Introduction

A McKinsey Quarterly Reader
www.mckinseyquarterly.com
The 21st-century corporation
The McKinsey Quarterly on leadership
Introduction
The 21st-century organization
Big corporations must make sweeping organizational
changes to get the best from their professionals.
Lowell L. Bryan and Claudia Joyce
August 2005
Competitive advantage from better interactions
Tacit interactions are becoming central to economic
activity. Making those who undertake them more effective
isn’t like tweaking a production line.
Scott C. Beardsley, Bradford C. Johnson, and
James M. Manyika
May 2006
The new metrics of corporate performance:
Proft per employee
Most measurements of performance are geared to
the needs of 20th-century manufacturing companies.
Times have changed. Metrics must change as well.
Lowell L. Bryan
February 2007
Mapping the value of employee collaboration
As collaboration within and among organizations becomes
increasingly important, companies must improve their
management of the networks where it typically occurs.
Robert L. Cross, Roger D. Martin, and Leigh M. Weiss
August 2006
Harnessing the power of informal employee
networks
Formalizing a company’s ad hoc peer groups can spur
collaboration and unlock value.
Lowell L. Bryan, Eric Matson, and Leigh M. Weiss
November 2007
The next revolution in interactions
Successful efforts to exploit the growing importance
of complex interactions could well generate durable
competitive advantages.
Bradford C. Johnson, James M. Manyika, and
Lareina A. Yee
November 2005
3
4
14
6
36
50
6
Table of contents
Copyright © 007 McKinsey & Company. All rights reserved.
The McKinsey Quarterly has been published since 1964
by McKinsey & Company, 55 East 5nd Street, New York,
New York 100.
Introduction
In about 1760, a few entrepreneurs in the north of England
had the idea of using steam engines to drive machines that
spun cotton thread—the germ of the frst industrial revolution.
A bit more than a century after that, some European and
American companies launched the second industrial revolution
by embracing innovations such as electricity and electric
motors, internal-combustion engines, production lines,
interchangeable parts, and hierarchical, vertically integrated
corporations. Today our whole planet is being transformed by a
dramatic, nonindustrial revolution based on intangibles such
as knowledge workers, intellectual capital, collaborative networks,
low-cost interactions (particularly tacit ones), and globalization.
Your business, your life, and your career will all change
profoundly. This collection shows what companies and the people
who work for them must do to succeed in that new world.
These articles illustrate the Quarterly’s best traditions: written
by McKinsey consultants, they offer new ways of thinking
about the challenges that companies and managers face today
by combining the practical bent of client work with the
fruits of extensive research. They represent just a small sampling
of the knowledge and ideas available to you as a member of
mckinseyquarterly.com.
Introduction
The McKinsey Quarterly 2005 Number 3
The 21st-century organization 5
About half a century ago, Peter Drucker coined the term “knowledge
worker” to describe a new class of employee whose basic means of
production was no longer capital, land, or labor but, rather, the productive
use of knowledge. Today, these knowledge workers, who might better
be called professionals, represent a large and growing percentage of the
employees of the world’s biggest corporations. In industries such as
fnancial services, health care, high tech, pharmaceuticals, and media and
entertainment, professionals now account for 25 percent or more of the
workforce and, in some cases, undertake most typical key line activities.
These talented people are the innovators of new business ideas. They
make it possible for companies to deal with today’s rapidly changing and
uncertain business environment, and they produce and manage the
intangible assets that are the primary way companies in a wide array of
industries create value.
Productive professionals make big enterprises competitive, yet these
employees now increasingly fnd their work obstructed. Creating and
exchanging knowledge and intangibles through interaction with
their professional peers is the very heart of what they do. Yet most of them
squander endless hours searching for the knowledge they need—even if it
resides in their own companies—and coordinating their work with others.
The 21st-century
organization
Big corporations must make sweeping organizational changes to get the
best from their professionals.
Lowell L. Bryan
and Claudia Joyce
The McKinsey Quarterly 2005 Number 3
The ineffciency of these
professionals has increased along
with their prominence. Consider the
act of collaboration. Each upsurge
in the number of professionals who
work in a company leads to an
almost exponential—not linear—
increase in the number of potential
collaborators and unproductive
interactions. Many leading
companies now employ 10,000 or
more professionals, who have
some 50 million potential bilateral
relationships. The same holds
true for knowledge: searching for it
means trying to fnd the person
in whose head it resides, because
most companies lack working
“knowledge markets.” One measure
of the diffculty of this quest is the
volume of global corporate e-mail, up from about 1.8 billion a day in
1998 to more than 17 billion a day in 2004. As fnding people and knowl-
edge becomes more diffcult, social cohesion and trust among professional
colleagues declines, further reducing productivity.
A flawed organizational design
Today’s big companies do very little to enhance the productivity of their
professionals. In fact, their vertically oriented organizational structures,
retroftted with ad hoc and matrix overlays, nearly always make professional
work more complex and ineffcient. These vertical structures—relics of
the industrial age—are singularly ill suited to the professional work process.
Professionals cooperate horizontally with one another throughout a
company, yet vertical structures force such men and women to search across
poorly connected organizational silos to fnd knowledge and collaborators
and to gain their cooperation once they have been found.
Worse yet, matrix structures, designed to accommodate the “secondary”
management axes that cut across vertical silos, frequently burden
professionals with two bosses—one responsible for the sales force, say,
and another for a product line. Professionals seeking to collaborate
thus need to go up the organization before they can go across it. Effective
collaboration often takes place only when the would-be collaborators
enlist hierarchical line managers to resolve conficts between competing
Article at a glance
Professional employees, who create value through
intangible assets such as brands and networks, now
constitute up to 25 percent or more of the workforce
in financial services, health care, high tech,
pharmaceuticals, and media and entertainment.
Making professionals productive enables big
corporations to be competitive, yet most of them do
little to improve the productivity of these employees.
Corporate organizational structures—designed
vertically, with matrix and ad hoc overlays—make
professional work more complex and inefficient.
Companies must change their organizational
structures dramatically to unleash the power of their
professionals and to capture the opportunities of
today’s economy.
The 21st-century organization
organizational silos. Much time is lost reconciling divergent agendas and
fnding common solutions.
Other ad hoc organizational devices, such as internal joint ventures,
co-heads of units, and proliferating task forces and study groups, serve
only to complicate the organization further and to increase the amount
of time required to coordinate work internally. The result is endless
meetings, phone calls, and e-mail exchanges as talented professionals—line
managers or members of shared utilities—waste valuable time grappling
with the complexity of a deeply fawed organizational structure.
A new organizational model
To raise the productivity of professionals, big corporations must change
their organizational structures dramatically, retaining the best of the
traditional hierarchy while acknowledging the heightened value of the people
who hatch ideas, innovate, and collaborate with peers to generate revenues
and create value through intangible assets such as brands and networks.
Companies can achieve these goals by modifying their vertical structures
to let different groups of professionals focus on clearly defned tasks—
line managers on earnings, for instance, and off-line teams on longer-term
growth initiatives—with clear accountability. Then these companies should
create new, overlaid networks and marketplaces that make it easier for
professionals to interact collaboratively and to fnd the knowledge they need.
Companies can not only build this new kind of organization but also
reduce the complexity of their interactions and improve the quality of
internal collaboration by implementing four interrelated organizational-
design principles:
1. Streamlining and simplifying vertical and line-management
structures by discarding failed matrix and ad hoc approaches and
narrowing the scope of the line manager’s role to the creation
of current earnings
2. Deploying off-line teams to discover new wealth-creating
opportunities while using a dynamic management process to resolve
short- and long-term trade-offs
3. Developing knowledge marketplaces, talent marketplaces, and formal
networks to stimulate the creation and exchange of intangibles
4. Relying on measurements of performance rather than supervision
to get the most from self-directed professionals
The McKinsey Quarterly 2005 Number 3
The ideas underlying each of these policies may not be entirely new,
but we don’t know of any company that applies all of them holistically—
and this failure limits the ability to perform up to potential. A company
that tries to simplify its vertical organizational structure without helping
large numbers of self-directed professionals to collaborate more easily
might increase its effciency, for example. But that would be more than
offset by a decrease in its effectiveness.
Simplify the line structure
The frst design principle is to clarify the reporting relationships,
accountability, and responsibilities of the line managers, who make good
on a company’s earnings targets, for all other considerations will get
short shrift until short-term expectations are met. To achieve this goal, a
company must establish a clearly dominant axis of management—product,
functional, geographic, or customer—and eliminate the matrix and
ad hoc organizational structures that often muddle decision-making
authority and accountability. Dynamic management and improved collab-
oration, as we show later, are better ways of accomplishing the purposes
of these ad hoc structures.
A company that aims to streamline its line-management structures should
create an effective enterprise-wide governance mechanism for decisions
that cross them, such as the choices involved in managing shared IT costs.
These mechanisms are typically created by defning and clarifying the
decision-making authority of each member of the senior leadership team
and establishing enterprise-wide governance committees as required. It
may also be necessary to take important support functions, which demand
focused management, out of the line structure, so that specialized
professionals (rather than line managers, who are often, at best, gifted
amateurs) can run these functions as shared utilities.
Finally, to promote the creation of enterprise-wide formal networks,
parallel structures and parallel roles should be established across the whole
extent of the company. Defning the role of the comptroller or the
country manager consistently throughout it, for example, helps the people
in those roles to interact and collaborate.
Manage dynamically
Once the newly simplifed vertical structure allows line managers to limit
their attention to meeting the near-term earnings expectations of the
company, it has the luxury of focusing other professionals on the long-term
creation of wealth. The advantages of such a separation are obvious. As
one executive we know put it, you don’t want people who are engaged in
hand-to-hand combat to design a long-term weapons program.
The 21st-century organization
Ongoing multiyear tasks such as launching new products, building new
businesses, or fundamentally redesigning a company’s technology platform
usually call for small groups of full-time, focused professionals with
the freedom “to wander in the woods,” discovering new, winning value
propositions by trial and error and deductive tinkering. Few down-
the-line managers, who must live day to day in an intensely competitive
marketplace, have the time or resources for such a discovery process.
Not that companies should forgo discipline while undertaking such a project.
In fact, the portfolio-of-initiatives approach to strategy enables them to
“plan on being lucky” by using the staged-investment processes of venture
capital and principal investing frms, as well as the R&D processes of
leading industrial corporations.
1
Companies that take this approach devote
a fxed part of their budgets (say, 2 to 4 percent of all spending) and
some of their best talent to fnding and developing longer-term strategic
initiatives. Each major one usually has a senior manager as its sponsor
to ensure that resources are well invested. Once an initiative is ready to be
scaled up—when revenues and cost projections become clear enough to
appear in the budget—it can be placed in the line structure.
Of course, at the enterprise level, companies must manage their short- and
long-term earnings in a way that integrates their spending on strategic
initiatives with the overall budget, so they will need to adopt a systemic,
effective way of making the necessary trade-offs. What we call dynamic
management can help: a combination
of disciplined processes, decision-
making protocols, rolling budgets,
and calendar-management procedures
makes it possible for companies
to manage the portfolio of initiatives
as part of an integrated senior-
management approach to running the
entire enterprise. Dynamic management forces companies to make resource
allocation trade-offs, explicitly, at the top of the house rather than allowing
them to be made, implicitly, by down-the-line managers struggling to make
their budgets. This change further simplifes the line managers’ role.
Develop organizational overlays
Having stripped away unproductive matrix and ad hoc structures from the
vertical organization and clarifed the line structure, a company must develop
How can managers translate the concept
of corporate performance into an
operational reality? See “Managing for
improved corporate performance”
on mckinseyquarterly.com.
1
Lowell L. Bryan, “Just-in-time strategy for a turbulent world,” The McKinsey Quarterly, 2002 special edition:
Risk and resilience, pp. 16–27. The primary stages of such an investment process are diagnosing the problem
or opportunity, designing a solution, creating the prototype, and scaling it up, with natural stopping points,
midcourse corrections, or both at the end of each stage.
The McKinsey Quarterly 2005 Number 3 10
organizational overlays in the form of markets and networks that help its
professionals work horizontally across its whole extent. These overlays make
it easier for them to exchange knowledge, to fnd and collaborate with
other professionals, and to develop communities that create intangible assets.
Because these market and network overlays help professionals to interact
horizontally across the organization without having to go up or down
the vertical chain of command, they boost rather than hinder productivity.
Companies that establish such overlays are making investments not
only to minimize the search and coordination costs of professionals who
exchange knowledge and other valuable intangibles among themselves
but also to maximize the opportunities for all sorts of cost-effective,
productive interactions among them.
We believe that moving simultaneously into knowledge marketplaces, talent
marketplaces, and formal networks will make all three more effective.
A knowledge marketplace, for example, helps members of a formal network
to exchange knowledge, which in turn helps to strengthen the network.
A talent marketplace works better if the people who offer and seek jobs in
it belong to the same formally networked community. In combination,
these techniques can make it possible for companies to work horizontally
in a far more cost-effective way.
Knowledge marketplaces. For the better part of the past 15 years,
knowledge management has generated a good deal of buzz. Despite heavy
investment, the benefts have been limited. Real value comes less from
managing knowledge and more—a lot more—from creating and exchanging
it. And the key to meeting this goal is understanding that the most
valuable knowledge of a company resides largely in the heads of its most
talented employees: its professionals.
Exchanging knowledge on a company-wide basis in an effective way is much
less a technological problem than an organizational one. As we have
argued, to promote the exchange of knowledge, companies must remove
structural barriers to the interaction of their professionals. These
companies must also learn how to encourage people who may not know
each other—after all, big corporations usually have large numbers of
professionals—to work together for their mutual self-interest. What’s the
best way of encouraging strangers to exchange valuable things? The
well-tested solution, of course, is markets, which the economy uses for just
this purpose. The trick is to take the market inside the company.
How can companies create effective internal markets when the product
is inherently intangible? Among other things, working markets need objects
The 21st-century organization 11
of value for trading, to say nothing of prices, exchange mechanisms, and
competition among suppliers. In addition, standards, protocols, regulations,
and market facilitators often help markets to work better.
These conditions don’t exist naturally—a knowledge marketplace is an
artifcial, managed one—so companies must put them in place.
2
In particular,
the suppliers of knowledge must have the incentives and support to codify
it (that is, to produce high-quality “knowledge objects”). “Buyers” must be
able to gain access to content that is more insightful and relevant, as well
as easier to fnd and assimilate, than alternative sources are.
Knowledge marketplaces are a relatively new concept, so they are rare.
We have found that building an effective one in a large company requires
signifcant investments to get the conditions in place—but that such
a marketplace can indeed be built. A successful mechanism of this kind
substantially improves the ability to create and exchange knowledge
and dramatically cuts search and coordination costs.
Talent marketplaces. A company can create similar effciencies by developing
a talent marketplace that helps employees in a talent pool, either within a
single organizational unit or across the enterprise, to explore alternative
assignments varying from short-term projects to longer-term operating
roles. Simultaneously, anyone with assignments to offer can review all of the
people looking for new opportunities. As with marketplaces for knowledge,
companies must invest in their talent markets to ensure that gifted men and
women looking for new jobs hook up with managers seeking talent.
Companies must defne the talent marketplace by specifying standardized
roles, validating the qualifcations of candidates, determining how managers
receive the job seekers’ performance evaluations, and so forth. The other
requirements include pricing (the compensation for a particular role or
assignment), an exchange mechanism to facilitate staffng transactions, and
protocols and standards (how long assignments run, the mechanics of
reassignment, the process of conveying decisions to reassign employees).
Talent marketplaces do exist—particularly in professional organizations—
but like knowledge marketplaces they are at an early stage of development.
Formal networks. People with common interests—such as similar work
(industrial engineers, say), the same clientele (the automotive industry), or the
same geography (China)—naturally form social networks. These networks
lower the cost of interaction while increasing its value to all participants.
2
Lowell L. Bryan, “Making a market in knowledge,” The McKinsey Quarterly, 2004 Number 3, pp. 100–11.
The McKinsey Quarterly 2005 Number 3 12
A network often provides them with increasing returns to scale: the larger
it is, the more chances they have to fnd opportunities for collaboration.
Social networks do face problems. They often have limited reach (for example,
because they don’t extend to many potential members in far-fung units and
geographies). What’s more, they sometimes operate ineffciently (several
conversations might be required to reach the right person), may rely too much
on the participants’ goodwill, and, most particularly, can fail to attract
enough investment to serve the common good of all members effectively.
The solution, for a company, is to boost the value of the network by investing
in it and formalizing its role within the organization. One such move
is the designation of a network “owner” to build common capabilities
(for instance, by making investments to generate knowledge). Others include
developing incentives for membership, defning separate territories (the
existence of more than one social network may confuse would-be members),
establishing standards and protocols, and providing for a shared infra-
structure (say, a technology platform supporting the network’s activities).
In fact, a formal network with specifc areas of economic accountability
can undertake many of the activities that have inspired companies to use
matrix management structures. A formal network relies on self-directed
people who work together out of self-interest, while a matrix uses a hierarchy
to compel people to work together. In addition, a formal network
enables people who share common interests to collaborate with relatively
little ambiguity about decision-making authority—ambiguity that generates
internal organizational complications and tension in matrixed structures.
Although social networks fourish at many companies, only a few have
formalized them. That next step, though, is one of the most important things
a company can do, because it removes unnecessary complexity from
horizontal interactions among talented people across organizational silos.
Measure performance
The fnal set of ideas rounding out this new organizational model involves
relinquishing some level of supervisory control and letting people direct
themselves, guided by performance metrics, protocols, standards, values,
and consequence-management systems.
To be sure, accountable leaders must control large companies even as many
of their workers become more and more self-directed. But what’s needed
is inspired leadership, not more intrusive management. Of course, manage-
ment will continue to be vital—particularly to get value from the many
The 21st-century organization 13
employees who will go on laboring in “industrially engineered” processes
and to hold all of a company’s workers and managers accountable for
their performance.
But as the workforce increasingly comes to consist of self-directed
professionals, leaders will have to manage them by setting aspirations and
using performance metrics that motivate them to organize their work,
both individual and collective, to meet those aspirations. One successful CEO
once told us that to motivate behavior, measuring performance is more
important than providing fnancial incentives to reward it. The challenge is
that to measure it effectively, the metrics must be tailored to individual
roles and people. Get the metrics wrong and unintended behavior is the result.
To motivate the collaborative behavior that makes this new organizational
model work, companies must create metrics that hold employees
individually accountable for their contribution to collective success—an
idea we call holding people “mutually accountable.” Such metrics
are particularly important for senior and top managers but are required,
more broadly, for all self-directed workers. People who are great at
developing the abilities of other talented people or at contributing distinctive
knowledge, for example, should be more highly valued than those who
are equally good at doing their own work but not at developing talent or
contributing knowledge.
A new organizational model for today’s big corporations will not emerge
spontaneously from the obsolete legacy structures of the industrial
age. Rather, companies must design a new model holistically, using new
principles that take into account the way professionals create value.
Big companies that follow these principles will get more value, at less cost,
from the managers and the professionals they employ. In the process,
they can become fundamentally better at overcoming the challenges—
and capturing the opportunities—of today’s economy.
Q
Lowell Bryan is a director and Claudia Joyce is a principal in McKinsey’s New York offce.
Copyright © 2005 McKinsey & Company.
All rights reserved.
The McKinsey Quarterly 2006 Number 2 14
Competitive advantage from better interactions 15
For many employees today, collaborative, complex problem solving is
the essence of their work. These “tacit” activities—involving the exchange of
information, the making of judgments, and a need to draw on multifaceted
forms of knowledge in exchanges with coworkers, customers, and suppliers—
are increasingly a part of the standard model for companies in the developed
world. Many employees engage in activities of this kind to some extent;
production workers at Toyota Motor, for instance, collaborate continually
with engineers and managers to fnd new ways of reducing costs and solving
quality problems. But employees such as managers and salespeople, whose
jobs consist primarily of such activities, now make up 25 to 50 percent of
the workforce. They are typically a company’s most highly paid workers
and make huge contributions to its competitive prospects in a fast-changing
global business landscape.
1
During the next decade, companies that make
these activities—and the employees most involved in them—more productive
will not only raise the top and bottom lines but also build talent-based
competitive advantages that rivals will fnd hard to match.
But building these advantages won’t be easy: companies must alter the way
they craft strategies, design organizations, manage talent, and leverage
Competitive advantage from
better interactions
Tacit interactions are becoming central to economic activity.
Making those who undertake them more effective isn’t like tweaking
a production line.
Scott C. Beardsley,
Bradford C. Johnson, and
James M. Manyika
L
e
i
g
h

W
e
l
l
s
1
Lowell L. Bryan and Claudia Joyce, “The 21st-century organization,” The McKinsey Quarterly, 2005
Number 3, pp. 24–33.
The McKinsey Quarterly 2006 Number 2 16
technology. The best way for
executives to begin is to understand
the nature of what economists
call tacit interactions—the search-
ing, coordinating, and monitoring
activities required to exchange
goods, services, and information.
During the past half century, the
faster pace of specialization,
globalization, and technical change
has profoundly altered companies,
their customers, the supply chains
around them, and, consequently,
the nature of work within them and
at their borders. The result is a
dramatic increase in the volume and
value of interactions.
2
In most
developed economies today, four
out of fve nonagricultural jobs
involve them; only one in fve
involves extracting raw materials
or working on a production line.
A century ago, the proportions were reversed. (This shift is under way in
the developing world as well. For a comparison of selected developed and
developing countries, see Exhibit 1.) The number of jobs chiefy involv-
ing the most complex interactions—tacit ones—is growing faster than any
other type of job in developed nations. Examples include running supply
chains, managing the way customers buy and experience products and
services, reviving brands, and negotiating acquisitions.
Companies boost their productivity by improving the effciency of
transformational activities (such as the extraction of raw materials) or of
transactions (for instance, the work of the clerks in the accounts-payable
function). But the productivity of marketing managers and lawyers can’t
be raised by standardizing their work or replacing them with machines.
(Nor can companies boost the tacit component of other jobs in this way—
automation does nothing, for example, to help the production workers
on a Toyota assembly line collaborate with others.) The old strategies for
effciency improvements don’t apply to employees whose jobs mostly
involve tacit interactions; instead, a company must boost these workers’
Article at a glance
Companies are looking for ways to improve the
effectiveness of their top talent: workers who
interact with others and draw on experience and
judgment to solve the deepest business problems.
What makes these workers valuable is their ability
to work collaboratively, to leverage “relationship
capital,” and to improvise and improve new
solutions within an environment that fosters trust
and constant learning.
To put these workers to best use, companies
must change the way they organize, hatch their
strategies, and manage their talent and IT.
The levers that managers must pull to get this
job done—flattening hierarchies and creating an
environment for constant learning—are familiar.
But it will be critical to understand exactly what
a company must do to use these workers most
effectively and how such efforts differ from other
kinds of productivity programs.
2
Patrick Butler, Ted W. Hall, Lenny Mendonca, Alistair M. Hanna, Byron Auguste, James M. Manyika, and
Anupam Sahay, “A revolution in interactions,” The McKinsey Quarterly, 1997 Number 1, pp. 4–23;
and Bradford C. Johnson, James M. Manyika, and Lareina A. Yee, “The next revolution in interactions,”
The McKinsey Quarterly, 2005 Number 4, pp. 20–33.
Competitive advantage from better interactions 17
productivity by making them more effective at what they do. As a result,
the company will build talent-based competitive advantages that are diffcult
for rivals to duplicate.
The boundaries between these three categories of business activities—
transformational, transactional, and tacit—are not static; they change
constantly as a result of innovations in the way functions and tasks are
organized and the impact of technology. Although all three are important
in today’s developed economies, it will be necessary to make a real effort
to boost the productivity of tacit interactions. Even as they become more
and more dominant, the managerial science for boosting their effective-
ness remains less well understood than are ways of increasing the effciency
of transformational and transactional activities. But that must now
change. Executives will have to learn how to compete, innovate, and manage
in an era when tacit interactions dominate and drive performance. Early
innovators are emerging, and sectors where tacit interactions have been
dominant for some time offer useful lessons.
Tacit productivity
In work of any kind, variability is a sure sign that things could be better.
Manufacturers know how to reduce variability in production work and
have therefore greatly raised their operating productivity over the past two
decades. Aided by technology, companies have adroitly smoothed variance
in call centers and IT help desk operations by standardizing interactions—
writing scripts for call-center operators, for instance—thus making tasks
r Ā n i v i 1 ŋ
.PSFKPCTSFRVJSFUBDJUJOUFSBDUJPOT
ü2 200ë
Ioteractioos
Exhibit 1 oI 2
Clorce. lr |ut| Je.eluueJ orJ Je.eluuir¸ ecuruaies, au|e ju|s |e(ui|e tocit irte|octiurs.
uriteJ Kir¸Jua ßeveIoped couotries
ßeveIopiog couotries
IraosIoræatiooaI IraosactiooaI Iacit
Cuauusitiur uí ecuruaies, ZJJ+, ° uí wu||íu|ce |] ju| t]ue
1
uriteJ Stotes
Ce|aor]
C|iro
19 Jc +5
15 ++ +1
Z5 J8 J/
lrJio +1 JJ Zc
Russio J+ J/ Z9
++ J1 Z5
ŋ
Caregoiizeu Ly µieuominanr joL acriviry: racir = comµlex inreiacrions, riansacrional = iourine inreiacrions, riansíoimarional =
exriacrion oi conveision oí iav mareiials, ŒŊŊ occuµarions sruuieu.
Souice: Buieau oí LaLoi Srarisrics, GloLal Insiglrs, Inreinarional LaLoui Oiganizarion, Unireu Þarions, Woilu Bank,
NcKinsey analysis
The McKinsey Quarterly 2006 Number 2 18
into routines. But look at work involving tacit interactions in almost any
company today; performance always fuctuates wildly.
Variability often characterizes the performance of, for instance, the sales
force. In most high-tech companies, enterprise salespeople manage a
broad number of interactions and must constantly solve problems to get
the job done. In addition to interacting with existing and potential
customers, salespeople work with marketing staffers, draw on the services
of technical-support and customer service specialists, and sort out ship-
ping problems with supply chain supervisors. Often, the variance between
the highest- and lowest-performing sales teams is wide. Effective perfor-
mance isn’t simply about generating leads and closing deals; it’s also about
how well a salesperson manages the work. In fact, interactions drive cus-
tomer satisfaction and loyalty—and, ultimately, success in sales.
Companies can analyze work done in processes and root out wasteful
activities so that employees do more in less time. But companies don’t
improve tacit interactions by forcing salespeople (or other tacit workers) to
follow a uniform procedure. On the contrary, that approach can under-
mine their effectiveness—salespeople, for instance, generate more sales and
profts when they have better information at their fngertips; can engage
in value-adding interactions with customers; are better networked with
customers, suppliers, and organizational colleagues; collaborate to develop
the better ideas that emerge from iterative teamwork; and learn and grow
in deal after deal.
That is also true for other workers engaged primarily in tacit interactions,
including software engineers at Google, Microsoft, and Yahoo!; Cisco
Systems’ manufacturing managers, who direct the connections among the
company’s salespeople, suppliers, and contract manufacturers; fund
managers at Blackstone and Fidelity Investments; doctors and nurses at
Kaiser Permanente; movie producers; merger integration managers;
and insurance agents. In insurance companies, tacit interactions now consti-
tute the primary activities of 63 percent of the workforce. The propor-
tion is 60 percent in securities companies, 70 percent in health care, and
45 percent in retailing. Even in utilities, 30 percent of the employees under-
take tacit work.
Executives recognize that they must manage these workers differently.
Managing for effectiveness in tacit interactions is about fostering change,
learning, collaboration, shared values, and innovation. Workers engage
in a larger number of higher-quality tacit interactions when organizational
Competitive advantage from better interactions 19
barriers (such as hierarchies and silos) don’t get in the way, when people trust
each other and have the confdence to organize themselves, and when they
have the tools to make better decisions and communicate quickly and easily.
These aren’t new management ideas; indeed, companies have always
had workers involved in tacit interactions. But the ever-increasing growth
in their number and value is driving companies to adopt such ideas more
quickly and deeply.
The competitive imperative
The need to move forward is both substantial and urgent, as our study of
more than 8,000 US companies with a preponderance of tacit interac-
tions suggests. We found that the
performance of companies in rela-
tively tacit-interactive sectors varied
far more than that of other compa-
nies. The level of performance
variability (defned as the standard
deviation of performance divided by the mean level of performance) was
0.9 for companies in sectors with a low level of tacit interactions. Among
companies in sectors with a middling number of tacit interactions it was
5.5, rising to 9.4 in sectors with a high level of interactions.
This widening variability as the extent of tacit interaction increases refects
two things. First, companies have considerable competitive headroom for
improving the productivity of those who undertake tacit interactions and
less headroom for improving the productivity of other workers. Second, in
some highly tacit sectors, companies in the top quartile understand how to
make tacit workers more effective and now have a signifcant performance
lead on rivals that still manage for effciency (Exhibit 2).
The top performers have also fgured out that by managing their tacit
interactions more effectively, they can create competitive advantages that
rivals in their sectors fnd hard to match—in particular, because tacit
interactions are hard to specify in advance, or “prewire.” Such interactions
involve talented people armed with experience, judgment, creativity, facts,
and the ability to connect the dots in problem solving, all of which make
their work more effective and more likely to achieve desired outcomes.
The actions and innovations of these people are usually specifc to a par-
ticular business situation. Tacit capabilities do not resemble IT systems
or reengineered processes that can be copied easily. Their power lies in the
collective company-specifc knowledge that emerges over time.
In certain highly tacit sectors,
companies in the top quartile
understand how to make their tacit
employees more effective
The McKinsey Quarterly 2006 Number 2 20
New management science
Efforts to make tacit interactions (and hence the talent that undertakes
them) more effective require changes in every facet of a business, from
hatching strategies to organization to managing talent and leveraging IT.
Each of these is essentially a piece in a set of interconnected changes.
The focus of managerial action is to establish conditions that allow tacit
interactions to emerge and fourish rather than trying to engineer connec-
tions from the top down. Management’s job is to foster connectivity, remove
barriers, facilitate learning, and provide new tools that help workers col-
laborate and learn within an environment that demands more and more
complex and often decentralized decision making.
Strategy and innovation
Wherever groups of people collaborate to solve problems—in the feld, the
supply chain, operations, marketing—innovations are more likely to occur
1FSGPSNBODFWBSJFTGPSUIFIJHIMZUBDJU
ü2 200ë
Ioteractioos
Exhibit 2 oI 2
Clorce. Cuauories cor c|eote cuauetiti.e oJ.orto¸es |] eííecti.el] aoro¸ir¸ tocit
irte|octiurs.
Average EßIIßA
1

