these initial findings, we discussed them with groups of executives participating in development programs, asking them to share their own experiences. These executives had come from all over the world and from every level in their respective organizations, from junior to senior, from a single company to mixed company groups. To date, we have questioned more than 5,000 international executives.

Our understanding has been further enriched through coaching and consult-ing work with many executives and their entire teams of first reports, often including discussions of each individual’s 360-degree feedback results. We believe that the combined use of in-depth case studies, interviews, field obser-vation and group discussions can yield insights and causalities that generate a clearer picture of the evolution of dysfunctional workplace relationships.

46 MIT SLOAN MANAGEMENT REVIEW SUMMER 2009 SLOANREVIEW.MIT.EDU

be assumed to have created the situation knowingly

or else to be “sending a signal.” The decision process

may also come under scrutiny. The boss’s failure to

consult or inform one or more subordinates about a

decision may be enough to trigger an impression that

the boss “doesn’t seem very open.” At the same time,

the new boss will be looking for areas of potential im-

provement. In some cases, the new leader will have

been appointed with an explicit mandate to make

changes. As a result, the new boss may be tempted to

move too fast, under the mistaken assumption that

members of the team are aware of problems and

agree on the level of urgency. Very quickly, the new

boss can elicit a negative label, branded as the “clue-

less” newcomer who wants to “change everything”

and does not understand or respect the culture.

Established Bosses: While the mislabeling of a

boss is more common early on in the relationship, it

can also happen in the midst of previously satisfactory

relationships. Established bosses may reach a point

where their style starts to irritate one or more of their

subordinates, who may begin to feel as though the

boss has “overstayed his welcome.” A hard-charging

style, for example, is terrific when the company is in

turnaround mode, but can be tiring once the ship has

been righted. A subordinate’s negative reaction may

be the result of a boss’s particular decision, action or

comment that serves as a “final straw,” or it may sim-

ply be the result of residual wear and tear in the

relationship — a buildup of toxins that can no longer

be evacuated by the social system. In another case, the

leadership needs of the situation may be perceived by

some subordinates to have changed. Arguably, this is

what happened to Scott McNealy at Sun Microsys-

tems Inc. Well suited to the mobilizing role required

during the boom years, he lost the confidence of close

colleagues as the right person to lead Sun out of the

problems caused when the tech bubble burst. Prob-

lems also may arise when the boss has led the team at

an intense pace or pushed through a lot of painful

measures and has run out of good will with some or

all of the team. At this point, it’s as though a switch has

tripped in subordinates’ minds. They lose both empa-

thy for the boss and confidence that this is the right

person to lead them into the next phase. The pep talks

that used to create a buzz and re-energize employees

may begin to feel forced and formulaic, and innova-

tive propositions may come across as disconnected

from reality. Whether for an incoming boss or an es-

tablished one, once a negative label is activated, the

situation tends to go from bad to worse. A number of

reinforcing mechanisms kick in, making it unlikely

that the negative label will be overturned.

Perceiving the Boss’s BehaviorOrganizations are complex, ambiguous and information-

rich environments. To avoid being overwhelmed,

people simplify, quickly identifying the most salient

characteristics or most plausible explanations and

discarding others. While this allows them to reduce

information-processing requirements, it leaves them

open to well-established errors known as confirmatory

biases. These biases influence the information pro-

cessing of subordinates in four ways.

1. What They Notice. Organizational reality pro-

duces a constant flow of signals — some observed

directly (both verbal and nonverbal), some reported

by others and some inferred from data. With too

much information to process, individuals in organi-

zations have to focus on what matters most. Labels

help people to filter out data that seem “less relevant.”

Classic experiments in social psychology and visual

cognition show how people’s attention is guided by

their expectations — and to what extent they can re-

main oblivious to even the most discrepant

information (like the intrusion of someone in a gorilla

suit) when they’re concentrating on other things.7 If

such blindness can occur on straightforward, time-

bound and uninterrupted observation tasks, it’s all the

more likely in situations where people are under real

pressure, tired, distracted, and have an emotional his-

tory with the person they’re observing. Under these

conditions, it is hardly surprising that subordinates

who regard their boss as meddling or coercive will

tend to notice instances when the boss interferes or

speaks abrasively, but pay much less attention when

the boss acts in a more empowering fashion.

2. What They Make of It. Organizational reality is

not just rich but also highly ambiguous, leaving con-

siderable room for different readings of the same data.

Subordinates who have an unfavorable impression of

the boss are liable to interpret the boss’s actions or com-

ments negatively. For example, should the boss favor

them in some way, they will see it as having been forced

L E A D E R S H I P

SLOANREVIEW.MIT.EDU SUMMER 2009 MIT SLOAN MANAGEMENT REVIEW 47

upon the boss (“he had no choice”); if the boss does

them a disservice, they will exaggerate the intent (“he

had lots of alternatives”). One of the psychological traps

subordinates fall into is “over-intentionalizing” —

projecting hidden intent where there is none. The

same actions can be interpreted very differently de-

pending on how subordinates view their boss. (See

“Boss Behavior From Different Angles.”)

As a result of these attribution biases, bosses may

find that they are disproportionately blamed for ac-

tions or decisions that have a negative impact on

subordinates, while they don’t receive due credit for

positive initiatives. In one case we studied, the intro-

duction of a new working-from-home policy

intended to benefit employees was viewed by some

team members as yet another cost-cutting measure,

since it reduced the company’s need for office space.

3. What They Remember. The tendency to “over-

intentionalize” leads subordinates to make negative

inferences about boss behavior. Research on mem-

ory suggests that those inferences may be stored

away as likely causes and are later retrieved as actual

causes.8 In other words, suppositions gradually

harden up to become facts. Moreover, information

that is stored away does not remain uncorrupted. It

decays; it gets confused with other memories. It

even becomes contaminated by information re-

ceived after the behavior has occurred.

Research on false memories shows that people

can “remember” things that did not happen but are

generally consistent with their view of how things

are. Memory turns out to be a reconstruction of the

past, not simply a reproduction of it.

4. What They Discuss and With Whom. The pre-

ceding cognitive biases can be reinforced by

subordinates’ interaction patterns. What begins with a

single detractor or strained relationship can easily

spread to the rest of the team. Subordinates compare

notes. When they feel let down or disrespected by the

boss, one of the first things they do is to seek a “reality

check” with a colleague. Typically, to get such feed-

back, they turn first to those who may be least inclined

to defend the boss — rather than to the boss’s infor-

mal lieutenant(s). Research shows that members of

the boss’s “out-group” spend more time than the “in-

group” discussing and analyzing the leader’s

differential treatment of subordinates.9 Of course,

their choice of informants determines the kind of

feedback they get. Their views of the boss are not only

corroborated but also reinforced by further reports of

lapses or misdeeds. Discussions around the coffee ma-

chine or water cooler ensure that disgruntled team

members develop an even more negative view of the

boss. If the boss puts a foot wrong, they will be told

about it. Over time, even subordinates who try to

maintain a more balanced view may find themselves

caught up in the general negative mood.10

Driving the Boss’s BehaviorBeyond biasing their perceptions of the boss’s ac-

tions, labels also guide subordinates’ behavior toward

the boss — and hence the boss’s responses toward

them. Consciously or not, subordinates can provoke

their bosses in two ways.

BOSS BEHAVIOR FROM DIFFERENT ANGLESOrganizations are complicated, overwhelming environments. To survive, sometimes employees oversimplify — jumping to conclusions about a boss’s behavior based on whether they were already inclined to think the boss was “good” or “bad.”

OBSERVED BEHAVIOR SUBORDINATE INTERPRETATION

Good boss Bad boss

Gives critical feedback Honest Abrasive

Makes a unilateral decision Decisive Autocratic

Instructs work to be redone Demanding Impossible to please

Imposes performance metrics Disciplined Control freak

Works weekends Driven Obsessive

Sticks with a dubious

course of action

Persistent Stubborn

Ignores his boss’s advice Self-confident Arrogant

Gives unsolicited advice Helpful Meddlesome

Asks specific questions Informed Micromanager

Delays response to proposal Reflective Unsupportive

Does not punish a mistake Compassionate Spineless

Loses temper in public Passionate Temperamental

Manages by walking about Empathic Clueless

Breaks a promise Opportunistic Untrustworthy

Limits interaction times Structured Unapproachable

48 MIT SLOAN MANAGEMENT REVIEW SUMMER 2009 SLOANREVIEW.MIT.EDU

Passive Provocation. Subordinates who begin to

see their boss as incompetent (which incoming

bosses generally are for several weeks) or difficult

may be less forthcoming about problems and less

prepared to ask the boss for help. The lack of contact

lessens the boss’s sensitivity to either the difficulties

faced or the progress made by the subordinate. It

also diminishes the boss’s ability to contribute to the

decision-making process and to obtain reassurance

that the subordinate has things well in hand. This

lack of contact is likely to raise the boss’s anxiety

about being blindsided and frustration with the

subordinate, who comes across as evasive or unco-

operative. As a result, the boss may feel compelled to

intensify monitoring and to ask more pointed ques-

tions of the subordinate, thus confirming the

subordinate’s view that exchanges with the boss are

generally unpleasant. Similarly, subordinates who find

the boss “harsh” or “unfair” may decide to discount

the boss’s critical feedback on the grounds that it is

driven by the boss’s disposition rather than their

own performance flaws.11 Of course, on realizing

that the feedback has been ignored, the boss may be

tempted to turn up the volume and issue even more

forceful instructions — thus further confirming the

subordinates’ negative view of the boss.

Active Provocation. Subordinates who suspect

that the new boss may be “just like” the old one

may test out the new boss, perhaps by bringing up

past injustices, including those related to pay or

promotion decisions. Such behavior is clearly frus-

trating for the new boss. Why should he be held

responsible for or feel obliged to repair the alleged

misdeeds of others? Yet the aggrieved subordinate

is still living with the consequences — in terms of

missed training or career opportunities, dimin-

ished status, lower pay and so on. Even if the boss

sympathizes with the subordinate, it may be diffi-

cult to “put matters right.” Hence, the new boss

may be blamed not for what he has done, but for

what he “refuses” to do — while the subordinate

quickly acquires a reputation for being “whiny” or

“aggressive.” Going one step further, subordinates

who view their boss as unreasonable have ways of

eliciting reactions from the boss that could be con-

strued as unreasonable or unfair. They may choose

to raise issues that the boss does not want raised

(past hiring or promotion decisions) or that, in the

boss’s mind, have already been settled. Subordi-

nates may make proposals in the wrong forum or

at a time when the boss simply does not have the

mental bandwidth to process what he perceives as

a digression. Subordinates who think that their

boss “never listens” can prove it to themselves and

others by making a suggestion at a time when the

boss cannot listen. Although “irrational” from the

boss’s perspective, such behavior allows subordi-

nates to deflect the blame for this dysfunctional

relationship onto their “impossible” boss and to

lessen their own responsibility. This behavior can

be curiously reassuring for subordinates and is

consistent with research into self-handicapping,

which suggests that people’s motivation to protect

their self-esteem and sense of competence will

sometimes lead them deliberately to sabotage their

own chances of success.12

CAUSAL MAPIt’s true that subordinates will sometimes form an inaccurate, negative picture of their boss. But it’s also true that bosses can make the situation worse by the way they respond. Whether they withdraw or become more coercive, bosses typically end up behaving in ways that match their subordinates’ negative expectations. The ultimate result? The negative perception becomes a reality.

Boss’sperceptions

become biased

Preconceptionsbased on past bosses

Anxiety overincoming boss

Boss is“just like”

predecessors

Fuelssubordinate

anxiety

Boss forcedto be morecoercive Activates

subordinateconfirmatory

biases

Drivessubordinate

behavior

Promptssocial

corroboration

Boss’sgood will

balance drops

Boss feels he nevergets benefitof the doubt

Boss feels rift growing

Emotionalcontagion

Bosswithdraws

Boss develops negative view of subordinate(s)

Subordinate(s) develop negative view of new boss

L E A D E R S H I P

SLOANREVIEW.MIT.EDU SUMMER 2009 MIT SLOAN MANAGEMENT REVIEW 49

How Unwary Bosses Get TrappedIt’s easy for bosses to get drawn into this process. As

they start to sense that they’re not getting through to

some of their subordinates, they fall prey to the very

same biases and blinders that afflict their subordi-

nates — including labeling, selective observation

and recall and biased attributions. Research shows

that people are more inclined to make simplistic

cognitive judgments when they are overstretched,

distracted or pressed for time.13 New bosses want to

know who they can depend on for support and will

quickly pick up on signals that certain relationships

are going to prove complicated or “high mainte-

nance.” The propensity of bosses to start sorting their

subordinates into an “in-group” and an “out-group”

is highlighted by a body of research known as leader-

member exchange theory.14 Field studies suggest that

these distinctions emerge quickly and remain sur-

prisingly stable over time.15 Once a boss has begun

to doubt the attitude or motives of certain subordi-

nates, he will distort the incoming evidence to match

the existing label. The same behavior or actions from

subordinates will therefore trigger very different at-

tributions depending on what the boss thinks of the

individual(s) concerned. Working long hours, for

example, can be interpreted as a sign of dedication

from a person the boss regards as “smart,” but as evi-

dence of an inability to prioritize or a slow mind

from someone for whom the boss has less respect. As

bosses develop a negative image of part or all of their

team, they will tend to adjust their behavior accord-

ingly. New bosses who experience difficulty in

engaging with a mistrustful team may overcompen-

sate in other directions. They may invest their

energies with other stakeholders (their bosses, cli-

ents, suppliers, analysts) or may choose to focus on

facets of their job that they have mastered (strategy,

operations, marketing, finance). In other words, they

may retreat into a comfort zone and disconnect from

the team. Another common reaction when bosses

feel a lack of engagement from subordinates is to be-

come more forceful and coercive. Bosses may react

especially badly if the signals coming back are suffi-

ciently disconnected from their own self-concept.

