View the step-by-step solution to:







Chip Heath and Nancy Staudenmayer


We argue that organizations often fail to organize effectively because

individuals have lay theories about organizing that lead to coordination

neglect. We unpack the notion of coordination neglect and describe

specific cognitive phenomena that underlie it. To solve the coordination

problem, organizations must divide a task and then integrate the

components. Individuals display shortcomings that may create problems

at both stages. First, lay theories often focus more on division of labor

than on integration. We discuss evidence that individuals display partition

focus (i.e. they focus on partitioning the task more than on integration)

and component focus (i.e. they tend to focus on single components of a

tightly interrelated set of capabilities, particularly by investing to create

highly specialized components). Second, when individuals attempt to

integrate components of a task, they often fail to use a key mechanism for

integration: ongoing communication. Individuals exhibit inadequate

communication because the 'curse of knowledge' makes it difficult to take

the perspective of another and communicate effectively. More importantly,

because specialists find it especially difficult to communicate with

specialists in other areas, the general problem of communication will

often be compounded by insufficient translation.


Highly motivated individuals often fail in their attempts at organizing. At

Xerox during the 1970s, top managers feared that Xerox might falter when

business transactions shifted from paper to electronic forms. They were savvy

enough to create the group of researchers that invented the first desktop PCs

(complete with mouse graphical interface) and easy means of networking them

(complete with networked laser printers and e-mail). Then top managers failed

continually to organize a structure that could bring the new technologies to the

marketplace (Smith & Alexander, 1988). Why? Software developers who are

urgently trying to finish a major piece of software frequently organize

themselves in ways that actually slow themselves down (Brooks, 1979;

DeMarco, 1995). Top managers at the best firms of the era systematically

complicated their own jobs by diversifying into other lines of businesses, then

floundered when they tried to design organizational structures to repair the

complications they created (Chandler, 1962). Front-line workers who are

earnestly trying to communicate with each other choose methods of

communication that make it even more difficult to organize their efforts

(Dougherty, 1992; Bechky, 1999). Even when motivations far exceed the

typical range of motivations available in a business setting, highly motivated

individuals still fail to organize effectively. During World War II the American

Navy was suffering devastating attacks on its convoys by German submarines.

It spent months trying to copy the more successful British Navy, willingly

borrowing every possible aspect of the British system except for the method of

organizing that eventually, much later, allowed them to succeed (Cohen &

Gooch, 1990). Why?

The examples above share a common characteristic: Actors were highly

motivated to succeed in their task, yet they chose bad ways of organizing their

actions. We belabor this point with many examples to raise a puzzling issue;

although the social sciences today have much to say about the problem of

motivation (which is less of a problem in the examples above), they have less

to say about the problem of improper organizing (which is a much more serious

problem). In order to accomplish their work, organizations must solve two

problems: motivating individuals so that their goals are aligned (the agency

problem) and organizing individuals so that their actions are aligned (the

coordination problem). Almost all of the founding texts in organization studies

emphasize the importance of both problems: agency and coordination

(Barnard, 1938; Simon, 1947; March & Simon, 1958; Burns & Stalker, 1961;

Galbraith, 1973). Yet although agency and coordination are both central

problems for organizations, in recent years, the agency problem has receivedfar more attention from researchers. Agency theory, a popular topic in both

economics and organizational studies, examines how principals can design

optimal incentives to align the goals of their employees or agents (Eisenhardt,

1989; Jensen & Meckling, 1976). Although the agency problem has become

increasingly popular, the coordination problem has not seen an equivalent rise

in popularity, despite the fact that it is equally central for organizations. In fact,

in economics, the coordination problem predates the interest in agency (e.g.

Marschak & Radner, 1972), yet it has fallen out of favor while the agency

problem has become increasingly popular (Milgrom & Roberts, 1992).

In many situations, the agency problem is not the main barrier to organizing.

Even when organization members are motivated to work hard, they may find it

difficult to coordinate their actions. In organizational research, early researchers

recognized this fact and developed studies and theories of organizational

design (Thompson, 1967; Woodward, 1965; Perrow, 1967). Now commonly

thought of as an 'old' theory, organizational design research has gradually

fallen out of favor (Staudenmayer, 1997) albeit with some important exceptions

(e.g. Tushman, 1979; Wageman, 1995; Gresov & Drazin, 1997; Crowston,

1997). The little research that is being conducted today typically draws upon

the frameworks and concepts put forth by Thompson and others in the 1960s.

In fact, Thompson's 1967 book is still the most cited source in organizational

design, and its citation count is falling over time (Staudenmayer, 1997).

In this chapter we try to return attention to the coordination problem as a

central problem in organizational studies. However, in contrast to the research

on organizational design, we explore the coordination problem by exploring the

cognitive problems that individuals face when they attempt to coordinate with

others. We argue that people have inadequate lay theories of organizing, and

that their lay theories hinder them when they attack the problem of

coordination. Our level of analysis is the individual and the cognitive processes

that individuals use to approach the coordination problem. We argue that when

individuals design organizational processes or when they participate in them,

they frequently fail to understand that coordination is important and they fail to

take steps to minimize the difficulty of coordination. To summarize this

hypothesis, we say that individuals exhibit coordination neglect.

We do not claim that coordination neglect among individuals will always

produce coordination failures in organizations. Industries and organizations

may provide pre-packaged processes and procedures that can repair the

cognitive shortcomings of individuals (Heath, Larrick & Klayman, 1998).

However, if organizations exist in a dynamic environment where they must

continually adapt, then individuals within organizations will continually face

novel coordination problems that they cannot easily address with pre-packagedsolutions.

Below, we will present numerous case studies (including those

above) that document how coordination neglect produced coordination failures

in organizations that were confronting novel situations.

The goal of this chapter is to unpack the notion of coordination neglect and

to describe more specific cognitive phenomena that underlie it. However,

before we attribute specific problems to coordination neglect, we first need to

explain the approach we are taking. The process of organizing is indeed

difficult, and we don't want to take credit for any problem that arises when

people try to organize. In the next section, we sharpen the concept of

coordination neglect by stipulating what it is and what it is not.


In this chapter, we analyze coordination neglect as a cognitive problem that is

rooted in the lay theories people use to think about organizing and coordinating

with others. People have intuitive, lay theories about many things - social

interactions, statistical causality, economic markets (Kahneman, Slovic &

Tversky, 1982; Furnham, 1988) - and we suggest they also have lay theories

about organization. All theories are incomplete, particularly lay theories, and

we are interested in understanding the psychological blind spots in these lay

theories that may cause people to neglect to coordinate their actions with


In this chapter, we will focus on several facets of coordination neglect, all of

which can be located on Fig. 1. In the simplest version of the coordination

problem, an organization divides an overall problem into subtasks and assigns

the parts to individuals. We could imagine, for example, an organization that

divides the modules of a computer program among programmers with similarskill. Organizations undertake this division of labor because individuals have

limited information-processing abilities (Simon, 1962; March & Simon, 1958).

Eventually, however, the organization must re-integrate the tasks that it

originally divided. In the end, the modules of a computer program must work

together as a single program, so the programmers who develop individual

modules must integrate their efforts. Thus, the flip-side of division of labor is


The coordination problem becomes especially complicated when organizations

divide a task among specialists. Specialization reduces the problem of

bounded rationality because individuals can concentrate on a component of the

task that meets their unique skills, training, and abilities. Here, we move away

from the computer programming example and toward, say, an automobile firm

that hires a variety of people with special skills - good transmission engineers

to design the transmission, knowledgeable production people to manufacture

the car, and effective marketers to sell it. Now, however, the task of integration

is even more complicated because the organization must integrate the efforts of

specialists who speak different languages and perceive the world in different


Throughout this chapter, we will explore different parts of Fig. 1 that

highlight particular problems in the lay theories of individual organizers.

However, first it is important for us to distinguish our approach, which depends

on flaws in the lay theories of individuals, from other potential approaches. The

coordination problem is difficult to solve, so we do not want to accuse people

of coordination neglect any time they have difficulty solving this difficult


For example, in order to argue that coordination neglect is a cognitive

problem, we must distinguish it from motivational problems such as agency

problems. Thus, it is important for us to provide examples where people really

would prefer to coordinate so that agency problems are not an issue. Engineers

and marketers may not talk to each other to coordinate their efforts because

they: (1) have more fun interacting with others in their own department or (2)

because they don't anticipate how much they need to interact to create a

successful product. The first is not coordination neglect, it is an agency

problem; an organization could presumably solve it by introducing an incentive

scheme that encourages the marketers and engineers to interact. The second is

more interesting because it suggests a cognitive limitation. This kind of

cognitive limitation is unlikely to be solved by any of the standard incentive

solutions to problems of control or agency.