per eæpIoyee,
2002-04, S t|uusorJ
ŋ
LBI1DA = eainings Leíoie inreiesr, raxes, ueµieciarion, anu amoirizarion.
Ō
Denneu as iario oí sranuaiu ueviarion ro mean íoi LBI1DA µei emµloyee virlin eacl inuusriy.
ō
Level oí inrensiry uenneu as: lov÷Lelov ŋŎ° inreiacrions voikeis, ligl÷ŋŎ° ro ŐŌ° inreiacrions voikeis,
liglei÷aLove ŐŌ° inreiacrions voikeis.
Souice: Comµusrar, NcKinsey analysis
r Ā n i v i 1 Ō
||ei¸|t
cuauories
9J
/
l|ur orJ
steel aills
Retoil
|or|s
lr.estaert
|or|s
higher = 9.4 -PX high = 5.5
|uw |i¸|
Ioteosity oI tacit ioteractioos
J
1,JJJ
-1JJ
J
1JJ
ZJJ
JJJ
+JJ
5JJ
cJJ
/JJ
8JJ
9JJ
0is¸uiseJ cuauor]
Z+c
JJZ
8J5
-ZJ
-8Z
Z
St|ote¸ic
oJ.orto¸e
u.e|
irJust|]
uee|s
¥ariatioo io coæpaoy-IeveI perIoræaoce
Z
Competitive advantage from better interactions 21
at the front lines of interaction than at the corporate center. Furthermore,
innovations in tacit interaction are by nature usually the result of decen-
tralized experimentation, trials, and learning.
A company can boost the number and quality of the interactions likely to
promote innovation if it creates the conditions that allow them to emerge.
Google, for instance, encourages its software engineers to devote 20 percent
of their time to pursuing their own ideas for new and innovative products
and services. Google Earth, a next-generation mapping application, was
one such product. New ideas are exposed to the market through the Google
Labs Web site and tested both inside and outside the company, which
assesses the success of initiatives by gauging how much attention and
resources they attract. Pilots that catch on are adopted and those that don’t
are shut down, so the allocation of resources is more an emergent activity
(which isn’t centrally planned or predetermined) than a managed one.
To boost the effectiveness of tacit interactions, companies must also
upend their strategic decision-making processes. Managers today com-
monly believe that more and better ideas will follow when communication
and interaction increase inside a company and beyond—with its partners,
suppliers, customers, and communities of interest—and become “multi-
directional.” But few companies bake this understanding into the develop-
ment of strategy by altering traditional top-down processes to include
mechanisms and approaches that allow a portfolio of initiatives to emerge
from internal and external interactions.
3
Not that management should
abdicate its role in setting thematic strategic goals and the company’s
direction—quite the opposite, since these become crucial to providing a
“magnetic north” as innovations occur at the interaction interfaces.
Finally, managers should construct incentives that stimulate collaboration
by encouraging innovators to share their inventions and insights within
the organization. A trader who works for one of a company’s funds isn’t
likely to share ideas that would help the people who run the rest of them
if each fund’s employees are rewarded by its performance relative to that
of the others. Rewards for collaborating and for sharing knowledge,
by contrast, help the organization as a whole rise to the level of its best
innovations. Rewards could also refect an innovator’s ability to attract
resources and users, such as customers, or refect the breadth and depth
of their personal and professional networks. Bottom-up innovators usually
don’t have the structural authority to order people to join a team; instead,
the innovator succeeds by infuencing them and leveraging a personal
network—collaborating, sharing, inspiring, and leading.
3
Lowell L. Bryan and Ron Hulme, “Managing for improved corporate performance,” The McKinsey Quarterly,
2003 Number 3, pp. 94–105.
The McKinsey Quarterly 2006 Number 2 22
Organization
To encourage more interaction, innovation, and collaboration, companies
must become more porous by continuing to break down barriers to
interactions—barriers such as hierarchies and organizational silos. While
the command and control exerted by hierarchies help a company to
manage its routine processes and tasks effciently, they also short-circuit
tacit interactions: information moves up and down a hierarchy at defned
management levels. By contrast, to stimulate interactions, organizations
want whatever information is relevant for solving a particular problem to
be shared among teams laterally, in real
time, irrespective of reporting chan-
nels and silos. What’s more, organiza-
tional structures presuppose structures
for getting work done, but tacit work
is improvisational and diffcult to
defne in advance, for it follows the
problem being solved and the nature
of the opportunity at hand.
Tacit interactions reduce the importance of structure and elevate the
importance of people and collaboration. Some of these changes are already
under way. In many companies, people now come together in project
teams, address an issue, and then disassemble to start the process again by
joining other informal teams. In fact, this approach is common in certain
professional-services and engineering frms, so their organizational charts
rarely refect what is really happening within them. Hierarchy-busting
has been a theme in the business press for years, but the pace of change has
been slow and its effectiveness questionable.
Companies will face a real challenge when they need to balance old- and
new-school management sciences, particularly if, as is often the case, their
tacit interactions are evenly balanced and intermingled with the transac-
tional and transformational activities they undertake. They will still need to
manage workers who primarily undertake transformational (production)
or transactional tasks—that is, to manage these workers for effciency—
while simultaneously enabling tacit workers to interact in more fuid
structures. The necessary balance will require trade-offs between perfor-
mance norms, on the one hand, and cooperative norms, on the other.
People, knowledge, values, and learning
Culture, metrics, and incentives will need to change as well. The kind of
network building that tacit workers must do to boost their effectiveness
thrives in a culture built on trust. It will thrive, too, in companies that
Workers will exchange information if there
is a fair return on sharing it and a clear
value for seeking it. See “Making a market
in knowledge” on mckinseyquarterly.com.
Competitive advantage from better interactions 23
reward collaboration, dispense group-based incentives, and measure tacit
work by its impact and the relationships that those who engage in it forge.
Output measures alone are far less effective in the messy, “ineffcient” world
of tacit innovation.
Few of these “soft” managerial mechanisms have legs today. How do you
measure the contribution of an employee who is 1 of 20 people on a team?
What was that employee’s contribution to the outcome? Is it even clear
whether that contribution was positive or negative? How do team managers
measure the work of people who serve on more than one team? Moreover,
as decision making becomes more decentralized and organizations grow
in size, it will become critical to articulate clearly the corporate and profes-
sional values that will provide a “compass” and shared sense of purpose
and direction to tacit workers. Evaluation processes will need to evolve to
include more peer- and project-based reviews, as opposed to the lines of
traditional reporting, and must also assess softer aspects of work such as
values, nonhierarchical leadership abilities, and mentoring skills.
Wherever tacit interactions take place, so do learning and the creation of
new knowledge. The people involved become sources of and contributors
to institutional learning. Companies can’t manage this kind of knowledge
from the top down. Instead, managers must promote its capture and
sharing by developing the right infrastructure and incentives, as well as a
“market in knowledge.” Recently, blogs (online diaries), wikis (Web sites
where users can contribute and edit content), and the like have created new,
decentralized, and dynamic approaches to the capture and dissemination
of the knowledge critical for tacit interactions.
The focus of learning changes too. Organizations can use programs delivered
in classrooms or sites to train production workers to operate lathes or call-
center personnel to handle incoming calls. But learning in the tacit world is
based much more on experience and apprenticeship and on the ways
in which both are scaled across the networks of people who participate in
interactions: inexperienced managers learn from experienced ones. Also,
managers continually change their roles—they must constantly study
examples and analogies. Companies may even fnd it worthwhile to expose
tacit workers to totally new experiences to round out their capabilities.
Finally, even hiring profles will change—indeed, in some tacit-intensive
industries, such as software and hospitals, they already have. Managers in
these organizations have redefned their job descriptions and criteria in order
to hire people who can solve problems, work under ambiguous report-
ing relationships, and network. But the pool of experienced tacit workers is
The McKinsey Quarterly 2006 Number 2 24
fnite, and demand is increasing; companies already feel the pinch. In
reaction, they may cast a wider global net for tacit talent. One thing is
clear: for tacit interactions, selecting and motivating talent are core pro-
cesses that drive effective outcomes.
Technology
Clearly, technology will play a critical role in fostering tacit interactions
and making them more effective and valuable. Indeed, technology has in large
part been responsible for the acceleration of tacit interactions over the past
20 years. Two decades ago, international calls were costly and e-mail was a
novelty; today, global voice connections are cheap, people around the
world send about 30 billion e-mails a day, and entirely new technologies—
broadband Internet, search capabilities such as Google, mobile phones,
personal digital assistants such as BlackBerries and Treos, and video-
conferencing—make it possible for tacit interactions to happen more easily.
Companies will increasingly focus on these kinds of technologies to further
improve tacit work, thereby raising a host of new IT-management issues.
The bulk of corporate investment in technology has been devoted to improv-
ing transactional and even transformational activities. New investments
in PDAs, collaborative software, wiki tools, and other technologies that
improve tacit interactions will be far less costly than, say, enterprise
transactional systems. But they do require new IT architectures and skills.
Some companies are already getting it right. Tacit-dominated sectors in the
top quartile of labor productivity growth have armed their employees with
fve times more IT stock than sectors in the bottom quartile. Further, they
are increasing their IT base per employee 40 percent more rapidly (on an
annual basis).
The good news is that, with few exceptions, most enterprises now have an
underlying communications infrastructure, which is vital for extending
the reach of interactions. This infrastructure must go on evolving to provide
a foundation for ever-richer media. Companies will increasingly need
to deploy technology that makes shared data, information, and expertise
available in real time; to offer decision support tools that help workers
involved in tacit interactions create insights from data and analyses and that
enhance the context and information that interactions require; to improve
the ability of employees, customers, and suppliers to interact; and to offer
effective collaboration tools for multiparty work fows.
Many of the technologies and tools that tacit workers are going to use will
promote the collaborative and dynamic pursuit, capture, and sharing of
knowledge and will allow for more video, audio, and graphics to facilitate
remote interactions and broader access to scarce expertise. Tools based on
Competitive advantage from better interactions 25
search capacities, collaborative approaches to capturing and organizing
knowledge, and new digital-learning channels are likely to emerge.
Further, these new tools and approaches have broader implications for
the way companies manage their IT infrastructures and operations. The
new world won’t use either the practices and organizing principles that
production-support and transaction systems require or the big, rigid busi-
ness applications designed to automate transactions and make them
effcient. Not that enterprise applications won’t be required; rather, they
must evolve to make tacit interactions more effective. Executives will
have to focus on deploying work-group-centric tools that are easy to set
up and tear down as projects and strategic experiments come and go. They
will also have to fnd ways of connecting these tools easily to preexist-
ing interaction platforms. In addition, the issue of information overload
must be addressed: already, Microsoft and others are trying to help tacit
workers flter data from interactions more successfully and thereby reduce
the burden of the excess information created when companies manage
interactions (and their rate of increase) improperly.
Managing in an environment where most workers mainly participate in
interactions will upend the greater part of what senior management has
learned over the past half century. But the opportunity to create new forms
of competitive advantage is clear for companies that take a new approach.
The time to start is now.
Q
The authors wish to acknowledge the contributions of their colleagues
Lowell Bryan, Dan Ewing, Roger Roberts, and Emily Szydlowski, as well as of John Hagel III;
Professor Hal Varian, of the University of California, Berkeley; and the
Cisco Thought Leadership Team.

Scott Beardsley is a director in McKinsey’s Brussels offce,
Brad Johnson is an associate principal in the Silicon Valley offce, and James Manyika is
a principal in the San Francisco offce. Copyright © 2006 McKinsey & Company.
All rights reserved.
The new metrics of corporate performance: Profit per employee 27
The new metrics of corporate
performance: Profit per
employee
Most measurements of performance are geared to the needs of
20th-century manufacturing companies. Times have changed. Metrics
must change as well.
Lowell L. Bryan
Let’s get right to the point: companies focus far too much on measur-
ing returns on invested capital (ROIC) rather than on measuring the contri-
butions made by their talented people. The vast majority of companies
still gauge their performance using systems that measure internal fnancial
results—systems based on metrics that don’t take suffcient notice of the
real engines of wealth creation today: the knowledge, relationships, repu-
tations, and other intangibles created by talented people and represented
by investments in such activities as R&D, marketing, and training.
Increasingly, companies create wealth by converting these “raw” intangibles
into the institutional skills, patents, brands, software, customer bases,
intellectual capital, and networks that raise proft per employee and ROIC.
These intangibles are true capital, in the sense of delivering cash returns,
even though the sources of those returns are intangible. Indeed, the most
valuable capital that companies possess today is precisely intangible rather
than fnancial.
1
Companies should redesign their fnancial-performance
metrics for this new age.
1
Karl Erik Sveiby, The New Organizational Wealth: Managing and Measuring Knowledge-Based Assets,
San Francisco: Berrett-Koehler Publishers, 1997.
K
e
r
r
y

G
r
a
d
y
The McKinsey Quarterly 2007 Number 1 28
Consider a simple approximation
of intangible capital: the market
value of a company less its invested
fnancial capital. Using book
capital as a crude proxy for fnan-
cial capital, in 2005 the intan-
gible capital of the world’s largest
150 companies was $7.5 trillion,
versus $800 billion in 1985.
Despite the evidence that intan-
gibles are now the true source of
corporate wealth, companies
tightly control discretionary spend-
ing on them. Advertising, R&D,
new-product development, training,
knowledge creation, software
projects, and so forth are almost
always expensed on a “What can
we afford?” basis. Why?
One reason is that accounting for
intangibles is diffcult. In particular,
each intangible’s specifc contri-
bution is hard to assess; how, for
example, do you value a brand?
Intangibles are embedded in the
value chain of production, so it
generally isn’t clear which intan-
gibles are the sources of profts—
or what specifc balance of intangi-
ble and tangible assets should
get the credit (or blame) for results.
The bigger problem is that most companies gear the way they measure their
fnancial performance to the needs of an earlier industrial age, when
capital enjoyed pride of place in the minds of strategists and investors. Com-
panies fll their annual reports with information about how they use
capital but fail to refect suffciently on their use of the “thinking-intensive”
people who increasingly drive wealth creation in today’s digital econ-
omy. The development of external fnancial reports according to generally
accepted accounting principles (GAAP) ranks among the principal foun-
dations of our modern global capital marketplace. Financial performance
Article at a glance
Today’s approach to measuring financial
performance is geared excessively to the
capital-intensive operating styles of 20th-century
industrial companies. It doesn’t sufficiently
account for factors such as the contributions of
talented employees that, more and more, are the
basic source of wealth.
Financial performance—observed through
balance sheets, cash flow reports, and income
statements—is and always will be the principal
metric for evaluating a company and its
managers. But greater attention should be paid
to the role of intangible capital and the ways of
accounting for it.
The superior performance of some of the largest
and most successful companies over the past
decade demonstrates the value of intangible
assets.
Companies can redesign the internal financial-
performance approach and set goals for the
return on intangibles by paying greater attention
to profit per employee and the number of
employees rather than putting all of the focus on
returns on invested capital.

Related articles
on mckinseyquarterly.com
“The 21st-century organization,”
2005 Number 3
“Managing for improved corporate performance,”
2003 Number 3
“Why the biggest and best struggle to grow,”
Web exclusive, January 2004
The new metrics of corporate performance: Profit per employee 29
(seen through balance sheets, cash fow reports, and income statements) no
doubt is and will remain the principal metric for evaluating a company
and its management. But it’s time to recognize that fnancial performance
increasingly comes from returns on talent, not on capital.
GAAP accounting currently treats investments in intangibles conservatively,
compared with the way it treats capital investments in tangible assests. Intan-
gible investments are mostly expensed, not capitalized. This conservatism
isn’t necessarily bad but does inspire top managers to cut discretionary spend-
ing on intangibles in order to deliver quick earnings. That approach may
raise short-term profts but can also undermine a company’s long-term health.
To boost the potential for wealth creation, strategically minded executives
must embrace a radical idea: changing fnancial-performance metrics to
focus on returns on talent rather than returns on capital alone. This shift
in perspective would have far-reaching implications—for measuring
performance, for evaluating executives, even for the way analysts measure
corporate value. Only if executives begin to look at performance in this
new way will they change internal measurements of performance and thus
motivate managers to make better economic decisions, particularly about
spending on intangibles.
Measuring financial performance in the digital age
Before exploring the new metrics needed to achieve these goals, let’s
refect upon the way some companies have recently created great wealth by
using their thinking-intensive people rather than their capital.
In past articles, my colleagues and I have examined how, from 1995 to
2005, the top 30 of the very largest companies in the world (ranked by market
capitalization) have seen their proft per employee rise to $83,000, from
$35,000.
2
On average, the number of people these companies employ has
grown to 198,000, from 92,000, and their ROIC (or book value, in the
case of fnancial institutions) has increased to 23 percent, from 17 percent
(Exhibit 1). As a result, the median market cap of this group of companies
rose to $168 billion, from $34 billion, with total returns to shareholders
(TRS) at 17 percent a year. The driver of this dramatic rise in market cap
was a fvefold increase in average profts—an increase brought on in turn
by a more than 100 percent jump in proft per employee and a doubling in
the number of employees. By comparison, these companies’ ROIC increased,
over this same period, by only a third.
2
Lowell L. Bryan and Michele Zanini, “Strategy in an era of global giants,” The McKinsey Quarterly,
2005 Number 4, pp. 46–59.
The McKinsey Quarterly 2007 Number 1 30
It is hardly a surprise that growth in profts and market caps should be
closely correlated and that a fvefold increase in profts should lead to a
similar increase in market caps. But these results do suggest that com-
panies need to take a new approach to measuring fnancial performance—
an approach based on maximizing returns on people. Total proft, after
all, is the product of proft per employee and the total number of employees,
so maximizing both expressions increases total proft, which drives
market capitalization.
Concentrating on this formula (as opposed to returns on capital) offers
several advantages. For one, unlike ROIC, proft per employee is a good proxy
for earnings on intangibles, partly because the number of people a com-
pany employs is easy to obtain. Capital, perhaps surprisingly, is subject to
the vagaries of accounting defnitions and such corporate-fnance deci-
sions as debt-to-equity ratios, dividend policies, and liquidity preferences.
As we’ve noted, and as any executive will testify, talent—not capital—
is usually the scarcer resource.
Clearly, then, a new set of metrics could help companies gauge their perfor-
mance more effectively. Executives should home in, frst, on how much
proft per employee a company generates. They should make the number of
employees a key factor in strategic thinking. And they should keep a clear
eye on ROIC, but more as a way of ensuring that the company earns more
than the cost of that capital than as an aspiration in its own right. With
these metrics, the company can set its goals for the return on intangibles
Q1 2007
Profit per employee
Exhibit 1 of 4
Glance: From1995 to 2005, the 30 largest companies (by market capitalization) have seen their
profits per employee increase dramatically.
r x n i v i 1 i
Soaring profits
Drivers of growth for 30 largest companies,
1
1995–2005
i
US anu íoieign comµanies Ly Ameiican ueµosiraiy ieceiµrs, roµ jo Ly maiker caµiralizarion in zooj, excluues ourlieis anu
comµanies virl negarive ner incomes.
2
1oral ieruins ro slaielolueis.
3
Oi Look value, in rle case oí nnancial insrirurions.
34
1995
168
2005
Median market
capitalization, $ billion
35
1995
83
2005
Profits per employee,
$ thousand
92
1995
198
2005
Average number of
employees, thousands
17
1995
23
2005
Return on invested
capital (ROIC),
3
%
CAGR = 9% CAGR = 8% Compound annual growth
rate (CAGR) of TRS
2
= 17%
CAGR = 3%
The new metrics of corporate performance: Profit per employee 31
(that is, proft per employee) and growth (the number of employees), as
well as its return on capital, which is largely a sanity check. Together, these
three metrics squarely highlight—and drive—market caps.
Proft per employee
If a company’s capital intensity doesn’t increase, proft per employee is a
pretty good proxy for the return on intangibles. The hallmark of fnancial
performance in today’s digital age is an expanded ability to earn “rents”
from intangibles.
3
Proft per employee is one measure of these rents. ROIC is
another. If a company boosts its proft per employee without increasing
its capital intensity, management will increase its rents, just as raising ROIC
above the cost of capital would. The difference is that viewing proft per
employee as the primary metric puts the emphasis on the return on talent.
This approach focuses the minds of managers on increasing proft relative
to the number of people a company employs. It suggests that the most
valuable use of an organization’s talent is the creation and use of intangibles.
Fortunately, the opportunities to increase proft per employee are unprece-
dented in a digital economy, where intangible assets are a rich source
of value. Opportunities to improve ROIC to an equal extent are hardly
as plentiful.
Another advantage of proft per employee is that it requires no adjustment for
accounting conventions. Since companies expense their spending on intan-
gibles but not on capital investments (which are usually depreciated over
time), proft per employee is a conservative, output-based measure. And
since it is based on accounting conventions, companies can easily benchmark
it against the comparable results of competitors and other companies.
Proft per employee therefore focuses companies on intangible-intensive
value propositions and, in turn, on talented people—those who, with some
investment, can produce valuable intangibles.
Number of employees
One way to improve a company’s proft per employee is simply to shed
low-proft employees. But if they generate proft greater than the cost of the
capital used to support their work, shedding them actually reduces
the creation of wealth, unless management adds an offsetting number of
workers who produce a higher proft per employee.
The Walton family, remember, consistently sits atop the Forbes annual
wealth list. Why? Because Wal-Mart Stores, the company the family controls,
3
Economists defne rent as the proft earned after a company pays for all of the factor costs of production
(labor, raw materials, and so forth), including the cost of capital.
The McKinsey Quarterly 2007 Number 1 32
not only hires large numbers of employees who generate a relatively low
average proft
4
but also uses a business model that enables it to handle the
complexity involved in managing huge numbers of employees, without
incurring offsetting diseconomies.
Real wealth creation therefore comes from increasing either a company’s
proft per employee (without offsetting reductions in the number of
employees or offsetting increases in capital intensity) or the number of
employees who earn that level of proft—or both. We can observe this
dynamic on a simple grid that illustrates the source of the proft earned by
a company and a competitor (Exhibit 2). The grid also shows how total
employment can serve as a crude proxy for the internal complexity of any
organization, particularly when it is compared with companies in similar
industries that have a comparable employment mix. From this vantage
point, proft per employee becomes a proxy for how well a company man-
ages that complexity.
A company can, of course, streamline its organization and use tools such
as formal networks, talent marketplaces, and knowledge marketplaces
5

4
In 2004 Wal-Mart employed 1.7 million people, who generated an average proft of $6,200 each.
5
For more information on talent markets, see Lowell L. Bryan, Claudia I. Joyce, and Leigh M. Weiss, “Making
a market in talent,” The McKinsey Quarterly, 2006 Number 2, pp. 98–109. For more information on
knowledge markets, see Lowell L. Bryan, “Making a market in knowledge,” The McKinsey Quarterly, 2004
Number 3, pp. 100–11.
Q 1 2007
Profit per employee
Exhibit 2 of 4
Glance: Companies can create wealth either by increasing profit per employee, by increasing the
number of employees earning such profits, or both.
r x n i v i 1 z
Talent as profit driver
Drivers of profit for 30 largest companies,
1
2002–04 (average)
Employees, thousands
P
r
o

t

p
e
r

e
m
p
l
o
y
e
e
,

$

t
h
o
u
s
a
n
d
Total
profit,
$ billion
i
US anu íoieign comµanies Ly Ameiican ueµosiraiy ieceiµrs, roµ jo Ly maiker caµiralizarion in zoo¸, excluues ourlieis anu
comµanies virl negarive ner incomes.
z
Royal Bank oí Scorlanu.
NTT DoCoMo Exxon Mobil
Microsoft
Chevron
AIG
BP
Royal Dutch/Shell
GSK
Pfizer
RBS
2
Wells Fargo
Berkshire
Hathaway Verizon
Toyota
Citigroup
IBM
Wal-Mart
Bank of
America
Dell
Intel
250
200
150
100
50
0
0 50 100 150 200 250 300 350 400 450 500 1,533
6
12
18
3
Profits driven by
size of workforce
Profits driven by
productivity
GE
The new metrics of corporate performance: Profit per employee 33
to mobilize intangibles throughout the enterprise. To the extent that it
does so, its proft per employee should increase, even in the absence of proft-
able new value propositions, if it removes any unproductive complexity.
Returns on capital
A company can also improve its proft per employee by substituting capital
for labor costs. Of course, while capital is relatively inexpensive and readily
available, it demands a return and for this reason must be used care-
fully. But if the company uses total employment to drive its growth aspira-
tions, the amount of capital it requires will be a derivative of the capital
its employees need for their work, rather than an independent aspiration.
Executives should therefore look at ROIC mainly as a sanity check. So
long as the return exceeds the cost, proft per employee is the better metric
because it not only represents the scarcest resource but also refects
proft after the expensing of necessary investments. Capital investment,
meanwhile, is depreciated or amortized.
Using the total number of employees as a metric also allows companies to
avoid subjective accounting judgments.
6
Book capital, on the other
hand, is—surprisingly—relatively ambiguous, for it is subject to some-
what arbitrary accounting conventions that involve goodwill, depreciation
schedules, and the way companies expense stock options, among other
things. Calculations of a company’s ROIC have their own limitations, partic-
ularly for fnancial institutions, whose assets are mostly fnancial.
Invested capital is not only a meaningless concept for such companies but
also requires them to make some heroic assumptions.
7
Maximizing market capitalization
The goal of these efforts to reorient fnancial-performance metrics around
talent, of course, is to maximize a company’s market cap, perhaps the
most important single measure of size and economic relevance. The market
cap directly affects a company’s ability to control its own strategic destiny
and is highly correlated with its total net income; of the top 30 companies
by net income from 2002 to 2004, all but 5 were in the top 30 by market
6
According to some observers, the many temporary contractual workers that certain large companies use
should be counted as employees. I disagree. These workers may depend on the company for work, but they are
largely fungible labor and usually don’t undertake the intensive intangible work that drives a company’s
profts. This is exactly why companies choose to rely on contractual labor.
7
See Felix Barber and Rainer Strack, “The surprising economics of a ‘people business,’” Harvard Business
Review, June 2005, Volume 83, Number 6, pp. 80–90, in which the authors propose using economic proft
per employee to gauge the true performance of “people businesses.” Economic proft subtracts the cost
of capital from proft per employee. Proft per employee is a more practical metric, as it can be taken directly
from accounting statements and allows for straightforward comparisons of performance across companies.
(Calculating economic proft per employee often requires internal company data.) A related concept, economic
contribution per employee, can be a useful internal metric.
The McKinsey Quarterly 2007 Number 1 34
value. A company can expose this correlation by displaying its net income
as the return on book equity multiplied by book equity and then compar-
ing that relationship with its total market cap disaggregated (in a strategic-
control map) into its market-to-book ratio multiplied by book equity
(Exhibit 3). The company can also see this same correlation by disaggre-
gating net income, using proft per employee and the total number of
employees. Doing so displays the total market cap as a function of the latter
and the market cap per employee (Exhibit 4).
Net income and market cap can therefore be regarded as functions of the
return on either capital or talent. The point is that although the two metrics
produce similar results, return on talent is a more powerful model in a
competitive environment where the intangible assets that talented employees
create provide the greater part of new wealth.
Q1 2007
Profit per employee
Exhibit 3 of 4
Glance: The correlation of market capitalization to net income can be viewed in relation to
returns on invested capital (ROIC) . . .
r x n i v i 1 j
The return-on-capital lens
Net income and market capitalization shown as returns on invested capital (ROIC), 2002–04 (average)
Book value, $ billion
Net
income,
$ billion
R
e
t
u
r
n

o
n

e
q
u
i
t
y
,

%
1
US anu íoieign comµanies Ly Ameiican ueµosiraiy ieceiµrs, ;i° oí roµ jo Ly income aie also in roµ jo Ly maiker caµiralizarion.
Souice: GloLal Vanrage, NcKinsey analysis
70
50
60
40
20
30
10
0
20
30
10
5
0 20 40 60 80 100 120
Book value, $ billion
Market
value,
$ billion
R
a
t
i
o

o
f

m
a
r
k
e
t

v
a
l
u
e

t
o

b
o
o
k

v
a
l
u
e
7
5
6
10
8
9
2
3
4
1
0
200
300
100
50
0 20 40 60 80 100 120
Pfizer
Royal Dutch/Shell
Toyota Motor
Wal-Mart Stores
Wells Fargo
AIG
Bank of America
Berkshire Hathaway
BP
Chevron
Citigroup
NTT DoCoMo
Verizon Communications
Vodafone
AT&T
Freddie Mac
ING
PetroChina
Exxon Mobil
GE
GlaxoSmithKline
IBM
Intel
Microsoft
Overlap
GSK
IBM
Intel
AT&T
Chevron
PetroChina
Microsoft
AIG
Bank of America
Exxon
Mobil
Citigroup
BP
Toyota
Berkshire
Hathaway
Wal-Mart
Pfizer
ING
Freddie Mac
Wells Fargo
GSK
Toyota
NTT DoCoMo
Chevron
Wells Fargo
Verizon
AIG
BP
Berkshire
Hathaway
Exxon Mobil
Bank of
America
Citigroup
Wal-Mart
Pfizer
Microsoft
Intel
IBM
Top 30
1
by net income
Rest of top 150 by
market cap
Top 30
1
by market cap
Rest of top 150 by
net income
GE
GE
The new metrics of corporate performance: Profit per employee 35
Today’s annual reports are flled with information about how companies
use capital but offer little about the number of employees, the mix of
employees, or the different kinds of employees (beyond a simple expense
item on compensation and benefts). Yet it is thinking-intensive talent,
not capital, that now drives the creation of wealth and thus deserves to be
measured more precisely by strategically minded executives.
 Q
Lowell Bryan is a director in McKinsey’s New York offce.
This article is adapted from his forthcoming book, Mobilizing Minds: Creating Wealth from 
Talent in the 21st-Century Organization, McGraw-Hill, spring 2007. Copyright © 2007
McKinsey & Company. All rights reserved.
Q 1 2007
Profit per employee
Exhibit 4 of 4
Glance: . . . or in relation to return on talent.
r x n i v i 1 ¸
The return-on-talent lens
Income and market capitalization shown as returns on talent, 2002–04 (average)
Employees, thousand
N
e
t

i
n
c
o
m
e
,

$

t
h
o
u
s
a
n
d
1
US anu íoieign comµanies Ly Ameiican ueµosiraiy ieceiµrs. ;i° oí roµ jo Ly income aie also in roµ jo Ly ner income.
Souice: GloLal Vanrage, NcKinsey analysis
250
200
50
100
150
0
12
18
6
3
0 50 100 150 200 250 350 300
Employees, thousand
M
a
r
k
e
t