For example, if a boss is suspected of bias when she

has done everything possible to be transparent or is

reproached for not listening when he thinks he has

been listening, the boss’s response may be aggressive

rather than conciliatory: “You think I’m being harsh?

I’ll show you harsh!” At which point the subordinates

respond, “We rest our case.” Whether they withdraw

or become more coercive, bosses typically end up be-

having in ways that match their subordinates’

negative expectations. While a subordinate may be

primarily responsible for triggering the process, both

parties keep it going with their mutual biases — and

become locked in a double self-fulfilling process.

(See “Causal Map.”)

Because the resulting vicious circles are difficult

to interrupt and unlikely to self-correct, preventing

their occurrence must be a priority for bosses.

Implications for BossesOur research shows that bosses must be mindful of,

and should explicitly manage, their own labeling,

expectations and biases. However, that is not

enough. For boss-subordinate relationships to

work, bosses have to mind not only their own men-

tal processes but also those of their subordinates.

To avoid acquiring an undeserved negative label,

individual bosses must take four important steps.

1. Understand the Situation. New bosses need to be

aware of what they are walking into, both in business

and human terms. Is this a turnaround, a realignment

or an upholding-success challenge? Prior to taking

charge, bosses must find out whether there is a per-

ceived need for change in their organization and, if

so, how widely shared is that view. Prospective bosses

should also ask how their arrival is likely to be perceived.

What was their predecessor like, and what were the cir-

cumstances of the predecessor’s exit? Coming after a

weak boss, a tough boss or someone who has grown

stale, the newcomer is likely to be welcomed and to

start out with a positive “opening balance.” Conversely,

the new boss who comes after several bad bosses may

start out “in the red.” Similarly, when succeeding a well-

liked boss who was removed for not being hard-nosed

enough, the new boss can expect a cautious reception.

Incoming leaders need to figure out where they stand.

When Bob Eckert took over as CEO of Mattel Inc., he

asked his top team to submit anonymous questions.

One question read: “I’ve heard you are an in-the-

trenches manager who listens to the lower levels. Does

this mean you’ll go around us and make decisions with-

out involving us?” Eckert realized, “I had completely

50 MIT SLOAN MANAGEMENT REVIEW SUMMER 2009 SLOANREVIEW.MIT.EDU

underestimated their wariness … They had done as

much homework on me as I had done on Mattel.”16

Fortunately, a question-and-answer session allowed

him to dispel that preconception.

2. Invest Early in Subordinates. New bosses must

spend significant time one-on-one with subordi-

nates for three reasons.

To understand them. The new boss needs to

find out how subordinates see the company’s chal-

lenges; how they related individually to the previous

boss and what they expect of the new boss; and

what strengths, weaknesses and other characteris-

tics they bring to their jobs.

To get to know them. Incoming leaders can

choose to be either more or less involved in the cre-

ation of their subordinates’ first impressions.

Frequent contacts in the developmental stages of

the relationship help bosses to establish the rules of

engagement with their subordinates, clarify their

expectations and explain key aspects of their lead-

ership style. This kind of clarity goes a long way

toward preventing bad dynamics, which are so

often fueled by unstated expectations and misun-

derstandings over priorities.

To establish a rapport. Time is the new leader’s

scarcest resource. Investing precious time in individ-

uals signals the boss’s commitment to them.

Developing a rapport with subordinates reassures

them that employees will be respected as individuals

even if their performance falls short of their own and

their boss’s expectations. It also helps to decrease anxi-

ety and defensiveness associated with feedback, as well

as reluctance to approach the boss for advice.

3. Be Mindful of One’s Own Behavior. New lead-

ers often overestimate the extent to which their good

intentions and good character will shine through.

Demonstrating one’s “authentic self” does not mean

“being natural.” Rather, it requires managers to seize

everyday opportunities to demonstrate that they are

trustworthy, supportive and fair. The subordinate of

one outstanding leader in our study recalled: “When

I first started working with him, I had the feeling that

he was not very open to employee input. So I would

go: ‘Yes, yes, OK. I’ll do that. Whatever you say.’ And

the second or third meeting he said: ‘You can push

back on me, you know. I’m not always right.’ So I’ve

tested him on this over time, and he’s never reacted

badly.” Managers need to establish and maintain a

positive goodwill balance with their subordinates,

both individually and collectively. They draw on that

store of good will every time they ask subordinates

for special efforts, impose their will or make “un-

forced errors.” To be able to afford such withdrawals,

managers need to make regular deposits — espe-

cially by taking actions that help subordinates in

their work, demonstrating openness to subordinate

input or supporting them in their careers. The tim-

ing of these deposits also matters. A few well-chosen

efforts early on in the relationship will create mo-

mentum and help the boss to acquire a positive label.

Once team members see the leader through a benign

lens, everything becomes simpler for the boss: posi-

tive actions receive more goodwill credit, while

unexpected demands and errors incur smaller losses.

Conversely, as we have seen, early errors can prove

particularly damaging.

4. Intervene Early. When bosses feel that they’re

not getting the credit they deserve from subordinates

or sense that certain subordinates are not engaging,

they have to make an effort on two fronts:

Beware of labeling. Some subordinates may be

more marked by past experiences than others and

may take longer to come around to the new leader’s

approach. This puts the onus on the boss to maintain

sufficient mental bandwidth to react productively to

the reticence of these subordinates and resist jump-

ing to hasty conclusions about them or writing them

off prematurely. Bosses must remind themselves why

they want to give subordinates — particularly those

who had complicated interactions with prior leaders

— a real chance to connect, and must realize that this

process cannot be instantaneous.

Act quickly. If in spite of the leader’s proactive ef-

forts a malaise seems to be developing, it is critical to

act quickly. Managers often withhold negative feed-

back at the outset, assuming that it might spoil the

development of a working relationship. But if a sub-

ordinate is doing something that concerns the boss,

that employee needs to be told sooner rather than

later. While it may not be pleasant for either party,

corrective feedback delivered early can be accepted as

part of the normal adaptation process. Delayed inter-

vention only raises the threat and embarrassment

L E A D E R S H I P

SLOANREVIEW.MIT.EDU SUMMER 2009 MIT SLOAN MANAGEMENT REVIEW 51

attached to the issue. It gives the feedback a punitive

edge and reduces the chances that the subordinate

will react constructively. Often, what prevents bosses

from entering this discussion is the assumption that

they know what is driving the behavior or outcome

they perceive, be it lack of skill, judgment or effort.

When bosses make the effort to check, they often dis-

cover that they have misread the situation,

misunderstood the subordinate’s conduct or motives

or overlooked mitigating factors (e.g., lack of train-

ing, lack of time, too many responsibilities, unclear

instructions or personal circumstances, such as fam-

ily problems). The fact that the boss is willing to

double-check to see if his initial reaction is right gives

employees proof that they will be evaluated fairly.

Forewarned Is ForearmedWhile bosses certainly need to work on their own be-

havior, the other side of the equation is equally

important: subordinates are not blameless or power-

less victims. The fact is that they have collective

histories and individual sensitivities, as well as anxi-

eties and mental biases, and that these sometimes

can make it extremely difficult for bosses to be effec-

tive leaders. Bosses, therefore, must try to understand

the interpersonal context they are walking into be-

fore they get sucked into a vicious circle with one or

more subordinates. As Mahatma Gandhi famously

put it: “The moment there is suspicion about a per-

son’s motives, everything he does becomes tainted.”

Jean-François Manzoni is a professor of leadership and organizational development at IMD in Lausanne, Switzerland. Jean-Louis Barsoux is a senior research fellow at IMD. Comment on this article or contact the authors at smrfeedback@mit.edu.

REFERENCES

1. For example, J.F. Manzoni and J.L. Barsoux, “The Set-Up-to-Fail Syndrome,” Harvard Business Review 76, no. 2 (March-April 1998): 101-113.

2. R.C. Liden, S.J. Wayne and D. Stilwell, “A Longitudinal Study on the Early Development of Leader-Member Exchanges,” Journal of Applied Psychology 78, no. 4 (1993): 662-674.

3. L.M. Andersson, “Employee Cynicism: An Examination Using a Contract Violation Framework,” Human Relations 49, no. 11 (1996): 1395-1418.

4. K.J. Dunegan, “Leader-Image Compatibility: An Image Theory View of Leadership,” Journal of Business and Management 9; no. 1(winter 2003): 61-77.

5. J.P. Forgas, “Affect and Information Processing Strate-gies: An Interactive Relationship,” in “Feeling and Thinking: The Role of Affect in Social Cognition,” ed. J.P. Forgas (Cambridge: Cambridge University Press, 2000), 253-280.

6. L. Ross, “The Intuitive Psychologist and His Shortcom-ings: Distortions in the Attribution Process,” in “Advances in Experimental Social Psychology,” ed. L. Berkowitz (New York: Academic Press, 1977), 173-220.

7. For example, H. Kelley, “The Warm-Cold Variable in First Impressions of Persons,” Journal of Personality 18 (1950): 431-439; and D.J. Simons and C.F. Chabris, “Goril-las in Our Midst: Sustained Inattentional Blindness for Dynamic Events,” Perception 28 (1999): 1059-1074.

8. S.L. Hannigan and M.T. Reinitz, “A Demonstration and Comparison of Two Types of Inference-Based Memory Errors,” Journal of Experimental Psychology 27, no. 4 (2001): 931-940.

9. P.M. Sias, “Constructing Perceptions of Differential Treatment: An Analysis of Coworker Discourse,” Com-munication Monographs 63 (1996): 171-187.

10. S.G. Barsade and D.E. Gibson, “Why Does Affect Matter in Organizations?” Academy of Management Perspectives 21, no.1 (2007): 36-59.

11. K. Leung, S. Su and M. Morris, “When Is Criticism Not Constructive? The Roles of Fairness Perceptions and Dis-positional Attributions in Employee Acceptance of Critical Supervisory Feedback,” Human Relations 54, no. 9 (2001): 1155-1187.

12. E.R. Hirt, R.K. Deppe and L.J. Gordon, “Self-Reported Versus Behavioral Self-Handicapping: Empirical Evidence for a Theoretical Distinction,” Journal of Personality and Social Psychology 61, no. 6 (1991): 981-991.

13. D.T. Gilbert, B.W. Pelham and D.S. Krull, “On Cogni-tive Busyness: When Person Perceivers Meet Persons Perceived,” Journal of Personality and Social Psychology 54, no. 5 (1988): 733-740.

14. R.M. Dienesch and R.C. Liden, “Leader-Member Exchange Model of Leadership: A Critique and Further Development,” Academy of Management Review 11 (1986): 618-634; and, more recently, M.C. Bolino, “What About Us? Relative Deprivation Among Out-Group Members in Leader-Member Exchange Relationships,” Academy of Management Proceedings (2007): 1-5.

15. F. Dansereau, G. Graen and W. J. Haga, “A Vertical Dyad Linkage Approach to Leadership Within Formal Organizations: A Longitudinal Investigation of the Role Making Process,” Organizational Behavior and Human Performance 13 (1975): 46-78.

16. R.A. Eckert, “Where Leadership Starts,” Harvard Business Review 79, no. 10 (November 2001): 53-60.

i. J.F. Manzoni, “Use of Quantitative Feedback by Superiors: Causes and Consequences” (D.B.A. diss., Harvard University Graduate School of Business Admin-istration, 1993).

Reprint 50434.

Copyright © Massachusetts Institute of Technology, 2009.

All rights reserved.

Bridging Faultlines in Diverse Teams

S U M M E R 2 0 0 7 V O L . 4 8 N O. 4

R E P R I N T N U M B E R 4 8 4 1 1

Lynda Gratton, Andreas Voigt and Tamara Erickson

Please note that gray areas reflect artwork that has been intentionally removed. The substantive content of the article appears as originally published.

22 MIT SLOAN MANAGEMENT REVIEW SUMMER 2007

O r g a n i z at i O n

ompanies create diverse teams to take on their most complex challenges —

tasks across boundaries, functions and geographies that no single department

or function could accomplish. Yet guiding these diverse teams to success re-

quires some counterintuitive management practices. In particular, team

leaders should focus on tasks at the early stages, rather than on interpersonal relationships,

and then switch to relationship building when the time is right.

In a recent study of teams at large companies, we found that diverse membership of

teams and task forces is becoming the order of the day. Take Nokia Corp., which frequently

brings together disparate talent from different departments among its businesses around

the globe, while at the same time partnering with many external suppliers. Or consider the

British Broadcasting Corporation, which routinely creates huge teams for events, such as

the production and broadcast teams for the 2006 FIFA World Cup and the 2008 Olympic

Games. These typically involve groups of more than 100 people, a high proportion of whom

are not full-time employees. Team members often represent more than 15 different nation-

alities, with skill sets ranging from electrical work to intellectual property, from scheduling

to production. The BBC’s teams also face the daunting challenge of a one-shot deal for

which execution has to be right the first time.