Furthermore, to document that people are neglecting coordination, it is also

important that we consider situations where people are thinking about how toalign action. We are not concerned with situations where systems evolve

organically over time in a way that produces hidden interdependencies

(Staudenmayer & Desanctis, 1999). In organizations that evolve complicated

processes over time (e.g. consider the process of acquiring parts at a large

manufacturing firm), a procedure may span multiple people and departments

and it may grow and change over time in spontaneous ways. In organizations,

such processes may be completely revealed only when teams attempt to

'reengineer' them (Hammer & Champy, 1993). Although this kind of hidden

interdependency is important, we want to focus on interdependencies that are

more obvious. Coordination neglect is clearest when people consciously try to

design or alter a process, yet they neglect to consider obvious issues of

coordination, e.g. Chandler's (1962) managers who explicitly grappled with

how to design their organizational structure.

In order to document coordination neglect, we not only require that people

be thinking consciously about coordination, we also require that the act of

thinking does not exceed their computational abilities. When we claim that

people exhibit coordination neglect, we don't want to reiterate that people have

difficulty performing difficult tasks. Herb Simon explained such difficulties

many years ago as a product of 'bounded rationality' (Simon, 1947). In order

to make the case that people exhibit coordination neglect, we should point out

situations where people neglect alternative ways of coordinating that are

equally or less cognitively demanding.

In sum, to provide examples of coordination neglect, we should point to

situations where people do not coordinate even though they prefer to

coordinate, where they are consciously considering how to coordinate, and

where there is a means of coordination that does not exceed their computational



In order to accomplish a complex task, organizations typically divide up the

task and assign components of the task to different people. However, whenever

the organization divides a complete task into components, the people who

perform the component tasks are interdependent and they must integrate their

efforts (Thompson, 1967, ch. 5). As Fig. 1 indicates, division of labor compels


In this section, we introduce two related aspects of coordination neglect:

partition focus and component focus. Partition focus refers to people's

tendency to neglect coordination because they focus more on dividing and

partitioning the task than on integrating the components they create. People notonly focus on partitioning the task into components, they also, when they are

trying to diagnose or improve a process, tend to focus on individual

components rather than the system as a whole. Component focus refers to this

tendency for people to focus on components, particularly by investing in

components so that specialists can become even stronger. The next two sections

explore these aspects of coordination neglect.

Partition Focus

It is completely reasonable for people to divide or partition tasks carefully; this

is one half of the equation for success in organizations (Simon, 1962; Lawrence

& Lorsch, 1967). However, the other half of the equation for success is

integration, and partition focus may cause people to neglect integration. When

people partition the world, they may tend to treat small interactions between

components as zero when they actually need to invoke a range of mechanisms

to integrate the components (March & Simon, 1958).

The most direct evidence of partition focus is provided by situations where

people consciously design a process from scratch, for example in the software

industry. Coordination is a central issue in software design because software

programs, which are constructed in segments or modules by individual

programmers, must work seamlessly as a whole. One prominent consultant for

the software industry has described the typical process that happens when

software designers are given a new project (DeMarco, 1995; DeMarco &

Lister, 1998). In a passage directed at team managers, DeMarco describes the

typical process of design: "You make a crude division of the whole into five or

ten pieces so you can put five or ten design teams to work. That crude divisionis a design step, but you don't approach it as such . . . That initial crude division

is the heart of the design, and since there is no one directly responsible for

revisiting its logic, it remains the heart of the design. The result is no design"

(p. 251). In software engineering, the division of labor is less successful when

there are more interfaces between modules and when the interfaces are more

complex. By partitioning the work immediately, rather than thinking through

the interfaces and patterns of interdependence among modules, design teams

"guarantee the interfaces among people are more complex than they need to be

. . . People are forced to interact with more of their teammates in order to get

anything done and the interactions are more complex. The result is less

possibility of independent work, more telephone tag, more meetings and more

frustration" (DeMarco & Lister, 1998, p. 255; Perry, Votta & Staudenmayer,

1994; 1996).

In this example, software engineers partition the project crudely and then

proceed immediately into implementation. Unfortunately, this inevitably

results in greater integration problems later on because they must continuously

loop back and make unanticipated changes to the original inadequate design, a

phenomenon that has been documented repeatedly by researchers in software

engineering (Kemerer, 1997; Boehm, 1984; Boehm & Papaccio, 1988). To

combat this habitual tendency to partition the task prematurely, DeMarco

(1995) argues that in a project designed to last a year, no coding should be done

until the last two months. According to his recommendation, designers should

spend 10 months selecting the right modules by minimizing the coordination

that must take place among modules. Coding and testing will then proceed

much more smoothly, requiring two months as opposed to 10 or more. Other

researchers have pointed out that not all firms can afford the luxury of 10

months of design (Cusumano & Selby, 1995; Yoffie & Cusumano, 1999; Iansiti

& Clark, 1994), and have suggested ways of integrating continuously. We argue

that all these solutions are essentially cognitive repairs for partition focus in the

lay theories of software design teams.

Partition focus runs sufficiently deep that, at times, it has become embodied

in the institutional language and planning procedures of the software industry.

In one of the most famous books on software design, Frederick Brooks (1979)

discusses common flaws that cause large software projects to fail. Brooks is a

credible observer of large software projects - he was the 'father' of the very

successful System/360 project at IBM; at its time, it was the largest software

project in the computer industry, comparable in size and expenditure to the

NASA space program. The first flaw listed by Brooks is institutionalized in the

very unit of effort that software managers used to estimate effort in the 50s and

60s: the man-month. Brooks notes that "men and months are interchangeablecommodities only when a task can be partitioned among many workers with no

communication among them." While this may be true of reaping wheat or

picking cotton, "it is not even approximately true of systems programming"

(p. 16). Brooks regarded this problem as so fundamental he chose it as the title

of his book: The Mythical Man-Month. Brooks argued that it is counterproductive

to focus only on partitioning the task into 'man months' because the

different partitions are not interchangeable. Among software programmers,

Brooks' warning against partition focus has been generalized in the wellknown

aphorism: "Bearing a child takes nine months, no matter how many

women are assigned".

One general explanation for partition focus is that it simplifies computational

costs and saves time. This does not mean that boundedly rational individuals

could not calculate a more elaborate set of interactions than they do (indeed

DeMarco, 1995, recommends that software managers spend the time to

consider such interactions), however 'cognitive misers' might try to shortcut

the process. In order to reduce computation costs, people may prefer to

categorize ambiguous information into relatively crisp, well-bounded categories

(March & Simon, 1958). People may think that it will be easier to plan

when they foreclose on a particular way of partitioning the task. Partition focus

may be related to a tendency to do what is most well-learned (Staw, Sandelands

& Dutton, 1981), particularly in situations where partition focus is enhanced by

specialization. Typically, specialists are highly trained in their own specialty

but know less about the specialties of others. Furthermore, this specialized

training is often exacerbated in organizations by reward systems that

inappropriately emphasize individual performance.

However, there is evidence that partition focus is not produced by

environmental rewards, or even by training, but by lay theories. For example,

in one experiment, groups of MBA students were given a bag of Lego blocks

and were asked to assemble their blocks to match a model (a man with arms,

legs, head and torso). They were given a long time to plan the exercise, but their

goal was to assemble the model in the least amount of time possible. Teams

could do a number of things to speed assembly both by partitioning the task and

by integrating it. In terms of partitioning, groups could appoint 'experts' for the

different body parts (arms, legs, torso); the individual body part experts could

draw a diagram to show how their body part was constructed, and they could

develop a specific plan for how to assemble it. In terms of integration, groups

could develop a master diagram of how all the body parts fit together, they

could talk about how to integrate the different body parts (e.g. how the arms

would attach to the torso), and they could appoint an 'overseer' to guide the

assembly process. Regressions showed that each partition and each integrationbehavior reduced assembly time by approximately the same amount. On the

other hand, the partition and integration behaviors were not performed equally

often; groups performed about 75% of the possible partitioning behaviors, but

only about 50% of a parallel set of integration behaviors (Heath, Jost & Morris,


The Lego exercise suggests that partition focus is a problem of lay theories

because the experimental procedure effectively rules out other potential

explanations of coordination problems. The procedure contains no external

incentives to create agency problems. There is no a priori specialization that

would complicate the task of integration. There are no strong cognitive

constraints that would prevent people from thinking through the problems of

integration. Finally, in contrast with most coordination situations, the

coordination required by this task was obvious and visible; people could see

that the arms needed to join the torso, so it was obvious that the person

assembling the arms should coordinate with the person assembling the torso.

