Beyond boundary spanning : managing external dependence in product development teams

/^^^^'^

<o'

'"-f o, •rec*'

"%

APR 181989

14
3-

WORKING PAPER
ALFRED P. SLOAN SCHOOL OF MANAGEMENT

Beyond Boundary Spanning: Managing External
Dependence in Product Development
Teams

By

Deborah G. Ancona
David Caldwell

WP 2103-89

March 1989

MASSACHUSETTS

INSTITUTE OF TECHNOLOGY

50 MEMORIAL DRIVE

CAMBRIDGE, MASSACHUSETTS 02139

Beyond Boundary Spanning: Managing External
Dependence in Product Development
Teams

By

Deborah G. Ancona
David Caldwell

WP 2103-89 March 1989

Beyond Boundary Spanning: Managing External Dependence in
Product Etevelopment Teams

Deborah Gladstein Ancona

Sloan School of Management

M.I.T.

Cambridge, MA. 02139

David Caldwell

Leavey School of Business and Administration

Santa Clara University

Santa Clara, CA. 95053

Beyond Boundary Spanning: Managing E>±emal Dependence in
Product Development Teams

ABSTRACT

Data from 409 members of 45 new product teams in five hi^-
technology conpanies indicate the existence of four sets of boundary
management activities which can be labeled Ambassador, Task Coor-
dinator, Scout, and Guard. These activities are related to the
frequency, type, and destination of communications between team members
and others. This paper describes these activities and examines the
irrpact of individual and environmental variables on the frequency in
vAiich individual team members engage in them.

Introduction

The development of new products is a critical challenge for many
organizations. For conpanies in highly conpetitive technology-based
industries, the issue is not solely the introduction of new products
but also how to accelerate the product development process (David,
1984) . One popular technique for shortening product development cycles
is the use of teams. Teams are recommended as a way to improve
coordination within the organization (Hackman & Walton, 1986; Kanter,
1983; Kazanjian & DrBzin, 1987) and to speed product development by
promoting the closer coupling of organizational functions necessary to
move from a sequential to a parallel development process (Van de Ven,
1986) . For these advantages to be realized, new product team members
must effectively interact with other functional areas and h i.erarchical
levels in the organization. In this paper, we attempt to describe the
pattern of activities in which team members engage in dealing with
outsiders and identify some of the factors that contr.ibute to the
frequency in vAiich individual team members engage in these activities.
Boundary Management in New Product Teams

To be successful, new product teams must obtain information,
resources, and support from others, both inside and outside of the
organization, use that information to create a viable product, and
finally transfer the technology and enthusiasm for the product to those
who will bring it to market (Ancona & Caldwell, 1987; Burgelman, 1983;
Quinn & Mueller, 1963) . This makes the new product team highly
dependent upon others, and suggests that an important way of under-

2

standing the performance of these teams lies in examining how they
manage relationships with other groups.

Some previous researchers have partially taken this approach and
examined the pattern and flov; of scientific communication into R&D
teams (Allen, 1984; Ebadi & Dilts, 1986; Ebadi & Uttertoack, 1984; Katz,
1982; Tushman, 1977, 1979) . These studies have documented a relation-
ship between the acquisition and transmission of information by
boundary-spanning individuals and performance of the group. For
exanple, greater communication with organizational colleagues outside
the group occurred in high performing R&D project teams than in low
performing teams (Allen, 1984) . Tushman (1979) found that communica-
tion in high-performing development teams followed a two-step pirocess,
with communication "stars" first obtaining information from outside the
group and then translating that information and transmitting it to the
group.

Although extremely useful, this research has focused on the
inportation of technical information into the group and not on the full
range of interactions between a team and its environment. A broader
conceptualization of boundary activities would include the ccnplete set
of activities necessary to build support for the embryonic product,
shape the demands of others, and coordinate the product's development
with other groups. As Bagozzi (1975) has observed, exchange theory
suggests that when a system is characterized by an interconnecting web
of relationships, such as those necessary for product development,
broad conceptualizations of the boundary management process are useful.