The challenges that Nokia and the BBC face are by no means unique. Between 2004 and

2006, we partnered with executives from 15 large European and American companies to study

55 of their teams. What was most striking about these teams was their sheer size and complex-

ity. The diversity of Nokia’s design team — with men and women of many nationalities and

with a wide range of ages, representing multiple functions from many different businesses

— was repeated in companies in many different industries from across the globe. Companies

in the media industry (such as Reuters Group PLC and the BBC), in telecommunications

(such as France Telecom and Canadian wireless giant Rogers Communications Inc.) and in

banking (such as Royal Bank of Scotland and Lehman Brothers Inc.) all employ large and

diverse teams to attack some of their most difficult problems. Many of the teams in the study

numbered more than 50 people, all had more than three nationalities represented and most

brought in people from several functions and businesses. (See “About the Research,” p. 24.)

Bridging Faultlines in Diverse Teams

Lynda Gratton is professor of management practice at London Business School. She is the author of Hot Spots: Why Some Teams, Workplaces and Organizations Buzz With Energy — and Others Don’t (San Fran-cisco: Berrett-Koehler Publishers, 2007). Andreas Voigt is a research assistant in organizational behavior at London Business School. Tamara Erickson is president of the Concours Institute, the research and education arm of the Concours Group, a professional services company headquartered in Kingwood, Texas, and a coau-thor of Workforce Crisis: How to Beat the Coming Shortage of Skills and Talent (Boston: Harvard Business School Press, 2006). Comment on this article or contact the authors through smrfeedback@mit.edu.

CProject teams can fly

or founder on the

demographic attributes

of team members and

the fractures they can

create. Here’s how to

recognize the potential

for division, and how

to respond in time when

team fractures do arise.

Lynda Gratton,

Andreas Voigt and

Tamara Erickson

SLOANREVIEW.MIT.EDU

SUMMER 2007 MIT SLOAN MANAGEMENT REVIEW 23

Companies form these teams precisely because they bring to

bear the range of experiences and attitudes that will ensure that

the final product or service will be market-sensitive and innova-

tive — more so than an offering designed by a group of people

with similar characteristics.1 Paradoxically, however, the very

nature of team diversity often creates challenges that reduce the

team’s innovative capacity and even significantly lessen its overall

effectiveness. So while business heads may wish for innovation

through diversity, what they sometimes achieve is reduced pro-

ductivity and efficiency.2

How Diverse Teams Fail, and How Some SucceedThe teams in our study failed in many different ways. Some

could not deliver on time; others fell short of the hoped-for

productivity; still others were unable to produce innovative re-

sults. Some teams broke up in acrimony and

bad feeling; some foundered in incompe-

tence. They were a litany of what could go

wrong. Many executives had been trained to

manage rather simple teams but were sur-

prised at just how hard it is to create a

high-performing team of diverse people.

Analyzing these problems uncovered two

root failures. The first was a failure of collabo-

ration, in which team members did not

develop trust and goodwill among them-

selves.3 The second was a failure of knowledge

sharing, in which team members withheld

their individual knowledge from other team

members or from other teams.4

Why is it so fiendishly difficult to enable

high-performing, diverse teams? To under-

stand the nature of the collaborative and

knowledge-sharing failures, we looked closely

at the demography of the teams. Using com-

plex statistical modeling, we created a unique

demographic profile for each team that took

into account team members’ ages, genders,

nationalities, education levels, functions and

tenures within the company. The model ex-

amined the configurations as a whole and the

interplay of different demographic attributes.

Close examination of these team dem-

ographic profiles revealed that in many cases

the failures in collaboration and knowledge

sharing were a direct result of faultlines —

subgroups or coalitions that emerge naturally

within teams, typically along various dem-

ographic lines. (See “An Overview of Faultline

Theory,” p. 25 for causes and characteristics

of faultlines.) These faultlines split the teams into subgroups

that were based on shared demographic characteristics. When

faultlines emerge, subgroups rarely collaborate with other sub-

groups, instead tending to share knowledge only within their

subgroup. Yet knowledge sharing across subgroups is critical for

complex teams to operate effectively.5

Despite the emergence of faultlines in many of the 55 teams,

some teams were able to work collaboratively and share knowl-

edge. Deeper analysis of these productive and innovative teams

showed that a defining factor was the behavior of the team leader

and the way in which she or he structured the leadership role.

Thus, although faultlines are a common hazard, some executives

are able to reduce the problems associated with diverse teams

and, indeed, to enhance the benefits.

In particular, the executives’ leadership style and the manner

24 MIT SLOAN MANAGEMENT REVIEW SUMMER 2007

in which they prioritized actions significantly reduced the extent

to which faultlines hindered collaboration and the flow of knowl-

edge. By studying the way these executives behaved, we are able

to make recommendations about leading diverse teams. How-

ever, while some of the recommendations are straightforward,

others are deeply counterintuitive and defy the received wisdom

about good management practice.

The Emergence of FaultlinesTo illustrate this seeming paradox, consider a team from a tele-

com company in our study. Think of the faultlines in this team

as analogous to geologic fractures in the Earth’s crust. Like geo-

logic fractures, faultlines can remain dormant and invisible for

some time. Geologic fractures explode as earthquakes when put

under immense pressure. The same is true of team faultlines,

which cross multiple layers of demography. In many cases, the

tensions of the faultlines emerge under the pressure of a com-

plex, time-dependent task.

The stated goal of the telecom team was to bring together

several parts of the business to build an innovative product and

service offering for one of the company’s major multinational

clients. Many of the client’s needs were standard, but meeting

their expectations would require the creation of a rather complex

service and delivery process. A business unit head was assigned

leadership of the new project and given charge of a core group of

22 people. She had to bring together the remainder of the team,

which was expected to total 48 people, from the four countries in

which the client had significant operations (Germany, the United

States, Japan and France), drawing from three functions (tech-

nology, marketing and operations). Although this may seem a

rather large and diverse team, many of the teams studied were of

a comparable size and complexity.

Initial Faultline Formation Is Based On Surface-Level Attributes The telecom team members began to get to know one another

through face-to-face encounters and e-mails. Within a short

time, however, the initial faultlines began to emerge. These fault-

lines were drawn around readily detectable demographic

attributes — for example, the team members’ age, gender and

functional background. Team members typically use obvious

characteristics to assign themselves and others to subgroups. In

this team, an early faultline emerged between a subgroup of male

technical engineers and a subgroup of female marketing special-

ists. Other subgroups rapidly emerged: for example, between

people of French nationality (many of whom had been in the

company for years and were typically in their 40s and 50s) and

Americans (many of whom had joined the company recently and

were in their 20s and 30s). Note that none of these subgroups was

based on a unitary dimension but rather on the combination of

multiple dimensions (gender and function in the first case; na-

tionality, age and job tenure in the second). A strong faultline

emerged because the team members fell into distinct, nonover-

lapping subgroups based on demographic attributes. (See “The

Emergence of Faultlines in a Telecom Team,” p. 26.)

As a rule, when subgroups emerge within complex teams, each

tends to see itself as an “in-group” — people like us, who we like

because we have something in common — and those across the

boundary of the faultline as an “out-group” — people not like us,

whose interests we may find puzzling. Each subgroup within the

telecom team collaborated closely among its members, who all

got to know, like and trust one another. In a sense, what each

subgroup was doing was learning more about what its members

already knew. This deepening of knowledge can be crucial as a

team builds professional insight and muscle. However, for a team

to become truly innovative, the combination of knowledge across

subgroups is essential.6

As members of the telecom team’s subgroups began to iden-

tify more strongly with one another, the team leader came to

understand that the emerging faultlines had negative conse-

quences. The dealings across the subgroups became a source of

tension and conflict, particularly under the pressure of task dead-

O r g a n i z at i O n

SLOANREVIEW.MIT.EDU

We studied 55 workgroups in 15 European and American

companies (ABN AMRO, BBC, BP, Citigroup, France Telecom,

Lehman Brothers, Marriott, Nokia, PricewaterhouseCoopers,

Reuters, Rogers Communications, Royal Bank of Scotland,

Siemens AG, Standard Chartered Bank and XL Global Serv-

ices). The study was based on a quantitative survey of

1,543 members of the workgroups and their leaders. The

teams ranged in size from four to 184 people, with an av-

erage team size of 43 members. Most of the teams were

diverse in terms of gender (32% women), age (ranging

from 4% under 25 to 15% over 46), nationality (26% Brit-

ish, 23% American, 23% European, 15% Asian and 13% rest

of world) and education level (7% high school diploma,

50% undergraduate degree, 37% master’s degree, and 6%

doctorate). We achieved a 64% response rate.

We measured demographic faultline strength along six

demographic variables: gender, age, nationality, educa-

tion, work function and tenure in the company. We

applied a statistical procedure that produced an overall

quantitative index of faultline strength by combining the

effects of all individual attributes. We used multiple re-

gression analyses to identify the faultlines and to control

for the size of the group, the degree of task complexity,

the geographical distance between team members and

industry-level differences.

About the Research

SUMMER 2007 MIT SLOAN MANAGEMENT REVIEW 25

lines. For example, the predominately male engineers at one stage

refused to allow the predominately female marketing group ac-

cess to some of their product findings. The story was that this

information was too technically sophisticated for the marketing

function; the reality was that the engineering subgroup did not

want to let go of their valuable knowledge. The subgroups just

didn’t seem to understand one another — they did not know the

other’s interior language, and they did not understand the other’s

key concepts and models.

Deep-Level Attributes Come Into Play At Later Stages of Team Development As complex teams develop and group members get

to know more about one another, a deeper layer of faultlines

becomes visible — this time based not on surface-level attri-

butes but rather on subtler, deep-level attributes such as personal

values, dispositions and attitudes. At the time of a team’s forma-

tion, these deeper-level attributes are not visible. They emerge as

team members interact with one another and reveal themselves

through their actions, words and what they choose to disclose

about their personal lives.7

In the telecom company team, a later set of subgroups also

formed around values, disposition and

attributes. A faultline surfaced between

those members with cooperative values

and a subgroup of people with more

competitive values.

The Leader’s Role in FaultlinesIn the geological analogy to faultlines,

various external factors (such as pres-

sure) have an impact on how a fault

actually fractures. Similarly, many as-

pects of a team’s context can affect the

extent to which faultlines impact the

team’s performance. Two examples are

the extent of the cooperative culture in

which the team operates and the de-

gree to which team members believe

senior executives work across bound-

aries. The most important factor in

determining whether destructive fault-

lines emerged was the style of leader,

and in particular the extent to which

the group’s leader was task-oriented or

relationship-oriented.8 Some leaders in

the study were able to adjust their lead-

ership styles as the project progressed,

beginning with a task orientation and

then switching to a relationship orien-

tation, or beginning with a relationship

orientation and switching to a task orientation. (See “The Four

Paths of Leadership Style,” p. 27.)

Path 1: Task Orientation In this pathway, the team leader uses a

strong and consistent task-oriented leadership style, as perceived

by team members, during the entire life of the team or project.

The leader can do this by creating a detailed project plan, build-

ing tight schedules for the work and emphasizing performance

goals that are high but realistic. This type of leader places great

emphasis on the task at hand, so he or she strives to remain ac-

cessible at all times and provide information that team members

need to carry out their day-to-day work. Leaders who follow this

pathway are often technically proficient — and they see their role

as providing the team with the technical and specialist assistance

critical to the task.

Path 2: Relationship Orientation This type of leader places par-

ticular emphasis on the culture of the team and on the extent

and depth of relationships among team members. They do this

by treating team members with kindness and respect, encourag-

ing a climate of trust and cooperation and providing recognition

SLOANREVIEW.MIT.EDU

Faultline theory explains how a combination and the configuration of the attributes of

team members can influence the team’s behavior and ultimately its performance. The

attributes that drive faultlines can be surface-level or deep-level. Readily detectable at-

tributes such as gender, age, nationality and education are surface-level. Underlying,

or deep-level, attributes include values, personality and knowledge.

Strong faultlines emerge in a team when there are a few fairly homogeneous

subgroups that are able to identify themselves. Weak faultlines can emerge in two

rather different configurations. Faultlines generally do not emerge, or do so only

weakly, when all the members of the team are rather similar (for example, the same

age and job function). However, at the other end of the spectrum, faultlines also are

unlikely to emerge when there is heterogeneity — when the members of the team

are all very different from one another (for example, different ages and working in

different job functions).

Strong faultlines are particularly likely to emerge when all the demographic attrib-

utes of the members of the subgroups form distinct, nonoverlapping categories. For

example, a strong faultline will emerge if all women in a team are over 50 years old and

all the men are under 30. In this example, gender and age have formed a single, strong

faultline.

Whether a faultline emerges depends on how apparent the attribute is to members

of the team. As the team members get to know each other better and learn what is

similar and what is different, the possible sources of potential faultlines increase.