Yet teams often experienced problems at the 'joints' of the model; and such

problems occurred simply because team members did not coordinate about

how to integrate their subassemblies.

In this section we have considered the simplest version of the coordination

problem where people must divide a task among individuals and then integrate

the components. The examples from both organizational settings and

laboratory work provide evidence that people focus more on partitioning the

task than they do on integrating it, and this evidence suggests that partition

focus may contribute to coordination neglect.

Component Focus

People focus not only on the process of partitioning a task, but also on the

individual partitions or components they create. We will label this tendency as

'component focus' to signal that it is related to partition focus. People exhibit

component focus when they try to intervene in an interrelated process by

focusing only on one part of the process. When people exhibit component

focus, they neglect the interrelationships and interactions among components.

In lay theories that exhibit component focus, wholes are not 'the sum of their

parts', they are a function of one part.

The division of labor is useful because it allows people to specialize,

however specialists often ignore the task of integrating with others. One chief

engineer at Ford looked back on his early years and ruefully recalled his narrow

view of automobiles during the time he specialized in designing chassis:

"When I saw a car driving down the road, all the rest [other than the chassis]disappeared. All I could see were the suspension arms going up and down"

(Walton, 1997, p. 73). It is very easy for component focus to become embodied

in the structure of the firm. Traditionally, Ford, like other automotive

companies, organized around very narrow functional specialties, not just

'engineering' but subdivisions of engineering like "chassis, powertrain,

electronics, climate control, plastics, and glass" (p. 74). These narrow units did

not interact with each other except through the planners who were in charge of

a given vehicle. According to one senior engineer, "In the old days, Ford had

this attitude, 'You want a car, we'll give you these pieces'." (p. 73).

Component focus is often exacerbated because people focus on enhancing

the quality of an individual component by making it more specialized. If people

assume that components are the source of competitive advantage (and not the

interrelationships among components), then a simple strategy is to get the best

'part' you can. Although specialization is useful, at some point organizations

may face tradeoffs between enhancing specialization and promoting integration.

Component focus may cause people to neglect coordination through


One example of component focus is found in Xerox's experience with its

Xerox PARC unit of computer scientists. In a series of decisions that ranks

among the worst business blunders of the 20th century, Xerox created the

personal computer in the form we know it today, then failed to commercialize

it. Xerox began well by creating an appropriate group of specialists. It

assembled at its Palo Alto Research Center (PARC) the most creative group of

computer scientists the computer industry has ever seen. In the 70s, it was wellknown

in the computer industry that "58 of the top 100 computer scientists in

the world" worked at PARC. The common wisdom was somewhat exaggeratedsince PARC's employment roster never exceeded 50, but the general idea was

approximately right because these computer scientists created a working set of

networked personal computers that were far ahead of their time (Cringley,

1996). In the late 1970s, an in-house video at Xerox depicted a working

computer system that was more than a decade ahead of the industry. A man

enters his office, sits at his personal computer, checks his e-mail on a graphical

computer interface using a mouse, and prints out a document on a laser printer.

The in-house video appeared at least a decade before anyone had considered

e-mail, mouses, laser printers or the graphical computer interface (which Steve

Jobs later borrowed to make the Macintosh the most successful product at

Apple). By 1978, over 1500 of these computers were in active use within Xerox

(Smith & Alexander, 1988, p. 202).

In creating PARC, top managers at Xerox recognized the benefits of

specialization: they recruited the right cadre of computer scientists, selected the

right management team to lead them, and provided an effective campus-style

working environment (complete with bean bag chairs). Yet according to Steve

Jobs, Xerox "grabbed defeat from the greatest victory in the computer industry.

[They] could have owned the entire computer industry today" (Cringley, 1996).

Xerox had an extremely advanced personal computer and a sales force with

direct access to every major corporation. How could it have failed to become

the greatest company of the personal computer revolution?

A book titled Fumbling the Future documents a series of mistakes Xerox

made in capitalizing on the success of PARC (Smith & Alexander, 1988).

Although Xerox successfully created an effective group of specialists, top

managers failed to create coordination mechanisms that would integrate the

innovations of the specialists into the mainstream business of the company.

For example, because top managers at Xerox consistently focused on

specialization, they made disastrous decisions about physical locations, not

once, but twice in PARC's history. Throughout Xerox's experience with PARC,

coordination suffered because PARC's location on the west coast was so far

from Xerox's headquarters on the east coast. When Xerox was first deciding

where to locate PARC, some managers argued that coordination would be

enhanced if PARC were closer to the rest of Xerox. Jack Goldman, head of

research for Xerox and an inside board member, noted in a memo to the CEO

that "If the new research center is too isolated from a Xerox environment, the

chances of relevant coupling to Xerox's needs and practices will be severely

diminished." In a prescient sentence he said, "one area normally considered as

an ideal research environment, Palo Alto, is eliminated because of the absence

of any nearby major Xerox facility." But the head of the research lab argued

strongly for the Palo Alto location because it was close to the emerging SiliconValley area and he won the location battle with this argument in favor of

specialization (p. 56).

In the second poor location decision later in PARC's career, top management

at Xerox had realized that they had failed to commercialize PARC technology,

so they created a new group to do so. An outside firm recommended two sites

for this new group: the San Francisco Bay area (close to PARC and the rest of

Silicon Valley) and Dallas. Xerox chose Dallas. Instead of placing the new

group in Silicon Valley next to the PARC researchers with whom it was

supposed to coordinate, top managers at Xerox considered it as a unique,

specialized component. "An elaborate financial model of a factory in Texas

versus one in California conclusively proved, on the basis of labor,

transportation, taxes, and other cost indicators, that Dallas would save Xerox

money" (p. 162). By considering this group as a 'specialist' in manufacturing,

managers chose a location that almost ensured that it would fail in its role to

commercialize PARC's technology. The results were predictable: According to

the manager of corporate R&D, "Dallas turned out to grow a culture that was

completely orthogonal to, and independent of, the digital world in general and

PARC in particular" (p. 163).

Xerox emphasized specialization in personnel choices as well. It selected an

academic 'specialist' to head the research lab - George Pake, a former provost

of Washington University in St. Louis. Although Pake had been an effective

university provost, he was ill-equipped to be the chief integrator between PARC

and the rest of Xerox. He consistently bungled opportunities to create

excitement about PARC's technology within the rest of Xerox because he was

too accustomed to an academic style of interaction. When he was appointed to

a Xerox-wide strategy committee, Pake was given an ideal forum to proselytize

for PARC technology. "Yet from the outset, others noticed that Pake had no

commercial instincts . . . [He] spoke awkwardly about business, insisting that

the strategy committee reach its conclusions by the 'scientific method' . . .

When discussions turned to PARC's technologies, Pake emphasized the work

to be accomplished in the laboratory instead of the commercial opportunities

that might already exist" (p. 150).

In sum, Xerox focused on one component of a successful new product

introduction: research and development. It created a group of specialists who

did, in fact, live up to their billing as the greatest assemblage of computer talent

ever. However, by focusing every feature of PARC around creating a

specialized component, top managers at Xerox failed to integrate this

specialized component with the other activities like marketing, manufacturing,

and finance that were necessary to successfully commercialize the new

technology. Although Xerox provides a vivid example of component focus,researchers have documented a similar emphasis on creating groups of

specialists in many firms that are trying to commercialize technology (cf.,

Iansiti, 1995).

Component focus can be found, not only when top managers are trying to

create new markets from scratch, but even when they are trying to learn from

a successful past model. There is evidence that component focus may blind

managers from understanding the sources of their previous success. One such

example comes from a brilliant book by David Hounshell & John Smith (1988)

that chronicles the history of research and development at Du Pont from 1902

to 1980. During the 1930s, Du Pont had a blockbuster decade, "Quite

unexpectedly, [fundamental research] produced neoprene and nylon. Du Pont

was successful in commercializing these important discoveries because it

already had extensive commercial and technical capabilities in rubber

chemicals, organic synthesis, high-pressure reactions, and fibers" (Hounshell &

Smith, 1988, p. 596). Du Pont's managers had an opportunity to learn from

their successes in nylon and neoprene. What should they have learned, and

what, in fact, did they learn?

According to Hounshell & Smith (1988), Du Pont should have learned that

success requires a set of integrated capabilities. In the case of nylon, Du Pont's

capability in fundamental research was matched by a number of other

capabilities that allowed it to develop the product and market it successfully.