c
a
p
i
t
a
l
i
z
a
t
i
o
n

p
e
r

e
m
p
l
o
y
e
e
,

$

m
i
l
l
i
o
n
4
6
5
2
3
1
0
200
300
50
0 50 100 150 200 250 300 350
Pfizer
Royal Dutch/Shell
Toyota Motor
Wal-Mart Stores
Wells Fargo
AIG
Bank of America
Berkshire Hathaway
BP
Chevron
Citigroup
Coca-Cola
NTT DoCoMo
Verizon Communications
Vodafone
AT&T
Freddie Mac
ING
PetroChina
Exxon Mobil
GE
GlaxoSmithKline
IBM
Intel
Microsoft
Overlap
Top 30
1
by net income
Rest of top 150 by
net income
Top 30
1
by market
capitalization
Rest of top 150 by
market capitalization
Net
income,
$ billion
Market
value,
$ billion
Microsoft
NTT DoCoMo
Exxon Mobil
Vodafone
Chevron
Royal
Dutch/Shell
Intel
GSK
BP
Wells Fargo
Bank of America
Citigroup
Verizon
Toyota
IBM
Berkshire
Hathaway
AIG
Pfizer
Exxon Mobil
Microsoft
Chevron
BP
AIG
GSK
Intel
Pfizer
ING
AT&T
Wells Fargo
Berkshire
Hathaway
Royal Dutch/Shell
Bank of America
Citigroup
IBM
Toyota
GE
GE
The McKinsey Quarterly 2006 Number 3 36
Mapping the value of employee collaboration 37
Although collaboration is at the heart of modern business processes, most
companies are still in the dark about how to manage it. Linear, process-
based tools such as activity-based costing, business process reengineering,
and total quality management have long been effective at measuring
and improving the effciency of people and organizations in accomplishing
individual tasks. But they do a poor job of shedding light on the largely
invisible networks that help employees get things done across functional,
hierarchical, and business unit boundaries.
1
This blind spot has become problematic. Falling communications costs,
globalization, and the increasing specialization of knowledge-based work
have made collaboration within and among organizations more important
than ever. As “tacit” interactions replace more routine economic activity
and the scale and complexity of many corporations creep upward, the need
to manage collaboration is growing.
2
Nearly 80 percent of the senior
Mapping the value of
employee collaboration
As collaboration within and among organizations becomes increasingly
important, companies must improve their management of the networks
where it typically occurs.
Robert L. Cross, Roger D. Martin,
and Leigh M. Weiss
S
t
u
d
i
o
/
l
a
b
1
This article focuses on intracompany interactions and collaborations rather than those extending beyond
the boundaries of the enterprise. For more about the latter, see John Seely Brown and John Hagel III,
“Creation nets: Getting the most from open innovation,” The McKinsey Quarterly, 2006 Number 2,
pp. 40–51.
2
For more on tacit interactions, which involve the exchange of information, the making of judgments, and
a need to draw on multifaceted forms of knowledge in exchanges with coworkers, customers, and suppliers,
see Scott C. Beardsley, Bradford C. Johnson, and James M. Manyika, “Competitive advantage from
better interactions,” The McKinsey Quarterly, 2006 Number 2, pp. 52–63. For more on the evolution of large
corporations, see Lowell L. Bryan and Michele Zanini, “Strategy in an era of global giants,” The McKinsey
Quarterly, 2005 Number 4, pp. 46–59.
The McKinsey Quarterly 2006 Number 3 38
executives surveyed in a 2005 study
said that effective coordination
across product, functional, and geo-
graphic lines was crucial for growth.
Yet only 25 percent of the respon-
dents described their organizations
as “effective” at sharing knowledge
across boundaries.
3
Many companies have responded
by spending heavily on collabora-
tion software. In hopes of dissemin-
ating best practices and sharing
expertise, a few leaders (such as BP,
HP, IBM, P&G, and Xerox) have
even begun identifying networks of
employees doing similar work.
Technology, though, at best fails
to deal with the underlying problem
and at worst becomes a source of information overload that undermines
effective collaboration. And it’s often unclear whether efforts to enhance
networks promote productive collaboration or just consume money
and time.
What companies need in a collaborative age is the ability to map and
analyze the value created (or destroyed) deep within employee networks.
Sophisticated network analysis approaches have emerged from the
academic world during the past two decades. But they have tended to focus
more on individual than organizational effectiveness and on commu-
nications, work fows, and the exchange of resources rather than on the
value those interactions create.
4
To make these tools more useful, executives
must reorient them toward the revenue and productivity benefts that
collaborative interactions generate, the costs such interactions impose, and
opportunities to improve connectivity at the points that create the greatest
economic value.
Article at a glance
Falling communications costs, globalization, and
the increasing specialization of knowledge-based
work are making collaboration within and among
organizations increasingly important.
Yet few companies understand or know how
to manage the intracompany networks in which
collaboration typically occurs.
A few leading companies are beginning to map their
networks of relationships, to analyze the economic
costs and benefits that key interactions create, and
to identify value-creating interventions.
Successful interventions help companies to reduce
complexity, redefine roles, and allocate financial,
physical, and human resources more efficiently.
3
For the full survey results, see “The McKinsey Global Survey of Business Executives, July 2005,”
mckinseyquarterly.com, July 2005.
4
See Thomas J. Allen, Managing the Flow of Technology, Cambridge, MA: MIT Press, 1984; Ronald S. Burt,
Structural Holes: The Social Structure of Competition, Cambridge, MA: Harvard University Press, 1995;
David Krackhardt and Jeffrey R. Hanson, “Informal networks: The company behind the chart,” Harvard
Business Review, July 1993, Volume 71, Number 4, pp. 104–11; Wayne E. Baker, Achieving Success through
Social Capital: Tapping Hidden Resources in Your Personal and Business Networks, San Francisco:
Jossey-Bass, 2000; and Rob Cross and Andrew Parker, The Hidden Power of Social Networks: Understanding
How Work Really Gets Done in Organizations, Cambridge, MA: Harvard Business School Press, 2004.
Mapping the value of employee collaboration 39
Consider the experience of a leading biotechnology company that relied
on sharing best practices among quality control engineers to help its
manufacturing facilities rapidly ramp up the production of new products.
Network analysis showed the company which engineers took part in the
interactions that generated time savings and the greatest and lowest cost,
respectively. Aggregated to reveal the economic value created through
interactions across locations, these fgures identifed the places in the net-
work where collaborative breakdowns inhibited the transfer of proven
practices and showed how costly these breakdowns were. As a result, the
company knew exactly where it made economic sense to invest in tools,
training, and team-building efforts.
Organizations hoping to emulate the biotechnology company (and
other pioneers in a wide range of sectors) must frst map their collaborative
networks and then analyze the economic benefts and costs that key
interactions within those networks create. Once executives understand the
value that’s fowing across networks, they can intervene in straight-
forward, cost-justifed ways. Typical examples include replicating high-
performing networks, training workers to emulate the collaborative
approaches of successful colleagues, making valuable expertise and advice
more readily available, and revamping performance metrics to refect
mutual accountabilities better. These kinds of successful interventions can
help companies reduce complexity, redefne roles, serve customers
and clients more effectively, and allocate fnancial, physical, and human
resources more effciently.
Understanding how work really gets done
Three examples will show how traditional ways of mapping processes
and analyzing activities have limits when it comes to understanding the
performance of individuals, teams, and entire organizations.
• Individual performance. A nonproft wanted to boost its fund-raisers’
productivity. Conventional wisdom suggested targeting certain types of
donors, managing the sales process in a defned sequence, and per-
suading donors through appeals tailored to their interests. Yet some high
performers followed few of these practices; several low performers
embraced them all.
• Team performance. A program to improve the processes of a large
global construction company boosted the effciency of its employees. But
performance disparities remained across sales offces, even after control-
ling for the varying attractiveness of their markets.
The McKinsey Quarterly 2006 Number 3 40
• Organization-wide performance. An engineering company experienced
growing pains as international expansion made it increasingly diffcult
to bring together construction managers and engineers, whose objectives
were frequently in confict. (The former focused on cutting costs,
the latter on technical solutions.) Unfortunately, the company’s linear
view of the construction process—emphasizing the tasks performed
by each group and the handoffs between them—shed little light on
collaborative issues.
Network analysis can help companies in circumstances like these. The
frst step is identifying the functions or activities where connectivity seems
most relevant and then mapping relationships within those priority areas.
Options for obtaining the necessary information include tracking e-mail,
observing employees, using existing data (such as time cards and project
charge codes), and administering short (5- to 20-minute) questionnaires.
Organizations mapping their decision-making processes might ask their
employees, “Whom do you ask for advice before making an important
decision?” Others targeting innovation might ask, “With whom are you
most likely to discuss a new idea?” Questions are posed bidirectionally:
if Joe says he was helpful to Jane, but she says she doesn’t know him,
his claim is disregarded. With the information in hand, companies can use
standard software to create network maps illustrating relationships
(Exhibit 1).
So far, so familiar. The real value comes when companies move from
mapping interactions to quantifying the benefts and costs of collaboration.
To do so, companies must assess the time employees spend on interactions
of various types, as well as the savings and sales contributions of specifc
collaborations. Key inputs to this analysis include fully loaded compen-
sation fgures for network participants and detailed survey results (for
example, the responses to queries such as, “How much time did working
with employee X save you?” or “On how many deals in the following
revenue bands did you work with employee Y?”).
Network analysis helped the companies described earlier address their
individual, team, and organization-wide performance issues. It turned out,
for example, that high-performing fund-raisers not only had strong
relationships with donors but also accounted for a disproportionate share
(25 percent) of the connections within the fund-raising group. Tenure
and experience were key reasons for the high performers’ strong networks,
so the organization was caught in a vicious cycle: low-tenure fund-
raisers got stuck on the fringes of both their internal and external networks,
became dissatisfed, and quit before they became productive. By helping
Mapping the value of employee collaboration 41
new fund-raisers rapidly replicate the high performers’ networks, the
nonproft expected to increase its revenue from employees with no more
than two years’ tenure by nearly 200 percent.
For the construction company, network analysis revealed that a key
distinction between the strongly and poorly performing offces was the
percentage of collaborative time (68 percent for the former, 50 percent
for the latter) that account managers spent with customers. By getting at
the roots of these issues—which further analysis attributed to hierarchy,
organizational design, and project-management processes—the construc-
tion company replicated the network orientation of high-performing
offces in poorly performing ones.
Finally, an analysis of one of the engineering company’s high-performing
groups showed that a small number of construction managers and engineers
single-handedly accounted for 35 percent of all the collaboration occur-
ring within it. This kind of collaboration dramatically enhanced the group’s
ability to deliver expertise. Identifying and building connectivity between
specialists in other groups helped the frm to raise its construction revenue
to $275 million, from $80 million, in a single year.

Reading a network map
Q3 2006
Network mapping
Exhibit 1 of 4
Glance: A network map illustrates relationships.
Represents membership in or affiliation with group
or category
1
Color
Shape Represents different types of nodes (eg, one shape
for individual, entity; another shape for projects,
articles, laboratories)
Node = individual or entity in network
Larger size represents greater degree of an assigned
value—eg, degree of connectedness, centrality
Size
Tie or link = connection between 2 nodes;
represents an assigned value—eg, decision-
making input, information transmission
Thicker line indicates more frequent interactions Line weight
Shows direction of interaction, eg, Kristina
provides James with decision-making input
Arrow
Spatial relationships
Nodes with highest number of connections are
located in center of map; those on periphery have
fewest connections
Location
of nodes
Patrick
Anya
Cho
Kristina
Project B
Federic
Project A
James
Selected elements of a sample network map

Group or category such as business unit, gender, geography, role, tenure.
Arun
Tanaka
Jon
The McKinsey Quarterly 2006 Number 3 42
Creating relational value
The powerful results of identifying and replicating high-performing
networks represent only a small part of the potential of network analysis.
It’s also possible to promote specifc interactions that help generate revenue
and boost productivity. Targeted
action is dramatically more effective
than promoting connectivity
indiscriminately, which typically
burdens already-overloaded employees
and yields network diseconomies.
A more informed network perspective
helps companies to identify the few
critical points where improved
connectivity creates economic value by cutting through business unit and
functional silos, physical distance, organizational hierarchies, and a scarcity
of expertise.
Generating revenue
A network view often uncovers “hidden” people whose contribution to
cross-selling or closing deals is far greater than individually focused
performance metrics might imply. It can also suggest where to replicate
collaborative behavior, when to draw in valuable experts from the
network’s fringe, and how to eliminate obstacles to collaborative sales
efforts—obstacles that include time, skills, personalities, incentives,
and ignorance of which colleagues have expertise. The experiences of a
global technology company and a consulting frm illustrate how these
issues play out in practice.
Improving cross-selling. A leading technology company used network
analysis during an effort to become more responsive to customers and
marketplace shifts. The analysis not only helped the company’s leaders fnd
out where collaboration generated revenue but also proved useful for
reframing the roles of key players in the network.
The company, for example, broke out collaborative contributions by bands
of revenue and learned that the most and least valuable interactions (those
generating more than $2,000,000 and less than $250,000, respectively)
invariably involved different people. What’s more, a network perspective
helped the company identify which colleagues knew about one other’s
expertise but didn’t draw on it. (Exhibit 2 shows how many people said
they were acquainted but saw no possibility of collaborating in a sales
effort.) In our experience this very real but usually invisible barrier to cross-
selling and account penetration is common in organizations.
Developing a network perspective can
help the 21st-century organization retain
the best of its traditional organizational
structures, while simultaneously acknowl-
edging the heightened value of innovations,
collaborators, and intangible assets.
See “The 21st-century organization” on
mckinseyquarterly.com.
Mapping the value of employee collaboration 43
To boost sales the company sought both to replicate the major contributors’
behavior and to help several key salespeople understand how collaboration
could make them more successful. It quickly became clear that the success
of the high-performing collaborators resulted from more than just expertise
or affability. When the company compared them with its other salespeople,
it found that they were accommodating, more responsive to requests,
fexible, amenable to constructive criticism, enthusiastic team players, and
effective confict managers. The importance of these traits caused the
company to overhaul its incentive program and to launch an effort to build
collaborative skills throughout the sales network.
Enhancing career paths. When a global consulting frm used network tools
to analyze the sales efforts of a group of roughly 80 partners, it identifed
two crucial categories of people who weren’t recognized by its performance-
management processes, which emphasized individual revenue production.
By making joint sales calls, sharing experiences, and the like, 10 partners
supported collaborative efforts yielding 60 percent of this group’s revenue;
the top 5 accounted for 38 percent. A completely different subset of partners
made an enormous contribution to the execution of projects by helping
others to save time and generate high-quality work; this second group, for
r x n i v i 1 z
A map of wasted opportunities
Q3 2006
Network mapping
Exhibit 2 of 4
Glance: A network perspective can identify the hidden barriers to collaboration.
Souice: Þervoik RounuraLle ar rle Univeisiry oí Viiginia, NcKinsey analysis
Individuals were asked to identify people whose expertise they knew about but
with whom they saw no opportunity to collaborate. The resulting network map
illustrates those potential but missed opportunities.
Developing collaboration opportunities for . . .
• . . . application services, architecture services, and
testing during the sales cycle can create
opportunities to sell packaged or bundled services.
* . . . global sales and application services can
create opportunities to generate innovative ideas,
hypotheses, and solutions.
Application services
Architecture services
Global sales
Global sales management
Testing, solutions
Other
A
No tie = individual is not aware
of other’s expertise or is aware,
has already collaborated
Tie = individual is aware of
other’s expertise but sees no
opportunity to collaborate
Node = individual in network
Color = individual’s department
Tie = connection between 2 nodes
B
A
B
The McKinsey Quarterly 2006 Number 3 44
example, contributed expertise on the problems of clients, visited them,
and helped with analyses. The contributions of these partners, too, were
highly concentrated: the top 10 people were responsible for 48 percent
of the value generated through time savings, and the top 5 for 32 percent.
The consulting frm used this knowledge to end a long-simmering disagree-
ment about dual career paths for partners. There was no longer any
question about the need to recognize the contributions of partners whose
expertise or experience played a key role in winning many new clients
and of those whose work improved the effectiveness and effciency of the
frm’s efforts to serve them.
Boosting productivity
Most companies—even high-performing ones—can fnd opportunities
to boost their collaborative productivity. Sometimes, network analysis
shows them that they can generate savings by facilitating the transfer of
advice and information from colleagues. In other cases, a network
perspective isolates unseen collaborative ineffciencies resulting from poor
job design, an ineffective allocation of the right to make decisions,
and outdated role defnitions, process steps, or organizational designs.
The specifc issues and interventions vary considerably across industries.
But some general themes emerge. Often, companies that operate without a
network perspective allocate resources ineffciently, manage talent blindly,
and experience large disparities in the effectiveness of collaboration within
and across units. Scrutinizing the time savings that relationships generate
helps companies to isolate what’s working; to decide what, where, and how
to invest in additional connectivity; and to redefne roles and staffng
levels. Examples from three very different industries illustrate the range
of possibilities.
Validating the effectiveness of networks, sharing good ideas. In the
petrochemical business, avoiding downtime is critical given the magni-
tude of its investment in fxed assets. Solving problems quickly often
requires collaboration across disciplines such as drilling, geology, physics,
and production. So one leading petrochemical company formed more
than 20 networks (ranging in size from 50 to several hundred employees)
and focused on work areas where people could beneft from sharing
best practices.
Having taken the unusual step of engineering these networks, the
petrochemical company was particularly eager to measure their impact.
Network analysis showed that the effort, which previously had
been operating largely on faith, was generating substantial, shareable
Mapping the value of employee collaboration 45
productivity benefts. One 60-person network alone
contributed $5 million in savings. A typical story
involves engineers and an out-of-commission oil well.
Engineers used their network to identify an expert who
had no relationship with the well but did have critical
knowledge that helped them fx it in two days instead
of the expected four. Network analysis thus allowed
the company to validate the effcacy of its networks.
The company then began taking steps to pass lessons
among networks. A knowledge-sharing team
interviewed the leaders of networks to collect and
disseminate best practices. Training sessions allowed
the leaders of the most successful networks to share what they had
learned. The keys to success included forming networks carefully around
focused topic areas closely related to the way work was actually done,
giving network members the leadership and training for success (rather
than merely dumping collaborative tools on them), and continually
tracking and measuring success to encourage participation and inform
decision making about when (and when not) to fnance incremental
network improvements.
Improving the allocation of resources. A global fnancial-services organiza-
tion mapped and calculated the time its key employees saved by sharing
information and resources with their colleagues. This effort helped the
company to make better decisions about how much to invest in its collabora-
tive relationships, whether to focus on collaboration within or across
groups, and what role collaboration should play in its human-resources
(HR) strategy. The success of a pilot effort led the company to replicate it
widely, yielding savings that should ultimately dwarf the initial benefts.
Network analysis, for example, allowed executives to prioritize the
company’s investment in collaboration by helping them to model the
fnancial benefts of improving the network and to weigh the anticipated
returns against the costs. After recognizing that a set of key brokers
occupied central positions in the network, for instance, the company
realized that connecting all of these people with each other and with just
one person on the network’s fringe would yield $140,000 a year in
savings within business units and $865,000 across them. Facilitating these
interactions would be far less costly than buying the group another
unused collaborative tool or holding an off-site meeting.
In addition, network analysis showed the company how to focus its
collaborative efforts within and across groups, since aggregating results by
The McKinsey Quarterly 2006 Number 3 46
business units, roles, projects, and hierarchical levels showed executives
where to direct the relevant investments. One division’s global network
of technical project managers generated monthly savings of 3,383 hours
(which translated into roughly $215,000). When the fnancial institution
realized that about 70 percent of these savings resulted from collabora-
tion within divisions, it began focusing more heavily on collaboration
among them to reduce the number of redundant efforts and to promote
the exchange of expertise in project-management tools, methodologies,
and technologies.
Another beneft to the company was an improved ability to measure and
manage talent. Executives were surprised to learn how much relational
value was created by people they hadn’t recognized as central contributors
and how little by others they had regarded as more infuential. The
company responded by fnancially rewarding the key collaborators (many
of whom had previously been frustrated by the failure to recognize their
effectiveness), redefning roles and performance metrics to promote
collaboration, and in some cases elevating (or demoting) the role of the
network’s central (or peripheral) fgures in the company’s succession plans.
r x n i v i 1 j
Uncovering the value
Q3 2006
Network mapping
Exhibit 3 of 4
Glance: Employees’ interaction times can be converted into an average monthly cost of
collaboration per employee.
Application
architect
Labor cost of time spent in or preparing for interactions with others (example of major US utility)
Application
architect
Total monthly interaction
labor cost, $
Number of
employees
Internal External
Average cost of interaction time
per employee per month, $
Business unit
architect
Business unit
architect
Data
architect
Data
architect
Infrastructure
architect
Infrastructure
architect
Systems
analyst
Systems
analyst
Project
manager
Project
manager
Other
Other
17
5
6
16
2
5
7
2,126
3,750
3,600
1,406
1,125
1,680
1,714
1,121
2,460
950
656
0
210
471
715
1,110
2,800
375
150
1,050
193
1,597
1,560
2,225
5,588
975
3,180
1,843
476
30
225
56
0
1,680
364
626
210
975
1,013
1,500
1,470
1,971
1,103
660
650
1,069
1,275
3,450
2,057
7,764
9,780
11,425
10,163
5,025
12,720
8,613
132,000
48,900
68,550
162,600
10,050
63,600
60,300
2,126
2,460
2,800
5,588
0
1,470
2,057
5,638
7,320
8,625
4,575
5,025
11,250
6,556
$1,000–$1,999 >$2,000 <$1,000
Souice: Þervoik RounuraLle ar rle Univeisiry oí Viiginia, NcKinsey analysis
Mapping the value of employee collaboration 47
Managers were also encouraged to coach a few people who didn’t know
how to build networks.
Eliminating inefficiencies. The chief information offcer (CIO) of a major
utility calculated collaboration’s average monthly cost per employee—both
within roles (for example, collaboration among all data architects) and
between them (between, say, data architects and other members of the IT
staff)—throughout the IT organization (Exhibit 3). This analysis helped
the company to root out collaborative ineffciencies. By comparing the
colleagues who were generally considered effective communicators, for
instance, the CIO was able to identify outliers: a small number of employees
who actually were signifcantly less effective than the rest (Exhibit 4).
Consequently, the company focused personalized coaching efforts on
collaborative issues that were unique to each of the unexpected low
performers.
Network analysis also helped to clarify roles by showing that the utility’s
data architects and project managers spent more than half of their time
collaborating. Yet these demands had never before infuenced the hiring,
r x n i v i 1 j
Uncovering the value
Q3 2006
Network mapping
Exhibit 3 of 4
Glance: Employees’ interaction times can be converted into an average monthly cost of
collaboration per employee.
Application
architect
Labor cost of time spent in or preparing for interactions with others (example of major US utility)
Application
architect
Total monthly interaction
labor cost, $
Number of
employees
Internal External
Average cost of interaction time
per employee per month, $
Business unit
architect
Business unit
architect
Data
architect
Data
architect
Infrastructure
architect
Infrastructure
architect
Systems
analyst
Systems
analyst
Project
manager
Project
manager
Other
Other
17
5
6
16
2
5
7
2,126
3,750
3,600
1,406
1,125
1,680
1,714
1,121
2,460
950
656
0
210
471
715
1,110
2,800
375
150
1,050
193
1,597
1,560
2,225
5,588
975
3,180
1,843
476
30
225
56
0
1,680
364
626
210
975
1,013
1,500
1,470
1,971
1,103
660
650
1,069
1,275
3,450
2,057
7,764
9,780
11,425
10,163
5,025
12,720
8,613
132,000
48,900
68,550
162,600
10,050
63,600
60,300
2,126
2,460
2,800
5,588
0
1,470
2,057
5,638
7,320
8,625
4,575
5,025
11,250
6,556
$1,000–$1,999 >$2,000 <$1,000
Souice: Þervoik RounuraLle ar rle Univeisiry oí Viiginia, NcKinsey analysis
The McKinsey Quarterly 2006 Number 3 48
staffng, or performance evaluations of such workers. As a result, they
lacked collaborative skills, their job functions and role descriptions were
incomplete, and the internal cost allocations used to establish transfer
prices for IT projects underestimated the total cost of certain programs and
thereby distorted the company’s resource allocation. Addressing these
issues not only improved the execution of projects and the company’s ability
to price them but also made internal customers more satisfed.
Finally, the company used network analysis to set appropriate staff levels.
It knew that it could avoid certain problems by involving its infrastructure
architects (who design and maintain major applications supporting vital
business functions) in key decisions at the right time. But often this didn’t
happen. Traditional budgeting and cost allocation processes might have
suggested hiring more infrastructure architects. Network analysis, however,
showed that they interacted less than most other employees of the IT
organization. The frst step for the CIO, therefore, was breaking down the
barriers that inhibited collaboration.
r x n i v i 1 ¸
Who needs help?
Q3 2006
Network mapping
Exhibit 4 of 4
Glance: Network analysis can identify those employees who would benefit from personalized
coaching.
Average number of
times individual was
cited as a source
of useful information
Average interaction time
Application architecture Infrastructure architecture
+
– +
Analysis of interaction times and
usefulness for a small number
of employees can lead to improved
organizational performance.
Souice: Þervoik RounuraLle ar rle Univeisiry oí Viiginia, NcKinsey analysis
Mapping the value of employee collaboration 49
Collaboration is an increasingly vital feature of business life. But when
companies just promote collaboration indiscriminately, they create
bottlenecks and diminish their organizational effectiveness. A network
perspective gives executives the information they need to foster
collaboration at the points where it delivers an economic return.
Q


Rob Cross, an adviser to McKinsey, is an associate professor and director of the
Network Roundtable at the University of Virginia’s McIntire School of Commerce, where
Roger Martin is an assistant professor;
Leigh Weiss is an associate principal in McKinsey’s Washington, DC, offce.
Copyright © 2006 McKinsey & Company.
All rights reserved.
Harnessing the power of informal employee networks 51
Harnessing the power
of informal employee
networks
Formalizing a company’s ad hoc peer groups can spur collaboration
and unlock value.
Lowell L. Bryan, Eric Matson,
and Leigh M. Weiss
In any professional setting, networks fourish spontaneously: human
nature, including mutual self-interest, leads people to share ideas and
work together even when no one requires them to do so. As they connect
around shared interests and knowledge, they may build networks that
can range in size from fewer than a dozen colleagues and acquaintances to
hundreds. Research scientists working in related felds, for example,
or investment bankers serving clients in the same industry frequently create
informal—and often socially based—networks to collaborate.
Most large corporations have dozens if not hundreds of informal networks,
which go by the name of peer groups, communities of practice, or
functional councils—or have no title at all. These networks organize and
reorganize themselves and extend their reach via cell phones, Black-
berries, community Web sites, and other accessories of the digital age. As
networks widen and deepen, they can mobilize talent and knowledge
across the enterprise. They also help to explain why some intangible-rich
companies, such as ExxonMobil and GE, have increased in scale and
scope and boast superior performance.
1
1
Lowell L. Bryan and Michele Zanini, “Strategy in an era of global giants,” The McKinsey Quarterly, 2005
Number 4, pp. 46–59.
J
e
r
r
y

S
t
e
n
b
a
c
k
The McKinsey Quarterly 2007 Number 4 52
As we studied these social and informal networks, we made a surprising
discovery: how much information and knowledge fows through them and
how little through offcial hierarchical and matrix structures. As we used
surveys and e-mail analysis to map the way employees actually exchange
information and knowledge, we concluded that the formal structures of
companies, as manifested in their organizational charts, don’t explain how
most of their real day-to-day work gets done.
So it’s unfortunate, at a time when the ability to create value increasingly
depends on the ideas and intangibles of talented workers, that corporate
leaders don’t do far more to harness the power of informal networks.
Valuable as they are, these ad hoc communities clearly have shortcomings:
they can increase complexity and confusion, and since they typically fy
under management’s radar, they elude control.
But companies can design and manage new formal structures that boost
the value of networks and promote effective horizontal networking across
the vertical silos of so many enterprises today. By building network
infrastructures, assigning “leaders” to focus discussion, and combining
hierarchy and collaboration to bring together natural professional
communities, formalized networks serve as an organizing structure for
collaborative professional work. They can replace cumbersome and
outdated matrix structures, facilitate the creation and sharing of proprietary
information and knowledge, and help build more and better personal
relationships among the members of a community. Most important, they
can enable leaders to apply the energy of diverse groups of professionals
and managers to realize collective aspirations.
The long and short of informal networks
Personal social networks, both within and outside of companies, increase
the value of collaboration by reducing the search and coordination costs of
connecting parties who have related knowledge and interests. They don’t
necessarily ft into the organizational chart. Consider the case of an energy
company staffer we’ll call Cole (Exhibit 1). Although he sits relatively far
down in the formal company structure, he acts as the hub in an informal
network because he has knowledge that others fnd valuable. Without
him, the production group would be cut off from the rest of the organiza-
tion. His boss Jones, the unit’s senior vice president, is connected
in the informal network to only two people, both in Exploration. This is
increasingly typical in today’s large, sprawling corporations. Informal
networks, slipping through the back channels, cross the lines of geography,
products, customer groups, and functions—where the action is—
and even through the thick silo walls of vertically oriented organizations.
Harnessing the power of informal employee networks 53
But though informal networks help many of the largest companies
capture wealth, they also cause severe headaches. As tens of thousands of
individuals search for knowledge and productive personal relationships
in social networks, they generate much of the overload of e-mails, voice
mails, and meetings that make today’s large companies more complex
and ineffcient. At one large company, we conducted a network analysis of
more than 1,000 people across a number of business units around the
world to gauge the frequency and quality of the interactions that generated
decisions about business planning and other processes. Nearly half
of the interactions were not central to making decisions. This analysis
suggested that redesigning the processes to eliminate or reduce the
noncore interactions could result in savings of tens of millions of dollars
and shorten the time to make the decisions.
Part of the problem is that informal networks, as ad hoc structures,
essentially rely on serendipity, so their effectiveness varies considerably.
Formal structure Informal structure
Senior vice president
(SVP) Jones
Exploration
Williams
Reservoir
Smith
Drilling
Taylor
Production
Stock
Network mapping reveals informal structure—
who interacts with whom to get work done.
Smith
Cohen
Hughes
Ramirez
Bell
Cole
Hussain
Kelly
G&G
1
Cross
Andrews
Petro-
physical
Production
Moore
Sen
Miller
Cohen Shapiro
O’Brien
Hussain
Taylor
Stock
Shapiro Paine
Kelly
Hughes
Ramirez
Miller
Williams
Smith
Andrews
Cohen
Cross
Moore Bell Sen
Jones Cole
• Jones, SVP, is connected to only 2 people,
both in Exploration
• Cole is at center of information exchange
network despite being relatively low in
organization chart
• Without Cole, production group would be
cut off from rest of organization
i
Geological anu geoµlysical.
Souice: RoL Cioss anu Anuiev Paikei, The Hidden Power of Social Networks, Bosron: HBS Piess, zoo¸
r x n i v i 1 i
A revealing map
The McKinsey Quarterly 2007 Number 4 54
In large companies a number of informal networks may form on related
topics but never integrate. People with valuable knowledge or skills may
not join the most appropriate network, belong to other informal networks,
or fail to discover that a network exists. What’s more, companies typi-
cally underinvest in the capabilities needed to make networks function effec-
tively and effciently. An informal network often has crucial members,
such as Cole, who serve as hubs to connect participants, but such members
can hobble or even undermine the network if they become overloaded,
act as gatekeepers, horde knowledge to gain power, or leave the company
(Exhibit 2).
The greatest limitation of these ad hoc arrangements is that they can’t be
managed. They may not be visible to management, and even when they
are it’s hard for corporations to take full advantage of them. Unintended
barriers, corporate politics, and simple neglect can keep natural net-
works from fourishing. At worst, informal networks can make dysfunc-
tional organizations even more so by adding complexity, muddling
roles, or increasing the intensity of corporate politics.
Formal networks
The specifc objective of designing and establishing formal networks is to
increase the value and lower the costs of collaboration among profes-
sionals. Since formal networks stimulate interactions that the organization
sponsors and encourages, they can be managed.
A leading petrochemical company, for example, recently designed more
than 20 formal networks, ranging in size from 50 to several hundred people,
to focus on specifc work areas so that employees could share best prac-
tices. This was critical, because the networks could minimize downtime in
these areas. In one case the company measured the impact of networks
Informal network with linchpin
Linchpin
Informal network if linchpin is removed
Removing the
linchpin leaves
the network
fragmented