Strong faultlines can create a fracture in the social fabric of the team. This fracture

can become a source of tension and a barrier to the creation of trust and goodwill and

to the exchange of knowledge and information. Inevitably, therefore, when strong

faultlines emerge, the team’s creative and innovative capacity is severely limited.

An Overview of Faultline Theory

26 MIT SLOAN MANAGEMENT REVIEW SUMMER 2007

and appreciation for individual and group accomplishments.

These leaders are often skilled communicators and listeners.

Path 3: Task Orientation, Switching to Relationship Orientation In this

pathway, the leader begins with a strong task-oriented style by

setting targets and scheduling work. As the project progresses,

these leaders encourage team members to collaborate with one

another and work to increase the general trust and goodwill

within the team.

Path 4: Relationship Orientation, Switching to Task Orientation This

type of leader begins by creating a feeling of trust within the

team, putting an emphasis on socialization and meetings. As the

project progresses, these leaders move to a more task-oriented

approach by setting clear goals and standards and carefully

monitoring the group’s progress.

Which, if any, of these leadership paths is most appropriate

when there are strong faultlines in a team? Where it is likely

strong faultlines will emerge, the natural tendency of many lead-

ers is to encourage team members to come together through

meetings and socializing. In effect, these leaders take Path 2.

However, this leadership action actually increases the likelihood

that faultlines will strengthen: when team members simply so-

cialize, their differences become more apparent, and the fractures

in the team can solidify.

In fact, a leader can significantly mitigate faultlines — but not

in the most obvious manner. To increase collaboration and

knowledge sharing across teams with strong faultlines, leaders

need to vary their leadership style according to how long the

team has been together. There are times when a task-oriented

style works very well and other times when a relationship orien-

tation would be more appropriate.

Recommendations For Leading Diverse Teams The leaders of complex teams should take four actions:

1. Diagnose the Probability of Faultlines Emerging At the outset of

a project, team leaders should think very carefully about the

diversity in their team and strive to predict as accurately as pos-

sible the probability of faultlines emerging. (See “The

Probability of Strong Faultlines Emerging in a Team,” p. 28 for

a short questionnaire that provides a way of gauging this likeli-

hood.) It is important to remember that faultlines are not a

natural result of diversity per se but are found in situations of

moderate diversity, when a team is neither very homogeneous

nor very heterogeneous in member attributes. A medium de-

gree of diversity leads to the emergence of only a

few fairly homogeneous subgroups. The divides

between these subgroups create the tensions that

can impede the team’s functioning.

If there is a high probability of subgroups and

faultlines emerging, then the leader should empha-

size a task-oriented leadership style in the early

stages of the project.

2. Focus on Task Orientation When a Team Is Newly Formed As previously mentioned, the inclination

for many leaders when they saw faultlines emerg-

ing in a new team was to focus on the relationships

between members of the subgroups. The leaders

created opportunities for people to get to know

one another better, hoping that socializing would

cause the faultlines to be bridged. Yet this exacer-

bates the problem. Simply put, in a team’s early

going, the more people interact with one another,

the more likely they are to make snap judgments

and to emphasize their differences.

A better strategy for leaders of teams with poten-

tial faultlines is to create energy around the task

itself. This was clearly visible in teams at the Royal

Bank of Scotland. In these teams — even those with

strong faultlines — collaboration and knowledge

sharing were strengthened through a host of task-

O r g a n i z at i O n

SLOANREVIEW.MIT.EDU

Faultlines quickly emerged around obvious demographic attributes with

the formation of a team in a telecommunications company. None of the

subgroups was based on a unitary dimension but rather on the combina-

tion of multiple dimensions (gender and function in the first case;

nationality, age and job tenure in the second).

The Emergence of Faultlines in a Telecom Team

Male Engineers

Female Marketers

American

French

SUMMER 2007 MIT SLOAN MANAGEMENT REVIEW 27

oriented characteristics. At the beginning of team formation,

leaders created very detailed descriptions of realistic perfor-

mance goals. Next, the work was planned and scheduled with

precision: Every project was on a 30-day, 60-day or 90-day time-

table. At this early stage, much effort was focused on providing

the necessary resources and coordinating team members’ activi-

ties. Team members learned about one another’s skills and

competencies rather than about their personalities and lives.

This task orientation focused the attention of team members on

performance and requirements. Subgroups did emerge, but they

revolved around task-oriented characteristics such as functional

expertise and education, rather than personality differences. By

learning who they could go to for particular types of informa-

tion, the subgroups at Royal Bank of Scotland were able to move

swiftly into the task itself

However, while this approach increases the early effectiveness

of teams, it is not as useful for dealing with some of the tensions

that later emerge, such as around deeper personality traits and

differences in values. To do this, team leaders must learn how and

when to switch leadership styles.

3. Learn When to Make the Switch Focusing on the task is crucial

to the early effectiveness of a team in which strong faultlines are

expected to emerge. However, if the team is to be effective in the

longer term, then the leader has to switch styles from task ori-

entation to relationship orientation. If the leader fails to make

this switch, the team will slowly become less effective, as ob-

served in one of the media teams in our study. The team leader

began with a task orientation, coordinating the

team’s activities, creating clear schedules and

providing technical support. This support from

the top initially ensured that the subgroups

learned more about each other and began the

process of sharing knowledge. At the same

time, the leader’s capacity to clearly state the

team’s mission and create a common goal en-

sured that the subgroups became strongly

aligned to a common purpose despite the fault-

lines already running through the team. The

members of the team began to feel that they

were united in a goal that was greater than the

differences among the subgroups.

Over time, however, deep-level faultlines in

this complex and diverse team began to emerge

and become increasingly important. The major

faultlines initially had formed around func-

tional specialization and nationality. The

media team’s creative designers came mostly

from the West Coast of the United States, while

the production teams were located on the East

Coast and in Germany. These differences were bridged by the

clear sense of a shared task and goals created by the team

leader. Later, though, subgroups began to form around the at-

tributes and personalities of group members. A subgroup of

high-energy, highly competitive people (known as Type A per-

sonalities) began to form. These are people who like to work

under pressure and create pressure for others, who enjoy the

rush of adrenaline and love competing.9 Members of this sub-

group were drawn from the West and East coasts of the United

States and from Germany; some were men, some were women;

some old, some young. Their deep-level Type A characteristic

cut across the surface-level characteristics.

As the original faultlines began to be bridged through a

shared task, this new personality-based, deep-level faultline be-

came an increasing source of conflict and tension. Those outside

the Type A category created names for them (“the crazies,” “the

no-lifers”), while the Type A subgroup became frustrated by

what they saw as slowness and a lack of focus from the others.

The task-oriented leader — who also had a Type A personality

— failed at this point to switch styles, instead continuing to plan,

organize, schedule and create tasks. Since the Type A subgroup

was the most vocal (and the team leader was a member of the

subgroup), the schedules simply became faster- and faster-paced

and the task demands more and more frequent. Conflicts arose

often, and those outside the Type A subgroup became more iso-

lated, demotivated and unhappy.

This discord came to a head during a performance meeting.

The business unit head overseeing the project criticized the team

SLOANREVIEW.MIT.EDU

The most important factor in determining whether destructive faultlines

emerged in a team was the style of its leader and, in particular, the extent

to which the leader acted along a continuum of task orientation and rela-

tionship orientation.

The Four Paths of Leadership Style

Time

Task-Oriented

Relationship-Oriented

Path 1

Path 4

Path 3Path 2

28 MIT SLOAN MANAGEMENT REVIEW SUMMER 2007

leader for the lack of innovative ideas coming from the team.

The team was fast-paced, but the results were often boring and

predictable. Looking back, this outcome should have been no

surprise: Some of the most creative members of the team were

not among the Type A subgroup, whose members were calling

all the shots. Yet those outside the dominant subgroup lacked the

power to slow down the program, argue for time for reflection

or get their creative ideas heard and discussed.10

The team leader had failed to read the signs that called

for switching from task-oriented leadership to relationship-

oriented leadership. The Type A subgroup was able to dominate

by overly influencing the agenda and taking most of the

resources.

4. Switch to Relationship Building When the Time Is Right In the

media company example, the team leader had built an effective

team but not an innovative team. The leader had failed to make

the switch that characterizes Path 3 of leadership.

A team leader did follow the third path in one of the fi-

nance companies studied. At one point, tensions were coming

to the fore around different values and personalities among

team members. The team leader — who previously had been

predominately task-oriented — was able to switch to a more

relationship-oriented style. Over the course of a few weeks, the

leader brought the team together for several social activities,

surfaced and talked openly about the tensions the members of

the team felt and showed respect to the various work styles and

values in the group.

Developing Better Teams and Team LeadersSo how does a team leader know when to switch from task ori-

entation to relationship orientation? The switch will be

successful only at the point at which the team has sufficient

shared experience to have developed a clear protocol for com-

munication and coordination of activities and an established

operational structure.11

As a guideline, when all members of a team have developed

specific expectations for the project and have negotiated a

O r g a n i z at i O n

SLOANREVIEW.MIT.EDU

For team leaders: This short survey will show you the probability

of a strong faultline emerging in your team. Rate your team

members against these four elements:

1. The number of nationalities in the team

a. Team members are all of the same nationality.

b. There are two nationalities.

c. There are three to five nationalities.

d. There are six to 10 nationalities.

e. There are more than 11 nationalities.

2. The current age, education and gender of the team members

a. The majority are the same gender and about the same age

and have the same education level.

b. The majority are the same gender and have the same educa-

tion level but are of different ages.

c. The majority are of the same age and have the same educa-

tion level and are both men and women.

d. The majority are the same age and gender and have differ-

ent education levels.

e. The team contains both men and women of different ages

and education levels.

3. The current business location of team members

a. They are all from the same function and the same business.

b. They are all from different functions in the same businesses.

c. They are all from different functions and different busi-

nesses within the company.

d. They are from different businesses and functions in the

company and from longstanding partners from outside the

company.

e. They are from different parts of the company and include

longstanding partners of the company and new partners

and customers of the company.

4. The values and aspirations of the team

a. All members of the team have very similar values, disposi-

tions and attitudes.

b. Many members of the team share values, dispositions and

attitudes.

c. There is a clear divide between groups with regard to values,

dispositions and attitudes.

d. Many people have different values, dispositions and attitudes.

e. There is a great deal of variety in values, dispositions and

attitudes.

Low probability of faultlines emerging:

Scoring mostly a’s and b’s — The team is relatively simple and

homogeneous, so it is unlikely that faultlines will emerge.

Scoring mostly e’s — The team is heterogeneous. Therefore, it is

unlikely that faultlines will emerge.

High probability of faultlines emerging:

Scoring mostly c’s and d’s — This team has a high potential for fault-

lines to emerge because a medium degree of diversity tends to lead

to the development of only a few fairly homogeneous subgroups.

The Probability of Strong Faultlines Emerging in a Team

SUMMER 2007 MIT SLOAN MANAGEMENT REVIEW 29

widely accepted influence structure, then the time is right to

switch to a relationship-oriented leadership style. Instilling

confidence in the team and creating opportunities to socialize

at that point helps the development of new abilities and allows

the team to grow. However, if the team is still trying to learn

the specifics of the project, clarify people’s roles and negotiate

members’ status and authority, then the switch would come

too early and would only amplify the underlying tensions be-

tween subgroups.

This provides important guidelines for managers who head

diverse teams, which are tasked with some of the most impor-

tant and difficult challenges that companies face. But it also

points to an important new development challenge for execu-

tives. Not only do future leaders need to develop solid program

management skills and confident interpersonal skills but they

also need to learn which leadership style to emphasize, based on

their team’s needs and characteristics. Providing team leaders

with a framework for assessing faultlines will give them the in-

sights needed to overcome the divisions and move teams toward

achieving their, and the company’s, ultimate goals.

REfEREncEs

1. The power of integration on innovation has been described in L. Gratton, “Managing Integration Through Cooperation,” Human Re-source Management 44, no. 2 (2005): 151-158; and S. Ghoshal and L. Gratton, “Integrating the Enterprise,” MIT Sloan Management Review 44, no. 1 (fall 2002): 31-38.

2. Richard Hackman’s work on group dynamics provides early in-sights into the formation and maturation of teams; see, for example, J.R. Hackman, ed., “Groups That Work (and Those That Don’t): Cre-ating Conditions for Effective Teamwork” (San Francisco: Jossey-Bass, 1989). The difficulties of managing complex teams are described in L. Gratton, “Hot Spots: Why Some Teams, Workplaces and Organizations Buzz With Energy — And Others Don’t” (San Francisco: Berrett Koehler Publishers, 2007).

3. The way in which teams collaborate with each other is increasingly seen as central to their effectiveness. An overview of this argument is provided in S. Alper, D. Tjosvold and K.S. Law, “Interdependence and Controversy in Group Decision Making: Antecedents to Effective Self-Managing Teams,” Organizational Behavior and Human Decision Processes 74, no. 1 (1998): 33-52.

4. Knowledge sharing has been argued to be central to the innovative capacity of a company. See, for example, I. Nonaka and H. Takeuchi, “The Knowledge-Creating Company: How Japanese Companies Create the Dynamics of Innovation” (New York: Oxford University Press, 1995).

5. The challenge of team faultlines is currently being studied by several scholars. For an academic description of faultline theory, see D.C. Lau and J.K. Murnighan, “Demographic Diversity and Faultlines: The Com-positional Dynamics of Organizational Groups,” Academy of Management Review 23, no. 2 (1998): 325-340.