Du Pont had previously produced rayon, which gave it specific expertise in

chemical engineering for artificial fabrics - the ability to scale up a clean,

precise laboratory process to an industrial-scale plant (p. 259-73). In its work

on rayon, Du Pont found that 25 variables had to be precisely controlled to

produce a uniform final product (p. 165). Du Pont also had an unusual

capability to manage high-pressure catalytic reactions because of its ammonia

business. Although Du Pont's ammonia business was a money-loser if it was

considered separately, it produced strong returns when Du Pont made nylon

because nylon production required similar high-pressure catalytic reactions

(p. 258), an unexpected example of economies of scope. Du Pont also had

expertise marketing fabrics to industrial customers; for example, it marketed

rayon as a substitute for silk in light fabrics during the fashion boom years after

World War I (p. 164-67) and as a basis of tire cords that improved the life of

heavy-duty tires (p. 169). In sum, Du Pont's success with nylon was produced

by a range of complementary capabilities and assets spread across the firm

(Teece, 1986).

Not surprisingly, Du Pont was pleased with its success in nylon, and

dedicated itself to discovering "new nylons" - proprietary products that would

produce the same high rates of return as nylon. Although nylon succeededbecause of a wide range of complementary capabilities, when company

executives talked about 'new nylons', they did not acknowledge these complex

interrelationships. Instead, they primarily focused on one component of their

success: fundamental research.

Over the next 30 years, Du Pont, in searching for new nylons, placed greater

and greater emphasis on fundamental research. As a consequence, research

programs were "pushed away from the company's commercial interests and the

nylon model became skewed" (p. 597). While the company's executives

believed that fundamental research would produce new nylons, the research

department instead produced fundamental research in areas where Du Pont

lacked complementary capabilities. The fundamental research group "lost

contact with many of the industrial departments and took on the trappings of a

high-quality scientific research establishment" (p. 597). During the three

decades from 1940 through 1960, Du Pont continued to search for "new

nylons" by investing only in its capability in fundamental research, despite the

fact that the company's only two real successes during this period, Orlon and

Dacron, took advantage of the same interrelated set of capabilities as nylon: not

only fundamental research, but also engineering, manufacturing, and marketing

of artificial fabrics. According to the historians, "In developing a mentality of

'new nylons', executives and research managers alike had forgotten why the

company had so easily and swiftly developed nylon. The pioneering work on

polymers had fitted neatly into the company's existing businesses, technologies,

and expertise" (p. 597).

Teece (1986) notes that in order for firms to commercialize an innovation

successfully, they must combine fundamental research with other capabilities

or assets. Research alone is not enough, firms must integrate it with other

capabilities like marketing, competitive manufacturing, and support. According

to Teece, firms succeed when they have not one single competency, but when

they own a set of assets that are complementary. We interpret Teece's

observations as a useful corrective for approaches that primarily emphasize a

single core competency (e.g. Hamel & Prahalad, 1990). Such approaches, in

their extreme could enhance the kind of component focus exhibited by top

managers at Du Pont.

Component focus may be exacerbated because, if people are focusing on

only one component at a time, they may preserve the illusion that resources are

being used efficiently. Redundancy conflicts with people's desires to avoid

'waste' (Arkes & Blumer, 1985). One consistent factor in examples of

component focus is that people seem to be trying to use resources to their

fullest capacity, whether personnel (DeMarco, 1995; 1997) or other resources

(Chandler 1963). In large software projects, "early overstaffing tends to forcepeople into shortcutting the key design activity (to give all those people

something to do)" (DeMarco, 1995, p. 260).

This variety of component focus fooled many of the top management teams

chronicled by Chandler (1962) in his famous account of the development and

diffusion of the multidivisional form. In a number of Chandler's case studies,

top managers diversified their product lines in an attempt to create economies

of scale in some component of their business: at General Foods, it was an

attempt to effectively use their central sales organization (p. 347); at the major

oil companies, it was an attempt to use residual petrochemicals from oil

refining (p. 361); at meat packing firms, it was an attempt to use their

refrigerated distribution network to carry, not only meat, but eggs, milk, and

poultry (p. 391). Although these firms sought to create economies by more

efficiently using one component of their business, they almost inevitably failed

to predict the administrative and coordination costs that they would incur by

using the excess resources in this component. The new strategy did not

typically increase the total output or size of operations, "but it quickly enlarged

the number and complexity of both tactical and strategic administrative

decisions" (p. 362). Chandler notes that even "a small amount of diversification

in relation to total production" sufficed to create enough complexity to warrant

a different structure (p. 362). We argue that many of their problems were

produced by component focus. By trying to create economies of scale in one

component of their organization, top managers dramatically complicated their

job because they failed to anticipate how much they would have to increase

coordination to use 'spare' capacity in that component.

Chandler's firms eventually solved their coordination problems by adopting

the multidivisional form, but they adopted this solution slowly and only after

much internal struggle. It would be fair to argue that such problems might be

less likely in today's environment where the multidivisional form is a common

solution to the kinds of integration problems that these managers faced.

However, the historical example is useful because it points out the problems

that may be caused by component focus in other situations where managers

confront a novel problem of coordination but do not have automatic access to

a prepackaged integration mechanism like the multidivisional form.

In many examples of component focus, managers seem to focus on

technology rather than on broader issues of organization. One example of this

occurred in the battle of the Atlantic during World War II. During WWII, the

Atlantic Ocean was the site of a protracted struggle between the American and

British ships that were trying to keep supplies flowing between the United

States and Great Britain and the German submarines that were trying to sink

them (Cohen & Gooch, 1990, p. 59-94). During the early stages of thisstruggle, the Americans were much less effective than the British in combating

German subs. According to military historians Cohen and Gooch, the U.S.

Navy made "a serious and protracted effort to learn from British experience"

(p. 87), but it borrowed only components of the British process, particularly

those involving technology (e.g. British ship designs for destroyer escorts or

British sonar). On the other hand, it neglected the organizational structure that

the British used to integrate the components. The British had a central

Intelligence Center with a small staff who collected all incoming information

(e.g. decrypted radio intercepts, photographs, prisoner interviews) and

communicated this information directly with field commanders (p. 76). Field

commanders could then divert their convoys away from German U-boat packs,

and could concentrate their scarce escort vehicles on protecting convoys that

were most endangered. This centralized operation also allowed the British to

test and deploy new tactics to combat subs. Because any combat unit on the

ocean might have only one or two chances to engage an enemy submarine, it

was important that the combat units use the right tactics the first time (p. 83).

Only after the Navy borrowed many components of British tactics without

success did they finally get around to borrowing the organizational mechanisms

the British used to integrate their efforts. The U.S. Navy created an unusual

military organization, the Tenth Fleet, to command all anti-submarine warfare.

It could even override the positions of naval commanders who were not under

its direct control. Interestingly, in the initial stages of the war, the Germans had

noted the absence of coordination among American forces at sea: "enemy air

patrols heavy but not dangerous because of inexperience," "the American

airmen see nothing; the destroyers and patrol vessels proceed at too great a

speed to intercept U-boats; likewise having caught one they do not follow up

with a tough enough depth charge attack" (p. 75). This changed after the Tenth

Fleet was created: "In the eighteen months before the creation of the Tenth

Fleet, the U.S. Navy sank 36 U-boats. In the six months after, it sank 75"

(p. 91).

Gooch and Cohen argue that many military historians who have tried to

explain the early failure of American submarine warfare, have suffered from

the same kind of component focus as the U.S. Navy. According to Gooch and

Cohen, many historians have blamed either a single individual (the commander

in chief of the U.S. Fleet) or a single cause like the absence of a coastal

blackout or of convoy support, or a "single problem" such as a missing piece

of technology that may have turned the tide (radar, more destroyers, etc.).

However, Gooch and Cohen argue that none of these components is sufficient

to explain American failures. For example, American success did not increase

even after the coastal blackout was imposed in 1942; and before thereorganization disseminated effective tactics, the Americans failed even when

destroyers were available and they actually spotted a U-boat (p. 79). In this

case, sophisticated historians joined the sophisticated military personnel in

WWII in their vulnerability to component focus.

Component focus on technology can also be seen in other organizations. For

example, General Motors engaged in the famous NUMMI joint venture with

Toyota, hoping they would be able to borrow some applications of technology

for their other factories. Instead they were confronted by relatively low-tech

machinery, but a new system for how workers interacted with each other and

integrated their actions. GM failed to understand or capitalize on this different

style of coordinating on the factory floor (Keller, 1989, pp. 124-144).