The missing link
Harnessing the power of informal employee networks 55
on engineers at an oil well, who used them to fnd experts with the
knowledge needed to get the well back into production in two days rather
than the anticipated four.
These networks succeeded because the company formed them around
focused topics closely related to the way work was carried out at the
wells. Management also appointed the networks’ leaders, gave the mem-
bers training, carefully identifed the members of each network across
the geographically dispersed company, made technology investments, and
sponsored a knowledge-sharing team that collected and disseminated
best practices.
2
Matrix decoded
Because formal networks cross line structures, they can easily be mistaken
for matrix organizational entities. But the differences are signifcant and
start with the organizing principles that underlie each (Exhibit 3). A matrix
organizes work through authority and is therefore principally based on
management hierarchy. A formal network organizes work through mutual
self-interest and is therefore principally based on collaboration.
In classic matrix organizations, managers and professionals have two
or more bosses who have authority over their work; the chief fnancial
offcer of a business unit, for example, might report both to its line
manager and to the corporate fnance chief. These matrixes represent dif-
ferent axes of management, such as products, geography, customers,
or functions. Hierarchical direction comes from two different sources,
and the person in the middle of the matrix must resolve any conficts.
In hierarchically organized companies, matrix management became popu-
lar because no matter how well organized their line structures may
have been by functions, geography, customers, or products, they felt they
needed secondary axes of management to stretch horizontally across
the enterprise and thus make it possible to integrate other work activities.
Matrix management worked reasonably well from its advent in the 1960s
until the late 1980s, particularly because it enabled limited collabora-
tion to take place within companies as they became increasingly aware
that hierarchical managers sometimes had to coordinate their activi-
ties. Matrix structures made sense because they were used sparingly and
therefore didn’t greatly confuse the hierarchical vertical line structures
responsible for much of the success of large 20th-century companies.
2
Robert L. Cross, Roger D. Martin, and Leigh M. Weiss, “Mapping the value of employee collaboration,”
The McKinsey Quarterly, 2006 Number 3, pp. 28–41.
The McKinsey Quarterly 2007 Number 4 56
But when globalization took hold, companies were forced to adapt to an
increasingly fuid and uncertain competitive environment, so more
work from different perspectives had to be integrated. As the number
of professionals needing to direct much of their own work has risen,
matrixed roles have proliferated. This increased the need for more interac-
tions, and decision making now swamps the time available for matrix
managers to coordinate the work personally. Furthermore, the amount of
knowledge and information that must be absorbed and exchanged often
exceeds the personal capacities of any individual, no matter how talented,
to deal with them in a matrix structure.
Professionals who want to work horizontally across an organization cur-
rently fnd themselves forced to search though poorly connected
organizational silos for the knowledge and collaborators they need. In
many companies these matrix and other hybrid organizations have
become dysfunctional. The symptoms include endless meetings, phone calls,
and e-mail exchanges, as well as confused accountability for results.
A new model
Companies need to build infrastructures to create and support formal
networks. Such well-designed and well-supported formal networks remove
bottlenecks and take much of the effort out of networking. Rather than
forcing employees to go up and down hierarchical chains of command, for-
mal networks create pathways for the natural exchange of information
and knowledge. Individual members of networks don’t have to fnd one
another through serendipity.
Q <_> <_>
<Article slug>
Exhibit <_> of <_>
Glance:

Hierarchical vs. collaborative
Matrix structure Formal network
Organizing principle
Mode of influence
Number of bosses
Implications
Organizes work through authority
Management hierarchy
Two or more representing different
axes of management, eg, product,
geography, customer, function
• Proliferation of matrixed roles
and complex hierarchical structure
• Excessive interactions
• Decision-making bottlenecks
• Difficulty finding knowledge
• Conflicts arising from different bosses
Organizes work through mutual
self-interest
Collaboration and leadership
One manager for each person;
one formal network leader for each
networked community
• Simple hierarchical structure
• Streamlined interactions
• Streamlined decision making
• Easier and faster to find knowledge using
network’s resources
• Single boss reduces number of conflicts
that need to be resolved
Harnessing the power of informal employee networks 57
Consider the US retail unit of a fnancial institution we’ll call Global Bank,
which was organized as a matrix. Its retail-marketing managers, forced
to report to a regional as well as a functional manager, often didn’t know
whose authority to recognize and had little opportunity to share
best practices with other marketing professionals across the organization.
In the new model (Exhibit 4) regional marketing managers still report to
the branch network’s regional managers but no longer have a second
boss in marketing. Instead, a branch-based formal network leader for mar-
keting facilitates their interactions with other marketing professionals.
The leader can’t give them orders but can encourage them to work for the
network’s beneft (for example, by asking them to help develop new

Formalizing a network
CEO
Executive vice president
(EVP), Global Bank
Example of global financial institution’s US retail operation
East Coast
retail division
South
division
Midwest
division
Credit card
division
Other
divisions
US branch-based
marketing leader
Marketing
Marketing
Marketing
network
Clear objectives
• Marketing network is
expected to show
measurable results
in key metrics
(eg, brand awareness)
Network owners
• Network is ‘owned’ by
US retail-marketing
executive (senior manager)
who reports to EVP
of Global Bank (top executive)
Flexible membership
• 30 core members, including
all those reporting to frontline
branch executives
• Extended members, including
people working on related
initiatives (eg, managers from
credit card division
distributing cards through
branches)
Network coordinators
• US branch-based leader
facilitates interactions
Standard roles, parallel
structures
• Marketing managers
at every branch have same
role description
US retail-marketing
leader
4 other frontline
managers
Bronx, Queens,
Staten Island
Manhattan,
Brooklyn
Northern NJ,
Long Island
Marketing
Marketing
The McKinsey Quarterly 2007 Number 4 58
promotional materials or to fnd better ways of using local-advertising
budgets). The marketing leader’s boss, the US retail-marketing executive, is
a senior manager who owns the formal network and mobilizes market-
ing talent for special projects, identifes candidates for marketing positions,
oversees the maintenance of the domain’s knowledge (for instance, branch
signage or promotional materials), and stimulates its creation. The company,
which expects the network to show measurable results in key metrics
(such as brand awareness), evaluates the owner by taking into account quali-
tative assessments of how well this formal network operates as compared
with others, as well as the expectations of corporate leaders.
Formalizing a network
To formalize a network, the company must defne who will lead it—that
is, the network owner—and make that leader responsible for investing
in the network to build its collective capabilities, such as knowledge that
is valuable for all members. The company can facilitate the development
of a formal network by providing incentives for participating in it (such as
community building off-sites) and for contributing to it (such as recogni-
tion for people who contribute distinctive knowledge).
Network owners facilitate interactions between members, stimulate the
creation of knowledge, maintain the network’s knowledge domain, and
help members do their jobs more effectively and effciently. For a formal
network to work effectively, its territory must be defned—informal
networks sometimes make overlapping claims on the same activities. Fur-
thermore, the network must have standards and protocols that describe
how it should work.
Another difference between a formal network and a matrix is that the
network owner isn’t a boss but rather a “servant leader.” The owner of a
network doesn’t oversee its work or personally manage or evaluate
the performance of individual members (except for direct reports) but may
provide input to the evaluation process.
The responsibilities of the formal leader of a network are primarily
limited to its activities, such as organizing the infrastructure supporting
it, developing an agenda for maintaining its knowledge domain, build-
ing a training program, holding conferences, and qualifying members as
professionally competent.
Despite this limited hierarchical authority, a formal network’s leader should
be held accountable (together with line management) for the network’s
Harnessing the power of informal employee networks 59
performance. After all, the leader has great ability to help its members
improve their performance and in this way can shape the organization. Much
of the leader’s impact comes from controlling the investments and activ-
ities that make the members individually—and the network collectively—
more effective, and much from the ability to inspire and persuade.
In professional frms, which have long used formal networks called practices,
it is always possible to tell the difference between talented and average
leaders. While the responsibilities might be the same, talented ones create
far more vibrant, exciting practices than their average counterparts do.
It is therefore entirely appropriate to hold the leader of a formal network
accountable for its performance, even if the leader has no direct author-
ity over individual members.
Connecting members to the network
In the model we propose, companies should design formal networks to
extend the reach of professional work throughout the organization but not
to interfere with its hierarchical decision-making processes. The idea is
to achieve this extended reach by adding value for the networks’ members,
not by exercising authority through hierarchical leaders.
To undertake the appropriate roles,
a formal network’s leader should
have a discrete budget to fnance net-
work investments, which give
the leader the muscle to offer the
members added value. These
investments might include infrastruc-
ture, both human and technolog-
ical, to support network interactions;
codifed knowledge in forms such
as documents, internal blogs,
and “networkpedias”; training for members; and activities such as confer-
ences to build a social community. Companies can evaluate the leader’s
performance by using some quantitative measures, such as the level of partici-
pation in conferences, e-mail volumes, standard measures from network
analysis (for example, the number of steps it takes for any person in the net-
work to reach anyone else), density, knowledge documents produced,
and downloads. Management can also track and test the effectiveness of a
network by assessing the satisfaction of its members, the effectiveness
of responses to inquiries, and the ability to fnd appropriate partners for
dialogue quickly.
The McKinsey Quarterly 2007 Number 4 60
But the real measure of the network’s success would be qualitative assess-
ments, made by members and senior leaders, of its effectiveness in
realizing its mission. These assessments might come in the form of stories
or case studies illustrating improvements in professional productivity.
Providing structure to professional work
Just as formal hierarchical structures defne management roles, formal
network structures defne collaborative professional ones. In this way
such networks can enable large companies to overcome the problems of
very large numbers by creating small, focused communities of interest
integrated within larger, more wide-ranging communities—for instance,
subcommunities focused on different aspects of fnancial services, such
as wholesale and retail banking.
The number of networks employees participated in would be up to them,
unless they were core members, for whom participation would be a job
requirement. In other words each member would build a personal social
network within the formal networks, depending on that member’s roles
and professional interests. The limits of network participation are largely
a function of time and interest; members would join networks that had
more value to them than the opportunity costs of their time and would
leave when those networks no longer had that much value.
By participating in more
than one network at
a time, talented workers
would gain the ability
to integrate knowledge
and access to talent
across a number of com-
munities. A person
in the retail-banking community could also be a member of a branding
community, for example, and members could bring knowledge gained
there into other communities.
The number of formal networks a company could create is limited
only by how much management chooses to invest in them. Their number
and size could vary with how well each of them serves its members—
effective networks would grow in membership and interactions; ineffective
ones would lose both. In this way formal networks regulate themselves.
Rapid growth proves the value of a network, its leader, and the money
invested in it.
Related articles on mckinseyquarterly.com
The 21st-century organization
The role of networks in organizational change
Just-in-time strategy for a turbulent world
Better strategy through organizational design
Harnessing the power of informal employee networks 61
Today’s mega-institutions have room for thousands of formal networks.
A company with 100,000 professional and managerial employees,
for example, could have 2,000 networks with 100 people apiece if each
professional and manager was a member of just 2 networks. Broad net-
works (in felds such as fnance or IT) might have thousands of members;
specialized ones (say, a Turkish interest group) might have only a
few dozen. Formal networked communities could form around not just
customer groups, products, geography, and functional lines but also
in conjunction with integrative crosscutting themes, such as risk manage-
ment and global forces (nanotechnology and changing demographics,
for instance).
Formal network structures can mobilize employees to generate value by
propagating knowledge and its creators all over the enterprise. Rather
than pushing knowledge and talent through a hierarchical matrix, formal
networks let employees pull these necessities toward them.
Q
Lowell Bryan is a director in McKinsey’s New York offce;
Eric Matson is a consultant and Leigh Weiss is an associate principal in the Boston offce.
Copyright © 2007 McKinsey & Company. All rights reserved.
The McKinsey Quarterly 2005 Number 4 62
G
é
r
a
r
d

D
u
B
o
i
s
The next revolution in interactions 63
Like vinyl records and Volkswagen Beetles, sustainable competitive
advantages are back in style—or will be as companies turn their attention
to making their most talented, highly paid workers more productive.
For the past 30 years, companies have boosted their labor productivity by
reengineering, automating, or outsourcing production and clerical jobs.
But any advantage in costs or distinctiveness that companies gained in this
way was usually short lived, for their rivals adopted similar technologies
and process improvements and thus quickly matched the leaders.
But advantages that companies gain by raising the productivity of their
most valuable workers may well be more enduring, for their rivals will fnd
these improvements much harder to copy. This kind of work is under-
taken by, for example, managers, salespeople, and customer service reps,
whose tasks are anything but routine. Such employees interact with other
employees, customers, and suppliers and make complex decisions based on
knowledge, judgment, experience, and instinct.
New McKinsey research reveals that these high-value decision makers
are growing in number and importance throughout many companies. As
businesses come to have more problem solvers and fewer doers in their
ranks, the way they organize for business changes. So does the economics
The next revolution in
interactions
Successful efforts to exploit the growing importance of
complex interactions could well generate durable competitive
advantages.
Bradford C. Johnson,
James M. Manyika, and Lareina A. Yee
The McKinsey Quarterly 2005 Number 4 64
of labor: workers who undertake
complex, interactive jobs typically
command higher salaries, and their
actions have a disproportionate
impact on the ability of companies
to woo customers, to compete, and
to earn profts. Thus, the potential
gains to be realized by making these
employees more effective at what
they do and by helping them to do
it more cost effectively are huge—
as is the downside of ignoring
this trend.
But to improve these employees’
labor performance, executives
must put aside much of what they
know about reengineering—and
about managing technology, organizations, and talent to boost productivity.
Technology can replace a checkout clerk at a supermarket but not a
marketing manager. Machines can log deposits and dispense cash, but they
can’t choose an advertising campaign. Process cookbooks can show how
to operate a modern warehouse but not what happens when managers band
together to solve a crisis.
Machines can help managers make more decisions more effectively and
quickly. The use of technology to complement and enhance what talented
decision makers do rather than to replace them calls for a very different
kind of thinking about the organizational structures that best facilitate
their work, the mix of skills companies need, hiring and developing
talent, and the way technology supports high-value labor. Technology and
organizational strategies are inextricably conjoined in this new world of
performance improvement.
1

Raising the labor performance of professionals won’t be easy, and it
is uncertain whether any of the innovations and experiments that some
pioneering companies are now undertaking will prove to be winning
formulas. As in the early days of the Internet revolution, the direction is
clear but the path isn’t. That’s the bad news—or, rather, the challenge
(and opportunity) for innovators.
Article at a glance
As more 21st-century companies come to specialize
in core activities and outsource the rest, they have
greater need for workers who can interact with other
companies, their customers, and their suppliers.
Thus the traditional organization, where a
few top managers coordinate the pyramid below
them, is being upended.
Raising the productivity of employees whose jobs
can’t be automated is the next great performance
challenge—and the stakes are high.
Companies that get it right will build complex talent-
based competitive advantages that competitors won’t
be able to duplicate easily—if at all.
1
Lowell L. Bryan and Claudia Joyce, “The 21st-century organization,” The McKinsey Quarterly, 2005
Number 3, pp. 24–33; and Lowell L. Bryan, “Getting bigger,” The McKinsey Quarterly, 2005 Number 3,
pp. 4–5.
The next revolution in interactions 65
The good news concerns competitive advantage. As companies fgure out
how to raise the performance of their most valuable employees in a range
of business activities, they will build distinctive capabilities based on a mix
of talent and technology. Reducing these capabilities to a checklist of
procedures and IT systems (which rivals would be able to copy) isn’t going
to be easy. Best practice thus won’t become everyday practice quite as quickly
as it has in recent years. Building sustainable advantages will again be
possible—and, of course, worthwhile.
The interactions revolution
Today’s most valuable workers undertake business
activities that economists call “interactions”:
in the broadest sense, the searching, coordinating,
and monitoring required to exchange goods
or services. Recent studies—including landmark
research McKinsey conducted in 1997
2
—show
that specialization, globalization, and technology
are making interactions far more pervasive in
developed economies. As Adam Smith predicted,
specialization tends to atomize work and to
increase the need to interact. Outsourcing, like
the boom in global operations and marketing, has dramatically increased
the need to interact with vendors and partners. And communications
technologies such as e-mail and instant messaging have made interaction
easier and far less expensive.
The growth of interactions represents a broad shift in the nature of
economic activity. At the turn of the last century, most nonagricultural
labor in business involved extracting raw materials or converting them
into fnished goods. We call these activities transformational because they
involve more than just jobs in production.
3
By the turn of the 21st century,
however, only 15 percent of US employees undertook transformational work
such as mining coal, running heavy machinery, or operating production
lines—in part because in a globalizing economy many such jobs are shifting
from developed to developing nations. The rest of the workforce now
consists of people who largely or wholly spend their time interacting.
2
Patrick Butler, Ted W. Hall, Alistair M. Hanna, Lenny Mendonca, Byron Auguste, James Manyika, and
Anupam Sahay, “A revolution in interaction,” The McKinsey Quarterly, 1997 Number 1, pp. 4–23.
3
Douglass C. North, “Institutions, Transaction Costs, and Productivity in the Long Run,” Washington
University at St. Louis economics working paper, economic history series, number 9309004, September 1993;
Douglass C. North, “Transaction Costs Through Time,” Washington University at St. Louis economics working
paper, economic history series, number 9411006, November 1994; and Douglass C. North, “Institutions and
Productivity in History,” Washington University at St. Louis economics working paper, economic history series,
number 9411003, November 1994.
The McKinsey Quarterly 2005 Number 4 66
Within the realm of interactions, another shift is in full swing as well, and
it has dramatic implications for the way companies organize and compete.
Eight years after McKinsey’s 1997 study, the frm’s new research on job
trends in a number of sectors fnds that companies are hiring more workers
for complex than for less complex interactions. Recording a shipment
of parts to a warehouse, for example, is a routine interaction; managing a
supply chain is a complex one.
Complex interactions typically require people to deal with ambiguity—
there are no rule books to follow—and to exercise high levels of judgment.
These men and women (such as managers, salespeople, nurses, lawyers,
judges, and mediators) must often draw on deep experience, which
economists call “tacit knowledge.” For the sake of clarity, we will therefore
refer to the more complex interactions as tacit and to the more routine
ones as transactional. Transactional interactions include not just clerical
and accounting work, which companies have long been automating or
eliminating, but also most of what IT specialists, auditors, biochemists, and
many others do (see sidebar, “About the research”).
Most jobs mix both kinds of activities—when managers fll out their
expense reports, that’s a transaction; leading workshops on corporate
About the research
The next wave of performance improvements—
to raise the effectiveness of tacit workers—will
be far more difficult than the improvement efforts
of the past. But companies that can innovate to
make their complex, higher-value business activities
deliver what their customers care about most will
probably gain significant (and not easily duplicated)
advantages in distinctiveness, quality, and cost.
We looked at the range of business activities
involved in more than 800 occupations in the
United States. Building on McKinsey’s 1997 study,
we placed every job in one of three categories:
transformational (extracting raw materials or
converting them into finished goods), transactional
(interactions that unfold in a generally rule-based
manner and can thus be scripted or automated),
and tacit (more complex interactions requiring
a higher level of judgment, involving ambiguity,
and drawing on tacit, or experiential, knowledge).
While any kind of work clearly involves activities


in all three of our categories, we placed each
job by determining its predominant activity. This
occupational segmentation allowed us to develop
a macroeconomic view of employment and wage
shifts and to isolate trends in tacit interactions.
We cross-checked the results with the 1997 activity-
level analysis and with other economists’ findings
on interactions.
Then we linked the occupational analysis to the
US government’s industry classifications and
quantified the mix of tacit, transactional, and trans-
formational activities within and across industries.
In addition, we used data from the International
Labour Organization, the World Bank, and other
sources to analyze these trends on a global basis.
Finally, interviews with economists and with
functional and industry experts throughout McKinsey
helped us to identify and understand the key
enablers of tacit and transactional interactions in
today’s companies.
The next revolution in interactions 67
strategy with their direct reports is tacit work. But what counts in a job are
its predominant and necessary activities, which determine its value added
and compensation.
During the past six years, the number of US jobs that include tacit
interactions as an essential component has been growing two and a half
times faster than the number of transactional jobs and three times faster
than employment in the entire national economy. To put it another way,
70 percent of all US jobs created since 1998—4.5 million, or roughly the
combined US workforce of the 56 largest public companies by market
r Ā n i v i 1 ŋ
"OFXLJOEPGXPSLFS
ü4 2005
Ioteractioos
Exhibit 1 oI 4
Clorce. Srous|ut uí ¸|uwt| |] ju| t]ue
ŋ
Caregoiizeu Ly µieuominanr joL acriviry: racir = comµlex inreiacrions, riansacrional = iourine inreiacrions, riansíoimarional =
exriacrion oi conveision oí iav mareiials, ŒŊŊ occuµarions sruuieu.
Ō
Lxcluues µair-rime anu selí-emµloyeu voikeis.
ō
Comµounu annual giovrl iare.
Ŏ
Quairiles Laseu on numLei oí emµloyees virlin given caregoiy.
Souice: US Buieau oí LaLoi Srarisrics, NcKinsey analysis
IotaI ü8 eæpIoyæeot,
rua|e| uí eaulu]ees ir
ailliurs,
Z
°
uS eaulu]aert orJ wo¸es |] ju| t]ue
1

Tocit
T|orsoctiurol
T|orsíu|aotiurol
J9
1Z5 1JJ° = J.5
+.5
1.9
-Z.9
1998
+J
18
1Z8
ZJJ+
+1
++
15
J.5
1.5
J.c
-Z.J
6Aß8,
J

1998-2004,
°
6Aß8,
J

1998-2004,
°
6haoge,
1998-2004,
ailliurs uí ju|s
Average wage
Tocit
AoouaI wage per eæpIoyee
io üoited 8tates (ooæioaI},
S t|uusorJs
ßistributioo oI average wage by guartiIe,
+
S t|uusorJs
J./ /.J
+.J 9.9
Z.9 +.8
J.1 +.5
Z9./
J/.J
+/.+
JJ.c
Z/.J
1998 ZJJ+
6haoge,
1998-2004,
S t|uusorJs
JJ°
/J°
Tocit ju|s
T|orsoctiurol
ju|s
New jobs io üoited 8tates, 1998-2004
1JJ° = c.+ ailliur
T|orsoctiurol
T|orsíu|aotiurol
J/.5
Z5.8
ZZ.5
|i¸|e| wo¸e
C|uuu aeor
|uwe| wo¸e
J
1J
ZJ
JJ
+J
5J
cJ
/J
8J
9J
Co|uerte|
T|orsíu|aotiurol T|orsoctiurol Tocit
/.e|o¸e wo¸e
wit|ir (uo|tile íu|
soaule uccuuotiur
/ccuurtort
Cere|ol
aoro¸e|
Re¸iste|eJ
ru|se
Custuae|
se|.ice
|eu|esertoti.e
Retoil soles
ue|sur Cos|ie|
ûííce cle||
T|uc| J|i.e|
Curst|uctiur
wu||e|
Teoa ossea|le|
|ostíuuJ cuu|
The McKinsey Quarterly 2005 Number 4 68
capitalization—require judgment and experience. These jobs now make up
41 percent of the labor market in the United States (Exhibit 1). Indeed,
most developed nations are experiencing this trend.
The balance is tipping toward complexity, in part because companies
have been eliminating the least complex jobs by streamlining processes,
outsourcing, and automating routine tasks. From 1998 to 2004, for
example, insurance carriers, fund-management companies, and securities
frms cut the number of trans-
actional jobs on their books
by 10 percent, 6.5 percent, and
2.7 percent a year, respectively.
Likewise, a more automated check-
in process at airports makes for
smaller airline check-in staffs, automated replenishment systems reduce the
need for supply chain bookkeepers, and outsourcing helps companies shed
IT help desk workers. Manufacturers too have eliminated transactional jobs.
Meanwhile, the number of jobs involving more complex interactions
among skilled and educated workers who make decisions is growing at
a phenomenal rate. Salaries refect the value that companies place on
these jobs, which pay 55 and 75 percent more, respectively, than those of
employees who undertake routine transactions and transformations.

Demand for tacit workers varies among sectors, of course. The jobs of
most employees in air transportation, retailing, utilities, and recreation
are transactional. Tacit jobs dominate felds such as health care and many
fnancial-services and software segments (Exhibit 2). But all sectors employ
tacit workers, and demand for them is growing; most companies, for
example, have an acute need for savvy frontline managers.
A new path to better performance
The demand for tacit employees and the high cost of employing them are
a clear call to arms. Companies need to make this part of the workforce
more productive, just as they have already raised the productivity of
transactional and manufacturing labor. Unproductive tacit employees will
be an increasingly costly disadvantage.
The point isn’t how many tacit interactions occur in a company—what’s
important is that they ought to add value. This shift toward tacit
interactions upends everything we know about organizations. Since the
days of Alfred Sloan, corporations have resembled pyramids, with a limited
number of tacit employees (managers) on top coordinating a broad span
The number of jobs that involve
relatively complex interactions
is growing at a phenomenal rate
The next revolution in interactions 69
of workers engaged in production and transactional labor. Hierarchical
structures and strict performance metrics that tabulate inputs and outputs
therefore lie at the heart of most organizations today.
But the rise of the tacit workforce and the decline of the transformational
and transactional ones demand new thinking about the organizational
structures that could help companies make the best use of this shifting
blend of talent. There is no road map to show them how to do so. Over
time, innovations and experiments to raise the productivity of tacit
employees (for instance, by helping them collaborate more effectively inside
and outside their companies) and innovations involving loosely coupled
teams will suggest new organizational structures.
r Ā n i v i 1 Ō
5IFKPCNJY
ü4
Ioteractioos
Exhibit 2 oI 4
Clorce. /ll sectu|s eaulu] tocit wu||e|s
ßistributioo oI job type
1
by iodustry,
2004, ° uí wu||íu|ce
6haoge io distributioo oI tacit
workers, 1998-2004,
ue|certo¸e uuirts
Tocit T|orsíu|aotiurol T|orsoctiurol
Tertiles
5+ JZ 1+
+.J
Mirir¸ J9 +/ 1+ -J.5
Retoil t|oJe
J 5 5J +5
|oue| aoruíoctu|ir¸
1.+
+c J8 1c
Curst|uctiur J.8 cZ Z1 1/
/ccuaauJotiurs ,e¸, |utel,
autel, |eJorJ||eo|íostì
Z.+ 9 /J Z1
Elect|icol e(uiuaert J.Z
5Z J1 1/
Reol estote -1.+
J c1 Jc
|usuitols
Z
Z.c + Zc c9
Cuauute| aoruíoctu|ir¸
/.J Zc +8 Zc
/i| t|orsuu|totiur ZJ.1 cc J+
Secu|ities 8.1 +1 59
ŋ
Caregoiizeu Ly µieuominanr joL acriviry: racir = comµlex inreiacrions, riansacrional = iourine inreiacrions, riansíoimarional =
exriacrion oi conveision oí iav mareiials, ŒŊŊ occuµarions sruuieu.
Ō
ÞumLeis uo nor sum ro ŋŊŊ°, Lecause oí iounuing.
Souice: US Buieau oí LaLoi Srarisrics, NcKinsey analysis
The McKinsey Quarterly 2005 Number 4 70
The two critical changes that executives must take into account as they
explore how to make tacit employees more productive are already clear,
however. First, the way companies deploy technology to improve the
performance of the tacit workforce is very different from the way they have
used it to streamline transactions or improve manufacturing. Machines
can’t recognize uncodifed patterns, solve novel problems, or sense
emotional responses and react appropriately; that is, they can’t substitute
for tacit labor as they did for transactional labor. Instead machines will
have to make tacit employees better at their jobs by complementing and
extending their tacit capabilities and activities.
Second, a look back at what it took to raise labor
productivity over the past ten years shows that the
overall performance of sectors improves when the
companies in them adopt one another’s managerial
best practices, usually involving technology. In
retailing, for instance, Wal-Mart Stores was a
pioneer in automating a number of formerly manual
transactional activities, such as tracking goods,
trading information with suppliers, and forecasting
demand. During the 1990s, most other general-
merchandise retailers adopted Wal-Mart’s innovations,
boosting labor productivity throughout the sector.
4