3

This broad view of boundary activity is much closer to that
frequently taken by organization theorists. Those drawing on a
resource dependence paradigm, open systems theory, or a strategic
management approach view the organization not simply as an adaptor to
environmental contingencies, but also as an entity that can mold,
enact, and manage its dependence on outsiders (Adams, 1976; Astley &
Van de Ven, 1983; Pennings, 1980; Pfeffer, 1972; Van de Ven & Walker,
1984) . For exairple, a more inclusive description of organizational
boundary activities was proposed by Adams (1980) . He outlined five
classes of activities: acquisition of organizational inputs, the
disposal of outputs, searching for and collecting information, repre-
senting the organization to outsiders, and buffering it from external
threat and pressure. Such a typology seems to reflect many of the
boundary activities in which a new product team must engage during the
product development process.

The first purpose of this study is to describe the range of
boundary activities i". v/nich nsw product team members engage. Our
genei-al hypothesis is that due to the conplexities of the new product
team's dependence on other parts of the organization, team members will
display complex patterns of boundary activity similar to those of
organizations .
Factors Associated With Boundary Activities

The second purpose of this study is to investigate the factors
which influence the type and frequency of team member boundary
activities. Specifically, we hypothesize that the types and frequency
of boundary activities in which individual team members engage will be

influenced by both the nature of the task environment and the charac-
teristics of the individual.

Task Environinent Characteristics. Although a great deal of
research has chronicled the relationship betv;een environmental condi-
tions and internal team structure (c.f. Gresov, 1988; Tushman & Nadler,
1978; Tushman, 1977, 1979), little work has examined the relationship
between the environment of a task group and its external initiatives
(see Allen, 1984 as an exception) . Further, much of this work has used
environmental uncertainty has the sole measure of the conditions a
group faces (Gresov, 1988) . However, reviews of the research and
development and organizational literatures indicate that more specific
descriptions of environments are both possible and desirable. For
example, uncertainty can be broken down into technical, organiza-
tional, and market uncertainty (Abemathy & Clark, 1985; Argote, 1982;
Galbraith, 1982). At the organizational level, Aldrich (1979) des-
cribes six characteristics of an environment: rich or lean, homo-
geneous or heterogeneous, stable or unstable, concentrated or dis-
persed, consensus or dissension, and the extent of turbulence.

These more detailed descriptions of environments are useful in
understanding tlie circumstances in which highly interdependent teams
must operate. Partially based on them, we propose a set of charac-
teristics that can influence the particular boundary activities in
which product team members engage.

The first task environment characteristic is the extent to v^ich
the new product is revolutionary as opposed to an incremental improve-
ment over an existing product. This variable corresponds to technical

5

uncertainty, or lack of infonnation about how to do the task (Crawford,
1974; Daft & Macintosh, 1981; Houston & Holmes, 1974). As Dewar and
EXitton (1986) have observed, different models may be necessary to
explain radical and incremental innovations. When the product is
revolutionary, the team therefore may need to go outside its borders
more frequently to get technical or conceptual assistance than if the
product uses a known technology (Brown & Utterback, 1985) .

The second variable which may influence the boundary activities of
the team members is the experience of the organization in developing
similar products or using similar technologies. Low levels of ex-
perience correspond to organizational uncertainty, or lack of standard
operating procedures and information about cause-effect relationships
(Galbraith, 1977; Thonpson & Tuden, 1959). When standard operating
procedures do not exist, informal boundary activity may be necessary to
aid coordination.

The third environmental characteristic that may influence external
activity is the e.±ent of competition the new p-^oduct v/ill face. High
levels of conpetition may reduce the predictability of others in the
environment and increase uncertainty (Abemathy & Clark, 1985; EXincan,
1972; Leblebici & Salancik, 1981) , thereby prompting team members to
broaden their attempts to collect information from others.

The fourth environmental characteristic is the stability of the
market for the new product. This is similar to the rate of environmen-
tal change (Aldrich (1979) . Products developed for stable markets may
have larger lives and more predictable development cycles than those
developed for changing markets. High rates of change may also focus

6

the group on external scanning for information that v;ill keep it
updated on current trends.

A final environmental variable is the availability of resources to
the team. Personnel, budget, and equipment allocations detennine
v^ether the team faces a rich or lean environment (Aldrich, 1979) . A
rich environment, where resources are both abundant and easily
available may reduce the necessity of certain boundary activities.
Similarly, lean environments may prompt the team to search for
resources throughout the organization.