6. The study of in-groups and out-groups has been a central theme of group analysis. For an overview of some of the key concepts, see, for example, H. Tajfel and J.C. Turner, “The Social Identity Theory of Inter-Group Behavior,” in “Psychology of Intergroup Relations,” ed. S. Worchel and L.W. Austin (Chicago: Nelson-Hall, 1986), 7-24.

7. The distinction between surface-level and deep-level attributes has been explored in more detail in S.E. Jackson, K.E. May and K. Whitney, “Understanding the Dynamics of Diversity in Decision-Mak-ing Teams,” in “Team Effectiveness and Decision Making in Organizations,” ed. R.A. Guzzo and E. Salas (San Francisco: Jossey-Bass, 1995), 204-261. For a more academic treatment of the theory, see D.A. Harrison, K.H. Price and M.P. Bell, “Beyond Rela-tional Demography: Time and the Effects of Surface- and Deep-Level Diversity On Work Group Cohesion,” Academy of Management Jour-nal 41, no. 1 (1998): 96-107.

8. Task orientation and relationship orientation are a continuum of leadership styles and have been examined in leadership re-search for more than a decade. For an overview of both, see G.A. Yukl, “Leadership in Organizations,” 6th ed. (Englewood Cliffs, New Jersey: Prentice-Hall, 2005). The distinction was first made in the 1950s in the seminal research by E.A. Fleishman, “The Descrip-tion of Supervisory Behavior,” Journal of Applied Psychology, 37, no. 1 (1953): 1-6.

9. For a more detailed overview of Type A characters, see, for exam-ple, M.A. Chesney and R.H. Rosenman, “Type A Behavior in the Work Setting,” in “Current Concerns in Occupational Stress,” ed. C.L. Cooper and R. Payne (London: John Wiley, 1980), 187-212; and M. Friedman and R.H. Rosenman, “Type A Behavior and Your Heart” (New York: Knopf, 1974), which provides a detailed description of the psychological construct.

10. The potentially negative impact of speed on creativity has been explored by several scholars. See, for example, C. Mainemelis, “When the Muse Takes It All: A Model For the Experience of Time-lessness in Organizations,” Academy of Management Review 26, no. 4 (2001): 548-565. For an overview on time, see D.G. Ancona, P.S. Goodman, B.S. Lawrence and M.L. Tushman, “Time: A New Re-search Lens,” Academy of Management Review 26, no. 4 (2001): 645-663.

11. For an overview of studies on the impact of time on group develop-ment and processes, see S. Blount, ed., “Time in Groups,” vol. 6, “Research on Managing Groups and Teams” (Greenwich, Connecticut: JAI Press, 2004).

Reprint 48411. For ordering information, see page 1.Copyright © Massachusetts Institute of Technology, 2007. All rights reserved.

SLOANREVIEW.MIT.EDU

A leader can significantly mitigate faultlines. To increase collaboration and knowledge sharing across teams with strong faultlines, leaders need to vary their leadership style according to how long the team has been together.

Reprint 48411. Copyright © Massachusetts Institute of Technology, 2007. All rights reserved.

Leading at the Enterprise Level

SPRING 2004 VOL.45 NO.3

REPRINT NUMBER 45315

Douglas A. Ready

MITSloanManagement Review

or the past couple of decades, companies have focused on creating

strong leaders of business units and influential heads of functions

— men and women responsible for achieving results in one corner of

an organization. But they have not paid as much attention to a more

important challenge: developing leaders who see the enterprise as a

whole and act for its greater good. And that perspective has become

increasingly necessary as companies seek to provide not just products

but broad-based customer solutions.

It is easy to understand why companies emphasized the development

of strong unit and functional leaders. Since the 1980s, the dominant view

has been that effective organizations are highly decentralized and therefore,

in theory, able to respond more quickly to customers. As business units

became increasingly autonomous, the leaders who reaped the most praise and

attention were bold, independent business builders. The challenge for those in

the corporate center was to manage, not to lead, and to be as hands-off as pos-

sible in their dealings with those running the businesses or the regions.

But left unchecked, such autonomy can produce an organizational culture

defined by a take-it-or-leave-it, product-push mentality. The demise of com-

panies that were icons of their industries, such as Digital Equipment Corp. or

Polaroid Corp., is evidence of this danger. These companies were once tech-

nology leaders with great products, but the spirit of autonomy that they fos-

tered eventually led to the creation of product silos overseen by powerful

leaders. Customers had a choice: They could buy the products on offer or go

elsewhere. When other companies began to offer better technologies, the two

giants’ products were no longer competitive, and customers fled.

Today, even cutting-edge products are not enough to ensure customer loy-

alty. Customers are stepping up demands for integrated solutions to their prob-

lems. While companies have developed new strategies for meeting those

demands, many have yet to change their thinking about what constitutes effec-

tive leadership in this environment. Decentralization itself is not the issue. The

main point is that solutions strategies cannot be implemented by leaders who

have a product-push silo mentality. Until companies recognize this fact, they will

be frustrated in their efforts to deliver integrated customer solutions by a gap

between their strategies and their capabilities.

SPRING 2004 MIT SLOAN MANAGEMENT REVIEW 87

Douglas A. Ready is a senior lecturer on leadership development at the MIT SloanSchool of Management. He also founded and leads the International Consortium forExecutive Development Research in Lexington, Massachusetts. He can be reached atinfo@icedr.org.

F

Leading at theEnterprise Level

Many companies

have outstanding

business-unit and

functional leaders

but far fewer who

can act in the

interests of the

corporate enterprise

as a whole.

Douglas A. Ready

88 MIT SLOAN MANAGEMENT REVIEW SPRING 2004

It is imperative, then, for companies to be able to identify and

develop enterprise leaders — people who can deliver differenti-

ated value by bringing the total resources of their companies to

their customers. In order to link strategy to leadership develop-

ment, they must be able to answer three questions: What are the

key elements of the enterprise leader’s job? Why is learning to

lead at the enterprise level such a difficult challenge? And what

can companies do to identify and develop enterprise leaders?

The Enterprise Leader’s JobOn a simple level, an enterprise leader is anyone accountable for

the economic and social welfare of the total enterprise, across

divisions, businesses, functions and locations. An enterprise

leader might run a business unit or oversee a major function but

will make decisions with the entire corporation in mind. In other

words, “enterprise leader” is not a job title — the term represents

a way of thinking and behaving. Effective enterprise leaders excel

at four tasks that may be difficult to quantify but are essential to

any company’s ability to compete.

First, enterprise leaders focus organizational attention on the

customer, setting priorities and driving out distractions. Second,

they build multiple organizational capabilities simultaneously,

especially in the areas of strategic competence and organizational

character. Third, they reconcile the tensions — between growth

and stability, for example — which are embedded in any organi-

zation. Finally, they create alignment by building consistency

between an organization’s statements of purpose, its processes,

and the skills and behaviors required of its people. These four

broad tasks encapsulate some of the hardest work that a manager

will ever attempt. A closer look at each aspect reveals the chal-

lenges facing the enterprise leader.

Focusing Organizational Attention. Massive complexity is a

fact of life today, but successful enterprise leaders don’t ignore it

or run from it. Instead, they search it out and reduce it so that

others in the organization can focus on the firm’s customers.

At PricewaterhouseCoopers, enterprise leaders brought order

to chaos in order to successfully complete the merger that created

the world’s largest professional services firm. In the late 1990s,

when Price Waterhouse merged with Coopers & Lybrand, the

newly combined company had 165,000 employees and con-

ducted business in more than 120 countries. Demand for the new

firm’s services required the company to hire 1,000 professionals

per week, on average, in order to keep up. But there was a down-

side, as within six months clients began complaining to senior

executives that PwC’s client knowledge was slipping. The firm’s

people were operating without the strategic purpose and cultural

compass that they needed to be effective.

PwC’s top 10 executives (the Global Leadership Team) then met

with the top 50 senior managers at the next tier to reduce the com-

plexity brought on by so much opportunity. The group’s mantra

for the next year was “ruthless prioritization.” PwC’s then-CEO,

James Shiro, was advocating an integration strategy to the firm’s

partners, but employees at the grass roots didn’t know what “inte-

gration” meant for them or what they were expected to do to bring

it about. Absent a clear explanation, a generic term like integration

could be taken to mean “slow, bureaucratic decision making.”

The new team of senior leaders focused PwC’s businesses on

areas in which they could add value that would differentiate the

firm from others in the industry. Given PwC’s emphasis on client

service, the team made small but important changes to the strat-

egy: PwC would commit to becoming “intelligently integrated,”

meaning it would integrate only if and when doing so would pro-

vide more value to its clients. The team also worked to repair the

organization’s cultural compass. Although the two companies

had well-articulated values before the merger, PwC’s senior lead-

ers created a compelling new set of core values that would serve

as a code of employee conduct for the merged firm. They rallied

the entire organization behind the new direction. Clients stopped

grumbling about the lack of focus and client knowledge, and

PwC, despite the slumping economy, emerged as the market force

envisioned by its leaders when they negotiated the merger.

Building Capabilities. Companies can be thought of as what

they do — as seen in their strategic vision and the way they exe-

cute it — and as who they are, as revealed by their organization’s

code of conduct and cultural norms. The “doing” aspects of the

enterprise leader’s job can be thought of as building strategic

Organizational Character:

“Being”

StrategicCompetence:

“Doing”

Building

Capabilities

and

Reconciling

Tensions

Strategic Vision

Competitiveness

ContinuousReinvention

CohesiveIdentity

Strategic competence and organizational character represent

what the company does and what the company is. Along these

two axes, enterprise leaders must articulate a compelling long-

term strategic vision while inspiring people to execute in the

short term; they must also create the cultural glue that bonds

the company while serving as the catalysts of radical change.

The Enterprise Leader’s Tasks

SPRING 2004 MIT SLOAN MANAGEMENT REVIEW 89

competence, while the “being” aspects are a matter of building

organizational character. Leaders with an enterprise perspective

understand the importance of building strategic competence and

organizational character simultaneously.

Consider how the enterprise leadership team at Continental

AG in Hanover, Germany, identified and built new organizational

capabilities when the company embarked on a large-scale strate-

gic and cultural transformation.

In the mid-1990s, Continental was the world’s fourth-largest

tire manufacturer. It was an OEM supplier to many prestigious

auto manufacturers and had a vibrant replacement business

throughout Europe. But in a fiercely competitive market known

for its tight margins, it was only modestly profitable. And the com-

pany’s newly appointed CEO, Hubertus von Grünberg, knew that

globalization would disrupt regional players such as Continental.

He and his enterprise leadership team of 50 decided that the com-

pany must transform itself from a European tire manufacturer into

something much more complex: a tier-one supplier of high-tech-

nology integrated wheel, brake and chassis systems to the global

automobile industry. Cross-border strategic alliances and acquisi-

tions would be the key to the accomplishment of this objective.

The obstacles to success were large. Continental had not grown

through partnerships or acquisitions, and thus its leadership

lacked both the mind-set and skills to carry out the new strategic

direction. Recognizing this reality, von Grünberg decided to

embark on a massive team-building and capability-building

effort. In conflict with his personal style (and level of comfort), he

engaged his enterprise leadership team in open dialogue on the

challenges facing the company. This was not as simple as it may

sound: The monthly meetings were at first very tense — the exec-

utives had to learn to trust von Grünberg, and the future direction

of the company was at stake. Over time, as it became clear that no

one would be punished for being honest, the tension abated and

the group’s members came to trust one another.

The leadership group also worked intensively on practical

matters, learning how to execute cross-border partnerships and

acquisitions, to work across cultural boundaries and to lead

cross-border project teams. The company brought in some new

executives who had such capabilities and dismissed others who

demonstrated that they were unwilling or unable to change.

Within two years, Continental had achieved what it set out to

do. After making several critical acquisitions from ITT Automo-

tive and undertaking a series of successful cross-border joint ven-

tures in Indonesia and Eastern Europe, Continental continued to

make high-quality tires but was also well positioned to provide

systems to the world’s automakers. A central ingredient of the

transformation’s success was the willingness and capacity of Con-

tinental’s leaders to subordinate the narrower interests of unit

and functional leaders and focus attention on doing what was

right for the company as a whole and its customers.

Reconciling Tensions. It is all too common for a company’s top

executives to play up the importance of managing innovation

and growth in order to remain competitive. Yet often in that same

firm, the environment feels anything but innovative and the

focus of attention is on cost containment, not growth. Does that

mean the company’s top team is deliberately deceptive or disin-

genuous? No, but the problem reveals one of the enterprise

leader’s most difficult jobs: dealing with the inherent ambiguities

that stem from the notion of building strategic competence and

organizational character simultaneously.

A couple of years ago, a large manufacturer with a well-devel-

oped culture of paternalism and a belief in lifetime employment

for its employees spotted an acquisition opportunity — a perfect

fit for its growth strategy. The economic benefits of the deal,

however, would bear fruit only if the company shed 10,000 jobs

soon after the deal was struck. There was no chance to grow

organically after the acquisition to salvage those jobs. The ana-

lysts eyeing the deal, well aware of the paternalistic nature of the

acquiring firm, signaled their uneasiness with the transaction in

the absence of major cost cutting beforehand. Even though the

manufacturer was healthy and growing, its top management

team decided the deal was critical to its future. The company

then cut 5,000 of its own employees in order to win the analysts’

approval of the purchase, guarding against a serious erosion of

its stock price.