In the product development literature, Marco Iansiti and others (1993; 1995;

Iansiti & Clark, 1994) have argued that ineffective development teams are

'element focused' while effective ones are 'system focused'. Iansiti notes that

most development processes in traditional companies are sequential and

element-focused. Basic researchers explore a new concept and hand it off to

other scientists; these scientists elaborate the concept until they discover a new

application, then they hand it off to engineers and manufacturing people; the

engineers and manufacturing people prototype and produce the new product

. . . This linear approach by element focused firms "tends to compartmentalize

specific knowledge" (Iansiti, 1993, p. 138). In contrast, system-focused

companies form a core group of managers, scientists, and engineers early in the

process, and this integration team modifies and adapts the new concepts from

fundamental research so that they mesh with the current capabilities of the

company. Iansiti shows that systems focus is much more effective than element

focus: system-focused companies solved 77% of major problems early on,

while element-focused companies, which were not as attuned to integration,

solved only 40% (Iansiti, 1993, see also Henderson & Clark, 1990).

Some modern organizational theorists have likewise criticized earlier

scholars for component focus on technology. For example, Perrow (1984)

conducted a famous analysis where he argued that some technologies are so

complex that they inevitably lead to 'normal accidents'. Weick & Roberts

(1993) argue that Perrow focused too much on technology rather than on the

dynamics of social coordination: "We suspect that normal accidents represent

a breakdown of social processes and comprehension rather than a failure of

technology. Inadequate comprehension can be traced to flawed [shared] mind

rather than flawed equipment" (p. 378).

In part, component focus may represent overgeneralizations of theories of

organizing that are plausible in a more sophisticated form. As people learn

rules, they frequently generalize rules too much before they learn theappropriate exceptions (Anderson, 1995). When lay theorists of organizing are

confronted with an unfamiliar situation, they may take basic principles of

organizing, like the value of specialization, and overextrapolate them. Because

an organization must eventually coordinate the actions of specialists, it should

not encourage specialization past a certain point, yet when managers focus only

on a component of the broader organization (such as auto engineers on the

chassis, Xerox managers with PARC, or Du Pont managers with 'fundamental'

R&D), they may enhance specialization in a way that detracts from


Summary: Partition Focus and Component Focus

In this section, we have argued that people exhibit coordination neglect, in part,

because of partition focus and component focus. People focus on the division

of labor rather than on the equally important process of integration, and when

they try to intervene in an ongoing process of coordination, they tend to focus

on specialized components of the process rather than attending to the

interrelationships as a whole.

In a clever paper, Weick & Roberts (1993) argue that in order to coordinate

effectively, people must do so with heed: "carefully, critically, consistently,

purposefully, attentively, studiously, vigilantly, conscientiously" (p. 361).

"Heedless performance . . . is a failure to see, to take note of, to be attentive to"

(p. 362). Weick and Roberts are quite correct to emphasize that it is important

to understand and manage attention, but the examples above seem to indicate,

not that people are inattentive, but that they neglect to pay attention to the right

things. People's lay theories lead them to heed some things (division of labor,

components, specialization) while simultaneously remaining heedless of others

(integration, the importance of complementary capabilities).



Organizations can integrate their efforts in many ways (March & Simon, 1958;

Thompson, 1967); they may establish routines or rules that standardize the

action of different units or they may establish plans or schedules that govern the

actions of independent units. But the most important means for units to

integrate, particularly in complex or uncertain environments, is for them to

communicate with each other on an ongoing basis. According to March &

Simon (1958), 'the greater the efficiency of communication within the

organization, the greater the tolerance for interdependence' (p. 183).In this section, we examine how coordination neglect may be enhanced

because individuals fail to take advantage of this key mechanism for

integration. Although good communication is fundamental to integration, we

argue that people often exhibit coordination neglect because of inadequate

communication and insufficient translation. Inadequate communication is

likely because a number of psychological processes make it difficult for

individuals to take the perspective of another when they are trying to

communicate. However, these standard problems of communication are

compounded in organizations because specialists must communicate with other

specialists who speak different languages. If organization members don't

anticipate the need to translate across specialists, then insufficient translation

will compound the basic problems of inadequate communication. Together,

both processes may cause people to fail to integrate their efforts and may result

in coordination neglect.

Inadequate Communication

Because organizations are filled with constrained information processors,

communication will always be incomplete. Organizations develop filters that

allow them to identify the most crucial information in the complex stream of

information that enters from the outside world and flows through the

organization (March & Simon, 1958; Arrow, 1974; Daft & Weick, 1984).

However, communication, although incomplete, need not be inadequate. Recall

that the British Navy during World War II created a successful submarine

warfare unit, but it did so only after it failed in a similar task during World War

I. In the successful WWII unit, analysts collected information and communicated

it directly with field commanders at sea; they could not onlycommunicate "hard intelligence" about the location and activity of enemy

units, they could also use their judgment to communicate "working fictions"

about possible movements and tactics (Cohen & Gooch, 1990, p. 77).

Interestingly, this successful Intelligence Center structure was developed only

after Britain failed to capitalize on their "brilliant cryptanalysts" during WWI

because of inadequate communication. In WWI, the cryptanalysts were "cut off

from non-cryptologic sources of intelligence, allowed to communicate only

with the Admiralty in London rather than operational commanders at sea, and

discouraged by admirals from offering educated hypotheses about likely enemy

behavior" (p. 77). In one spectacular failure, the British Navy failed to cripple

the exposed German Fleet at the Battle of Jutland precisely because these

organizational barriers prevented the Navy from adequately integrating and

communicating information. By WWII, the British Navy arranged an

organizational structure and procedures that allowed it to communicate more


Evidence indicates that managers systematically underestimate the importance

of communication when they are planning important tasks. Recall

Brook's (1979) famous book on software design, The Mythical Man-Month. In

his list of the common flaws in large software projects, one major flaw relates

to situations where managers underestimate the importance and difficulty of

communication. Brooks observes that when a project falls behind, managers

tend to add people to the project in hopes of delivering it more quickly.

Unfortunately, this tactic compounds the difficulty of communication. When

engineers are added to a project, two kinds of communication automatically

increase. First, the existing engineers must communicate with new engineers to

train them. Every new engineer must understand something about the

technology, the project's goals, the overall strategy, and the work plan. This

slows down the project because the existing engineers, who should be

producing code, instead spend their time training new ones. Second, the pattern

of communication becomes more complicated because the new engineers must

be integrated into existing communication flows. Brooks argues that because

software involves complex interrelationships, the project quickly loses more

time by increasing communication costs than it reduces time by partitioning

tasks among more people. This argument has been enshrined in the software

industry as 'Brook's Law': "Adding people to a late software project makes it


In the American automobile industry, it took half a century for managers to

realize that design times are cut in half when you make communication easier

by putting functional representatives on a single cross-functional team in the

same location. In the 1980s, research found that Japanese manufacturersroutinely beat American firms in terms of both cycle time and quality; this

prompted a large scale study of the world-wide auto industry that attempted to

understand the reasons behind this difference (Womack & Jones, 1996;

Wheelwright & Clark, 1992). Numerous studies contributed to this research

effort, and, in general, they attributed a large part of the Japanese success to

relatively simple structures and communications practices (e.g. forming a

cross-functional team and locating them in a single site). The results have since

been replicated in other industries and settings (Ancona & Caldwell, 1992;

Brown & Eisenhardt, 1995; Eisenhardt & Tabrizi, 1995).

Although the experience of the American auto industry suggests that

practicing managers may sometimes neglect the importance of adequate

communication, perhaps academic researchers sometimes do so as well. Hinds

& Kiesler (1995) note that although hierarchy and informal networks have

always existed side by side, researchers and managers have historically

emphasized the structural and operational significance of hierarchy, while

downplaying informal networks as "ad hoc" linkages that are created through

accidents of physical proximity, personal history, or demographic similarity.

"When informal networks were seen as 'the grapevine'—unplanned, personal,

and casual - neither managers nor researchers viewed them as integral to

formal organization or crucial to accomplishing work" (p. 388).