But in the world of tacit work, it’s less likely that companies will succeed
in adopting best practices quite so readily. Capabilities founded on talented
people who make smarter decisions about how to deploy tangible and
intangible assets can’t be coded in software and process diagrams and then
disseminated throughout a sector.
Tacit technology
Companies have three ways of using technology to enhance and extend
the work of tacit labor. First, and most obviously, they can use it to
eliminate low-value-added transactional activities that keep employees
from undertaking higher-value work. Pharmacies, for example, are using
robots to fll prescriptions in an effort to maximize the amount of time
pharmacists can interact with their customers. Meanwhile, The Home
Depot is trying out automated self-checkout counters in some stores.
4
Brad Johnson, James Manyika, and Lenny Mendonca, US Productivity Growth 1995–2000: Understanding
the Contributions of Information Technology Relative to Other Factors, McKinsey Global Institute,
October 2001 (www.mckinsey.com/mgi); Diana Farrell, Terra Terwilliger, and Allen P. Webb, “Getting IT
spending right this time,” The McKinsey Quarterly, 2003 Number 2, pp. 118–29; and Diana Farrell,
“The real new economy,” Harvard Business Review, October 2003, Volume 81, Number 10, pp. 104–12.
The next revolution in interactions 71
The retailer isn’t just automating and eliminating transactional tasks; its
chairman and CEO, Robert Nardelli, believes that automated counters can
reduce by as much as 40 percent the time customers spend waiting at cash
registers. Just as important, the new counters mean that people who used to
operate the old manual ones can be deployed in store aisles as sales staff—a
much higher-value use of time.
Furthermore, technology can allocate activities more effciently between
tacit and transactional workers. At some companies, for example,
technology support—traditionally, tacit work undertaken by staff experts
on PCs and networks—has been split into tacit and transactional roles.
Transactional workers armed with scripts and some automated tools handle
the IT problems of business users; only when no easy solution can be found
is a tacit employee brought in.
Second, technology makes it possible to boost the quality, speed, and
scalability of the decisions employees make. IT, for instance, can give
them easier access to fltered and structured information, thereby helping
to prevent such time wasters as volumes of unproductive e-mail. Useful
databases could, say, provide details about the performance of offshore
suppliers or expanded lists of experts in a given feld. Technology tools can
also help employees to identify key trends, such as the buying behavior of a
customer segment, quickly and accurately.
Kaiser Permanente is one of the organizations now pioneering the use of
such technologies to improve the quality of complex interactions. The health
care provider has developed not only unifed digital records on its patients
but also innovative decision-support tools, such as programs that track the
schedules of caregivers for patients with diabetes and heart disease. Although
it is hard to determine quantitatively whether physicians are making better
judgments about medical care, data suggest that Kaiser has cut its patients’
mortality rate for heart disease to levels well below the US national average.
Finally, new and emerging technologies will let companies extend the
breadth and impact of tacit interactions. Loosely coupled systems are
more likely than hard-coded systems and connections to be adapted
successfully to the highly dynamic work of tacit employees. This point will
be particularly critical, since tacit interactions will occur as much within
companies as across them.
5
Broadband connectivity and novel applications
(including collaborative software, multiple-source videoconferencing, and
5
John Seely Brown and John Hagel III, “Flexible IT, better strategy,” The McKinsey Quarterly, 2003 Number 4,
pp. 50–9.
The McKinsey Quarterly 2005 Number 4 72
IP telephony) can facilitate, speed up, and progressively cut the cost of such
interactions as collaboration among communities of interest and build
consensus across great distances. Companies might then involve greater
numbers of workers in these activities, reach rural consumers and suppliers
more effectively, and connect with networks of people and specialized talent
around the world.
6

Competitive advantage redux
Technology itself can’t improve patient care or customer service or make
better strategic decisions. It does help talented workers to achieve these ends,
but so, for example, do organizational
models that motivate tacit employees
and help them spot and act on ideas.
These kinds of models usually involve
environments that encourage tacit
employees to explore new ideas, to
operate in a less hierarchical (that is,
more team-oriented and unstructured)
way, and to organize themselves for work. Most of today’s organizational
models, by contrast, aim to maximize the performance of transactional or
transformational workers. Tacit models are new territory.
As a result, it won’t be easy for companies to identify and develop
distinctive new capabilities that make the best use of tacit interactions—
new ways to speed innovations to market, to make sales channels more
effective, or to divine customer needs, for instance. But at least such
capabilities will also be diffcult for competitors to duplicate. Best practices
will be hard to transplant from one company to another if they are based
on talented people supported by unique organizational and leadership
models and armed with a panoply of complementary technologies. If it
becomes harder for performance innovations to spread through a sector
and thereby to boost the performance of all players, it will once again be
possible to build operating-cost advantages and distinctive capabilities
sustainable for more than a brief moment.
During the past few years, advantages related to costs and distinctiveness
have rarely lasted for long: they eroded quickly when companies built them
from innovations in the handling of what are essentially transactional
interactions. E*Trade Financial, for instance, gained tactical advantages
by optimizing transactional activities to create more effcient and less
The rigidity of traditional organizational
models too often limits innovation
and learning. See “From push to pull:
The next frontier of innovation” on
mckinseyquarterly.com.
6
Scott Beardsley, Luis Enriquez, Carsten Kipping, and Ingo Beyer von Morgenstern, “Telecommunications
sector reform—A prerequisite for networked readiness,” Global Information Technology Report 2001–2002:
Readiness for the Networked World, World Economic Forum, Oxford University Press, June 2002, pp. 118–37.
The next revolution in interactions 73
expensive ways of making trades but then watched its unique position
evaporate when other discount brokers and fnancial advisers embraced
the new technology and cut their trading fees. Cheap trades were no
longer a suffcient point of differentiation.
By contrast, advantages built on tacit interactions might stand. A company
could, for example, focus on improving the tacit interactions among
its marketing and product-development staff, customers, and suppliers to
better discern what customers want and then to provide them with
more effective value-added products and services. That approach would
create a formidable competitive capability—and it is diffcult to see
how any rival could easily implement the same mix of tacit interactions
within its organization and throughout its value chain.
Looking forward
As companies explore how to expand the potential of their most valuable
employees, they face more than a few challenges. For one thing, they
will have to understand what profle of interactions—transactional and
tacit—is critical to their business success and to allocate investments for
improving the performance of each. Some companies will have to redeploy
talent from transactional to tacit activities, as Home Depot did. Others,
following the example of companies such as Toyota Motor and Cisco
Systems, may fnd it necessary to redeploy their available tacit capacity to
transformational and transactional activities, thus bringing a new level
of problem solving to many kinds of transformational jobs. At the same
time, it will be necessary to guard against becoming overly reliant on
r Ā n i v i 1 ō
*OWFTUJOHJOUFDIOPMPHZ
ü4
Ioteractioos
Exhibit 3 oI 4
Clorce. Tec|rulu¸] ir.estaert |] ju| t]ue
ŋ
Quairiles Laseu on giovrl in laLoi µiouucriviry, ŋœœŒ÷ŌŊŊŎ.
Ō
Comµounu annual giovrl iare.
ō
Caregoiizeu Ly µieuominanr joL acriviry: racir = comµlex inreiacrions, riansacrional = iourine inreiacrions, riansíoimarional =
exriacrion oi conveision oí iav mareiials, ŒŊŊ occuµarions sruuieu.
Souice: US Buieau oí LaLoi Srarisrics, NcKinsey analysis
lT stuc| ue| eaulu]ee |] ju| t]ue, ZJJ+, S t|uusorJs ,ruairolì
Ju| t]ue
J
6Aß8,
Z

1998-ZJJ+, °
6Aß8,
Z

1998-ZJJ+, ° Iop guartiIe
1
/.+
J8.Z
c.Z
11./
1Z.c
/.c
+.1
8.c
5.c
ßottoæ guartiIe
1
+.J
J./
1.J
Tocit
T|orsíu|aotiurol
T|orsoctiurol
The McKinsey Quarterly 2005 Number 4 74
a few star tacit employees
and to manage critical tacit or
transactional activities
undertaken by partners or
vendors.
On the human-resources side,
companies will need a better
understanding of how they
can hire, develop, and manage
for tacit skills rather than
transactional ones—something
that will increasingly determine
their ability to grow. Certain
organizations must therefore
learn to develop their tacit skills
internally, perhaps through
apprenticeship programs, or
to provide the right set of
opportunities so that their
employees can become more
seasoned and knowledgeable.
What’s more, performance is
more complex to measure and
reward when tacit employees
collaborate to achieve results.
How, after all, do you measure
the interactions of managers?
7

Companies will also have to think
differently about the way they
prioritize their investments in
technology. On the whole, such
investments are now intended
largely to boost the performance
of transformational activities—
manufacturing, construction, and
so on—or of transactional ones. Companies invest far less to support tacit
tasks (Exhibit 3).
So they must shift more of their IT dollars to tacit tools, even while they
still try to get whatever additional (though declining) improvements can
r Ā n i v i 1 Ŏ
1FSGPSNBODFWBSJFT
ü4 2005
Ioteractioos
Exhibit 4 oI 4
Clorce. |e|íu|aorce .o|ies t|e aust ir tocitirtersi.e irJust|ies
0isue|siur uí o.e|o¸e EBlT0/ ue| eaulu]ee íu| cuauories |] irJust|]
t]ue,
1
|otiu uí storJo|J Je.iotiur tu aeor
T|orsíu|aotiurol
T|orsoctiurol
Tocit
Z.J
Z.Z
J.J
ßy iodustry subsector
üveraII
IraosIoræatiooaI
ûil, ¸os e(uiuaert
|uuse|ulJ ouuliorces
1.9
1.J
Curst|uctiur, ío|a
aoc|ire|]
|uae curst|uctiur
J.8
J.c
|uae íu|ris|ir¸s J.5
IraosactiooaI
Restou|orts
/i|lires
J.c
Z.9
Cuauute| |o|Jwo|e
Sueciolt] stu|es
J
1.1
J./
/uuo|el, |etoil J.5
|uuJ, |etoil J.5
0euo|taert stu|es J.5
Iacit
/uulicotiur suítwo|e
S]steas suítwo|e
/./
+.J
lrsu|orce
|eolt| co|e se|.ices
J.8
J.c
Sueciolt] lerJir¸
Z
Z.5
|u|lis|ir¸ 1.J
ŋ
LBI1DA = eainings Leíoie inreiesr, raxes, ueµieciarion, anu amoiriza-
rion, caregoiizeu Ly µieuominanr joL acriviry virlin inuusriy: racir =
comµlex inreiacrions, riansacrional = iourine inreiacrions,
riansíoimarional = exriacrion oi conveision oí iav mareiials, uara
ielecr inuusriy veiglreu-aveiage GDP Ly joL ryµe.
Ō
Incluues cieuir caiu issueis, moirgage lenueis, euucarion lenueis.
ō
Incluues kirclen ierail, movie ienral, oínce suµµly sroies.
7
Lowell L. Bryan, “Making a market in knowledge,” The McKinsey Quarterly, 2004 Number 3, pp. 100–11.
The next revolution in interactions 75
be had, in particular, from streamlining transactions. The performance
spread
8
between the most and least productive manufacturing companies
is relatively narrow. The spread widens in transaction-based sectors—
meaning that investments to improve performance in this area still make
sense. But the variability of company-level performance is more than
50 percent greater in tacit-based sectors than in manufacturing-based ones
(Exhibit 4). Tacit activities are now a green pasture for improvement.
Q
8
As measured by revenue or EBITDA (earnings before interest, taxes, depreciation, and amortization)
per employee.
The authors wish to acknowledge the contributions of their colleagues
Scott Beardsley, Lowell Bryan, Luis Enriquez, Dan Ewing, Diana Farrell, Sumit Gupta,
Lenny Mendonca, Navin Ramachandran, and Roger Roberts, as well as of John Hagel,
Professor Hal Varian of the University of California, Berkeley, and the
Cisco Thought Leadership Team.

Brad Johnson is an associate principal in McKinsey’s Silicon Valley offce, and
James Manyika is a principal in the San Francisco offce, where Lareina Yee is a consultant.
Copyright © 2005 McKinsey & Company. All rights reserved.



The McKinsey Quarterly on leadership Table of contents

3

Introduction

4

The 21st-century organization Big corporations must make sweeping organizational changes to get the best from their professionals. Lowell L. Bryan and Claudia Joyce August 2005

14

Competitive advantage from better interactions Tacit interactions are becoming central to economic activity. Making those who undertake them more effective isn’t like tweaking a production line. Scott C. Beardsley, Bradford C. Johnson, and James M. Manyika May 2006 

6

The new metrics of corporate performance: Profit per employee Most measurements of performance are geared to the needs of 20th-century manufacturing companies. Times have changed. Metrics must change as well. Lowell L. Bryan February 2007

36

Mapping the value of employee collaboration As collaboration within and among organizations becomes increasingly important, companies must improve their management of the networks where it typically occurs. Robert L. Cross, Roger D. Martin, and Leigh M. Weiss August 2006

50

Harnessing the power of informal employee networks Formalizing a company’s ad hoc peer groups can spur collaboration and unlock value. Lowell L. Bryan, Eric Matson, and Leigh M. Weiss November 2007

6

The next revolution in interactions Successful efforts to exploit the growing importance of complex interactions could well generate durable competitive advantages. Bradford C. Johnson, James M. Manyika, and Lareina A. Yee November 2005

Copyright © 007 McKinsey & Company. All rights reserved. The McKinsey Quarterly has been published since 1964 by McKinsey & Company, 55 East 5nd Street, New York, New York 100.

Introduction 

Introduction

In about 1760, a few entrepreneurs in the north of England

had the idea of using steam engines to drive machines that spun cotton thread—the germ of the first industrial revolution. A bit more than a century after that, some European and American companies launched the second industrial revolution by embracing innovations such as electricity and electric motors, internal-combustion engines, production lines, interchangeable parts, and hierarchical, vertically integrated corporations. Today our whole planet is being transformed by a dramatic, nonindustrial revolution based on intangibles such as knowledge workers, intellectual capital, collaborative networks, low-cost interactions (particularly tacit ones), and globalization. Your business, your life, and your career will all change profoundly. This collection shows what companies and the people who work for them must do to succeed in that new world. These articles illustrate the Quarterly’s best traditions: written by McKinsey consultants, they offer new ways of thinking about the challenges that companies and managers face today by combining the practical bent of client work with the fruits of extensive research. They represent just a small sampling of the knowledge and ideas available to you as a member of mckinseyquarterly.com.

 The McKinsey Quarterly 2005 Number 3 .

Lowell L. . health care. yet these employees now increasingly find their work obstructed. who might better be called professionals. They make it possible for companies to deal with today’s rapidly changing and uncertain business environment. and they produce and manage the intangible assets that are the primary way companies in a wide array of industries create value. pharmaceuticals. These talented people are the innovators of new business ideas. professionals now account for 25 percent or more of the workforce and. in some cases. the productive use of knowledge. In industries such as financial services. Creating and exchanging knowledge and intangibles through interaction with their professional peers is the very heart of what they do. these knowledge workers. or labor but. Productive professionals make big enterprises competitive. Peter Drucker coined the term “knowledge worker” to describe a new class of employee whose basic means of production was no longer capital. land.The 21st-century organization 5 The 21st-century organization Big corporations must make sweeping organizational changes to get the best from their professionals. undertake most typical key line activities. Bryan and Claudia Joyce About half a century ago. rather. Today. represent a large and growing percentage of the employees of the world’s biggest corporations. high tech. Yet most of them squander endless hours searching for the knowledge they need—even if it resides in their own companies—and coordinating their work with others. and media and entertainment.

their vertically oriented organizational structures. Effective collaboration often takes place only when the would-be collaborators enlist hierarchical line managers to resolve conflicts between competing . Professionals seeking to collaborate thus need to go up the organization before they can go across it. The McKinsey Quarterly 2005 Number 3 The inefficiency of these professionals has increased along with their prominence. most companies lack working “knowledge markets. Consider the Professional employees. frequently burden professionals with two bosses—one responsible for the sales force. designed to accommodate the “secondary” management axes that cut across vertical silos. yet most of them do little to improve the productivity of these employees. and another for a product line. Each upsurge constitute up to 25 percent or more of the workforce in the number of professionals who in financial services.000 or Corporate organizational structures—designed more professionals. with matrix and ad hoc overlays—make some 50 million potential bilateral professional work more complex and inefficient. The same holds true for knowledge: searching for it Companies must change their organizational structures dramatically to unleash the power of their means trying to find the person professionals and to capture the opportunities of in whose head it resides. work in a company leads to an pharmaceuticals. nearly always make professional work more complex and inefficient. high tech.” One measure of the difficulty of this quest is the volume of global corporate e-mail. because today’s economy. say. relationships. who create value through intangible assets such as brands and networks. These vertical structures—relics of the industrial age—are singularly ill suited to the professional work process. Many leading companies now employ 10. Worse yet. now act of collaboration. social cohesion and trust among professional colleagues declines. Article at a glance A flawed organizational design Today’s big companies do very little to enhance the productivity of their professionals. yet vertical structures force such men and women to search across poorly connected organizational silos to find knowledge and collaborators and to gain their cooperation once they have been found.8 billion a day in 1998 to more than 17 billion a day in 2004. almost exponential—not linear— increase in the number of potential Making professionals productive enables big collaborators and unproductive corporations to be competitive. further reducing productivity. who have vertically. and media and entertainment. In fact. As finding people and knowledge becomes more difficult. Professionals cooperate horizontally with one another throughout a company. interactions. retrofitted with ad hoc and matrix overlays. up from about 1. matrix structures. health care.

Companies can not only build this new kind of organization but also reduce the complexity of their interactions and improve the quality of internal collaboration by implementing four interrelated organizationaldesign principles: 1. overlaid networks and marketplaces that make it easier for professionals to interact collaboratively and to find the knowledge they need. Then these companies should create new. Other ad hoc organizational devices. A new organizational model To raise the productivity of professionals. Deploying off-line teams to discover new wealth-creating opportunities while using a dynamic management process to resolve short. and proliferating task forces and study groups. serve only to complicate the organization further and to increase the amount of time required to coordinate work internally. and collaborate with peers to generate revenues and create value through intangible assets such as brands and networks. such as internal joint ventures. Relying on measurements of performance rather than supervision to get the most from self-directed professionals . for instance. and e-mail exchanges as talented professionals—line managers or members of shared utilities—waste valuable time grappling with the complexity of a deeply flawed organizational structure. phone calls. big corporations must change their organizational structures dramatically.The 21st-century organization  organizational silos. Streamlining and simplifying vertical and line-management structures by discarding failed matrix and ad hoc approaches and narrowing the scope of the line manager’s role to the creation of current earnings 2. and formal networks to stimulate the creation and exchange of intangibles 4. The result is endless meetings. talent marketplaces.and long-term trade-offs 3. innovate. Developing knowledge marketplaces. Much time is lost reconciling divergent agendas and finding common solutions. and off-line teams on longer-term growth initiatives—with clear accountability. Companies can achieve these goals by modifying their vertical structures to let different groups of professionals focus on clearly defined tasks— line managers on earnings. retaining the best of the traditional hierarchy while acknowledging the heightened value of the people who hatch ideas. co-heads of units.

at best. Finally. and responsibilities of the line managers. for all other considerations will get short shrift until short-term expectations are met. such as the choices involved in managing shared IT costs. Defining the role of the comptroller or the country manager consistently throughout it. helps the people in those roles to interact and collaborate. As one executive we know put it. Simplify the line structure The first design principle is to clarify the reporting relationships. to promote the creation of enterprise-wide formal networks. are better ways of accomplishing the purposes of these ad hoc structures. it has the luxury of focusing other professionals on the long-term creation of wealth. you don’t want people who are engaged in hand-to-hand combat to design a long-term weapons program. but we don’t know of any company that applies all of them holistically— and this failure limits the ability to perform up to potential. out of the line structure. These mechanisms are typically created by defining and clarifying the decision-making authority of each member of the senior leadership team and establishing enterprise-wide governance committees as required. To achieve this goal. A company that aims to streamline its line-management structures should create an effective enterprise-wide governance mechanism for decisions that cross them. Dynamic management and improved collaboration. geographic. The advantages of such a separation are obvious. for example. accountability. functional. a company must establish a clearly dominant axis of management—product. Manage dynamically Once the newly simplified vertical structure allows line managers to limit their attention to meeting the near-term earnings expectations of the company. It may also be necessary to take important support functions. who make good on a company’s earnings targets. But that would be more than offset by a decrease in its effectiveness. as we show later. A company that tries to simplify its vertical organizational structure without helping large numbers of self-directed professionals to collaborate more easily might increase its efficiency. The McKinsey Quarterly 2005 Number 3 The ideas underlying each of these policies may not be entirely new. who are often. which demand focused management. so that specialized professionals (rather than line managers. for example. or customer—and eliminate the matrix and ad hoc organizational structures that often muddle decision-making authority and accountability. parallel structures and parallel roles should be established across the whole extent of the company. gifted amateurs) can run these functions as shared utilities. .

Once an initiative is ready to be scaled up—when revenues and cost projections become clear enough to appear in the budget—it can be placed in the line structure. decisionHow can managers translate the concept making protocols. at the top of the house rather than allowing them to be made. What we call dynamic management can help: a combination of disciplined processes. or both at the end of each stage. The primary stages of such an investment process are diagnosing the problem or opportunity. midcourse corrections. companies must manage their short. Of course. “Just-in-time strategy for a turbulent world. . implicitly. rolling budgets. 2 to 4 percent of all spending) and some of their best talent to finding and developing longer-term strategic initiatives. focused professionals with the freedom “to wander in the woods. This change further simplifies the line managers’ role.” discovering new. a company must develop 1 Lowell L. 2002 special edition: Risk and resilience. who must live day to day in an intensely competitive marketplace. Few downthe-line managers. Bryan. to manage the portfolio of initiatives as part of an integrated seniormanagement approach to running the entire enterprise. Develop organizational overlays Having stripped away unproductive matrix and ad hoc structures from the vertical organization and clarified the line structure. In fact. by down-the-line managers struggling to make their budgets. building new businesses.and long-term earnings in a way that integrates their spending on strategic initiatives with the overall budget. as well as the R&D processes of leading industrial corporations. at the enterprise level. designing a solution. Not that companies should forgo discipline while undertaking such a project. 16–27. the portfolio-of-initiatives approach to strategy enables them to “plan on being lucky” by using the staged-investment processes of venture capital and principal investing firms.The 21st-century organization  Ongoing multiyear tasks such as launching new products. pp. explicitly. winning value propositions by trial and error and deductive tinkering. Each major one usually has a senior manager as its sponsor to ensure that resources are well invested. of corporate performance into an and calendar-management procedures operational reality? See “Managing for improved corporate performance” makes it possible for companies on mckinseyquarterly. creating the prototype. Dynamic management forces companies to make resource allocation trade-offs. have the time or resources for such a discovery process. with natural stopping points. effective way of making the necessary trade-offs.” The McKinsey Quarterly. or fundamentally redesigning a company’s technology platform usually call for small groups of full-time.1 Companies that take this approach devote a fixed part of their budgets (say. so they will need to adopt a systemic. and scaling it up.com.

What’s the best way of encouraging strangers to exchange valuable things? The well-tested solution. and formal networks will make all three more effective. of course. How can companies create effective internal markets when the product is inherently intangible? Among other things. helps members of a formal network to exchange knowledge. talent marketplaces. Companies that establish such overlays are making investments not only to minimize the search and coordination costs of professionals who exchange knowledge and other valuable intangibles among themselves but also to maximize the opportunities for all sorts of cost-effective. And the key to meeting this goal is understanding that the most valuable knowledge of a company resides largely in the heads of its most talented employees: its professionals. is markets. As we have argued. they boost rather than hinder productivity. These companies must also learn how to encourage people who may not know each other—after all. and to develop communities that create intangible assets. these techniques can make it possible for companies to work horizontally in a far more cost-effective way. Exchanging knowledge on a company-wide basis in an effective way is much less a technological problem than an organizational one. big corporations usually have large numbers of professionals—to work together for their mutual self-interest. working markets need objects . Because these market and network overlays help professionals to interact horizontally across the organization without having to go up or down the vertical chain of command.10 The McKinsey Quarterly 2005 Number 3 organizational overlays in the form of markets and networks that help its professionals work horizontally across its whole extent. to promote the exchange of knowledge. In combination. to find and collaborate with other professionals. the benefits have been limited. which in turn helps to strengthen the network. Knowledge marketplaces. These overlays make it easier for them to exchange knowledge. A knowledge marketplace. We believe that moving simultaneously into knowledge marketplaces. Despite heavy investment. A talent marketplace works better if the people who offer and seek jobs in it belong to the same formally networked community. The trick is to take the market inside the company. Real value comes less from managing knowledge and more—a lot more—from creating and exchanging it. companies must remove structural barriers to the interaction of their professionals. which the economy uses for just this purpose. knowledge management has generated a good deal of buzz. For the better part of the past 15 years. productive interactions among them. for example.

100–11. A successful mechanism of this kind substantially improves the ability to create and exchange knowledge and dramatically cuts search and coordination costs. than alternative sources are. standards. “Making a market in knowledge. exchange mechanisms. and market facilitators often help markets to work better.” The McKinsey Quarterly. Simultaneously. In addition. the same clientele (the automotive industry). A company can create similar efficiencies by developing a talent marketplace that helps employees in a talent pool. . or the same geography (China)—naturally form social networks. pp. say).2 In particular. to say nothing of prices. an exchange mechanism to facilitate staffing transactions. These networks lower the cost of interaction while increasing its value to all participants.The 21st-century organization 11 of value for trading. People with common interests—such as similar work (industrial engineers. to produce high-quality “knowledge objects”). Talent marketplaces do exist—particularly in professional organizations— but like knowledge marketplaces they are at an early stage of development. anyone with assignments to offer can review all of the people looking for new opportunities. Bryan. regulations. Talent marketplaces. As with marketplaces for knowledge. validating the qualifications of candidates. Companies must define the talent marketplace by specifying standardized roles. to explore alternative assignments varying from short-term projects to longer-term operating roles. Formal networks. “Buyers” must be able to gain access to content that is more insightful and relevant. the mechanics of reassignment. These conditions don’t exist naturally—a knowledge marketplace is an artificial. so they are rare. the suppliers of knowledge must have the incentives and support to codify it (that is. 2 Lowell L. and so forth. We have found that building an effective one in a large company requires significant investments to get the conditions in place—but that such a marketplace can indeed be built. the process of conveying decisions to reassign employees). and protocols and standards (how long assignments run. 2004 Number 3. companies must invest in their talent markets to ensure that gifted men and women looking for new jobs hook up with managers seeking talent. protocols. and competition among suppliers. as well as easier to find and assimilate. managed one—so companies must put them in place. The other requirements include pricing (the compensation for a particular role or assignment). determining how managers receive the job seekers’ performance evaluations. either within a single organizational unit or across the enterprise. Knowledge marketplaces are a relatively new concept.

a formal network with specific areas of economic accountability can undertake many of the activities that have inspired companies to use matrix management structures. a technology platform supporting the network’s activities). Measure performance The final set of ideas rounding out this new organizational model involves relinquishing some level of supervisory control and letting people direct themselves. the more chances they have to find opportunities for collaboration. A formal network relies on self-directed people who work together out of self-interest. is one of the most important things a company can do. But what’s needed is inspired leadership. Of course.12 The McKinsey Quarterly 2005 Number 3 A network often provides them with increasing returns to scale: the larger it is. though. guided by performance metrics. only a few have formalized them. Others include developing incentives for membership. because they don’t extend to many potential members in far-flung units and geographies). management will continue to be vital—particularly to get value from the many . by making investments to generate knowledge). The solution. protocols. establishing standards and protocols. accountable leaders must control large companies even as many of their workers become more and more self-directed. most particularly. not more intrusive management. They often have limited reach (for example. Although social networks flourish at many companies. and. for a company. values. they sometimes operate inefficiently (several conversations might be required to reach the right person). a formal network enables people who share common interests to collaborate with relatively little ambiguity about decision-making authority—ambiguity that generates internal organizational complications and tension in matrixed structures. defining separate territories (the existence of more than one social network may confuse would-be members). is to boost the value of the network by investing in it and formalizing its role within the organization. In addition. standards. can fail to attract enough investment to serve the common good of all members effectively. That next step. One such move is the designation of a network “owner” to build common capabilities (for instance. may rely too much on the participants’ goodwill. In fact. and consequence-management systems. Social networks do face problems. because it removes unnecessary complexity from horizontal interactions among talented people across organizational silos. and providing for a shared infrastructure (say. What’s more. To be sure. while a matrix uses a hierarchy to compel people to work together.

Copyright © 2005 McKinsey & Company. But as the workforce increasingly comes to consist of self-directed professionals. from the managers and the professionals they employ. for example. both individual and collective. measuring performance is more important than providing financial incentives to reward it. . should be more highly valued than those who are equally good at doing their own work but not at developing talent or contributing knowledge. Get the metrics wrong and unintended behavior is the result. To motivate the collaborative behavior that makes this new organizational model work. Rather. leaders will have to manage them by setting aspirations and using performance metrics that motivate them to organize their work. In the process. Q Lowell Bryan is a director and Claudia Joyce is a principal in McKinsey’s New York office. companies must create metrics that hold employees individually accountable for their contribution to collective success—an idea we call holding people “mutually accountable. People who are great at developing the abilities of other talented people or at contributing distinctive knowledge. using new principles that take into account the way professionals create value. to meet those aspirations. All rights reserved. more broadly. they can become fundamentally better at overcoming the challenges— and capturing the opportunities—of today’s economy. Big companies that follow these principles will get more value. at less cost. The challenge is that to measure it effectively.” Such metrics are particularly important for senior and top managers but are required. companies must design a new model holistically. the metrics must be tailored to individual roles and people. One successful CEO once told us that to motivate behavior. A new organizational model for today’s big corporations will not emerge spontaneously from the obsolete legacy structures of the industrial age.The 21st-century organization 13 employees who will go on laboring in “industrially engineered” processes and to hold all of a company’s workers and managers accountable for their performance. for all self-directed workers.

14 The McKinsey Quarterly 2006 Number 2 .

Manyika For many employees today. Bryan and Claudia Joyce. Scott C. complex problem solving is the essence of their work. These “tacit” activities—involving the exchange of information. for instance. and leverage 1 Leigh Wells Lowell L. pp. customers. collaborate continually with engineers and managers to find new ways of reducing costs and solving quality problems. “The 21st-century organization. But employees such as managers and salespeople. now make up 25 to 50 percent of the workforce. manage talent. the making of judgments. Johnson. Bradford C. Beardsley. production workers at Toyota Motor. But building these advantages won’t be easy: companies must alter the way they craft strategies.” The McKinsey Quarterly. and a need to draw on multifaceted forms of knowledge in exchanges with coworkers. . design organizations.1 During the next decade. 24–33. whose jobs consist primarily of such activities. collaborative. Making those who undertake them more effective isn’t like tweaking a production line. companies that make these activities—and the employees most involved in them—more productive will not only raise the top and bottom lines but also build talent-based competitive advantages that rivals will find hard to match. They are typically a company’s most highly paid workers and make huge contributions to its competitive prospects in a fast-changing global business landscape. 2005 Number 3.Competitive advantage from better interactions 15 Competitive advantage from better interactions Tacit interactions are becoming central to economic activity. Many employees engage in activities of this kind to some extent. and suppliers— are increasingly a part of the standard model for companies in the developed world. and James M.

has profoundly altered companies. But the productivity of marketing managers and lawyers can’t be raised by standardizing their work or replacing them with machines.2 In most environment for constant learning—are familiar. the capital. managing the way customers buy and experience products and services. Johnson. Hall. solutions within an environment that fosters trust globalization. for example. the work of the clerks in the accounts-payable function). Manyika. and Anupam Sahay. involves extracting raw materials or working on a production line. four But it will be critical to understand exactly what out of five nonagricultural jobs a company must do to use these workers most effectively and how such efforts differ from other involve them. and monitoring judgment to solve the deepest business problems. The best way for executives to begin is to understand Companies are looking for ways to improve the the nature of what economists effectiveness of their top talent: workers who call tacit interactions—the searchinteract with others and draw on experience and ing. (Nor can companies boost the tacit component of other jobs in this way— automation does nothing. instead. only one in five kinds of productivity programs.” and to improvise and improve new faster pace of specialization. For a comparison of selected developed and developing countries. the nature of work within them and at their borders. and information. and Bradford C. James M. and. (This shift is under way in the developing world as well. The result is a The levers that managers must pull to get this dramatic increase in the volume and job done—flattening hierarchies and creating an value of interactions. Yee. What makes these workers valuable is their ability to work collaboratively. Lenny Mendonca. and Lareina A. must change the way they organize. companies around them. developed economies today. Manyika. and technical change and constant learning. a company must boost these workers’ 2 Patrick Butler. Ted W. pp. Alistair M. see Exhibit 1. 2005 Number 4. consequently. reviving brands. Byron Auguste. . A century ago. 4–23.” The McKinsey Quarterly. hatch their strategies. Hanna. to help the production workers on a Toyota assembly line collaborate with others. the supply chains To put these workers to best use. their customers. services. and negotiating acquisitions. activities required to exchange goods. and manage their talent and IT. 1997 Number 1. Article at a glance Companies boost their productivity by improving the efficiency of transformational activities (such as the extraction of raw materials) or of transactions (for instance. coordinating.) The old strategies for efficiency improvements don’t apply to employees whose jobs mostly involve tacit interactions. 20–33. pp. James M. to leverage “relationship During the past half century.16 The McKinsey Quarterly 2006 Number 2 technology. the proportions were reversed. “The next revolution in interactions.” The McKinsey Quarterly. “A revolution in interactions. Examples include running supply chains.) The number of jobs chiefly involving the most complex interactions—tacit ones—is growing faster than any other type of job in developed nations.