This example highlights the built-in tensions that top execu-

tives must wrestle with. In developing strategic capabilities, they

have to balance their need to develop a vision of tomorrow with

the importance of making decisions that will satisfy customers

today. In managing organizational character, they need to build

strong cohesive cultures while being prepared to lead dramatic

change and reinvention. For the manufacturer, the success of its

long-term strategy depended in part on a short-term path of

cost cutting and job reductions — a move that was in direct

conflict with the company’s longstanding organizational cul-

ture. Although the company’s leaders went ahead with the

reduction in head count, they eliminated as many jobs as possi-

ble through attrition and created outplacement centers for laid-

off employees.

Such ambiguities are never easily resolved, but they must not

get swept under the rug. The former CEO of British Petroleum

(now BP Plc), Lord David Simon, once explained how he thought

of these tensions: “It’s very important for me and the other sen-

ior leaders here to behave in a manner that is deeply representa-

tive of BP’s culture, yet we must also be strangers in our own

company.” Only then would he and other top executives be able

to make necessary decisions in the company’s long-term inter-

ests, even when they were at odds with the organization’s culture.

More recently, BP’s current CEO, John Browne, has been quite

vocal about reinventing BP as an environmentally friendly com-

90 MIT SLOAN MANAGEMENT REVIEW SPRING 2004

pany while it continues to explore aggressively for oil. Leadership

at the top today requires people who can handle tensions of the

type that are embedded in BP’s position.

Creating Alignment. One can find useful examinations of differ-

ent aspects of the leader’s role in a variety of books and articles.

Some stress how important it is for those at the top to provide

their organizations with purpose; others emphasize the need for

leaders to form culture or to stimulate innovation or growth. But

problems arise if companies treat such topics as stand-alone issues

rather than as interlocking components of a system. Effective

enterprise leaders are skilled at simultaneously managing five

processes in order to create organizational alignment. This sys-

tems view of the enterprise leader’s role in creating organizational

alignment can be thought of as managing the five M’s: meaning,

mind-set, mobilization, measurement and mechanisms for renewal.

In 2002, Canada’s RBC Financial Group had to realign the

organization when it added a major new initiative to its shortlist

of strategic priorities. Having learned through research that its

customers felt it was too focused on products, RBC planned to

transform itself into a firm that provided integrated solutions for

all its clients’ financial needs. CEO Gordon Nixon referred to the

initiative as “implementing cross-platform leverage.” He con-

vened RBC’s Group Council, made up of the company’s top 125

leaders, and asked for their help in creating the organizational

alignment needed to turn the new strategy into a reality.

The enterprise leaders first had to acknowledge that there was

no shared sense of meaning in the company about the cross-plat-

form strategic initiative. There was no passion for the term and

little understanding of what cross-platform leverage would look

and feel like if it were operational throughout RBC. The enter-

prise leadership team met this challenge by shaping a collective

definition of what cross-platform leverage would mean for every

business and in every function in the company. Next, the team

conducted an audit of the company’s culture and mind-set and

articulated the behavioral changes that would be needed, first by

the team itself and then across the entire organization, to imple-

ment the new strategy. The company would have to abandon the

strong unit-oriented culture that had been the norm at RBC for

decades. The profile of an effective executive at RBC would have

to change from that of the strong individual performer to one

who mobilized resources across boundaries.

The enterprise leaders also examined which resources they

would need to mobilize so that RBC could excel at the new strat-

egy. Processes and IT systems that didn’t support integration had

to be replaced with knowledge-sharing processes that facilitated

the improvement of client knowledge. The team also took a close

look at metrics and incentives. Rewards that had been focused on

individual and unit performance were replaced by those that

focused on enterprise performance and improvements in client

satisfaction. Finally, the team identified mechanisms for renewal

— committing to meet every few months, for example, to review

progress — which would help RBC avoid complacency.

In less than one year, RBC’s enterprise leadership team reshaped

the company’s strategy, changed relationships across divisions and

revamped its performance and talent management processes. It

successfully brought about the organizational alignment that was

needed to change RBC’s capabilities, culture and mind-set to one

that was focused on delivering integrated customer solutions.

An Unnatural ActHaving considered the key elements of the enterprise leader’s job,

companies then need to ask why developing such leaders is so dif-

ficult. Indeed, there are three major obstacles built in to many

organizations that make developing enterprise leadership virtu-

ally an unnatural act.

The first obstacle can be found in organizational culture,

which often favors unit over enterprise leadership. Senior leaders

are role models, and the stories they tell and promotions they

make build an implicit cultural image of what effective leadership

means. In other words, when top teams make icons out of suc-

cessful unit performers and reward them for their behaviors, they

create a powerful disincentive for managers to aspire to behave as

enterprise leaders. The cultural imagery of the unit leader as hero

makes it very difficult for a new model of leader to emerge.

It takes leadership from the top to break the mold of cultural

norms. One of the first things Sam Palmisano did as IBM Corp.’s

new chairman and CEO was to change the name of IBM’s Enter-

prise Leadership Group to the Enterprise Leadership Team. While

that may sound trivial, this seemingly minor change sent a clear

message that IBM’s culture of celebrating individual contributors

and rewarding hero-leaders was coming to an end and would be

replaced by a strong emphasis on teamwork.

A second obstacle to developing leaders with an enterprise

perspective is the emphasis on specialized expertise in many

companies. Several consulting firms contend that the identifica-

tion, articulation and development of competencies should be

the cornerstone of a company’s leadership development efforts.

This view is so prevalent that it would be difficult to find a hand-

ful of major companies whose leadership initiatives do not have

competency frameworks as their foundation. And such frame-

works have much to recommend in them. The logic of compe-

tencies is linear, illustrating how company strategy is linked to

certain organizational capability requirements, which are con-

nected to a set of skills and leadership behaviors that need to be

developed in order to execute the strategy.

Yet while competency frameworks provide anchors for under-

standing leadership development requirements, they have tended

to focus on individual rather than team development. The focus

on individual development can prevent people from taking the

SPRING 2004 MIT SLOAN MANAGEMENT REVIEW 91

broader view and seeing the company as an interconnected whole,

with challenges that cannot be met by expertise in one area or suc-

cess in one unit. Moreover, an excessive reliance on competencies

tends to drive individuals to strengthen their particular expertise

rather than to stretch themselves for new challenges, such as mak-

ing the transition from unit to enterprise leadership. People end

up being overwhelmed by an ever-growing list of tasks and skills

in which they must demonstrate their competency.

The third obstacle is that contrary to strategic goals, reward

systems — like many organizational cultures — often favor the

accomplishments of individual units at the expense of corporate

success. Imagine a company in which the CEO and top team have

articulated an integration-based strategy that will require com-

pany leaders to think and act across boundaries and on behalf of

the entire enterprise. Now imagine that the company’s reward

system offers leaders incentives for achieving unit success, even

when they fail to behave as enterprise leaders. Which will win out

in the end — a theoretical statement about integration or a pow-

erful incentive system that rewards unit performance?

Senior executives create rewards to motivate employees to

pursue the business models established at the top. When those

models change, changes to the reward structure often lag behind

changes to the strategy. Ironically, under those circumstances, the

major problem with the system is that it is effective. Executives

who have been rewarded for behaving as unit leaders for years

will naturally struggle to behave as enterprise leaders.

Overcoming the ObstaclesFortunately, companies are not helpless in the face of these

impediments to enterprise leadership development. The

approach to this issue taken by PwC, Continental, BP, RBC and

IBM holds lessons for other organizations. These companies,

which differ by industry, size and strategic intent, have at least

one thing in common: They focus attention on three ways of

grooming individuals to be enterprise leaders.

They create cultures in which leaders are held accountable for

having an enterprise perspective. Top-team commitment is

critically important to the development of leaders with an enter-

prise orientation. CEOs and top executive teams must not only

believe in the importance of developing enterprise leaders, they

must also do whatever it takes to create a culture of accountabil-

ity — one that ensures that emerging leaders possess a cross-

boundary enterprise perspective.

To create that culture, top executives must develop a well-

thought-out point of view on the importance and role of enter-

prise leadership in the company. The IBM Leadership

Framework is a prime example. Accountability-based cultures

reward individuals for results, so it is important for companies to

make clear that they expect all leaders in the organization to

develop future enterprise leaders. This policy should also include

consequences for those who fail to cultivate such leaders.

They create opportunities for next-generation enterprise lead-

ers to emerge. Top executives must reinforce the culture of

accountability by giving potential leaders the chance to develop

an enterprise view. While that may seem obvious, this critically

important step is often avoided on one pretext or another. At

RBC, however, division leaders engaged in a widespread crossfer-

tilization of talent to show they supported the company’s strat-

egy. They made key jobs available to high-potential senior

executives in order to develop their enterprisewide perspectives.

They install HR processes that support the company’s leader-

ship development philosophies. Companies that effectively

develop enterprise leaders create process infrastructures that sup-

port their leadership frameworks and leadership development

activities. Such infrastructures combat the “cream will rise to the

top” philosophy, in which leaders simply emerge because of their

innate talent. They minimize the risk that their pool of enterprise

leaders will be too small for lack of enough cream.

These companies differentiate between the skills and behav-

iors of unit and enterprise leaders. They have talent-tracking

processes that spot “high potentials” as quickly as possible. They

provide multiple opportunities for those people to receive feed-

back early in their careers, enabling them to develop as unit or

enterprise leaders. And they have career and performance man-

agement processes that don’t punish individuals for taking risks.

This is important because companies often use the criterion of

repeated success as a screen in determining who is, and who isn’t,

a high-potential leader, making it risky for individuals to take on

roles that might lead to their derailment.

Bridging the Strategy-Capabilities GapMany companies are populated with senior executives who are

graduates of the leader-as-hero school of leadership. But as cus-

tomers step up their demands for solutions, those companies will

need to change their views on what constitutes effective leadership.

Increasingly, tomorrow’s senior executives will need to lead with an

enterprise perspective. It will take time to develop these leaders and

even more time for them to learn how to come together in teams.

Companies that fail to develop leaders who can work effec-

tively across unit, functional and geographic boundaries will be

unable to close the strategy-capabilities gap. But those that work

on identifying and developing their next generation of enterprise

leaders will be well prepared to deliver on the promise of provid-

ing innovative solutions for their customers.

Reprint 45315. For ordering information, see page 1.Copyright Massachusetts Institute of Technology, 2004. All rights reserved.

How To Lead a Self-Managing Team

SUMMER 2004 VOL.45 NO.4

REPRINT NUMBER 45410

Vanessa Urch Druskat and Jane V. Wheeler

MITSloanManagement Review

Please note that gray areas reflect artwork that hasbeen intentionally removed. The substantive contentof the article appears as originally published.

SUMMER 2004 MIT SLOAN MANAGEMENT REVIEW 65

o get work done, many companies organize employees into self-managing teams

that are basically left to run themselves with some guidance from an external

leader. In fact, comprehensive surveys report that 79% of companies in the Fortune

1,000 and 81% of manufacturing organizations currently deploy such “empowered,”

“self-directed” or “autonomous” teams.1 Because of their widespread use, much

research has been devoted to understanding how best to set up self-managing teams to

maximize their productivity and effectiveness. Interestingly, though, relatively little

attention has been paid to the leaders who must oversee such working groups.

At first, it seems contradictory: Why should a self-managing team require any lead-

ership? After all, doesn’t the group manage itself? In actuality, though, self-managing

teams require a specific kind of leadership. Even a team that is autonomous in terms

of its activities and decision making must still continually receive direction from

higher levels in the organization. And it also must report to that hierarchy through a

person who is ultimately held accountable for the group’s performance. Many man-

agers today are expected to fulfill the role of external leader,

but most receive conflicting signals regarding how to go

about it.2 Should they, for instance, be involved in their

team’s decision-making process? If so, how should they par-

ticipate without detracting from the group’s autonomy?

To investigate such issues, we conducted a study of 300

self-managing teams at a large manufacturing plant of a For-

tune 500 corporation. (See “About the Research.”) We investi-

gated both average- and superior-performing external leaders

at that site to determine the behaviors that separated one

group from the other. Our research has shown that, contrary

to common perception, the best external leaders were not

necessarily the ones who had adopted a hands-off approach,

nor were they simply focused on encouraging team members

in various ways.3 Instead, the external leaders who had con-

tributed most to their team’s success excelled at one skill:

managing the boundary between the team and the larger

organization. That process required specific behaviors that

How To Lead a

Self-Managing TeamTeams that are basically

left to run themselves can

be highly efficient and

productive. To be successful,

though, such autonomous

groups require a specific

type of external leadership.

By Vanessa Urch Druskat and

Jane V. Wheeler

Vanessa Urch Druskat is a professor of organizational behavior atthe Whittemore School of Business and Economics at the Universityof New Hampshire. Jane V. Wheeler is professor of management atBowling Green State University’s College of Business Administration.They can be reached at vanessa.druskat@unh.edu and jvwheel@cba.bgsu.edu.