In addition to the specific characteristics of the task environ-
ment, we propose a final factor which may influence the pattern of
individual boundary activity. Ancona and Caldwell (1987) argue that
the demands of the product development task change from: 1) exploring
technical ideas and product potential; 2) to exploiting the information
and resources the team has garnered and moving to efficient development
of the product; 3) to exporting the expertise and enthusiasm for the
product to others. The stage of the product development process may
pose a specific set of task demands that may shift team member's
external communications. Such changes may translate into shifting
boundary activities.

Individual Characteristics. We propose that two general classes
of individual variables will be related to boundary activities. The
first of these is the individual's role in the organization. Following
Allen (1970) , we propose that formal team leaders will be most likely
to engage in boundary management activities. This will be particularly
pronounced for those activities that involve dealing with upper

7
management .

The second class of individual variables relates to the product
team member's experience. Taylor (1972) observes that individuals
become communication gatekeepers only when they have sufficient
ejqserience in the organization and when they are engaged in technical
work. We propose a similar hypothesis, namely that individuals with
longer experience in the company will display higher levels of boundary
activity since these individuals have had more of an opportunity to
develop contacts and to understand the corporate culture.

In addition to the years of experience, we propose that boundary
activity will be related to the nature of experiences an individual has
had over his or her career. Substantial research suggests that
similarity in background or personal characteristics influences
attraction and frequency of communication (Good & Nelson, 1971;
Newcomb, 1961) . This suggests that individuals will more frequently
interact with people from the same functional area. Based on this
reasoning, it is li>:ely that individuals who have had experience in
particular functional areas will be more likely to engage in boundary
activities than individuals ^o do not have the same experiences.

A final variable that may influence the boundary activities of an
individual is the extent to which that person is integrated into the
project. Since individuals are frequently assigned to more than one
project or have multiple responsibilities, there is variability in the
amount of time people spend on a project. Some people are assigned
nearly full time to the project; others have more diverse respon-
sibilities. The amount of time the individual spends on the project

8

may influence the type of boundary activities in v^ich the individual

engages.

Summary and Hypotheses

In summary, this research has two goals. First, we seek to
identify the range of boundary activities that new product team members
use to manage their dependence on those outside the team's boundaries.
Second, we hypothesize that particular task environment and individual
variables will be related to the frequency in v^ich product development
team members engage in the pattern of boundary activities. More
specifically, we hypothesize that when the new product is revolution-
ary, new to the company, and in a conpetitive and highly unstable
market, individuals will engage in a different pattern of boundary
activities than wtien the product and its market are more stable and
predictable. In addition, the stage of the product development process
may influence the frequency of boundary activity. Similarly, we
hypothesize that individuals v;ho are team leaders, have served on many
teams, have substantial experience in the company, spend a large
percentage of time on the project, and have experience in functions
other than engineering and research will be more active in boundary
management than individuals who do not have these characteristics.

Methods

Two separate data collections were undertaken. The first in-
cluded: 1) set of interviews with 38 new product managers; and 2) logs
kept by all members of two teams for a two-week period. The purpose of

m

9

these interviews and logs was to derive a reasonably conplete set of
boundary activities and to provide an initial test of some of the
measures used in the second study. The second study involved the
leaders and members of 45 new product teams. In the second study
questionnaires and interviews were used to investigate the structure of
boundary activities and the factors that contribute to them. In
addition, all team leaders were interviewed as were senior managers in
each conpany who provided data on product characteristics.
Study 1

Subjects. Thirty-eight leaders of new product teams in seven
corporations in the conputer, integrated circuit, and analytic in-
strumentation industries were interviewed. Interviews ranged between
one and eight hours, with an average duration of approximately three
hours. All of taie leaders had at least three years experience in high-
technology organizations and all but seven had managed new product
teams prior to the one they were currently describing. Six of the
leaders were higher level managers vdio were responsible for multiple
teams.

Identification of Boundary Activities. During part of the
interview, each leader was asked to describe the interactions that he
or his team members had with other individuals outside the new product
team. We asked leaders to be as inclusive as possible in their
descriptions and to include all forms of communication including
meetings, one-on-one discussions, telej±ione calls, and conputer
messages. Also noted were those actions aimed at controlling the
team's boundary, such as avoiding meeting with others until a par-

10

ticular decision had been made or preventing outsiders from meeting
with the team.