Early innovators are emerging. for instance—thus making tasks . Even as they become more and more dominant. innovate. Aided by technology. Tacit productivity In work of any kind. companies have adroitly smoothed variance in call centers and IT help desk operations by standardizing interactions— writing scripts for call-center operators. and manage in an era when tacit interactions dominate and drive performance. Although all three are important in today’s developed economies. The boundaries between these three categories of business activities— transformational. But that must now change. they change constantly as a result of innovations in the way functions and tasks are organized and the impact of technology. Manufacturers know how to reduce variability in production work and have therefore greatly raised their operating productivity over the past two decades.Competitive advantage from better interactions 17 productivity by making them more effective at what they do. variability is a sure sign that things could be better. As a result. and tacit—are not static. the managerial science for boosting their effectiveness remains less well understood than are ways of increasing the efficiency of transformational and transactional activities. the company will build talent-based competitive advantages that are difficult for rivals to duplicate. Executives will have to learn how to compete. transactional. it will be necessary to make a real effort to boost the productivity of tacit interactions. and sectors where tacit interactions have been dominant for some time offer useful lessons.

and lowest-performing sales teams is wide. that approach can undermine their effectiveness—salespeople. But companies don’t improve tacit interactions by forcing salespeople (or other tacit workers) to follow a uniform procedure. can engage in value-adding interactions with customers. Executives recognize that they must manage these workers differently. Variability often characterizes the performance of. generate more sales and profits when they have better information at their fingertips. tacit interactions now constitute the primary activities of 63 percent of the workforce. The proportion is 60 percent in securities companies. In addition to interacting with existing and potential customers. Workers engage in a larger number of higher-quality tacit interactions when organizational . 30 percent of the employees undertake tacit work. enterprise salespeople manage a broad number of interactions and must constantly solve problems to get the job done. But look at work involving tacit interactions in almost any company today. In insurance companies. interactions drive customer satisfaction and loyalty—and. and Yahoo!. 70 percent in health care. it’s also about how well a salesperson manages the work. learning. Companies can analyze work done in processes and root out wasteful activities so that employees do more in less time. merger integration managers. Often. are better networked with customers. and 45 percent in retailing. and insurance agents. ultimately. movie producers. Microsoft. for instance. salespeople work with marketing staffers. doctors and nurses at Kaiser Permanente. On the contrary. and sort out shipping problems with supply chain supervisors. the sales force. In most high-tech companies. and learn and grow in deal after deal. In fact. the variance between the highest. success in sales. performance always fluctuates wildly. who direct the connections among the company’s salespeople. suppliers. including software engineers at Google. suppliers. collaboration. Even in utilities. and organizational colleagues. fund managers at Blackstone and Fidelity Investments. draw on the services of technical-support and customer service specialists. for instance. and innovation. shared values. Managing for effectiveness in tacit interactions is about fostering change. That is also true for other workers engaged primarily in tacit interactions.18 The McKinsey Quarterly 2006 Number 2 into routines. and contract manufacturers. collaborate to develop the better ideas that emerge from iterative teamwork. Effective performance isn’t simply about generating leads and closing deals. Cisco Systems’ manufacturing managers.

performance of companies in relacompanies in the top quartile tively tacit-interactive sectors varied understand how to make their tacit far more than that of other compaemployees more effective nies.4 in sectors with a high level of interactions. First.9 for companies in sectors with a low level of tacit interactions. all of which make their work more effective and more likely to achieve desired outcomes. Second. Tacit capabilities do not resemble IT systems or reengineered processes that can be copied easily. in some highly tacit sectors. The actions and innovations of these people are usually specific to a particular business situation. companies have always had workers involved in tacit interactions. facts. because tacit interactions are hard to specify in advance. These aren’t new management ideas. But the ever-increasing growth in their number and value is driving companies to adopt such ideas more quickly and deeply. they can create competitive advantages that rivals in their sectors find hard to match—in particular. and the ability to connect the dots in problem solving. We found that the In certain highly tacit sectors. Among companies in sectors with a middling number of tacit interactions it was 5.” Such interactions involve talented people armed with experience. and when they have the tools to make better decisions and communicate quickly and easily. as our study of more than 8. when people trust each other and have the confidence to organize themselves. . companies have considerable competitive headroom for improving the productivity of those who undertake tacit interactions and less headroom for improving the productivity of other workers. Their power lies in the collective company-specific knowledge that emerges over time. or “prewire. companies in the top quartile understand how to make tacit workers more effective and now have a significant performance lead on rivals that still manage for efficiency (Exhibit 2). indeed.5. The level of performance variability (defined as the standard deviation of performance divided by the mean level of performance) was 0. The top performers have also figured out that by managing their tacit interactions more effectively. rising to 9. This widening variability as the extent of tacit interaction increases reflects two things. The competitive imperative The need to move forward is both substantial and urgent. judgment. creativity.Competitive advantage from better interactions 19 barriers (such as hierarchies and silos) don’t get in the way.000 US companies with a preponderance of tacit interactions suggests.

the supply chain. operations. Strategy and innovation Wherever groups of people collaborate to solve problems—in the field. Management’s job is to foster connectivity. Each of these is essentially a piece in a set of interconnected changes. facilitate learning.20 The McKinsey Quarterly 2006 Number 2 New management science Efforts to make tacit interactions (and hence the talent that undertakes them) more effective require changes in every facet of a business. remove barriers. from hatching strategies to organization to managing talent and leveraging IT. and provide new tools that help workers collaborate and learn within an environment that demands more and more complex and often decentralized decision making. marketing—innovations are more likely to occur . The focus of managerial action is to establish conditions that allow tacit interactions to emerge and flourish rather than trying to engineer connections from the top down.

3 Lowell L. instead. encourages its software engineers to devote 20 percent of their time to pursuing their own ideas for new and innovative products and services. companies must also upend their strategic decision-making processes. for instance. Managers today commonly believe that more and better ideas will follow when communication and interaction increase inside a company and beyond—with its partners. New ideas are exposed to the market through the Google Labs Web site and tested both inside and outside the company.3 Not that management should abdicate its role in setting thematic strategic goals and the company’s direction—quite the opposite. A trader who works for one of a company’s funds isn’t likely to share ideas that would help the people who run the rest of them if each fund’s employees are rewarded by its performance relative to that of the others. To boost the effectiveness of tacit interactions. and leading. managers should construct incentives that stimulate collaboration by encouraging innovators to share their inventions and insights within the organization. Finally. Rewards could also reflect an innovator’s ability to attract resources and users. was one such product. 94–105. “Managing for improved corporate performance. a next-generation mapping application.Competitive advantage from better interactions 21 at the front lines of interaction than at the corporate center. Bryan and Ron Hulme. sharing. Pilots that catch on are adopted and those that don’t are shut down. Google Earth. trials. the innovator succeeds by influencing them and leveraging a personal network—collaborating.” But few companies bake this understanding into the development of strategy by altering traditional top-down processes to include mechanisms and approaches that allow a portfolio of initiatives to emerge from internal and external interactions. inspiring. Bottom-up innovators usually don’t have the structural authority to order people to join a team. since these become crucial to providing a “magnetic north” as innovations occur at the interaction interfaces. and learning.” The McKinsey Quarterly. pp. such as customers. . A company can boost the number and quality of the interactions likely to promote innovation if it creates the conditions that allow them to emerge. innovations in tacit interaction are by nature usually the result of decentralized experimentation. customers. so the allocation of resources is more an emergent activity (which isn’t centrally planned or predetermined) than a managed one. Rewards for collaborating and for sharing knowledge. suppliers. Google. help the organization as a whole rise to the level of its best innovations. Furthermore. which assesses the success of initiatives by gauging how much attention and resources they attract. and communities of interest—and become “multidirectional. by contrast. or reflect the breadth and depth of their personal and professional networks. 2003 Number 3.

to manage these workers for efficiency— while simultaneously enabling tacit workers to interact in more fluid structures. but tacit work value for seeking it. organizaWorkers will exchange information if there tional structures presuppose structures is a fair return on sharing it and a clear for getting work done. and incentives will need to change as well. particularly if. In fact. on the one hand. It will thrive. people now come together in project teams. this approach is common in certain professional-services and engineering firms. In many companies. Some of these changes are already under way. is improvisational and difficult to define in advance.com. Tacit interactions reduce the importance of structure and elevate the importance of people and collaboration. in real time. metrics. The necessary balance will require trade-offs between performance norms. What’s more. to stimulate interactions. and cooperative norms. While the command and control exerted by hierarchies help a company to manage its routine processes and tasks efficiently.22 The McKinsey Quarterly 2006 Number 2 Organization To encourage more interaction. but the pace of change has been slow and its effectiveness questionable. and then disassemble to start the process again by joining other informal teams. for it follows the problem being solved and the nature of the opportunity at hand. The kind of network building that tacit workers must do to boost their effectiveness thrives in a culture built on trust. too. and learning Culture. knowledge. values. address an issue. Hierarchy-busting has been a theme in the business press for years. By contrast. in companies that . organizations want whatever information is relevant for solving a particular problem to be shared among teams laterally. their tacit interactions are evenly balanced and intermingled with the transactional and transformational activities they undertake. so their organizational charts rarely reflect what is really happening within them. and collaboration. They will still need to manage workers who primarily undertake transformational (production) or transactional tasks—that is. on the other. Companies will face a real challenge when they need to balance old.and new-school management sciences. they also short-circuit tacit interactions: information moves up and down a hierarchy at defined management levels. People. as is often the case. See “Making a market in knowledge” on mckinseyquarterly. irrespective of reporting channels and silos. companies must become more porous by continuing to break down barriers to interactions—barriers such as hierarchies and organizational silos. innovation.

Competitive advantage from better interactions

23

reward collaboration, dispense group-based incentives, and measure tacit work by its impact and the relationships that those who engage in it forge. Output measures alone are far less effective in the messy, “inefficient” world of tacit innovation. Few of these “soft” managerial mechanisms have legs today. How do you measure the contribution of an employee who is 1 of 20 people on a team? What was that employee’s contribution to the outcome? Is it even clear whether that contribution was positive or negative? How do team managers measure the work of people who serve on more than one team? Moreover, as decision making becomes more decentralized and organizations grow in size, it will become critical to articulate clearly the corporate and professional values that will provide a “compass” and shared sense of purpose and direction to tacit workers. Evaluation processes will need to evolve to include more peer- and project-based reviews, as opposed to the lines of traditional reporting, and must also assess softer aspects of work such as values, nonhierarchical leadership abilities, and mentoring skills. Wherever tacit interactions take place, so do learning and the creation of new knowledge. The people involved become sources of and contributors to institutional learning. Companies can’t manage this kind of knowledge from the top down. Instead, managers must promote its capture and sharing by developing the right infrastructure and incentives, as well as a “market in knowledge.” Recently, blogs (online diaries), wikis (Web sites where users can contribute and edit content), and the like have created new, decentralized, and dynamic approaches to the capture and dissemination of the knowledge critical for tacit interactions. The focus of learning changes too. Organizations can use programs delivered in classrooms or sites to train production workers to operate lathes or callcenter personnel to handle incoming calls. But learning in the tacit world is based much more on experience and apprenticeship and on the ways in which both are scaled across the networks of people who participate in interactions: inexperienced managers learn from experienced ones. Also, managers continually change their roles—they must constantly study examples and analogies. Companies may even find it worthwhile to expose tacit workers to totally new experiences to round out their capabilities. Finally, even hiring profiles will change—indeed, in some tacit-intensive industries, such as software and hospitals, they already have. Managers in these organizations have redefined their job descriptions and criteria in order to hire people who can solve problems, work under ambiguous reporting relationships, and network. But the pool of experienced tacit workers is

24

The McKinsey Quarterly 2006 Number 2

finite, and demand is increasing; companies already feel the pinch. In reaction, they may cast a wider global net for tacit talent. One thing is clear: for tacit interactions, selecting and motivating talent are core processes that drive effective outcomes. Technology Clearly, technology will play a critical role in fostering tacit interactions and making them more effective and valuable. Indeed, technology has in large part been responsible for the acceleration of tacit interactions over the past 20 years. Two decades ago, international calls were costly and e-mail was a novelty; today, global voice connections are cheap, people around the world send about 30 billion e-mails a day, and entirely new technologies— broadband Internet, search capabilities such as Google, mobile phones, personal digital assistants such as BlackBerries and Treos, and videoconferencing—make it possible for tacit interactions to happen more easily. Companies will increasingly focus on these kinds of technologies to further improve tacit work, thereby raising a host of new IT-management issues. The bulk of corporate investment in technology has been devoted to improving transactional and even transformational activities. New investments in PDAs, collaborative software, wiki tools, and other technologies that improve tacit interactions will be far less costly than, say, enterprise transactional systems. But they do require new IT architectures and skills. Some companies are already getting it right. Tacit-dominated sectors in the top quartile of labor productivity growth have armed their employees with five times more IT stock than sectors in the bottom quartile. Further, they are increasing their IT base per employee 40 percent more rapidly (on an annual basis). The good news is that, with few exceptions, most enterprises now have an underlying communications infrastructure, which is vital for extending the reach of interactions. This infrastructure must go on evolving to provide a foundation for ever-richer media. Companies will increasingly need to deploy technology that makes shared data, information, and expertise available in real time; to offer decision support tools that help workers involved in tacit interactions create insights from data and analyses and that enhance the context and information that interactions require; to improve the ability of employees, customers, and suppliers to interact; and to offer effective collaboration tools for multiparty work flows. Many of the technologies and tools that tacit workers are going to use will promote the collaborative and dynamic pursuit, capture, and sharing of knowledge and will allow for more video, audio, and graphics to facilitate remote interactions and broader access to scarce expertise. Tools based on

Competitive advantage from better interactions

25

search capacities, collaborative approaches to capturing and organizing knowledge, and new digital-learning channels are likely to emerge. Further, these new tools and approaches have broader implications for the way companies manage their IT infrastructures and operations. The new world won’t use either the practices and organizing principles that production-support and transaction systems require or the big, rigid business applications designed to automate transactions and make them efficient. Not that enterprise applications won’t be required; rather, they must evolve to make tacit interactions more effective. Executives will have to focus on deploying work-group-centric tools that are easy to set up and tear down as projects and strategic experiments come and go. They will also have to find ways of connecting these tools easily to preexisting interaction platforms. In addition, the issue of information overload must be addressed: already, Microsoft and others are trying to help tacit workers filter data from interactions more successfully and thereby reduce the burden of the excess information created when companies manage interactions (and their rate of increase) improperly.

Managing in an environment where most workers mainly participate in interactions will upend the greater part of what senior management has learned over the past half century. But the opportunity to create new forms of competitive advantage is clear for companies that take a new approach. The time to start is now.

Q

The authors wish to acknowledge the contributions of their colleagues Lowell Bryan, Dan Ewing, Roger Roberts, and Emily Szydlowski, as well as of John Hagel III; Professor Hal Varian, of the University of California, Berkeley; and the Cisco Thought Leadership Team. • Scott Beardsley is a director in McKinsey’s Brussels office, Brad Johnson is an associate principal in the Silicon Valley office, and James Manyika is a principal in the San Francisco office. Copyright © 2006 McKinsey & Company. All rights reserved.

.

and other intangibles created by talented people and represented by investments in such activities as R&D. intellectual capital. companies create wealth by converting these “raw” intangibles into the institutional skills. relationships. customer bases. Metrics must change as well. The New Organizational Wealth: Managing and Measuring Knowledge-Based Assets. the most valuable capital that companies possess today is precisely intangible rather than financial. 1997. Indeed.The new metrics of corporate performance: Profit per employee 27 The new metrics of corporate performance: Profit per employee Most measurements of performance are geared to the needs of 20th-century manufacturing companies. The vast majority of companies still gauge their performance using systems that measure internal financial results—systems based on metrics that don’t take sufficient notice of the real engines of wealth creation today: the knowledge. Increasingly. 1 Kerry Grady Karl Erik Sveiby. and networks that raise profit per employee and ROIC . patents. San Francisco: Berrett-Koehler Publishers. marketing.1 Companies should redesign their financial-performance metrics for this new age. software. even though the sources of those returns are intangible. in the sense of delivering cash returns. . brands. reputations. Bryan Let’s get right to the point: companies focus far too much on measuring returns on invested capital (ROIC) rather than on measuring the contri- butions made by their talented people. Lowell L. These intangibles are true capital. and training. Times have changed.

Consider a simple approximation of intangible capital: the market value of a company less its invested financial capital. January 2004 The bigger problem is that most companies gear the way they measure their financial performance to the needs of an earlier industrial age. Financial performance .5 trillion. The development of external financial reports according to generally accepted accounting principles (GAAP) ranks among the principal foundations of our modern global capital marketplace. so it generally isn’t clear which intangibles are the sources of profits— or what specific balance of intangible and tangible assets should get the credit (or blame) for results. do you value a brand? Intangibles are embedded in the value chain of production. knowledge creation. But greater attention should be paid to the role of intangible capital and the ways of accounting for it. are the basic source of wealth. and income statements—is and always will be the principal metric for evaluating a company and its managers. Companies can redesign the internal financialperformance approach and set goals for the return on intangibles by paying greater attention to profit per employee and the number of employees rather than putting all of the focus on returns on invested capital. training. for example. Advertising. Financial performance—observed through balance sheets.” 2003 Number 3 “Why the biggest and best struggle to grow. Despite the evidence that intangibles are now the true source of corporate wealth.” Web exclusive. Why? One reason is that accounting for intangibles is difficult. companies tightly control discretionary spending on them.” 2005 Number 3 “Managing for improved corporate performance. each intangible’s specific contribution is hard to assess. Companies fill their annual reports with information about how they use capital but fail to reflect sufficiently on their use of the “thinking-intensive” people who increasingly drive wealth creation in today’s digital economy. The superior performance of some of the largest and most successful companies over the past decade demonstrates the value of intangible assets. new-product development.com “The 21st-century organization. cash flow reports. software projects. Related articles on mckinseyquarterly. more and more. how. versus $800 billion in 1985. in 2005 the intangible capital of the world’s largest 150 companies was $7. when capital enjoyed pride of place in the minds of strategists and investors. R&D. Using book capital as a crude proxy for financial capital.28 The McKinsey Quarterly 2007 Number 1 Article at a glance Today’s approach to measuring financial performance is geared excessively to the capital-intensive operating styles of 20th-century industrial companies. In particular. It doesn’t sufficiently account for factors such as the contributions of talented employees that. and so forth are almost always expensed on a “What can we afford?” basis.

The driver of this dramatic rise in market cap was a fivefold increase in average profits—an increase brought on in turn by a more than 100 percent jump in profit per employee and a doubling in the number of employees. Intangible investments are mostly expensed. This conservatism isn’t necessarily bad but does inspire top managers to cut discretionary spending on intangibles in order to deliver quick earnings. 2005 Number 4. from 1995 to 2005. with total returns to shareholders (TRS) at 17 percent a year. Bryan and Michele Zanini. the median market cap of this group of companies rose to $168 billion.” The McKinsey Quarterly. in the case of financial institutions) has increased to 23 percent. . the top 30 of the very largest companies in the world (ranked by market capitalization) have seen their profit per employee rise to $83. these companies’ ROIC increased. strategically minded executives must embrace a radical idea: changing financial-performance metrics to focus on returns on talent rather than returns on capital alone. That approach may raise short-term profits but can also undermine a company’s long-term health. for evaluating executives. This shift in perspective would have far-reaching implications—for measuring performance. 46–59. my colleagues and I have examined how. let’s reflect upon the way some companies have recently created great wealth by using their thinking-intensive people rather than their capital. In past articles. Only if executives begin to look at performance in this new way will they change internal measurements of performance and thus motivate managers to make better economic decisions. By comparison.000. compared with the way it treats capital investments in tangible assests. not capitalized. and income statements) no doubt is and will remain the principal metric for evaluating a company and its management. cash flow reports. over this same period. the number of people these companies employ has grown to 198. even for the way analysts measure corporate value. from $35. pp. from $34 billion. from 92. As a result. particularly about spending on intangibles. GAAP accounting currently treats investments in intangibles conservatively.2 On average. But it’s time to recognize that financial performance increasingly comes from returns on talent.000. and their ROIC (or book value. “Strategy in an era of global giants.The new metrics of corporate performance: Profit per employee 29 (seen through balance sheets. by only a third. from 17 percent (Exhibit 1). 2 Lowell L. To boost the potential for wealth creation. Measuring financial performance in the digital age Before exploring the new metrics needed to achieve these goals.000. not on capital.000.

then. thousands CAGR = 8% Return on invested capital (ROIC). top companies with negative net incomes. dividend policies. first. profit per employee is a good proxy for earnings on intangibles.3 % CAGR = 3% 1995 2005 1995 2005 1995 2005 1995 . Clearly. perhaps surprisingly. US and foreign companies by American depositary receipts. With these metrics. and liquidity preferences. a new set of metrics could help companies gauge their performance more effectively. on how much profit per employee a company generates. in the case of nancial institutions. And they should keep a clear eye on ROIC . talent—not capital— is usually the scarcer resource. Soaring profits Drivers of growth for 30 largest companies. Capital. Executives should home in. 3 Or book value. is subject to the vagaries of accounting definitions and such corporate-finance decisions as debt-to-equity ratios. so maximizing both expressions increases total profit. but more as a way of ensuring that the company earns more than the cost of that capital than as an aspiration in its own right. But these results do suggest that companies need to take a new approach to measuring financial performance— an approach based on maximizing returns on people.1 1995–2005 Median market capitalization. by market capitalization in It is hardly a surprise that growth in profits and market caps should be closely correlated and that a fivefold increase in profits should lead to a similar increase in market caps. $ billion Compound annual growth rate (CAGR) of TRS2 = 17% 168 198 83 23 35 34 92 17 Profits per employee. the 30 largest companies (by market capitalization) have seen their profits per employee increase dramatically. partly because the number of people a company employs is easy to obtain. after all. $ thousand CAGR = 9% Average number of employees. As we’ve noted. and as any executive will testify. They should make the number of employees a key factor in strategic thinking. Concentrating on this formula (as opposed to returns on capital) offers several advantages. For one. unlike ROIC . is the product of profit per employee and the total number of employees. which drives market capitalization. Total profit. the company can set its goals for the return on intangibles . excludes outliers and 2005 2 Total returns to shareholders.Q1 2007 Profit per employee The McKinsey Quarterly 2007 Number 1 30 Exhibit 1 of 4 Glance: From 1995 to 2005.

and so forth). Since companies expense their spending on intangibles but not on capital investments (which are usually depreciated over time).3 Profit per employee is one measure of these rents. consistently sits atop the Forbes annual wealth list. remember. This approach focuses the minds of managers on increasing profit relative to the number of people a company employs. It suggests that the most valuable use of an organization’s talent is the creation and use of intangibles. profit per employee) and growth (the number of employees). profit per employee is a conservative. Profit per employee If a company’s capital intensity doesn’t increase. the company the family controls. . If a company boosts its profit per employee without increasing its capital intensity. Another advantage of profit per employee is that it requires no adjustment for accounting conventions. ROIC is another. Opportunities to improve ROIC to an equal extent are hardly as plentiful. on talented people—those who. companies can easily benchmark it against the comparable results of competitors and other companies. Fortunately. output-based measure. Profit per employee therefore focuses companies on intangible-intensive value propositions and.The new metrics of corporate performance: Profit per employee 31 (that is. Why? Because Wal-Mart Stores. just as raising ROIC above the cost of capital would. Together. which is largely a sanity check. where intangible assets are a rich source of value. But if they generate profit greater than the cost of the capital used to support their work. the opportunities to increase profit per employee are unprecedented in a digital economy. can produce valuable intangibles. The difference is that viewing profit per employee as the primary metric puts the emphasis on the return on talent. including the cost of capital. with some investment. unless management adds an offsetting number of workers who produce a higher profit per employee. as well as its return on capital. in turn. The hallmark of financial performance in today’s digital age is an expanded ability to earn “rents” from intangibles. 3 Economists define rent as the profit earned after a company pays for all of the factor costs of production (labor. shedding them actually reduces the creation of wealth. raw materials. Number of employees One way to improve a company’s profit per employee is simply to shed low-profit employees. management will increase its rents. these three metrics squarely highlight—and drive—market caps. The Walton family. profit per employee is a pretty good proxy for the return on intangibles. And since it is based on accounting conventions.

The grid also shows how total employment can serve as a crude proxy for the internal complexity of any organization. thousands US and foreign companies by American depositary receipts.200 each. profit per employee becomes a proxy for how well a company manages that complexity. We can observe this dynamic on a simple grid that illustrates the source of the profit earned by a company and a competitor (Exhibit 2). pp. A company can. 98–109. see Lowell L. by increasing the Number 3. From this vantage point. and Leigh M. Real wealth creation therefore comes from increasing either a company’s profit per employee (without offsetting reductions in the number of employees or offsetting increases in capital intensity) or the number of employees who earn that level of profit—or both. of course. top companies with negative net incomes.533 Employees. 5 For more information on talent markets. Bryan. “Making a market in knowledge. pp. excludes outliers and . number of employees earning such profits. streamline its organization and use tools such as formal networks. Bryan. particularly when it is compared with companies in similar industries that have a comparable employment mix. “Making Exhibit 2 of 4 a market in talent. Joyce. by market capitalization in . Royal Bank of Scotland. see Lowell L. 100–11. 2006 Number 2. For more information on knowledge markets. Weiss.1 2002–04 (average) 250 Profit per employee.7 million people. or both.” The McKinsey Quarterly. who generated an average profit of $6. without incurring offsetting diseconomies. 2004 Glance: Companies can create wealth either by increasing profit per employee. Talent as profit driver Drivers of profit for 30 largest companies. $ billion 18 12 6 3 0 0 50 100 150 200 250 300 350 400 450 500 1. Claudia I. $ thousand NTT DoCoMo Exxon Mobil 200 Microsoft Profits driven by productivity 150 100 50 Chevron AIG Royal Dutch/Shell Dell GSK BP Bank of America Pfizer RBS2 Wells Fargo Berkshire Hathaway Verizon Citigroup Toyota GE IBM Profits driven by size of workforce Wal-Mart Intel Total profit. and knowledge marketplaces5 Q 1 2007 4 Profit per employee In 2004 Wal-Mart employed 1. talent marketplaces.32 The McKinsey Quarterly 2007 Number 1 not only hires large numbers of employees who generate a relatively low average profit 4 but also uses a business model that enables it to handle the complexity involved in managing huge numbers of employees.” The McKinsey Quarterly.

pp. but they are largely fungible labor and usually don’t undertake the intensive intangible work that drives a company’s profits. whose assets are mostly financial. (Calculating economic profit per employee often requires internal company data. particularly for financial institutions. Of course. it demands a return and for this reason must be used carefully. This is exactly why companies choose to rely on contractual labor. is to maximize a company’s market cap. all but 5 were in the top 30 by market 6 According to some observers. Number 6. in which the authors propose using economic profit per employee to gauge the true performance of “people businesses. I disagree. among other things. while capital is relatively inexpensive and readily available. if it removes any unproductive complexity. can be a useful internal metric. Profit per employee is a more practical metric. depreciation schedules. To the extent that it does so. 80–90. meanwhile. profit per employee is the better metric because it not only represents the scarcest resource but also reflects profit after the expensing of necessary investments. for it is subject to somewhat arbitrary accounting conventions that involve goodwill. These workers may depend on the company for work. . Capital investment. the amount of capital it requires will be a derivative of the capital its employees need for their work. is—surprisingly—relatively ambiguous. rather than an independent aspiration. Returns on capital A company can also improve its profit per employee by substituting capital for labor costs. “The surprising economics of a ‘people business. its profit per employee should increase. is depreciated or amortized. on the other hand. and the way companies expense stock options. perhaps the most important single measure of size and economic relevance. 7 See Felix Barber and Rainer Strack. even in the absence of profitable new value propositions. June 2005.The new metrics of corporate performance: Profit per employee 33 to mobilize intangibles throughout the enterprise.7 Maximizing market capitalization The goal of these efforts to reorient financial-performance metrics around talent. Executives should therefore look at ROIC mainly as a sanity check.) A related concept. the many temporary contractual workers that certain large companies use should be counted as employees. economic contribution per employee. of the top 30 companies by net income from 2002 to 2004. But if the company uses total employment to drive its growth aspirations. Invested capital is not only a meaningless concept for such companies but also requires them to make some heroic assumptions. So long as the return exceeds the cost. Volume 83. of course. as it can be taken directly from accounting statements and allows for straightforward comparisons of performance across companies.” Economic profit subtracts the cost of capital from profit per employee.6 Book capital. Using the total number of employees as a metric also allows companies to avoid subjective accounting judgments. The market cap directly affects a company’s ability to control its own strategic destiny and is highly correlated with its total net income. Calculations of a company’s ROIC have their own limitations.’” Harvard Business Review.