T

66 MIT SLOAN MANAGEMENT REVIEW SUMMER 2004

can be grouped into four basic functions: relating, scouting,

persuading and empowering. (See “The Work of the External

Leader,” p. 69.) External leaders who excelled at those capabilities

were able to drive their teams to superior performance.

Managing in No-Man’s LandAs recently as the 1960s, self-managing teams were practically

unheard of. Indeed, an early experiment by General Foods Corp.

to deploy self-managing teams on its Gaines dog-food produc-

tion line more than 30 years ago was something of a sensation.

Even, the company’s own senior management expected the proj-

ect to fail.4 Since then, self-managing teams have become an

increasingly popular way for companies to get work done. Even

so, the role of the people in charge of those groups has remained

somewhat of a mystery.

To be sure, leading a team that needs to manage itself is inher-

ently tricky. The role is highly ambiguous by nature (and, on the

face of it, oxymoronic). In general, self-managing teams tend to

have well-defined job functions and are responsible for monitor-

ing and managing their own performance. Instead of managers

telling them what to do, these teams gather and synthesize infor-

mation, make important decisions, and take collective responsi-

bility for meeting their goals. What’s left for the external leader to

do? As one of the external leaders in our study remarked, “Maybe

we’re not needed, you know?”

But the reality is that the buck stops at the external leaders.

Specifically, companies hold them responsible for their team’s

performance. If the quality or productivity of a team is substan-

dard, its external leader is taken to task. As another team leader in

our study noted, “The hardest part is that I’m also held account-

able. If they make the wrong decision, it still comes back on me.”

In essence, the job of external leader exists squarely in the middle

of a managerial no-man’s land.

Back in 1977, a field study published about that early experi-

ment at General Foods revealed external leaders caught in the

middle: Their teams criticized them for being too controlling,

while their own managers complained that they were being too

lax.5 Not long after, another research study looked at teams in

other organizational settings and found that those three con-

stituencies — team members, external leaders and their upper-

We conducted a study of external lead-

ers at a Fortune 500 durable-consumer-

goods manufacturing plant in the

Midwest. The facility employs more than

3,000 people and is the world’s largest

manufacturer of its particular product.

We chose this site because it offered a

rich research sample: The plant had tran-

sitioned to self-managing work teams

almost five years prior to our data collec-

tion, and there were 300 such teams

reporting up through 66 external lead-

ers. This large pool of external leaders in

one location provided a wide range of

individuals in that role and allowed us to

control for organizational context.

Our method was to select samples of

high-performing and average external

leaders and to discover through inter-

views how their behaviors, strategies and

attitudes differed. We used three criteria

for selecting our key participants: nomi-

nations from team members, nomina-

tions from managers and objective

performance data for the leaders’ teams.

We then conducted intensive three-hour

interviews with each of the 19 individu-

als selected, regarding critical incidents

they recalled. The interviews were con-

ducted blind to the “superior” or “aver-

age” status of the leaders. A content

analysis of those interviews allowed us to

develop an exhaustive list of actions and

behaviors that were noted consistently

by superior leaders but not by average

ones. Initially, our exhaustive list con-

tained approximately 30 behaviors that

we hypothesized made a difference

between the “superior” and “average”

performers. Iterative readings of the

interview transcripts helped us to win-

now down that list to the 11 distinct

behaviors reported in this article. Two

expert coders, who were blind to the

“superior” and average” status of the

leaders, coded all transcripts for the

presence of the 11 behaviors and deter-

mined the reliability of the codes. The

mean reliability was 92%. Statistical

analyses also revealed that 10 of the

behaviors were demonstrated signifi-

cantly more often by the superior-per-

forming leaders. (The exception was

“diagnosing member behavior,” which

was demonstrated often by both average

and superior performers.)

To help interpret the findings and

to develop our model of how these 11

behaviors emerge as the process of

effective external leadership, we also

spoke with a total of 90 team members:

52 in focus groups and 38 in one-on-one

interviews. We supplemented that infor-

mation with interviews of the 10 man-

agers to whom the external leaders

reported, and we also collected informa-

tion via questionnaires from the broader

group of senior managers and directors

at the plant.

Furthermore, we later examined the

applicability of our findings by interview-

ing external leaders at other organiza-

tions and in different industries,

including those for product-development

groups in a design organization and

executive-level teams at a bank, a health-

care organization and a manufacturing

company. This subsequent work suggests

that our results have broader applicabil-

ity for any leader of an empowered team,

regardless of the group’s level or task.

About the Research

SUMMER 2004 MIT SLOAN MANAGEMENT REVIEW 67

level managers — all had different ideas about the role.6

Unfortunately, that lack of consensus generally remains

today. Prior to and since our study, we have found confusion

about the external-leader role in other companies and at all

organizational levels. Consider the senior executive of a large

Midwestern bank who desperately wanted to empower his team

of high-level bank executives but was unsure how to go about

that. When he attended meetings, he felt team members relied

too heavily on his opinion, but when he stopped going to meet-

ings, he felt stuck in an information black hole. Even with his

many years of experience, he really did not know how to man-

age that group.

The problem is that a self-managing team requires leadership

of a very different sort. Researchers agree that the external-leader

role is more complex than the traditional manager role.7 This is

true partly because the typical external leader is in charge of sev-

eral self-managing teams at any one time. (In our study, they

were responsible for as many as eight.)

More importantly, the external leader absolutely must avoid

any heavy-handed attempts at managing. Case studies have

shown that external leaders who struggle with their role usually

end up exerting too much control, which then undermines the

self-managing team’s ability to get work done.8 Because of that

and other issues, various researchers have labeled external lead-

ers as the most common impediment to the success of self-

managing teams.9

Confusion about the role of the external leader might be tol-

erable if that function weren’t so crucial, but a variety of studies

have shown that the success of a self-managing team greatly

depends on its external leader.10 A newly created group, for

example, often needs both effective coaching and a champion

who can represent its decisions to other executives in the organi-

zation. Teams also depend on external leaders for help in acquir-

ing resources. That’s why it’s common to see an underperforming

team successfully turn itself around with only the change of its

external leader — or to see a once-great team suffer after its

external leader departs.

Four Functions, 11 BehaviorsAlthough the essence of a self-managing team is autonomy, the

quality of its link to the organization is pivotal to success. Some

external leaders perform that role much better than others, with

the superior ones tending to excel at relating, scouting, persuad-

ing and empowering. Each function requires specific behaviors,

of which there are a total of 11. In the exhibit “The Work of the

External Leader,” the relating activities are shown at the begin-

ning of the process because, without the formation of those rela-

tionships, external leaders will find scouting difficult. Scouting,

in turn, equips external leaders with the information they need

to persuade. And various aspects of scouting and persuading

help pave the way for greater empowerment of a team, which

then contributes to the group’s ultimate effectiveness.

The well-known key to making self-managing teams work

is to delegate considerable authority to the group, granting it

tremendous flexibility in making its own decisions.11 What is

less well-known are the kinds of leadership activities and behav-

iors that are needed to build the foundation for team empower-

ment. Relating, scouting and persuading are those critical

building blocks.

Relating External leaders must continually move back and forth

between the team and the broader organization to build rela-

tionships. Success in this area requires three behaviors: being

socially and politically aware, building team trust and caring for

team members.

Being socially and politically aware. During our research study,

we heard stories of both team triumphs and irritations. In one

case, an external leader recounted the time he allowed one of his

teams a freedom that was inconsistent with the company’s infor-

mal policies. Later, he was taken aback by the backlash he experi-

enced from his peers in the organization. Recalls the leader, “This

person sends a note to my manager, to his manager and to my

colleagues … saying, ‘Since when did this policy change?’” The

leader was also vehemently attacked by another team leader for

his “ineffective supervision.”

What happened? Clearly, the incident suggests a lack of

political awareness: The leader simply didn’t anticipate the

impact of his decision on others. Nor did he perceive the need

to build a broader consensus. Incidentally, in our study the per-

formance of that leader had been categorized as average. By

contrast, the superior external leaders had consistently demon-

strated an understanding of the broader organization, including

the individual concerns and decision-making criteria of impor-

External leaders find themselves caught in the middle: Their teams criticize them for being too controlling, while their own managers complain they are being too lax.

68 MIT SLOAN MANAGEMENT REVIEW SUMMER 2004

tant constituencies, such as the engineering and human

resources groups. As one superior leader bluntly put it, “I’ve got

a good rapport with all those groups, and most things I ask for

get done. … You can get moved to the top of the list for things

in a hurry if you’re not [upsetting] people.”

Building team trust. Superior external leaders also consistently

recognized the value of building good relationships with their

teams, even to the point of achieving insider status. Given that

the leaders had little time to spend with any one team, such

acceptance was far from automatic. In fact, one leader in our

study was impeached by her team members, who did not trust

that she had their best interests in mind. Contrast that with the

experience of another leader who, when his team complained of

problems with new equipment, said, “Whatever it costs, if you

guys aren’t happy with it, we’ll get another system in here …

because you’re the ones who are gonna have to work with it.” The

team was incredulous. One member pointed out that the new

system had cost $23,000 to install.

Caring for team members. In our study, average leaders were

more likely to see the personal problems of the team members as

impediments to getting work done, whereas superior leaders

more often recognized them as opportunities to build relation-

ships. One superior leader described an incident in which one of

his employees had a problem with her disability leave. The leader

took it upon himself to call the insurance department in the

health center to clear the dispute, which had been an ongoing

source of worry for the worker. “It was … instant relief for that

person,” recalls the leader. “It was not a big problem, but it was

big to her.”

Scouting To scout effectively, external leaders must demonstrate

three behaviors: seeking information from managers, peers and

specialists; diagnosing member behavior; and investigating prob-

lems systematically.

Seeking information from managers, peers and specialists.

Superior leaders appeared to be significantly more likely to seek

information from others in the organization, whether as advice

or simply in response to technical questions from the team.

Sometimes the leaders used that information to influence team

decision making, especially when they wanted to persuade peo-

ple to take into account broader organizational considerations.

For example, when a team wanted to hire a colleague who had

filled in for absent members in the past, the group’s external

leader wisely decided to step in. The leader wanted his team

members to be able to make their own choice, but he also sensed

that their lack of knowledge and respect for a new hiring policy

at the company could create problems with management. “I

could tell this was going to be a sticky situation,” he recalls,

because the new policy was more formal and favored candidates

with greater seniority. So the leader talked to HR as well as other

external leaders as to how they filled their jobs. He then

informed his boss about the situation and invited him to the

team’s next meeting. At the meeting, the team members were

able to persuade the external leader’s boss to delay implement-

ing the new policy so that they could hire the person they

wanted. By seeking and relaying all the pertinent information,

the leader was able to enable his team to make its own decision

(without disrupting the organization), which helped boost the

team’s morale.

Not surprisingly, external leaders who routinely sought

information from the broader organization were able to advo-

cate more effectively for their teams. They were able to help

their teams gain valuable political awareness and build social

capital for them — all of which would often pay off in less obvi-

ous, but no less important, ways. The team members themselves

realized the importance of having an external leader who could

get information faster than they could, had access to data they

didn’t have and could easily make contact with upper manage-

ment. On the other hand, when team members felt that an

external leader was not keeping them adequately informed, they

tended to develop an “us against them” attitude. As one such

member put it, “[They] don’t want to give up information; it’s

their power.”

Diagnosing member behavior. Because external leaders are

typically responsible for the performance of several teams, they

are rarely on the scene when something critical occurs within a

specific group. As a result, they must often gain insight after the

fact. To do so, they frequently need to add to their incomplete

information by analyzing and making sense of verbal and non-

verbal cues from team members. One leader described an

Average leaders tend to see team members’ personal problems as impediments, whereas superior leaders view them as opportunities to build relationships.

SUMMER 2004 MIT SLOAN MANAGEMENT REVIEW 69

episode when his team had done months of benchmarking

research to suggest a policy change to the directors of the organ-

ization. But the directors quickly shot down the proposal and

instead told the team members that their time might be better

spent improving their quality and productivity. When the

leader heard the news he had mixed emotions. He was upset

because he knew how much time his team had spent on the pro-

posal, yet he also knew he had to support the directors’ decision

because, after all, he was part of management. So he went to the

team members and started to give them a pep talk but quickly

stopped himself. “I could tell they thought I was full of you-

know-what. You could see it in their eyes,” he recalls. So, he

instead offered people a sympathetic ear and acknowledged

their well-intentioned efforts.

Interestingly, both the superior and the average leaders in

our study talked about the need to read their team members

accurately. In fact, it was the leadership capability they most

commonly identified as important.

Investigating problems systematically. When superior leaders

got wind of a potential problem, they were significantly more

likely to deploy an inquisitive and systematic approach to inves-

tigate the matter. They would begin by asking the team members

myriad questions to collect data and identify the issues, after

which they might visit an external constituent to collect addi-

tional information. “You’ve got to be down on the floor,” one

superior leader asserted. “When the line is down, I’m over there

within a minute.” By collecting firsthand information from the

team members, superior leaders were able to fully understand

their group’s perspective, enabling them to make informed rec-

ommendations that would be acceptable to their teams as well as

to external constituents. In contrast, average leaders were more

likely to attempt problem solving with less data or input from the

team members.

Persuading With respect to external leadership, effective persua-

sion requires two behaviors: obtaining external support and

influencing the team.