Why is communication inadequate? One barrier to adequate communication is

that we must be able to take the perspective of others and understand what we

need to communicate to them. In psychology, there is a great deal of evidence

that people find this kind of perspective-taking to be difficult. For example,

children who are asked to direct another person around a landscape, will

assume that the listener sees the landscape exactly like they do, even if their

listener is in a different place (Piaget & Inhelder, 1967). Yet this phenomenon

is not limited to children. When adults in an experimental market were given

private information about the value of an experimental object in other markets,

they found it impossible to ignore this information even though the value of the

object in the experiment was determined by the people who were interacting in

their own market (Camerer, Loewenstein & Weber, 1989). Here, adults lost

money because they found it impossible to ignore their private information. In

this study the researchers labeled their phenomenon the 'curse of knowledge'

- people could not abandon their own perspective even though they should

have been highly motivated to do so.

Because people suffer from the curse of knowledge and have difficulty

taking another's perspective, they also overestimate how easy it would be to

communicate their knowledge to the other party (Keysar, 1994; Hinds, 1999In a striking demonstration of this, Newton (1990) asked people to participate

in an experiment in one of two roles: 'tappers' and 'listeners'. 'Tappers'

received a list of 25 well-known songs and they were asked to tap out the

rhythm of one of the songs; listeners tried to identify the song based solely on

the taps. Note that this design induces dramatic differences in the perspective

and information of the two participants. Although listeners heard only a series

of disconnected taps, the tappers, according to their own reports, 'heard' the

lyrics and complete musical accompaniment as they banged out the rhythm of

their song. This inside information made it hard for the tappers to anticipate the

states of their listeners. Although the tappers tapped out 120 songs during the

experiments, listeners only identified 3 (a rate of 2.5%). Tappers, however,

incorrectly predicted that listeners would identify 50%.

In the experiment above, the curse of knowledge makes communication

inadequate because people have a complete picture of what they intend to

convey, and this complete picture blinds them to gaps in the information they

actually convey to others. An engineer on a disk drive project described this

kind of communication problem on a new product team: "There were a lot of

specs, but these were only detailed conceptually. They wanted 'something like

this'. As a result the specs get interpreted widely. You end up delivering

something they didn't ask for . . . I was working with one or two people at the

customer organization, then they showed our design to fourteen others who

said: 'Oh My God! We didn't want that'!!" (Dougherty, 1992, p. 189). In

general, this kind of problem is exacerbated by experience and expertise

(Hinds, 1999).

Allen (1977) followed multiple R&D teams who were working to develop

new high-technology products for sophisticated customers using a matchedpair

design. The teams that performed better were more consistent about

communicating. Lower performers were more "irregular" about consulting

internal colleagues and in the middle of the project they "virtually cut off

contact with colleagues outside their project team." Allen notes that the high

performers stayed in closer touch with organizational colleagues throughout

the project, and thus "obtained the necessary information to prevent problems

from getting too far out of control" (p. 103-104). Recent research in network

theory by Burt (1997), Krackhardt (1996), and others re-emphasizes the value

of entrepreneurial networks for individuals and groups.

The problem of communication in organizations is much more formidable

than the normal problem of communication between two people. When two

individuals are communicating face to face, they can use a number of strategies

to repair breakdowns in communication as they occur (Clark, 1996).

Communication in organizations is a more formidable problem because itrequires individuals to communicate through formats (e.g. specifications,

blueprints, memos or budgets) that are relatively impoverished and that

separate the people who are communicating by time and space. Top managers

are particularly likely to fall prey to inadequate communication because their

inside information and their expertise in business may make them particularly

prone to the curse of knowledge; furthermore, they are isolated by structural

and social distance from feedback that might prompt them to repair inadequate

communication (Heath & Walston, 2000).

The problem of inadequate communication will be particularly difficult to

overcome when knowledge is tacit. Von Hippel (1990) notes that information

is often 'sticky'; hard to understand and interpret away from the specific,

applied context where it arose. For example, one production manager may have

difficulty telling another production manager at a different plant why their new

production line is successful because success may depend on a number of

subtle aspects of layout, staffing, and process flow that are hard to notice and

verbalize. Interestingly however, von Hippel implies that people can overcome

the problem of sticky information if they become aware of how much their

knowledge depends on a specific context. We interpret this to mean that

information is sticky and communication inadequate, in part, because of the

kinds of cognitive problems we have considered in this section.

Agency misattributions. Communication is often inadequate in organizations

because people attribute coordination problems, not to inadequate communication,

but to inadequate motivation on the part of their communication partner.

This may lead people to stop communicating prematurely because they think

their partner is not motivated to coordinate with them.

Recall the distinction earlier in this chapter between the two tasks of

organizations: aligning goals and aligning actions. The first is the agency

problem, the second is the coordination problem. We suggest that while both

problems are important in organizations, people are likely to interpret

coordination problems as agency problems, what we label an agency

misattribution. It is harder to imagine someone having different knowledge

than different motivations (Klayman, Loewenstein, Heath & Hsee, 2000). If the

curse of knowledge leads people to believe that they are communicating

something obvious to the other person (recall the tapping game), then when a

listener fails to understand the 'obvious' message, the communicator may be

less likely to assume the other person has different information (a communication

problem) than different motivations (an agency problem). Indeed, even in

situations where differences in knowledge are obvious, like the tapping gamedescribed above, people often attribute coordination failure to agency

problems. In one variant of the tapping game, when listeners failed to

understand the song tapped by the tapper, over 40% of tappers accused their

listeners of "not working very hard" to understand the song. When tappers

confronted a coordination failure, they failed to recognize the difficulty of the

task (because of the curse of knowledge) and instead accused their listeners of

lack of effort; a classic agency misattribution (Morris, Heath & Jost, 2000).

Interestingly, one could argue not only organizational participants, but also

organizational theorists are subject to agency misattributions. Indeed, examples

of potential agency misattributions can be found on both sides of the political/

theoretical spectrum. On the side of the spectrum that assumes individuals

rationally pursue their own self-interest, economists have spent much more

time pursuing the agency problem than considering the equally important

problem of coordination (Milgrom & Roberts, 1992). On the opposite side of

the spectrum, the human relations school has made a career arguing that nonmonetary

factors are important sources of motivations, yet it was famously

embarrassed to find that job satisfaction didn't predict job performance

(Iaffaldono & Murchinsky, 1985; Staw, 1986). Instead, performance is more

determined by coordination mechanisms like organizational routines and

procedures (Herman, 1973; Bhagat, 1982; Staw, 1986). Here, the human

relations school engaged in a kind of agency misattribution because it assumed

that performance problems were driven by dissatisfaction (an agency problem)

rather than organizational routines and procedures (a coordination problem).

These examples suggest that agency misattributions do not depend on a

particular theory of human motivation - whether intrinsic or extrinsic.

However, they do require people to emphasize motivational issues over the

knowledge-based issues associated with coordination.

If agency misattributions play an important role in preventing effective

communication, it may be particularly important that organizations use

techniques like physical location to bring different departments together. By

locating people together, people not only gain more opportunities to

communicate, they may also become more willing to communicate because

they may be less likely to engage in agency misattributions and assume that

coordination failures signal a lack of motivation by their partner. Rich, face-toface

communication may be particularly important in establishing trust (Daft &

Lengel, 1984; 1986). If we are communicating with another person who has

very different knowledge than we do, then we may need to see the puzzled

expressions on their face to understand their questions are motivated by

ignorance rather than spite, ill feelings, or petty resistance.Consider the earliest account of the dilemmas of specialization and integration

in large organizations:

And the whole earth was of one language, and of one speech. [. . . And the people] said,

let us build a city and a tower, whose top may reach unto heaven; and let us make us a name

for ourselves. And the LORD came down to see the city and the tower, which the children

of men builded. And the LORD said, . . . let us go down and confound their language that

they may not understand one another's speech. So the LORD scattered them abroad upon

the face of all the earth: and they did not build the city. Therefore the name of that place

is called Babel; because the LORD did there confound the language of all the earth, and

from thence did the LORD scatter them abroad upon the face of all the earth.

Genesis 11: 1-9

In the traditional theological interpretation, the story of Babel is about the

perils of hubris. However, it can also be read as a parable about the

communication difficulties introduced by division of labor and specialization.

Whenever organizations undertake a sufficiently large task (e.g. building a

large tower), they must partition it into components. However, when a task is

partitioned to form groups of specialists, each group of specialists tends to

develop a different language. As organizations become sufficiently specialized,

the specialists do "not understand one another's speech." Partitioning a task

leads to Babelization, and if the Babbelings are not translated sufficiently,

integration fails. In this section, we suggest that people will exhibit

coordination neglect, not only because of inadequate communication, but also

because of insufficient translation.

The organizational literature has long recognized the potential problems of

specialization. In the classic study by Dearborne & Simon (1958), a group ofbusiness executives were given a complex case study about a failing company.