. return on talent is a more powerful model in a competitive environment where the intangible assets that talented employees create provide the greater part of new wealth. 2002–04 (average) 70 60 Return on equity. Source: Global Vantage. Exxon Mobil $ billion Berkshire Hathaway GE Toyota Citigroup 10 0 Intel Toyota Citigroup Berkshire Hathaway NTT DoCoMo BP Chevron AIG Verizon Bank of Wells Fargo America 300 200 100 50 120 0 20 40 60 80 100 120 20 40 60 80 100 Book value. using profit per employee and the total number of employees. $ billion Overlap AT&T Freddie Mac ING PetroChina AIG Bank of America Berkshire Hathaway BP Chevron Citigroup Exxon Mobil GE GlaxoSmithKline IBM Intel Microsoft Book value. Doing so displays the total market cap as a function of the latter and the market cap per employee (Exhibit 4). % of top by income are also in top by market capitalization. McKinsey analysis value. The point is that although the two metrics produce similar results. A company can expose this correlation by displaying its net income as the return on book equity multiplied by book equity and then comparing that relationship with its total market cap disaggregated (in a strategiccontrol map) into its market-to-book ratio multiplied by book equity (Exhibit 3). The company can also see this same correlation by disaggregating net income.Q1 2007 Profit per employee Exhibit 3 of 4 The McKinsey Quarterly 2007 Number 1 34 Glance: The correlation of market capitalization to net income can be viewed in relation to returns on invested capital (ROIC) . Net income and market cap can therefore be regarded as functions of the return on either capital or talent. $ billion Pfizer Royal Dutch/Shell Toyota Motor Wal-Mart Stores Wells Fargo NTT DoCoMo Verizon Communications Vodafone 1 US and foreign companies by American depositary receipts. . The return-on-capital lens Net income and market capitalization shown as returns on invested capital (ROIC). % 50 40 30 ING Freddie Mac 20 Wells Fargo IBM Wal-Mart Bank of America Pfizer BP GE PetroChina Chevron AIG AT&T Microsoft GSK Top 301 by net income 10 Ratio of market value to book value 9 8 7 6 5 4 3 2 1 0 0 Intel GSK Top 301 by market cap Rest of top 150 by net income Rest of top 150 by market cap Net income. $ billion Exxon Mobil Microsoft 30 20 10 5 IBM Wal-Mart Pfizer Market value. .

$ billion 300 200 50 350 GSK 0 50 100 150 200 250 300 350 0 0 50 100 150 Berkshire Hathaway Toyota 200 250 300 Employees. thousand Overlap AT&T Freddie Mac ING PetroChina AIG Bank of America Berkshire Hathaway BP Chevron Citigroup Exxon Mobil GE GlaxoSmithKline IBM Intel Microsoft Employees. Copyright © 2007 McKinsey & Company. Source: Global Vantage. the mix of employees. not capital. $ million Top 301 by net income 6 5 4 3 Microsoft Top 301 by market capitalization Rest of top 150 by market capitalization NTT DoCoMo 150 Microsoft Chevron Exxon Mobil Vodafone AIG Pfizer BP Wells Fargo Bank of America GE Citigroup IBM Verizon 100 Intel AIG 50 0 BP Royal Dutch/Shell Bank of America GSK Citigroup Pfizer Wells Fargo GE ING Toyota AT&T IBM Berkshire Hathaway Net income. $ thousand Rest of top 150 by net income Market capitalization per employee.Q 1 2007 Profit per employee Exhibit 4 of 4 The new metrics of corporate performance: Profit per employee Glance: . .  Q Lowell Bryan is a director in McKinsey’s New York office. % of top by income are also in top by net income. . Yet it is thinking-intensive talent. spring 2007. 2002–04 (average) 250 Exxon Mobil 200 Net income. McKinsey analysis Today’s annual reports are filled with information about how companies use capital but offer little about the number of employees. thousand Pfizer Royal Dutch/Shell Toyota Motor Wal-Mart Stores Wells Fargo Coca-Cola NTT DoCoMo Verizon Communications Vodafone 1 US and foreign companies by American depositary receipts. Mobilizing Minds: Creating Wealth from  Talent in the 21st-Century Organization. This article is adapted from his forthcoming book. . or the different kinds of employees (beyond a simple expense item on compensation and benefits). McGraw-Hill. that now drives the creation of wealth and thus deserves to be measured more precisely by strategically minded executives. 35 The return-on-talent lens Income and market capitalization shown as returns on talent. or in relation to return on talent. $ billion 18 12 6 3 Royal Dutch/Shell Intel 2 Chevron 1 Market value. All rights reserved.

36 The McKinsey Quarterly 2006 Number 3 .

2005 Number 4. the need to manage collaboration is growing. Studio/lab . companies must improve their management of the networks where it typically occurs. 40–51. Johnson. “Competitive advantage from better interactions. most companies are still in the dark about how to manage it.” The McKinsey Quarterly. 2006 Number 2. Roger D. customers. 52–63. For more about the latter. 2 For more on tacit interactions. Robert L.2 Nearly 80 percent of the senior 1 This article focuses on intracompany interactions and collaborations rather than those extending beyond the boundaries of the enterprise. Linear. 2006 Number 2. and Leigh M. pp.” The McKinsey Quarterly. For more on the evolution of large corporations. Bradford C. which involve the exchange of information. As “tacit” interactions replace more routine economic activity and the scale and complexity of many corporations creep upward. the making of judgments. Weiss Although collaboration is at the heart of modern business processes. and James M. see Scott C. see Lowell L. hierarchical. globalization. 46–59.1 This blind spot has become problematic.” The McKinsey Quarterly. see John Seely Brown and John Hagel III. Bryan and Michele Zanini. business process reengineering. and the increasing specialization of knowledge-based work have made collaboration within and among organizations more important than ever. Martin. and suppliers. pp. processbased tools such as activity-based costing. pp. But they do a poor job of shedding light on the largely invisible networks that help employees get things done across functional. Beardsley. Falling communications costs.Mapping the value of employee collaboration 37 Mapping the value of employee collaboration As collaboration within and among organizations becomes increasingly important. “Creation nets: Getting the most from open innovation. and total quality management have long been effective at measuring and improving the efficiency of people and organizations in accomplishing individual tasks. Cross. and business unit boundaries. and a need to draw on multifaceted forms of knowledge in exchanges with coworkers. “Strategy in an era of global giants. Manyika.

functional. to analyze the economic costs and benefits that key interactions create. and the exchange of resources rather than on the value those interactions create. and Rob Cross and Andrew Parker. the costs such interactions impose. Burt. Cambridge. a few leaders (such as BP. . Number 4. MA: Harvard Business School Press.38 The McKinsey Quarterly 2006 Number 3 Article at a glance Falling communications costs. executives surveyed in a 2005 study said that effective coordination across product. Achieving Success through Social Capital: Tapping Hidden Resources in Your Personal and Business Networks. MA: Harvard University Press. pp. 1995.com. 104–11. HP. 4 See Thomas J. Baker. And it’s often unclear whether efforts to enhance networks promote productive collaboration or just consume money and time.3 Many companies have responded by spending heavily on collaboration software. redefine roles. Technology. though. Sophisticated network analysis approaches have emerged from the academic world during the past two decades. Cambridge. 1984. and the increasing specialization of knowledge-based work are making collaboration within and among organizations increasingly important. and to identify value-creating interventions. and geographic lines was crucial for growth. What companies need in a collaborative age is the ability to map and analyze the value created (or destroyed) deep within employee networks. P&G . and allocate financial. In hopes of disseminating best practices and sharing expertise. work flows. at best fails to deal with the underlying problem and at worst becomes a source of information overload that undermines effective collaboration.4 To make these tools more useful. July 1993.” mckinseyquarterly. Yet few companies understand or know how to manage the intracompany networks in which collaboration typically occurs. IBM. Managing the Flow of Technology. Volume 71. Yet only 25 percent of the respondents described their organizations as “effective” at sharing knowledge across boundaries. 2000. employees doing similar work. Ronald S. Allen. But they have tended to focus more on individual than organizational effectiveness and on communications. even begun identifying networks of physical. 3 For the full survey results. globalization. The Hidden Power of Social Networks: Understanding How Work Really Gets Done in Organizations. “Informal networks: The company behind the chart. and Xerox) have Successful interventions help companies to reduce complexity. executives must reorient them toward the revenue and productivity benefits that collaborative interactions generate. Cambridge. July 2005. 2004. Wayne E. July 2005.” Harvard Business Review. A few leading companies are beginning to map their networks of relationships. and human resources more efficiently. David Krackhardt and Jeffrey R. MA: MIT Press. and opportunities to improve connectivity at the points that create the greatest economic value. Hanson. San Francisco: Jossey-Bass. Structural Holes: The Social Structure of Competition. see “The McKinsey Global Survey of Business Executives.

serve customers and clients more effectively. A program to improve the processes of a large global construction company boosted the efficiency of its employees. and team-building efforts. redefine roles. making valuable expertise and advice more readily available.Mapping the value of employee collaboration 39 Consider the experience of a leading biotechnology company that relied on sharing best practices among quality control engineers to help its manufacturing facilities rapidly ramp up the production of new products. Typical examples include replicating highperforming networks. and persuading donors through appeals tailored to their interests. • Individual performance. Yet some high performers followed few of these practices. A nonprofit wanted to boost its fund-raisers’ productivity. These kinds of successful interventions can help companies reduce complexity. Once executives understand the value that’s flowing across networks. and human resources more efficiently. respectively. Network analysis showed the company which engineers took part in the interactions that generated time savings and the greatest and lowest cost. they can intervene in straightforward. training. • . Organizations hoping to emulate the biotechnology company (and other pioneers in a wide range of sectors) must first map their collaborative networks and then analyze the economic benefits and costs that key interactions within those networks create. cost-justified ways. teams. As a result. and entire organizations. even after controlling for the varying attractiveness of their markets. Aggregated to reveal the economic value created through interactions across locations. and revamping performance metrics to reflect mutual accountabilities better. the company knew exactly where it made economic sense to invest in tools. these figures identified the places in the network where collaborative breakdowns inhibited the transfer of proven practices and showed how costly these breakdowns were. several low performers embraced them all. training workers to emulate the collaborative approaches of successful colleagues. managing the sales process in a defined sequence. Understanding how work really gets done Three examples will show how traditional ways of mapping processes and analyzing activities have limits when it comes to understanding the performance of individuals. Team performance. Conventional wisdom suggested targeting certain types of donors. But performance disparities remained across sales offices. and allocate financial. physical.

but she says she doesn’t know him. team. It turned out. Organizations mapping their decision-making processes might ask their employees. “With whom are you most likely to discuss a new idea?” Questions are posed bidirectionally: if Joe says he was helpful to Jane. for example. Key inputs to this analysis include fully loaded compensation figures for network participants and detailed survey results (for example. his claim is disregarded. companies can use standard software to create network maps illustrating relationships (Exhibit 1). By helping . Tenure and experience were key reasons for the high performers’ strong networks.40 The McKinsey Quarterly 2006 Number 3 • Organization-wide performance. companies must assess the time employees spend on interactions of various types. so the organization was caught in a vicious cycle: low-tenure fundraisers got stuck on the fringes of both their internal and external networks. So far. and quit before they became productive. To do so. the responses to queries such as. Network analysis can help companies in circumstances like these. the company’s linear view of the construction process—emphasizing the tasks performed by each group and the handoffs between them—shed little light on collaborative issues. and organization-wide performance issues. The real value comes when companies move from mapping interactions to quantifying the benefits and costs of collaboration. as well as the savings and sales contributions of specific collaborations. became dissatisfied. that high-performing fund-raisers not only had strong relationships with donors but also accounted for a disproportionate share (25 percent) of the connections within the fund-raising group. using existing data (such as time cards and project charge codes). Options for obtaining the necessary information include tracking e-mail. and administering short (5. so familiar. observing employees. The first step is identifying the functions or activities where connectivity seems most relevant and then mapping relationships within those priority areas. Network analysis helped the companies described earlier address their individual. the latter on technical solutions.to 20-minute) questionnaires. “How much time did working with employee X save you?” or “On how many deals in the following revenue bands did you work with employee Y?”). An engineering company experienced growing pains as international expansion made it increasingly difficult to bring together construction managers and engineers.) Unfortunately. (The former focused on cutting costs. whose objectives were frequently in conflict. “Whom do you ask for advice before making an important decision?” Others targeting innovation might ask. With the information in hand.

one shape for individual. tenure. laboratories) Larger size represents greater degree of an assigned value—eg. decisionmaking input. an analysis of one of the engineering company’s high-performing groups showed that a small number of construction managers and engineers single-handedly accounted for 35 percent of all the collaboration occurring within it. centrality Arrow Tie or link = connection between 2 nodes. This kind of collaboration dramatically enhanced the group’s ability to deliver expertise. entity. articles. and project-management processes—the construction company replicated the network orientation of high-performing offices in poorly performing ones. organizational design. role. represents an assigned value—eg. By getting at the roots of these issues—which further analysis attributed to hierarchy. in a single year. eg. geography. information transmission Thicker line indicates more frequent interactions Shows direction of interaction. network analysis revealed that a key distinction between the strongly and poorly performing offices was the percentage of collaborative time (68 percent for the former. another shape for projects. Kristina provides James with decision-making input Spatial relationships Location of nodes Nodes with highest number of connections are located in center of map. Identifying and building connectivity between specialists in other groups helped the firm to raise its construction revenue to $275 million. Finally. 41 Reading a network map Selected elements of a sample network map Project A Federic Project B Arun Anya Cho Kristina Jon Tanaka Patrick James Node = individual or entity in network Color Represents membership in or affiliation with group or category1 Line weight Shape Represents different types of nodes (eg.Q3 2006 Network mapping Exhibit 1 of 4 Mapping the value of employee collaboration Glance: A network map illustrates relationships. degree of connectedness. 50 percent for the latter) that account managers spent with customers. the nonprofit expected to increase its revenue from employees with no more than two years’ tenure by nearly 200 percent. . gender. new fund-raisers rapidly replicate the high performers’ networks. For the construction company. those on periphery have fewest connections Size Group or category such as business unit. from $80 million.

organizational hierarchies. while simultaneously acknowlburdens already-overloaded employees edging the heightened value of innovations.) In our experience this very real but usually invisible barrier to crossselling and account penetration is common in organizations. and ignorance of which colleagues have expertise. What’s more. and yields network diseconomies. which typically structures.42 The McKinsey Quarterly 2006 Number 3 Creating relational value The powerful results of identifying and replicating high-performing networks represent only a small part of the potential of network analysis. and a scarcity of expertise. collaborators. The company. It can also suggest where to replicate collaborative behavior. helps companies to identify the few critical points where improved connectivity creates economic value by cutting through business unit and functional silos. It’s also possible to promote specific interactions that help generate revenue and boost productivity. a network perspective helped the company identify which colleagues knew about one other’s expertise but didn’t draw on it. The experiences of a global technology company and a consulting firm illustrate how these issues play out in practice. . A leading technology company used network analysis during an effort to become more responsive to customers and marketplace shifts. broke out collaborative contributions by bands of revenue and learned that the most and least valuable interactions (those generating more than $2. (Exhibit 2 shows how many people said they were acquainted but saw no possibility of collaborating in a sales effort.com. Targeted action is dramatically more effective Developing a network perspective can help the 21st-century organization retain than promoting connectivity the best of its traditional organizational indiscriminately. and intangible assets. The analysis not only helped the company’s leaders find out where collaboration generated revenue but also proved useful for reframing the roles of key players in the network. incentives. See “The 21st-century organization” on A more informed network perspective mckinseyquarterly. Generating revenue A network view often uncovers “hidden” people whose contribution to cross-selling or closing deals is far greater than individually focused performance metrics might imply.000 and less than $250. personalities. for example. when to draw in valuable experts from the network’s fringe. Improving cross-selling. and how to eliminate obstacles to collaborative sales efforts—obstacles that include time.000. physical distance. respectively) invariably involved different people.000. skills.

It quickly became clear that the success of the high-performing collaborators resulted from more than just expertise or affability. it identified two crucial categories of people who weren’t recognized by its performancemanagement processes. it found that they were accommodating. . solutions Tie = individual is aware of other’s expertise but sees no opportunity to collaborate No tie = individual is not aware of other’s expertise or is aware. global sales and application services can create opportunities to generate innovative ideas. The resulting network map illustrates those potential but missed opportunities. . application services. When a global consulting firm used network tools to analyze the sales efforts of a group of roughly 80 partners. flexible. amenable to constructive criticism. . 10 partners supported collaborative efforts yielding 60 percent of this group’s revenue. hypotheses. The importance of these traits caused the company to overhaul its incentive program and to launch an effort to build collaborative skills throughout the sales network. A B Developing collaboration opportunities for . this second group. Node = individual in network Tie = connection between 2 nodes Color = individual’s department Application services Architecture services Global sales Global sales management Testing. architecture services. 43 A map of wasted opportunities Individuals were asked to identify people whose expertise they knew about but with whom they saw no opportunity to collaborate. and solutions. the top 5 accounted for 38 percent. Enhancing career paths. . and the like. . By making joint sales calls. McKinsey analysis To boost sales the company sought both to replicate the major contributors’ behavior and to help several key salespeople understand how collaboration could make them more successful.Q3 2006 Network mapping Exhibit 2 of 4 Mapping the value of employee collaboration Glance: A network perspective can identify the hidden barriers to collaboration. . enthusiastic team players. more responsive to requests. which emphasized individual revenue production. and testing during the sales cycle can create opportunities to sell packaged or bundled services. has already collaborated Other Source: Network Roundtable at the University of Virginia. sharing experiences. A completely different subset of partners made an enormous contribution to the execution of projects by helping others to save time and generate high-quality work. A• . When the company compared them with its other salespeople. and effective conflict managers. B* . for .

Sometimes. The consulting firm used this knowledge to end a long-simmering disagreement about dual career paths for partners. avoiding downtime is critical given the magnitude of its investment in fixed assets. There was no longer any question about the need to recognize the contributions of partners whose expertise or experience played a key role in winning many new clients and of those whose work improved the effectiveness and efficiency of the firm’s efforts to serve them. sharing good ideas. The contributions of these partners. to decide what. were highly concentrated: the top 10 people were responsible for 48 percent of the value generated through time savings. Examples from three very different industries illustrate the range of possibilities. Scrutinizing the time savings that relationships generate helps companies to isolate what’s working. which previously had been operating largely on faith. In other cases. the petrochemical company was particularly eager to measure their impact. physics. shareable . and helped with analyses. and outdated role definitions. So one leading petrochemical company formed more than 20 networks (ranging in size from 50 to several hundred employees) and focused on work areas where people could benefit from sharing best practices. Often. The specific issues and interventions vary considerably across industries. and production. an ineffective allocation of the right to make decisions. In the petrochemical business. geology. and experience large disparities in the effectiveness of collaboration within and across units. manage talent blindly. contributed expertise on the problems of clients. Network analysis showed that the effort. a network perspective isolates unseen collaborative inefficiencies resulting from poor job design. Solving problems quickly often requires collaboration across disciplines such as drilling.44 The McKinsey Quarterly 2006 Number 3 example. or organizational designs. too. process steps. was generating substantial. and how to invest in additional connectivity. visited them. But some general themes emerge. and to redefine roles and staffing levels. and the top 5 for 32 percent. Having taken the unusual step of engineering these networks. where. network analysis shows them that they can generate savings by facilitating the transfer of advice and information from colleagues. companies that operate without a network perspective allocate resources inefficiently. Validating the effectiveness of networks. Boosting productivity Most companies—even high-performing ones—can find opportunities to boost their collaborative productivity.

Engineers used their network to identify an expert who had no relationship with the well but did have critical knowledge that helped them fix it in two days instead of the expected four. One 60-person network alone contributed $5 million in savings. Network analysis thus allowed the company to validate the efficacy of its networks. Training sessions allowed the leaders of the most successful networks to share what they had learned. network analysis showed the company how to focus its collaborative efforts within and across groups.Mapping the value of employee collaboration 45 productivity benefits. for example. yielding savings that should ultimately dwarf the initial benefits. Network analysis. This effort helped the company to make better decisions about how much to invest in its collaborative relationships. the company realized that connecting all of these people with each other and with just one person on the network’s fringe would yield $140. Improving the allocation of resources. After recognizing that a set of key brokers occupied central positions in the network.000 a year in savings within business units and $865. A knowledge-sharing team interviewed the leaders of networks to collect and disseminate best practices.000 across them. The company then began taking steps to pass lessons among networks. The keys to success included forming networks carefully around focused topic areas closely related to the way work was actually done. since aggregating results by . The success of a pilot effort led the company to replicate it widely. allowed executives to prioritize the company’s investment in collaboration by helping them to model the financial benefits of improving the network and to weigh the anticipated returns against the costs. whether to focus on collaboration within or across groups. In addition. A global financial-services organization mapped and calculated the time its key employees saved by sharing information and resources with their colleagues. A typical story involves engineers and an out-of-commission oil well. and continually tracking and measuring success to encourage participation and inform decision making about when (and when not) to finance incremental network improvements. Facilitating these interactions would be far less costly than buying the group another unused collaborative tool or holding an off-site meeting. for instance. and what role collaboration should play in its human-resources (HR) strategy. giving network members the leadership and training for success (rather than merely dumping collaborative tools on them).

000–$1.800 375 150 1. it began focusing more heavily on collaboration among them to reduce the number of redundant efforts and to promote the exchange of expertise in project-management tools.597 1.999 <$1.750 3.680 1.470 1. One division’s global network of technical project managers generated monthly savings of 3.971 Other Application architect Business unit architect Data architect Infrastructure architect Systems analyst Project manager Other 17 5 6 16 2 5 7 2.121 2.383 hours (which translated into roughly $215.843 476 30 225 56 0 1.275 3.680 364 Project manager 626 210 975 1.225 5.714 1.057 Source: Network Roundtable at the University of Virginia.069 1.000 Number of employees Average cost of interaction time per employee per month.46 The McKinsey Quarterly 2006 Number 3 business units.000 $1.406 1.103 660 650 1.588 975 3.180 1.000). and in some cases elevating (or demoting) the role of the Exhibit 3 of 4 network’s central (or peripheral)be converted the an average monthly cost of plans.125 1. and hierarchical levels showed executives where to direct the relevant investments. McKinsey analysis . projects.460 950 656 0 210 471 715 1. roles. The company responded by financially rewarding the key collaborators (many of whom had previously been frustrated by the failure to recognize their Q3 2006 effectiveness). methodologies. Another benefit to the company was an improved ability to measure and manage talent.500 1.050 193 Infrastructure Systems architect analyst 1. Executives were surprised to learn how much relational value was created by people they hadn’t recognized as central contributors and how little by others they had regarded as more influential. Glance: Employees’ interaction times can figures in into company’s succession collaboration per employee.600 1. $ Application architect Business unit Data architect architect 1. When the financial institution realized that about 70 percent of these savings resulted from collaboration within divisions.126 3.450 2.013 1. redefining roles and performance metrics to promote Network mapping collaboration. and technologies. Uncovering the value Labor cost of time spent in or preparing for interactions with others (example of major US utility) >$2.560 2.110 2.

025 11.764 9.800 5.556 7.780 11. the CIO was able to identify outliers: a small number of employees who actually were significantly less effective than the rest (Exhibit 4). $ Internal External 2. Yet these demands had never before influenced the hiring.470 2. Eliminating inefficiencies.126 2.025 5.300 1.000 48.050 12.900 68.575 5.425 10. say.Mapping the value of employee collaboration 47 Managers were also encouraged to coach a few people who didn’t know how to build networks. the company focused personalized coaching efforts on collaborative issues that were unique to each of the unexpected low performers. for instance.460 2.720 63.550 162. Consequently. The chief information officer (CIO) of a major utility calculated collaboration’s average monthly cost per employee—both within roles (for example. data architects and other members of the IT staff)—throughout the IT organization (Exhibit 3).320 8.625 4.250 6.600 60. Total monthly interaction labor cost.057 8. Network analysis also helped to clarify roles by showing that the utility’s data architects and project managers spent more than half of their time collaborating.163 132.600 10. By comparing the colleagues who were generally considered effective communicators. This analysis helped the company to root out collaborative inefficiencies.588 0 5. collaboration among all data architects) and between them (between.613 .638 7.

therefore. however. It knew that it could avoid certain problems by involving its infrastructure architects (who design and maintain major applications supporting vital business functions) in key decisions at the right time. showed that they interacted less than most other employees of the IT organization. The first step for the CIO. the company used network analysis to set appropriate staff levels. was breaking down the barriers that inhibited collaboration. or performance evaluations of such workers. Addressing these issues not only improved the execution of projects and the company’s ability to price them but also made internal customers more satisfied. Network analysis. . Who needs help? + Analysis of interaction times and usefulness for a small number of employees can lead to improved organizational performance. As a result. McKinsey analysis staffing. Traditional budgeting and cost allocation processes might have suggested hiring more infrastructure architects. Finally. their job functions and role descriptions were incomplete. Average number of times individual was cited as a source of useful information – Average interaction time + Application architecture Infrastructure architecture Source: Network Roundtable at the University of Virginia. But often this didn’t happen.Q3 2006 Network mapping The McKinsey Quarterly 2006 Number 3 48 Exhibit 4 of 4 Glance: Network analysis can identify those employees who would benefit from personalized coaching. they lacked collaborative skills. and the internal cost allocations used to establish transfer prices for IT projects underestimated the total cost of certain programs and thereby distorted the company’s resource allocation.

office. is an associate professor and director of the Network Roundtable at the University of Virginia’s McIntire School of Commerce. All rights reserved. an adviser to McKinsey. Q Rob Cross. . DC. A network perspective gives executives the information they need to foster collaboration at the points where it delivers an economic return. But when companies just promote collaboration indiscriminately. Leigh Weiss is an associate principal in McKinsey’s Washington.Mapping the value of employee collaboration 49 Collaboration is an increasingly vital feature of business life. they create bottlenecks and diminish their organizational effectiveness. where Roger Martin is an assistant professor. Copyright © 2006 McKinsey & Company.

.

These networks organize and reorganize themselves and extend their reach via cell phones. Lowell L. and other accessories of the digital age.Harnessing the power of informal employee networks 51 Harnessing the power of informal employee networks Formalizing a company’s ad hoc peer groups can spur collaboration and unlock value. Eric Matson. Most large corporations have dozens if not hundreds of informal networks. or functional councils—or have no title at all. Bryan. leads people to share ideas and work together even when no one requires them to do so. Research scientists working in related fields. they may build networks that can range in size from fewer than a dozen colleagues and acquaintances to hundreds. they can mobilize talent and knowledge across the enterprise. As networks widen and deepen. “Strategy in an era of global giants.1 1 Jerry Stenback Lowell L. have increased in scale and scope and boast superior performance. or investment bankers serving clients in the same industry frequently create informal—and often socially based—networks to collaborate. Weiss In any professional setting. networks flourish spontaneously: human nature. such as ExxonMobil and GE . including mutual self-interest. for example. Bryan and Michele Zanini. and Leigh M.” The McKinsey Quarterly. community Web sites. . which go by the name of peer groups. Blackberries. As they connect around shared interests and knowledge. They also help to explain why some intangible-rich companies. pp. 46–59. communities of practice. 2005 Number 4.

52 The McKinsey Quarterly 2007 Number 4 As we studied these social and informal networks. as manifested in their organizational charts. they elude control. cross the lines of geography. they can enable leaders to apply the energy of diverse groups of professionals and managers to realize collective aspirations. he acts as the hub in an informal network because he has knowledge that others find valuable. formalized networks serve as an organizing structure for collaborative professional work. . Without him. these ad hoc communities clearly have shortcomings: they can increase complexity and confusion. both in Exploration. They can replace cumbersome and outdated matrix structures. is connected in the informal network to only two people. we made a surprising discovery: how much information and knowledge flows through them and how little through official hierarchical and matrix structures. at a time when the ability to create value increasingly depends on the ideas and intangibles of talented workers. But companies can design and manage new formal structures that boost the value of networks and promote effective horizontal networking across the vertical silos of so many enterprises today. slipping through the back channels. Valuable as they are. and help build more and better personal relationships among the members of a community. the production group would be cut off from the rest of the organization. assigning “leaders” to focus discussion. both within and outside of companies. His boss Jones. and functions—where the action is— and even through the thick silo walls of vertically oriented organizations. and since they typically fly under management’s radar. By building network infrastructures. facilitate the creation and sharing of proprietary information and knowledge. They don’t necessarily fit into the organizational chart. don’t explain how most of their real day-to-day work gets done. increase the value of collaboration by reducing the search and coordination costs of connecting parties who have related knowledge and interests. Most important. The long and short of informal networks Personal social networks. the unit’s senior vice president. that corporate leaders don’t do far more to harness the power of informal networks. Although he sits relatively far down in the formal company structure. As we used surveys and e-mail analysis to map the way employees actually exchange information and knowledge. customer groups. products. sprawling corporations. This is increasingly typical in today’s large. we concluded that the formal structures of companies. and combining hierarchy and collaboration to bring together natural professional communities. Consider the case of an energy company staffer we’ll call Cole (Exhibit 1). So it’s unfortunate. Informal networks.

The Hidden Power of Social Networks. Part of the problem is that informal networks. As tens of thousands of individuals search for knowledge and productive personal relationships in social networks. essentially rely on serendipity. they generate much of the overload of e-mails. But though informal networks help many of the largest companies capture wealth. both in Exploration • Cole is at center of information exchange network despite being relatively low in organization chart • Without Cole.000 people across a number of business units around the world to gauge the frequency and quality of the interactions that generated decisions about business planning and other processes. so their effectiveness varies considerably. Boston: HBS Press. At one large company. voice mails. and meetings that make today’s large companies more complex and inefficient. . they also cause severe headaches.Harnessing the power of informal employee networks 53 A revealing map Formal structure Senior vice president (SVP) Jones Informal structure Network mapping reveals informal structure— who interacts with whom to get work done. is connected to only 2 people. O’Brien Exploration Williams Drilling Taylor Production Stock Stock G&G1 Petrophysical Sen Cross Andrews Moore Miller Shapiro Production Reservoir Cohen Cole Kelly Smith Hughes Paine Jones Andrews Cohen Smith Hughes Ramirez Bell Cole Hussain Kelly Cohen Smith Shapiro Miller Williams Jones. SVP. This analysis suggested that redesigning the processes to eliminate or reduce the noncore interactions could result in savings of tens of millions of dollars and shorten the time to make the decisions. Source: Rob Cross and Andrew Parker. as ad hoc structures. Nearly half of the interactions were not central to making decisions. production group would be cut off from rest of organization • Cross Hussain Ramirez Moore Bell Taylor Sen Geological and geophysical. we conducted a network analysis of more than 1.