Obtaining external support. Teams often need support from

the broader organization, and superior leaders are able to per-

form this advocacy role more effectively. One such leader

remembered a time when his team members were having

equipment problems. After talking with them, he used their

ideas to sketch a new piece of machinery and came in during

other shifts to gain additional input from other teams. He was

then able to obtain management’s authorization to build the

new piece of equipment. In our study, team members agreed

that leaders were most helpful when they were able to get the

attention of important external constituents. Average leaders

seemed to seek such external support less frequently, and when

they did they were less successful in obtaining it. One average

9 Delegatingauthority

10 Exercisingflexibilityregardingteamdecisions

11 Coaching

Team-

Focused

Behavior

Organization-

Focused

Behavior

Organization/Team BoundaryTeam

Effectiveness

Relating Scouting Persuading Empowering

1 Beingsocially andpoliticallyaware

2 Buildingteam trust

3 Caring forteammembers

4 Seeking infor-mation frommanagers,peers andspecialists

5 Diagnosingmemberbehavior

6 Investigatingproblemssystematically

7 Obtainingexternalsupport

8 Influencingthe team

External leaders must perform 11 behaviors that can be grouped into four categories: relating, scouting, persuading and empower-

ing. The behaviors are distinct but mutually reinforcing, and they require the external leader to continually cross the border between

the team and the broader organization. In this model of external leadership, social and political awareness of the broader organiza-

tion provides access to the individuals and groups that can help the leader best meet the team’s needs; strong relationships allow

the leader access to information in the team and the organization, which aids the leader in making sense of the needs of both par-

ties; good information enables the leader to encourage or persuade the team to behave in ways that facilitate the organization’s

effectiveness; and the sense of control afforded by that influence allows the leader to empower the team more fully, resulting in

greater team effectiveness.

The Work of the External Leader

70 MIT SLOAN MANAGEMENT REVIEW SUMMER 2004

leader remembered an incident when he pleaded with the

scheduling department to make a change for one of his teams.

But because he had not built relationships with that department

and had not expended the effort to obtain the information nec-

essary to support his case, his request was denied.

Influencing the team. Effective external leaders were also adept

at swaying their teams to decisions that best met the needs of

the organization. Keep in mind that prior to doing so, these

superior leaders had already established trust with their teams,

had systematically investigated the problem at hand and had

used their external contacts to obtain all necessary information.

One superior leader, for example, collected statistics from the

accounting department to persuade his struggling team to think

of ways to improve its productivity. Using that data, the leader

impressed upon his team how much the company lost in prof-

its every minute a manufacturing line was down. “This is the

money that we didn’t make,” the leader told the team. “If some-

body is cutting the line off just to eat a sandwich, it is costing

everybody.” Three months later, the group’s performance had

improved markedly, and whenever the team was about to fall

behind schedule, people would work a little longer to stay on

track. Interestingly, those team members were going the extra

distance without their leader having to ask (let alone demand)

that they do so.

Empowering External leaders of self-managing teams can

empower those teams by demonstrating three behaviors: dele-

gating authority, exercising flexibility regarding team decisions

and coaching.

Delegating authority. External leaders typically have great dis-

cretion over the amount and type of authority that they dele-

gate. In general, the superior leaders tended to empower their

teams with more responsibility. For example, when one team

was told by management that it had to move its assembly line,

its leader left it up to the team to work out the details. “They

very much had ownership,” remembers the leader. “They knew

what we needed to do and took it from there. This needed to be

their baby.” However, average leaders tended to be more fearful

about delegating authority, and they would often make deci-

sions for the team and solve its problems covertly. It is impor-

tant to note that the average leaders’ reluctance to delegate

authority was not because they were leading the poorer-per-

forming teams. In fact, many of the superior leaders in our

study had been transferred to a problem team and had worked

with it until they felt comfortable delegating authority. Eventu-

ally, as the team gained more and more ownership, its perform-

ance improved.

Exercising flexibility regarding team decisions. Sometimes a

team proposes something that is outlandish or that appears to

reflect poorly on its leader. In such situations, average leaders

were likely to express their reservations, whereas superior leaders

frequently replied with comments such as, “It’s not what I think;

it’s what you think.” One of the teams in our study, for example,

wanted to present its quarterly performance results at the depart-

ment meeting by performing a skit. At first, the team’s leader was

a little hesitant because nothing like that had been done before

and the meeting would be attended by all of the managers in the

organization. “I thought, ‘Oh, we could look so foolish on this,’ ”

she recalls. But she gave the team her approval and, as it turned

out, the skit went over well, which boosted the team’s sense of

cohesion and identity. Of course, teams don’t always make good

decisions, and a major responsibility of external leaders is to pre-

vent serious mistakes. But even when a team has to be reined in,

superior leaders did so only after considering the proposal as

open-mindedly as possible.

Coaching. Coaching involves a number of activities, including

working one-on-one with employees, giving feedback to the

team and demonstrating certain behaviors (such as effective

meeting facilitation) for others to model. Past research has

found that superior leaders tend to be active in educating and

coaching others,12 and our study confirmed that. In particular,

superior leaders focused on strengthening a team’s confidence,

its ability to manage itself and its contributions from individual

members. A common coaching behavior was for leaders to

work with people who had just taken on roles of greater respon-

sibility. For example, after a meeting in which people argued

for 20 minutes about coffee, one superior leader helped the

team member in charge to run more efficient meetings, first by

developing an agenda and then by keeping the group focused

on that agenda.

Superior leaders develop strong relationships inside the team and across the organization; average leaders tend to relate well to one of those parties but not to both.

SUMMER 2004 MIT SLOAN MANAGEMENT REVIEW 71

Self-Management in the Long RunFor external leaders to relate, scout, persuade and empower,

they must frequently — but quietly — cross the boundary

between the team and the broader organization. The idea that a

leader must manage this boundary is certainly not new; past

research has linked it to the effectiveness of traditionally man-

aged teams,13 and several studies of empowered teams have

proposed that it is the central focus of the external leader’s

role.14 Our own research supports those findings. Specifically,

the superior leaders in our study were able to develop strong

relationships both inside the team and across the organization.

In contrast, average performers tended to relate well to the team

or to those in the broader organization, but not to both.

The four functions — relating, scouting, persuading and

empowering — are important for the leader of any group but

particularly so for those in charge of self-managing teams.

(After all, if a team has not been empowered, how can it man-

age itself?) And the same process would be useful for leaders of

geographically dispersed teams, given their similar challenge of

having to rely on imperfect information to influence behavior

from a remote location.

An important thing to remember about team autonomy is

that self-management is not an either-or condition. Instead, it’s

a continuum, and external leaders should also be constantly

guiding and developing their teams so that they become

increasingly independent. In fact, the ultimate goal of the exter-

nal leader should be the delegation of all 11 of the behaviors

described. For example, external leaders should be developing

their teams’ relationships with relevant individuals and other

groups in the broader organization, eventually relinquishing

that primary activity to the teams themselves. And when teams

do become less dependent on their external leaders, the leaders

will theoretically be freer to assume the responsibility for an

even larger number of groups.

The teams in our study had made the transition to self-man-

agement five years before our research commenced. At that

time, the superior participants were already performing the 11

essential behaviors of external leadership. We would not be sur-

prised to learn today that those individuals have since moved

their teams further along the continuum of increased autonomy

and independence. In fact, we would be greatly surprised if that

were not the case.

REFERENCES

1. E.E. Lawler III, “High Involvement Management” (San Francisco:Jossey-Bass, 1986); E.E. Lawler III, “Strategies for High Perfor-mance Organizations” (San Francisco: Jossey-Bass, 1998); and G.Taninecz, T.H. Lee, A.V. Feigenbaum, B. Nagle and P. Ward, “BestPractices and Performances: Manufacturers Tackling Leading-EdgeInitiatives Generally Reap the Best Results,” Industry Week, Dec. 1,1997, 28-43.

2. M.M. Beyerlein, D.A. Johnson and S.T. Beyerlein, “Introduction,” inM.M. Beyerlein, D.A. Johnson and S.T. Beyerlein, eds., “Advances inInterdisciplinary Studies of Work Teams,” vol. 3 (Greenwich, Connecti-cut: JAI Press, 1996), ix-xv; C.C. Manz and H.P. Sims Jr., “Searchingfor the ‘Unleader’: Organizational Member Views on Leading Self-Managed Groups,” Human Relations 37 (1984): 409-424; and C.C.Manz and H.P. Sims Jr., “Leading Workers To Lead Themselves: TheExternal Leadership of Self-Managed Work Teams,” AdministrativeScience Quarterly 32 (1987): 106-128.

3. See C.C. Manz, “Leading Workers To Lead Themselves.”

4. R.E. Walton, “The Topeka Work System: Optimistic Visions, Pes-simistic Hypotheses and Reality,” in R. Zager and M.P. Rosow, eds.,“The Innovative Organization: Productivity Programs in Action” (Elms-ford, New York: Pergamon Press, 1982): 260-287.

5. R.E. Walton, “Work Innovations at Topeka: After Six Years,” Journalof Applied Behavioral Science 13 (1977): 422-433; and R.E. Walton,“The Topeka Work System.”

6. C.C. Manz, “Searching for the ‘Unleader.’ ”

7. M.M. Beyerlein, “Introduction”; and J.R. Hackman, “The Psychologyof Self-Management in Organizations,” in M.S. Pallack and R.O.Perloff, eds., “Psychology and Work: Productivity, Change andEmployment” (Washington, D.C.: American Psychological Association,1986), 89-136.

8. See, for example, E.E. Lawler III, “High Involvement Management”;R.E. Walton, “Work Innovations at Topeka”; and R.E. Walton, “TheTopeka Work System.”

9. R.I. Beekun, “Assessing the Effectiveness of Sociotechnical Inter-ventions: Antidote or Fad?” Human Relations 47 (1989): 877-897; andR.E. Walton and L.A. Schlesinger, “Do Supervisors Thrive in Participa-tive Work Systems?” Organizational Dynamics 8 (winter 1979): 24-39.

10. S.G. Cohen, L. Chang and G.E. Ledford Jr., “A Hierarchical Con-struct of Self-Management Leadership and Its Relationship to Qualityof Work Life and Perceived Work Group Effectiveness,” PersonnelPsychology 50, no. 2 (summer 1997): 275-308; B.L. Kirkman and B.Rosen, “A Model of Work Team Empowerment,” in R.W. Woodmanand W.A. Pasmore, eds., “Research in Organizational Change andDevelopment,” vol. 10 (Greenwich, Connecticut: JAI Press, 1997),131-167; and B.L. Kirkman and B. Rosen, “Beyond Self-Management:Antecedents and Consequences of Team Empowerment,” Academy ofManagement Journal 42 (1999): 58-74.

11. S.G. Cohen, G.E. Ledford Jr. and G.M. Spreitzer, “A PredictiveModel of Self-Managing Work Team Effectiveness,” Human Relations49 (1996): 643-676; B.L. Kirkman, “Beyond Self-Management”; andR.E. Walton, “The Topeka Work System.”

12. C.C. Manz, “Leading Workers To Lead Themselves”; R. Wageman,“Critical Success Factors for Creating Superb Self-Managing Teams,”Organizational Dynamics 26 (summer 1997): 49-61; and R. Wageman,“How Leaders Foster Self-Managing Team Effectiveness: DesignChoices Versus Hands-On Coaching,” Organization Science 12, no. 5(2001): 559-577.

13. D.G. Ancona and D.F. Caldwell, “Bridging the Boundary: ExternalActivity and Performance in Organizational Teams,” Administrative Sci-ence Quarterly 37, no. 4 (December 1992): 634-665.

14. See S.G. Cohen, “A Hierarchical Construct of Self-ManagementLeadership”; J.L. Cordery and T.D. Wall, “Work Design and Supervi-sory Practice: A Model,” Human Relations 38, no. 5 (1985): 425-441;and J.R. Hackman, “Leading Teams: Setting the Stage for Great Per-formances” (Boston: Harvard Business School Press, 2002).

Reprint 45410. For ordering information, see page 1.Copyright Massachusetts Institute of Technology, 2004. All rights reserved.

PDFs ■ Permission to Copy ■ Back Issues ■ Reprints

Articles published in MIT Sloan Management Review are copyrighted by the Massachusetts Institute of Technology unless otherwise specified at the end of an article.

MIT Sloan Management Review articles, permissions, and back issues can be purchased on our Web site: www.pubservice.com/msstore or you may order through our Business Service Center (9 a.m.-7 p.m. ET) at the phone numbers listed below. Paper reprints are available in quantities of 250 or more.

To reproduce or transmit one or more MIT Sloan Management Review articles by electronic or mechanical means (including photocopying or archiving in any information storage or retrieval system) requires written permission. To request permission, use our Web site (www.pubservice.com/msstore), call or e-mail: Toll-free: 800-876-5764 (US and Canada) International: 818-487-2064 Fax: 818-487-4550 E-mail: MITSMR@pubservice.com

Posting of full-text SMR articles on publicly accessible Internet sites is prohibited. To obtain permission to post articles on secure and/or password-protected intranet sites, e-mail your request to MITSMR@pubservice.com

Customer Service MIT Sloan Management Review PO Box 15955 North Hollywood, CA 91615