It was easy to predict how executives would assess the 'core problem' in the

case based on their functional background: the finance executives saw a

financial problem, the marketing executives saw a marketing problem, and the

production executives saw a production problem. Thus specialists see the world

in different ways and stress different content. Although the potential problems

of coordinating specialists are well-described in the organizational literature,

our question is whether lay theories acknowledge these problems and allow

people to overcome them through sufficient translation.

Some authors have been skeptical about whether people will always

recognize the problems of translation. Translation was a central feature of

Allen's (1977) study of gatekeepers in engineering firms. He notes that

translation problems are especially difficult to identify in organizations,

"Anyone who does not speak French knows his deficiency, but very often we

think we know what someone from another organization is saying when in fact

our understanding is very different" (p. 139). According to Allen, the "principal

contribution" of gatekeepers is translation, they convert information "into terms

that are relevant and understandable for the members of their organization"

(p. 166). In his book, Allen devotes separate chapters to the basic communication

problem (Chapter 5, The Importance of Communication Within the

Laboratory) and to the translation problem (Chapter 6, Communications

Among [Sub-] Organizations), yet in Chapter 7, he puzzles over the fact that

both of these problems are rarely solved, "Given the obvious benefits of

internal consulting, it is puzzling that it is so infrequently used" (p. 183).

Specialization would not create such serious problems if people only

realized they need to work differently to translate across specialist boundaries.

However, there is evidence that specialists, rather than understanding the

translation problem, respond to integration issues by reasserting the same

strategies that made them specialists in the first place. For example, specialists

often develop symbols that can convey large amounts of information in a

compact way (e.g. abbreviations, technical language, blueprints, the numerical

summaries of accounting systems; March & Simon, 1958). Unfortunately,

when specialists adopt these abstract symbols they often assume that they are

equally meaningful for other parts of the organization. Bechky (1999) studied

a manufacturing firm that designed and built the complicated pieces of

machinery that are used to produce silicon chips; this firm's competitive

advantage in the market depended on its capabilities in clever engineering and

precision manufacturing. Not surprisingly, the engineers and the manufacturing

people spoke very different languages. When engineers tried to communicate

with the assemblers, they communicated in ways that reinforced the translationproblem. For example, in an attempt to communicate better with the

assemblers, the engineers endeavored to make their engineering drawings as

comprehensible as possible: to "add intelligence to the drawing" (p. 83).

However, the assemblers in the production area did not share the specialized

training that allowed the engineers to read the drawings. Assemblers

"mistrusted" the drawings (p. 68). Bechky implies that this was a problem of

inadequate translation. "The drawing was clear to the [engineers] who created

it, because they worked in the context of engineering drawings all the time . . .

However, assemblers lacked this knowledge" (p. 68).

Indeed, because engineers could not abandon their perspectives, they often

compounded the problem of translation. Engineers made their drawings

"increasingly elaborate" in the hope that this would "clarify" the production

process for the assemblers. "This drove them to greater abstraction in the

documentation, which caused further communication problems . . ." (p. 94).

Here, the engineers behaved a little like the American tourist who tries to

translate his or her ideas in a foreign country by repeating the same English

phrase at a louder volume. The interesting part of the process was that the

engineers often neglected to use other means of translation that would have

been simpler. Bechky (1999) notes that the physical machine was the most

effective translation device. It was "the lowest common denominator" so it

worked "most effectively and quickly to resolve misunderstandings" when

individuals from different departments were trying to communicate. One

assembler noted that, "If we do it from the engineering drawing we can get

confused and make mistakes. Looking at [a physical machine] is easier and

better" (p. 83). Yet even though the physical machine improved translation, the

firm anointed the abstract engineering drawing as the "privileged" form of

communication (p. 93). Bechky's careful study as well as other studies in the

literature (e.g. Henderson, 1991) suggest that people may sometimes neglect

appropriate boundary objects because they fail to recognize the importance of


If people neglect the importance of translation, they may also undervalue

people who act as translators by spanning boundaries among groups within an

organization or outside. Anecdotes suggest that organizations often sack the

wrong people during mergers (Economist, January 9, 1999) - perhaps because

the first people who are fired during downsizings are those who are not clear

members of one department or another (i.e. the very people who are probably

bridging gaps between departments). For example, Dougherty & Bowman

(1995) found that downsizing hindered strategic problem solving because it

broke the networks of informal relationships that innovators use to work outstrategic problems - to acquire support and resources for new initiatives and to

translate the innovation into terms that will be accepted by senior managers.

Although the examples above indicate that the translation problem is

important, there is evidence that even sophisticated observers of organizations

may underestimate the difficulty of translation. Lawrence & Lorsch (1967),

who devoted much of their book to the problems of integration, critiqued

earlier organizational theorists for not realizing the inevitability of translation

problems. According to Lawrence & Lorsch (1967), the "major failing" of the

classical writers on organizations was that they did not recognize that

partitioning the organization into departments would lead each department to

"develop specialized working styles and mental processes," a process they

called 'differentiation' (i.e. "not just the simple fact of partition and specialized

knowledge," but fundamental "differences in attitude and behavior," p. 9). They

argued that differentiation would make it impossible for an organization to

coordinate itself using the simple coordination mechanisms recommended by

classical writers (primarily hierarchy). If Lawrence and Lorsch are correct, then

the problems of translation are not obvious even to many experts who are

thinking carefully about organizations.

Emotional barriers to translation across specialists. To this point, the idea of

translation has been considered at a fairly rational level. If people have trouble

communicating, it is because they don't think to give all the information to

another person, or because they translate insufficiently from one specialist's

language to another. However, this rational approach is insufficient to

understand the complete dynamics of communication and translation because

it ignores an important emotional component of the communication process.

Communication requires trust because both parties must assume that the other

is making good-faith efforts to coordinate (Grice, 1975; Clark, 1996). Here it

is useful to recall our earlier discussion of agency misattributions, because if

people are subject to agency misattributions, then they are likely to assume that

translation problems are a sign that the other person, with whom they are trying

to coordinate, is operating in 'bad faith'.

Agency misattributions are quite likely when specialists communicate.

Specialists may be somewhat suspicious of others' motives to begin with

because they come from different groups with distinctive backgrounds,

preferences, and language. Research in social psychology on 'minimal groups'

has shown that dividing a group based on even trivial distinctions (e.g. liking

abstract art by Klee vs. Kandinsky) has surprisingly quick effects on group

dynamics; people allocate more resources to their own group, talk up the

qualities of their own group, and denigrate those of the other group (Tajfel1970; Brewer, 1979). The minimal group experiments provide empirical

parables about how easy it is to produce ethnocentrism and emotional conflict,

but in organizations, differences among groups are anything but minimal.

When specialists try to coordinate with each other, their suspicions about others

may enhance agency misattributions; causing them to attribute integration

problems to bad motives by the other party. True, group dynamics may produce

real agency problems, but we predict that specialization may make groups

suspicious of one another's motives even in situations where the different

groups are actually quite motivated to work toward the same goal.

Consistent with the idea that agency misattributions are common among

specialists, Lawrence & Lorsch (1967) noted that people "personalize the

conflicts that arise with representatives of other organizational units. Of course

they know logically that an organization needs different kinds of specialists, but

they forget the full meaning of this when they run into a particular person who

is "impossible to work with." Then they all too readily turn to an explanation

based on personality traits that writes off the individual as an oddball and

justifies them in withdrawing from the conflict or forcing it" (p. 217).

One researcher who has studied new product introductions argues that

although outsiders may believe agency problems contribute to new product

failures, they are not, in fact, very common: "From the outside looking in, one

can see the conventional stereotypes for each department: technical people

never settle on a design, field people are short term, manufacturing people

always say no, and planning people are conceptual. But from the inside looking

out, each thought world is truly concerned with the successful development of

the product, and each has an important insight into the product or market that

is essential to a new product's development" (Dougherty, 1992, p. 191).

However, although each department is motivated to develop the product, the

departments may fail to coordinate because they translate their goals and

expectations insufficiently. "Technical people, for example, expect field people

to tell them exactly what customers want in the design. Field people, however,

cannot identify these 'specs' because [they think] product innovation involves

meeting shifts in customer needs, so they expect technical people to produce

alternative designs quickly" (p. 189). Agency misattributions may lead team

members to ignore the translation problem because they assume the other party

has the wrong incentives. According to Dougherty, differences in specialization

can preclude optimal integration "by producing severe frustrations and

withdrawals into separate thought worlds." If people suffer from agency

misattributions they are unlikely to take the time to address such translation

problems because they assume the other person is not truly motivated to

cooperate.Summary: Inadequate Communication and Insufficient Translation

Although organizations can integrate their efforts in many ways, the most

important mechanism of integration, particularly in complex or uncertain

environments, is for units to communicate with each other on an ongoing basis

(March & Simon, 1958; Thompson, 1967). In this section we have discussed

two problems that may hinder individuals from taking advantage of this

integration mechanism: inadequate communication and insufficient translation.