What’s more. Since formal networks stimulate interactions that the organization sponsors and encourages. they can be managed. informal networks can make dysfunctional organizations even more so by adding complexity. At worst. The greatest limitation of these ad hoc arrangements is that they can’t be managed. companies typically underinvest in the capabilities needed to make networks function effectively and efficiently. who serve as hubs to connect participants.54 The McKinsey Quarterly 2007 Number 4 In large companies a number of informal networks may form on related topics but never integrate. and simple neglect can keep natural networks from flourishing. An informal network often has crucial members. ranging in size from 50 to several hundred people. recently designed more than 20 formal networks. because the networks could minimize downtime in these areas. People with valuable knowledge or skills may not join the most appropriate network. They may not be visible to management. Formal networks The specific objective of designing and establishing formal networks is to increase the value and lower the costs of collaboration among professionals. corporate politics. such as Cole. In one case the company measured the impact of networks The missing link Informal network with linchpin Informal network if linchpin is removed Linchpin Removing the linchpin leaves the network fragmented . or leave the company (Exhibit 2). or increasing the intensity of corporate politics. muddling roles. belong to other informal networks. act as gatekeepers. A leading petrochemical company. but such members can hobble or even undermine the network if they become overloaded. for example. horde knowledge to gain power. This was critical. Unintended barriers. and even when they are it’s hard for corporations to take full advantage of them. or fail to discover that a network exists. to focus on specific work areas so that employees could share best practices.

made technology investments. Martin. Roger D. Hierarchical direction comes from two different sources. “Mapping the value of employee collaboration. the chief financial officer of a business unit. and sponsored a knowledge-sharing team that collected and disseminated best practices. geography. pp. These networks succeeded because the company formed them around focused topics closely related to the way work was carried out at the wells. for example. or products. geography. A matrix organizes work through authority and is therefore principally based on management hierarchy. Matrix structures made sense because they were used sparingly and therefore didn’t greatly confuse the hierarchical vertical line structures responsible for much of the success of large 20th-century companies.Harnessing the power of informal employee networks 55 on engineers at an oil well. or functions. Weiss. 2 Robert L. customers. they can easily be mistaken for matrix organizational entities. Cross. matrix management became popular because no matter how well organized their line structures may have been by functions. might report both to its line manager and to the corporate finance chief. customers. In hierarchically organized companies. A formal network organizes work through mutual self-interest and is therefore principally based on collaboration. who used them to find experts with the knowledge needed to get the well back into production in two days rather than the anticipated four. .2 Matrix decoded Because formal networks cross line structures. managers and professionals have two or more bosses who have authority over their work. such as products. they felt they needed secondary axes of management to stretch horizontally across the enterprise and thus make it possible to integrate other work activities. particularly because it enabled limited collaboration to take place within companies as they became increasingly aware that hierarchical managers sometimes had to coordinate their activities. Management also appointed the networks’ leaders. In classic matrix organizations. Matrix management worked reasonably well from its advent in the 1960s until the late 1980s. But the differences are significant and start with the organizing principles that underlie each (Exhibit 3). and the person in the middle of the matrix must resolve any conflicts.” The McKinsey Quarterly. 2006 Number 3. These matrixes represent different axes of management. gave the members training. and Leigh M. 28–41. carefully identified the members of each network across the geographically dispersed company.

no matter how talented. Rather than forcing employees to go up and down hierarchical chains of command. A new model Companies need to build infrastructures to create and support formal networks. customer. This increased the need for more interactions. Furthermore. Individual members of networks don’t have to find one another through serendipity. eg. Such well-designed and well-supported formal networks remove bottlenecks and take much of the effort out of networking. collaborative Matrix structure Organizing principle Organizes work through authority Formal network Organizes work through mutual self-interest Collaboration and leadership One manager for each person. and e-mail exchanges. matrixed roles have proliferated. Professionals who want to work horizontally across an organization currently find themselves forced to search though poorly connected organizational silos for the knowledge and collaborators they need.Q <_> <_> <Article slug> The McKinsey Quarterly 2007 Number 4 56 Exhibit <_> of <_> Glance: Hierarchical vs. and decision making now swamps the time available for matrix managers to coordinate the work personally. As the number of professionals needing to direct much of their own work has risen. phone calls. one formal network leader for each networked community Simple hierarchical structure Streamlined interactions Streamlined decision making • Easier and faster to find knowledge using network’s resources • Single boss reduces number of conflicts that need to be resolved • • • Mode of influence Number of bosses Management hierarchy Two or more representing different axes of management. product. . so more work from different perspectives had to be integrated. companies were forced to adapt to an increasingly fluid and uncertain competitive environment. geography. The symptoms include endless meetings. In many companies these matrix and other hybrid organizations have become dysfunctional. function Proliferation of matrixed roles and complex hierarchical structure • Excessive interactions • Decision-making bottlenecks • Difficulty finding knowledge • Conflicts arising from different bosses • Implications But when globalization took hold. formal networks create pathways for the natural exchange of information and knowledge. as well as confused accountability for results. to deal with them in a matrix structure. the amount of knowledge and information that must be absorbed and exchanged often exceeds the personal capacities of any individual.

Harnessing the power of informal employee networks 57 Consider the US retail unit of a financial institution we’ll call Global Bank. Queens. In the new model (Exhibit 4) regional marketing managers still report to the branch network’s regional managers but no longer have a second boss in marketing. often didn’t know whose authority to recognize and had little opportunity to share best practices with other marketing professionals across the organization. parallel structures • Marketing managers at every branch have same role description Marketing . The leader can’t give them orders but can encourage them to work for the network’s benefit (for example. brand awareness) Network owners Network is ‘owned’ by US retail-marketing executive (senior manager) who reports to EVP of Global Bank (top executive) • • CEO Executive vice president (EVP). forced to report to a regional as well as a functional manager. a branch-based formal network leader for marketing facilitates their interactions with other marketing professionals. Staten Island Manhattan. Global Bank US retail-marketing leader Flexible membership 30 core members. managers from credit card division distributing cards through branches) • East Coast retail division South division Midwest division Credit card division Other divisions 4 other frontline managers Northern NJ. including people working on related initiatives (eg. Long Island Bronx. which was organized as a matrix. Instead. including all those reporting to frontline branch executives • Extended members. Brooklyn Marketing Marketing • Network coordinators US branch-based leader facilitates interactions Marketing Marketing network US branch-based marketing leader Standard roles. Its retail-marketing managers. by asking them to help develop new Formalizing a network Example of global financial institution’s US retail operation Clear objectives Marketing network is expected to show measurable results in key metrics (eg.

the network must have standards and protocols that describe how it should work.” The owner of a network doesn’t oversee its work or personally manage or evaluate the performance of individual members (except for direct reports) but may provide input to the evaluation process. branch signage or promotional materials).58 The McKinsey Quarterly 2007 Number 4 promotional materials or to find better ways of using local-advertising budgets). such as organizing the infrastructure supporting it. Formalizing a network To formalize a network. The company can facilitate the development of a formal network by providing incentives for participating in it (such as community building off-sites) and for contributing to it (such as recognition for people who contribute distinctive knowledge). the US retail-marketing executive. Furthermore. its territory must be defined—informal networks sometimes make overlapping claims on the same activities. such as knowledge that is valuable for all members. Despite this limited hierarchical authority. is a senior manager who owns the formal network and mobilizes marketing talent for special projects. as well as the expectations of corporate leaders. The company. a formal network’s leader should be held accountable (together with line management) for the network’s . developing an agenda for maintaining its knowledge domain. identifies candidates for marketing positions. the network owner—and make that leader responsible for investing in the network to build its collective capabilities. The responsibilities of the formal leader of a network are primarily limited to its activities. The marketing leader’s boss. evaluates the owner by taking into account qualitative assessments of how well this formal network operates as compared with others. and qualifying members as professionally competent. oversees the maintenance of the domain’s knowledge (for instance. and stimulates its creation. the company must define who will lead it—that is. For a formal network to work effectively. Another difference between a formal network and a matrix is that the network owner isn’t a boss but rather a “servant leader. which expects the network to show measurable results in key metrics (such as brand awareness). stimulate the creation of knowledge. Network owners facilitate interactions between members. and help members do their jobs more effectively and efficiently. maintain the network’s knowledge domain. holding conferences. building a training program.

which have long used formal networks called practices. training for members. and activities such as conferences to build a social community. knowledge documents produced. Much of the leader’s impact comes from controlling the investments and activities that make the members individually—and the network collectively— more effective. companies should design formal networks to extend the reach of professional work throughout the organization but not to interfere with its hierarchical decision-making processes. it is always possible to tell the difference between talented and average leaders. . and the ability to find appropriate partners for dialogue quickly. and much from the ability to inspire and persuade. While the responsibilities might be the same. After all. Connecting members to the network In the model we propose. The idea is to achieve this extended reach by adding value for the networks’ members.Harnessing the power of informal employee networks 59 performance. e-mail volumes. Management can also track and test the effectiveness of a network by assessing the satisfaction of its members. such as the level of participation in conferences. exciting practices than their average counterparts do. the leader has great ability to help its members improve their performance and in this way can shape the organization. and “networkpedias”. not by exercising authority through hierarchical leaders. standard measures from network analysis (for example. It is therefore entirely appropriate to hold the leader of a formal network accountable for its performance. the effectiveness of responses to inquiries. To undertake the appropriate roles. both human and technological. even if the leader has no direct authority over individual members. talented ones create far more vibrant. internal blogs. codified knowledge in forms such as documents. to support network interactions. which give the leader the muscle to offer the members added value. density. In professional firms. and downloads. the number of steps it takes for any person in the network to reach anyone else). a formal network’s leader should have a discrete budget to finance network investments. Companies can evaluate the leader’s performance by using some quantitative measures. These investments might include infrastructure.

. its leader. and the money invested in it. and members could bring knowledge gained there into other communities. talented workers The role of networks in organizational change would gain the ability Just-in-time strategy for a turbulent world to integrate knowledge and access to talent Better strategy through organizational design across a number of communities. of its effectiveness in realizing its mission. members would join networks that had more value to them than the opportunity costs of their time and would leave when those networks no longer had that much value. The limits of network participation are largely a function of time and interest. In other words each member would build a personal social network within the formal networks. In this way such networks can enable large companies to overcome the problems of very large numbers by creating small. A person in the retail-banking community could also be a member of a branding community. Related articles on mckinseyquarterly. The number of networks employees participated in would be up to them.60 The McKinsey Quarterly 2007 Number 4 But the real measure of the network’s success would be qualitative assessments. By participating in more than one network at The 21st-century organization a time.com The number of formal networks a company could create is limited only by how much management chooses to invest in them. such as wholesale and retail banking. made by members and senior leaders. These assessments might come in the form of stories or case studies illustrating improvements in professional productivity. Providing structure to professional work Just as formal hierarchical structures define management roles. Their number and size could vary with how well each of them serves its members— effective networks would grow in membership and interactions. for whom participation would be a job requirement. focused communities of interest integrated within larger. In this way formal networks regulate themselves. ineffective ones would lose both. more wide-ranging communities—for instance. depending on that member’s roles and professional interests. for example. Rapid growth proves the value of a network. unless they were core members. formal network structures define collaborative professional ones. subcommunities focused on different aspects of financial services.

and functional lines but also in conjunction with integrative crosscutting themes. Formal network structures can mobilize employees to generate value by propagating knowledge and its creators all over the enterprise. Broad networks (in fields such as finance or IT) might have thousands of members. for instance). a Turkish interest group) might have only a few dozen. A company with 100. geography. Formal networked communities could form around not just customer groups. Q Lowell Bryan is a director in McKinsey’s New York office. such as risk management and global forces (nanotechnology and changing demographics. could have 2.Harnessing the power of informal employee networks 61 Today’s mega-institutions have room for thousands of formal networks. specialized ones (say. . for example. formal networks let employees pull these necessities toward them. All rights reserved.000 professional and managerial employees.000 networks with 100 people apiece if each professional and manager was a member of just 2 networks. Rather than pushing knowledge and talent through a hierarchical matrix. Copyright © 2007 McKinsey & Company. Eric Matson is a consultant and Leigh Weiss is an associate principal in the Boston office. products.

62 The McKinsey Quarterly 2005 Number 4 Gérard DuBois .

This kind of work is undertaken by.The next revolution in interactions 63 The next revolution in interactions Successful efforts to exploit the growing importance of complex interactions could well generate durable competitive advantages. Yee Like vinyl records and Volkswagen Beetles. companies have boosted their labor productivity by reengineering. salespeople. sustainable competitive advantages are back in style—or will be as companies turn their attention to making their most talented. So does the economics . and Lareina A. and suppliers and make complex decisions based on knowledge. and instinct. Such employees interact with other employees. the way they organize for business changes. Bradford C. highly paid workers more productive. As businesses come to have more problem solvers and fewer doers in their ranks. for their rivals will find these improvements much harder to copy. Johnson. for their rivals adopted similar technologies and process improvements and thus quickly matched the leaders. whose tasks are anything but routine. managers. or outsourcing production and clerical jobs. But any advantage in costs or distinctiveness that companies gained in this way was usually short lived. James M. customers. New McKinsey research reveals that these high-value decision makers are growing in number and importance throughout many companies. judgment. For the past 30 years. But advantages that companies gain by raising the productivity of their most valuable workers may well be more enduring. and customer service reps. automating. for example. experience. Manyika.

2005 Number 3. executives must put aside much of what they know about reengineering—and about managing technology. “The 21st-century organization. That’s the bad news—or. As in the early days of the Internet revolution. The use of technology to complement and enhance what talented decision makers do rather than to replace them calls for a very different kind of thinking about the organizational structures that best facilitate their work. Thus the traditional organization. and talent to boost productivity. Companies that get it right will build complex talentbased competitive advantages that competitors won’t be able to duplicate easily—if at all.” The McKinsey Quarterly. and Lowell L. Bryan and Claudia Joyce. the mix of skills companies need.64 The McKinsey Quarterly 2005 Number 4 Article at a glance As more 21st-century companies come to specialize in core activities and outsource the rest. But to improve these employees’ labor performance.” The McKinsey Quarterly. 2005 Number 3.1 Raising the labor performance of professionals won’t be easy. hiring and developing talent. and their actions have a disproportionate impact on the ability of companies to woo customers. pp. the challenge (and opportunity) for innovators. Bryan. Process cookbooks can show how to operate a modern warehouse but not what happens when managers band together to solve a crisis. rather. is being upended. the direction is clear but the path isn’t. 4–5. of labor: workers who undertake complex. they have greater need for workers who can interact with other companies. and their suppliers. 24–33. “Getting bigger. Thus. and it is uncertain whether any of the innovations and experiments that some pioneering companies are now undertaking will prove to be winning formulas. their customers. Raising the productivity of employees whose jobs can’t be automated is the next great performance challenge—and the stakes are high. and to earn profits. 1 Lowell L. pp. Technology can replace a checkout clerk at a supermarket but not a marketing manager. Machines can log deposits and dispense cash. interactive jobs typically command higher salaries. organizations. and the way technology supports high-value labor. Machines can help managers make more decisions more effectively and quickly. the potential gains to be realized by making these employees more effective at what they do and by helping them to do it more cost effectively are huge— as is the downside of ignoring this trend. but they can’t choose an advertising campaign. where a few top managers coordinate the pyramid below them. Technology and organizational strategies are inextricably conjoined in this new world of performance improvement. . to compete.

the searching. Hanna. North. and technology are making interactions far more pervasive in developed economies. coordinating. 1997 Number 1. like the boom in global operations and marketing. Hall. “Institutions and Productivity in History. . Best practice thus won’t become everyday practice quite as quickly as it has in recent years. North. and Anupam Sahay. Byron Auguste. pp. Reducing these capabilities to a checklist of procedures and IT systems (which rivals would be able to copy) isn’t going to be easy.” Washington University at St. number 9411003. Douglass C. And communications technologies such as e-mail and instant messaging have made interaction easier and far less expensive. has dramatically increased the need to interact with vendors and partners.” Washington University at St. number 9411006. globalization. At the turn of the last century. most nonagricultural labor in business involved extracting raw materials or converting them into finished goods. running heavy machinery. 2 Patrick Butler. and Douglass C. of course. 4–23.The next revolution in interactions 65 The good news concerns competitive advantage. Building sustainable advantages will again be possible—and. number 9309004. and monitoring required to exchange goods or services. Louis economics working paper. September 1993. worthwhile. economic history series. Outsourcing. We call these activities transformational because they involve more than just jobs in production. North. November 1994. Louis economics working paper. Ted W. As Adam Smith predicted. Louis economics working paper.3 By the turn of the 21st century. Recent studies—including landmark research McKinsey conducted in 19972—show that specialization. As companies figure out how to raise the performance of their most valuable employees in a range of business activities. James Manyika. The rest of the workforce now consists of people who largely or wholly spend their time interacting. Alistair M. “A revolution in interaction. however. The interactions revolution Today’s most valuable workers undertake business activities that economists call “interactions”: in the broadest sense. November 1994. “Institutions. or operating production lines—in part because in a globalizing economy many such jobs are shifting from developed to developing nations.” The McKinsey Quarterly. and Productivity in the Long Run. economic history series. economic history series. specialization tends to atomize work and to increase the need to interact. Transaction Costs. Lenny Mendonca. “Transaction Costs Through Time.” Washington University at St. 3 Douglass C. only 15 percent of US employees undertook transformational work such as mining coal. they will build distinctive capabilities based on a mix of talent and technology. The growth of interactions represents a broad shift in the nature of economic activity.

Most jobs mix both kinds of activities—when managers fill out their expense reports. and tacit (more complex interactions requiring a higher level of judgment. transactional (interactions that unfold in a generally rule-based manner and can thus be scripted or automated). but also most of what IT specialists. judges. and other sources to analyze these trends on a global basis. we placed every job in one of three categories: transformational (extracting raw materials or converting them into finished goods). We looked at the range of business activities involved in more than 800 occupations in the United States. Then we linked the occupational analysis to the US government’s industry classifications and quantified the mix of tacit. While any kind of work clearly involves activities in all three of our categories. for example. lawyers. we will therefore refer to the more complex interactions as tacit and to the more routine ones as transactional. and mediators) must often draw on deep experience. leading workshops on corporate About the research The next wave of performance improvements— to raise the effectiveness of tacit workers—will be far more difficult than the improvement efforts of the past. the firm’s new research on job trends in a number of sectors finds that companies are hiring more workers for complex than for less complex interactions. which companies have long been automating or eliminating. or experiential. These men and women (such as managers. This occupational segmentation allowed us to develop a macroeconomic view of employment and wage shifts and to isolate trends in tacit interactions. which economists call “tacit knowledge. and cost. Eight years after McKinsey’s 1997 study. and transformational activities within and across industries. managing a supply chain is a complex one. We cross-checked the results with the 1997 activitylevel analysis and with other economists’ findings on interactions. nurses. Complex interactions typically require people to deal with ambiguity— there are no rule books to follow—and to exercise high levels of judgment. biochemists. Building on McKinsey’s 1997 study. we used data from the International Labour Organization. is a routine interaction. higher-value business activities deliver what their customers care about most will probably gain significant (and not easily duplicated) advantages in distinctiveness.” For the sake of clarity. involving ambiguity. another shift is in full swing as well. Transactional interactions include not just clerical and accounting work.66 The McKinsey Quarterly 2005 Number 4 Within the realm of interactions. But companies that can innovate to make their complex. that’s a transaction. Recording a shipment of parts to a warehouse. salespeople. interviews with economists and with functional and industry experts throughout McKinsey helped us to identify and understand the key enablers of tacit and transactional interactions in today’s companies. . and many others do (see sidebar. Finally. and it has dramatic implications for the way companies organize and compete. “About the research”). In addition. transactional. quality. knowledge). the World Bank. we placed each job by determining its predominant activity. and drawing on tacit. auditors.

To put it another way.5 million. During the past six years. which determine its value added and compensation.The next revolution in interactions 67 strategy with their direct reports is tacit work. 70 percent of all US jobs created since 1998—4. But what counts in a job are its predominant and necessary activities. or roughly the combined US workforce of the 56 largest public companies by market . the number of US jobs that include tacit interactions as an essential component has been growing two and a half times faster than the number of transactional jobs and three times faster than employment in the entire national economy.

most developed nations are experiencing this trend. insurance carriers. retailing. and outsourcing helps companies shed IT help desk workers. respectively. with a limited number of tacit employees (managers) on top coordinating a broad span . Since the days of Alfred Sloan. Meanwhile. utilities. corporations have resembled pyramids. But all sectors employ tacit workers. a more automated checkin process at airports makes for smaller airline check-in staffs. These jobs now make up 41 percent of the labor market in the United States (Exhibit 1).7 percent a year. the number of jobs involving more complex interactions among skilled and educated workers who make decisions is growing at a phenomenal rate. automated replenishment systems reduce the need for supply chain bookkeepers. of course. Companies need to make this part of the workforce more productive. respectively. for example. The jobs of most employees in air transportation. fund-management companies. just as they have already raised the productivity of transactional and manufacturing labor. Unproductive tacit employees will be an increasingly costly disadvantage. which pay 55 and 75 percent more. and demand for them is growing.5 percent. is growing at a phenomenal rate Likewise. From 1998 to 2004. 6. have an acute need for savvy frontline managers. outsourcing. and relatively complex interactions 2.68 The McKinsey Quarterly 2005 Number 4 capitalization—require judgment and experience. and recreation are transactional. The point isn’t how many tacit interactions occur in a company—what’s important is that they ought to add value. than those of employees who undertake routine transactions and transformations. for example. most companies. Demand for tacit workers varies among sectors. and automating routine tasks. The balance is tipping toward complexity. This shift toward tacit interactions upends everything we know about organizations. and securities firms cut the number of transactional jobs on their books The number of jobs that involve by 10 percent. in part because companies have been eliminating the least complex jobs by streamlining processes. Manufacturers too have eliminated transactional jobs. Salaries reflect the value that companies place on these jobs. Indeed. Tacit jobs dominate fields such as health care and many financial-services and software segments (Exhibit 2). A new path to better performance The demand for tacit employees and the high cost of employing them are a clear call to arms.

by helping them collaborate more effectively inside and outside their companies) and innovations involving loosely coupled teams will suggest new organizational structures. Hierarchical structures and strict performance metrics that tabulate inputs and outputs therefore lie at the heart of most organizations today. innovations and experiments to raise the productivity of tacit employees (for instance. But the rise of the tacit workforce and the decline of the transformational and transactional ones demand new thinking about the organizational structures that could help companies make the best use of this shifting blend of talent. Over time. There is no road map to show them how to do so.The next revolution in interactions 69 of workers engaged in production and transactional labor. .

and most obviously. October 2003. pp.mckinsey. 2003 Number 2.70 The McKinsey Quarterly 2005 Number 4 The two critical changes that executives must take into account as they explore how to make tacit employees more productive are already clear. pp. McKinsey Global Institute. Meanwhile. usually involving technology. . that is. and Allen P. During the 1990s. and forecasting demand. Volume 81. and Lenny Mendonca. it’s less likely that companies will succeed in adopting best practices quite so readily. solve novel problems. Number 10. Diana Farrell. for instance. First. or sense emotional responses and react appropriately. The Home Depot is trying out automated self-checkout counters in some stores.” The McKinsey Quarterly. US Productivity Growth 1995–2000: Understanding the Contributions of Information Technology Relative to Other Factors. First. they can use it to eliminate low-value-added transactional activities that keep employees from undertaking higher-value work. such as tracking goods. boosting labor productivity throughout the sector. Machines can’t recognize uncodified patterns. Terra Terwilliger. Wal-Mart Stores was a pioneer in automating a number of formerly manual transactional activities. the way companies deploy technology to improve the performance of the tacit workforce is very different from the way they have used it to streamline transactions or improve manufacturing. “The real new economy. “Getting IT spending right this time. they can’t substitute for tacit labor as they did for transactional labor. 4 Brad Johnson. for example. 118–29. Instead machines will have to make tacit employees better at their jobs by complementing and extending their tacit capabilities and activities.4 But in the world of tacit work.” Harvard Business Review. most other generalmerchandise retailers adopted Wal-Mart’s innovations. Capabilities founded on talented people who make smarter decisions about how to deploy tangible and intangible assets can’t be coded in software and process diagrams and then disseminated throughout a sector. October 2001 (www. Tacit technology Companies have three ways of using technology to enhance and extend the work of tacit labor. James Manyika.com/mgi). In retailing. 104–12. a look back at what it took to raise labor productivity over the past ten years shows that the overall performance of sectors improves when the companies in them adopt one another’s managerial best practices. Pharmacies. and Diana Farrell. are using robots to fill prescriptions in an effort to maximize the amount of time pharmacists can interact with their customers. Webb. Second. trading information with suppliers. however.

such as the buying behavior of a customer segment. only when no easy solution can be found is a tacit employee brought in. pp. since tacit interactions will occur as much within companies as across them.” The McKinsey Quarterly. technology support—traditionally. and 5 John Seely Brown and John Hagel III. can give them easier access to filtered and structured information. Kaiser Permanente is one of the organizations now pioneering the use of such technologies to improve the quality of complex interactions. technology makes it possible to boost the quality. At some companies. 2003 Number 4. speed. new and emerging technologies will let companies extend the breadth and impact of tacit interactions. Finally. the new counters mean that people who used to operate the old manual ones can be deployed in store aisles as sales staff—a much higher-value use of time. This point will be particularly critical. Robert Nardelli. multiple-source videoconferencing. and scalability of the decisions employees make. The health care provider has developed not only unified digital records on its patients but also innovative decision-support tools. for example. better strategy. Loosely coupled systems are more likely than hard-coded systems and connections to be adapted successfully to the highly dynamic work of tacit employees. Although it is hard to determine quantitatively whether physicians are making better judgments about medical care. “Flexible IT. Useful databases could. tacit work undertaken by staff experts on PCs and networks—has been split into tacit and transactional roles. thereby helping to prevent such time wasters as volumes of unproductive e-mail. quickly and accurately. IT. Transactional workers armed with scripts and some automated tools handle the IT problems of business users. Just as important. .The next revolution in interactions 71 The retailer isn’t just automating and eliminating transactional tasks.5 Broadband connectivity and novel applications (including collaborative software. for instance. Second. Furthermore. Technology tools can also help employees to identify key trends. data suggest that Kaiser has cut its patients’ mortality rate for heart disease to levels well below the US national average. 50–9. technology can allocate activities more efficiently between tacit and transactional workers. believes that automated counters can reduce by as much as 40 percent the time customers spend waiting at cash registers. such as programs that track the schedules of caregivers for patients with diabetes and heart disease. provide details about the performance of offshore suppliers or expanded lists of experts in a given field. say. its chairman and CEO.

and to organize themselves for work. do organizational models that motivate tacit employees The rigidity of traditional organizational and help them spot and act on ideas. but so. Most of today’s organizational models.com. for instance. employees to explore new ideas. Companies might then involve greater numbers of workers in these activities. it won’t be easy for companies to identify and develop distinctive new capabilities that make the best use of tacit interactions— new ways to speed innovations to market. aim to maximize the performance of transactional or transformational workers. advantages related to costs and distinctiveness have rarely lasted for long: they eroded quickly when companies built them from innovations in the handling of what are essentially transactional interactions. If it becomes harder for performance innovations to spread through a sector and thereby to boost the performance of all players. pp. Oxford University Press. It does help talented workers to achieve these ends. But at least such capabilities will also be difficult for competitors to duplicate. Luis Enriquez. to operate in a less hierarchical (that is. E*Trade Financial. 118–37. June 2002. and connect with networks of people and specialized talent around the world. reach rural consumers and suppliers more effectively. it will once again be possible to build operating-cost advantages and distinctive capabilities sustainable for more than a brief moment. See “From push to pull: These kinds of models usually involve The next frontier of innovation” on environments that encourage tacit mckinseyquarterly. models too often limits innovation and learning.” Global Information Technology Report 2001–2002: Readiness for the Networked World. or to divine customer needs. for instance. and Ingo Beyer von Morgenstern. “Telecommunications sector reform—A prerequisite for networked readiness. speed up. to make sales channels more effective. As a result. Best practices will be hard to transplant from one company to another if they are based on talented people supported by unique organizational and leadership models and armed with a panoply of complementary technologies.72 The McKinsey Quarterly 2005 Number 4 IP telephony) can facilitate. more team-oriented and unstructured) way. for example. and progressively cut the cost of such interactions as collaboration among communities of interest and build consensus across great distances. During the past few years. Carsten Kipping. Tacit models are new territory. by contrast. .6 Competitive advantage redux Technology itself can’t improve patient care or customer service or make better strategic decisions. World Economic Forum. gained tactical advantages by optimizing transactional activities to create more efficient and less 6 Scott Beardsley.

may find it necessary to redeploy their available tacit capacity to transformational and transactional activities. following the example of companies such as Toyota Motor and Cisco Systems. as Home Depot did. they will have to understand what profile of interactions—transactional and tacit—is critical to their business success and to allocate investments for improving the performance of each. Cheap trades were no longer a sufficient point of differentiation. thus bringing a new level of problem solving to many kinds of transformational jobs. Looking forward As companies explore how to expand the potential of their most valuable employees. customers. and suppliers to better discern what customers want and then to provide them with more effective value-added products and services. Some companies will have to redeploy talent from transactional to tacit activities. they face more than a few challenges. A company could. By contrast. it will be necessary to guard against becoming overly reliant on . focus on improving the tacit interactions among its marketing and product-development staff. That approach would create a formidable competitive capability—and it is difficult to see how any rival could easily implement the same mix of tacit interactions within its organization and throughout its value chain. for example. At the same time.The next revolution in interactions 73 expensive ways of making trades but then watched its unique position evaporate when other discount brokers and financial advisers embraced the new technology and cut their trading fees. Others. advantages built on tacit interactions might stand. For one thing.

How. or to provide the right set of opportunities so that their employees can become more seasoned and knowledgeable. 100–11. and so on—or of transactional ones. Bryan. after all. Certain organizations must therefore learn to develop their tacit skills internally. perhaps through apprenticeship programs.” The McKinsey Quarterly. performance is more complex to measure and reward when tacit employees collaborate to achieve results. such investments are now intended largely to boost the performance of transformational activities— manufacturing. and manage for tacit skills rather than transactional ones—something that will increasingly determine their ability to grow. On the human-resources side. . So they must shift more of their IT dollars to tacit tools. do you measure the interactions of managers?7 Companies will also have to think differently about the way they prioritize their investments in technology. Companies invest far less to support tacit tasks (Exhibit 3). construction. “Making a market in knowledge. develop. On the whole. even while they still try to get whatever additional (though declining) improvements can 7 Lowell L.74 The McKinsey Quarterly 2005 Number 4 a few star tacit employees and to manage critical tacit or transactional activities undertaken by partners or vendors. pp. 2004 Number 3. What’s more. companies will need a better understanding of how they can hire.

Lowell Bryan. taxes. and the Cisco Thought Leadership Team. depreciation. The spread widens in transaction-based sectors— meaning that investments to improve performance in this area still make sense. in particular. Navin Ramachandran.The next revolution in interactions 75 be had. Professor Hal Varian of the University of California. The authors wish to acknowledge the contributions of their colleagues Scott Beardsley. and James Manyika is a principal in the San Francisco office. The performance spread8 between the most and least productive manufacturing companies is relatively narrow. . Sumit Gupta. Tacit activities are now a green pasture for improvement. Copyright © 2005 McKinsey & Company. and amortization) per employee. Dan Ewing. Diana Farrell. • Brad Johnson is an associate principal in McKinsey’s Silicon Valley office. Lenny Mendonca. But the variability of company-level performance is more than 50 percent greater in tacit-based sectors than in manufacturing-based ones (Exhibit 4). where Lareina Yee is a consultant. Q 8 As measured by revenue or EBITDA (earnings before interest. as well as of John Hagel. from streamlining transactions. All rights reserved. and Roger Roberts. Luis Enriquez. Berkeley.

Sign up to vote on this title
UsefulNot useful