Communication is difficult in general. People are prone to the curse of

knowledge which makes it difficult for them to take another person's

perspective well enough to communicate adequately. In face-to-face conversations,

people have a variety of means of repairing instances of inadequate

communication, but in organizations, where much communication takes place

across time between individuals who do not interact face-to-face, the problems

of inadequate communication become more significant (DeSanctis, Staudenmayer

& Wong, 1999).

Organizations complicate the basic problem of communication, because they

require people to communicate across differentiated groups of specialists.

Thus, if people translate insufficiently across specialized languages or 'thought

worlds', coordination will be further hindered. The problems of communication

and translation are magnified because differentiation and specialization

may leave people suspicious of one another so that they are more likely to make

agency misattributions when they encounter coordination problems; attributing

problems not, as they should, to lack of communication, but instead to

misaligned motivations.


In this chapter, we have argued that people in organizations often exhibit

coordination neglect. Even when they desire to coordinate with others, when

they are thinking actively about the problem, and when coordination does not

exceed their computational abilities, people may have blind spots that make

them likely to fail in their coordination attempts. In the paper, we have focused

on two different aspects of coordination neglect:

(1) Partition focus and component focus. People focus on the process of

partitioning a task more than the process of integration and they tend to

focus on individual components when they try to diagnose problems or

intervene to provide a solution.

Inadequate communication and insufficient translation. People do not

communicate adequately in general and they fail to realize the additional

problems of translating across differentiated specialists.

Although the figures in this paper indicate how both sets of problems arise from

the basic process of division of labor in organizations, it is possible to think of

these two sets of problems operating in a two-stage temporal sequence. In the

first stage, an organization must plan the division of labor along with any

integration mechanisms it deems necessary. In the second stage, the

organization must integrate its efforts in an ongoing basis. Presumably, if

people make errors in one stage they can offset them with superior performance

in the other; effective planning may reduce some of the demands of

implementation, and skilled implementation may overcome some of the

problems of inadequate planning. Yet the evidence above suggests that

organizations will experience predictable problems in both stages.

The examples in this chapter suggest that coordination neglect plays a role

in many important decisions. Individuals indeed have gaps in their lay theories

of organizing. Note, however, that these gaps are seen even in managers and

scholars who are quite sophisticated about the problems of organizing. This

observation is consistent with research in individual decision making that has

suggested that any bias that can be documented in naive individuals can also be

documented, in a more subtle form, in experts (Kahneman, Slovic & Tversky,

1982). It is interesting that many examples of partition focus or inadequate

communication are found with sophisticated managers in otherwise successful

companies. Even sophisticated organizational theorists have neglected these

problems at times (e.g. see critiques of the organizational literature by

Lawrence & Lorsch, 1967; Cohen & Gooch, 1990; Weick & Roberts, 1993;

Hinds & Kiesler, 1995). Such examples suggest that the problem of

coordination neglect is not trivial, particularly whenever people encounter

problems that are, to them at least, novel (Chandler, 1962; Brooks, 1979;

DeMarco, 1997).

Bad (and Good) Ways of Repairing Coordination Neglect

While we have emphasized the problem of coordination neglect, we also want

to highlight that organizations may create other problems if they adopt overly

simple cognitive repairs for coordination neglect. For example, if organizations

attempted to repair partition focus or component focus by decreasing the

division of labor, they could easily create other problems such as lack ofrequisite variety or expertise (Weick, 1983; Nemeth & Staw, 1989). Similarly,

people may overemphasize the ability of hierarchy to coordinate complex

processes. In group situations, students typically want to simplify coordination

by 'appointing a leader'. Hierarchy appeals to our fascination with people as

the source of action (Weber, Rottenstreich, Camerer & Knez, 1999; Meindl,

Erlich & Dukerich, 1985), but it is likely to be ineffective in complex, uncertain

environments (Lawrence & Lorsch, 1967; Hinds & Kiesler, 1995).

It's clear that organizations can over-apply the lessons they learn about

individual shortcomings. For example, thanks to the research attention on the

benefits of cross-functional teams, many organizations have tried to repair

inadequate communication by using team meetings as a generic repair for

every coordination problem. However, more communication is not always

better. Many managers in firms today complain of the length and frequency of

meetings (which interfere with their ability to get their own functional work

done) (Perlow, 1995; Staudenmayer, 1997).

In fact, thoughtful organizations solve problems of coordination by

integrating the efforts of their members in ways other than direct communication.

For example, software teams sometimes enforce integration by using

special processes that force integration among separate workers. In some firms,

software teams do a nightly 'build' to put together the modules for the entire

program (Staudenmayer, 1997; Cusumano & Selby, 1995). Individual developers

can decide whether or not to submit an updated version of their module

to the nightly build, but if they do so, they must take care that their module

doesn't 'break the build', i.e. produce problems for a module other than their

own, causing the overall program to crash. In this procedure, individual

programmers are not forced to communicate with others to ensure that their

module doesn't break the build, yet they are prompted to test their code and to

anticipate what problems they may create for other modules when they change

their own.

As another source of repairs for coordination neglect, Tom Malone of MIT

and a group of colleagues have developed a process to suggest alternative ways

of solving coordination problems (Malone et al, 1999). They collect examples

of how different organizations perform similar processes of coordination, and

organize these examples in an on-line 'process handbook'. They analyze

processes at various levels of abstraction, so they capture both the details of the

specific processes as well as the 'deep structure' of their similarities. As a

result, managers can explicitly represent the similarities and differences among

related processes and they can more easily generate alternatives to solve a

particular coordination problem.

Although there is ample evidence of coordination neglect, organizations that

attempt to repair coordination problems in an ad hoc way may find that their

would-be repairs create additional difficulties. There are solutions to these

problems, but they will require careful attention to the underlying requirements

of integration and to overcoming the cognitive barriers we have identified.


On a lighter note, perhaps the most extreme evidence for coordination neglect

is provided by conspiracy theories. In the typical conspiracy theory, a diverse

set of military, industrial, and government agencies coordinate seamlessly over

long periods of time despite organizational and geographic barriers. Extreme

versions of conspiracy theories feature coordination across planets and species

(a serious neglect of translation problems). If individuals fully understood the

difficulty of coordination, it seems unlikely that they would be quite so facile

in assuming the level of coordination present in the typical conspiracy theory.

Conspiracy theories also play a role in the day-to-day analysis of

sophisticated experts. In Robert Jervis' (1976) brilliant book on the psychology

of international relations, he devotes a set of chapters to 'Common

Misperceptions' in international relations. The first of these chapters is entitled

'Perceptions of Centralization' (pp. 319-342). According to Jervis, a common

misperception is to see the behavior of others as "more centralized, planned,

and coordinated than it is" (p. 319). For example, during World War II, "many

observers believed the German fifth column [espionage force] was largely

responsible for the Allies' difficulty in mobilizing and the swift German

victories . . . Later investigation showed that the fifth columnists had done very

little and that the incidents attributed to them were caused by Allied

disorganization" (Jervis, 1976, p. 322-323).

Such examples suggest that coordination neglect, even in its most extreme

forms, may play an active role in how we approach and interpret the

coordination problem in organizations.


1.      What is the article about?

2.     What is the gap the article wants to fill or the problem it wants to address?

3.     What is/are the basic argument(s)

4.     What is the conclusion?

5.     What do you learn from the article?

6.      Finally, your general reflection?

Recently Asked Questions

Why Join Course Hero?

Course Hero has all the homework and study help you need to succeed! We’ve got course-specific notes, study guides, and practice tests along with expert tutors.

  • -

    Study Documents

    Find the best study resources around, tagged to your specific courses. Share your own to gain free Course Hero access.

    Browse Documents
  • -

    Question & Answers

    Get one-on-one homework help from our expert tutors—available online 24/7. Ask your own questions or browse existing Q&A threads. Satisfaction guaranteed!

    Ask a Question
Let our 24/7 Management tutors help you get unstuck! Ask your first question.
A+ icon
Ask Expert Tutors You can ask You can ask ( soon) You can ask (will expire )
Answers in as fast as 15 minutes
A+ icon
Ask Expert Tutors