In addition, two teams of seven and eight members respectively,
were asked to keep logs of all of their external activities over a two-
week period. The logs required members to record all communications
with outsiders including the person with whom they communicated, the
hierarchical level of that person, his or her functional group, and the
purpose of the communication.

Based on an inspection of the interview transcripts and the logs,
a set of 24 boundary activities was identified for use in Study 2.
These items included actions such as persuading others to support the
team, attempt ing to acquire resources for the team, bringing technical
and political information into the group, and keeping progress of the
group secret.
Study 2

Subjects. Questionnaires were distributed to new product team
members and leaders of forty-five teams in five corporations in the
conputer, analytic instrumentation, and photographic industries. Of
the 450 questionnaires distributed, a total of 409 questionnaires were
returned, yielding a response rate of approximately 89 percent.
Response rates were approximately equal across the set of conpanies and
the final total of respondents per company varied from 39 to 129. In
addition to the data collected by questionnaire, each team leader was
interviewed for approximately one hour. The average age of the sanple
was 38.6 years, 88 percent were male and 75 percent possessed at least
a four-year college degree. Approximately 77 percent of the sample

11

were from the engineering or R & D function; the remaining were spread
across manufacturing, marketing, sales and service, and product
management.

Measurement of the Boundary Activities. The 24 boundary ac-
tivities identified in Study 1 were converted to questionnaire items.
Respondents were asked to indicate on five-point Likert scales the
extent to v^^ich they felt each of the items was part their respon-
sibility in dealing with people outside the team.

In addition to reports of boundary responsibilities, data were
collected on individual communication with outsiders in other function-
al and hierarchical areas in order to provide some validation of the
boundary activity scores. Individuals indicated the extent to v^iich
their time on the project was spent working alone (x = 48.25 percent;
s.d. = 28.37), working with others on the team (x = 37.71 percent; s.d.
= 26.47) . and working on the product with individuals not assigned to
the team (x = 13.90 percent; s.d. = 12.02) . A second set of communica-
tion measures asked the respondent to indicate on an anchored 6-point
scale (1 = Not at all; 6 = Several times per day) the frequency with
which the person communicated with people in other functional areas
including: manufacturing (x = 3.01; s.d. = 1.76), marketing (x = 2.08;
s.d. = 1.39), sales and service (x = 1.89; s.d. = 1.20), R & D (x =
2.27; s.d. = 1.35), the top management of R&D (x = 2.07; s.d. = 1.28),
top division management (x = 1.59; s.d. = 1.11) and top corporate
management (x = 1.30; s.d. = .75).

Measurement of Task Environment Characteristics. During the
interviews with the new product team leaders, a number of structured

12

questions were asked regarding the task environment of the new product
team. Team leaders were asked to indicate on a five-point Likert scale
the following: 1) the extent to v^ich the product used a revolutionary
technology (1 = Same as other products, 5 = revolutionary; x = 3.16,
s.d. = 1.10); 2) the experience of the company in developing similar
products (1 = Never done it before, 5 = Do it all the time; reversed
scoring, x = 3.09, s.d. = 1.47) ; 3) the amount of external competition
the product faces (1 = None, 5 = A great deal; x = 3.99, s.d. = .93) ;
4) the rate at which the targeted market is changing (1 = Not at all, 5
= very rapidly; x = 3.21, s.d. = .84) ; 5) the availability of person-
nel, equipment, and funds (three items are averaged to determine the
richness of the environment; 1 = rarely available, 5 = no problem
securing; reversed scored, x = 2.17, s.d. = .37).

In addition, the team leader indicated the stage of the product
development process. Teams were categorized as being in the creation,
development, or diffusion stage of product development (Ancona &
Caldwell, 1987) . Teams in the creation phase had not yet developed
complete product specifications. Teams in the developmient stage were
completing the technical tasks necessary to produce a working proto-
type. Teams in the diffusion stage had met project specifications and
were in the process of transferring the prototype to those v^o would
manufacture and distribute it.

Measurement of the Individual Correlates of Boundary Activity.
The questionnaire also contained items regarding individual background
and experience. Three of these were the number of new product teams on
which the respondent had previously served (x = 9.38, s.d. = 16.30),

13
the years that the respondent had been with the coirpany (x = 11.23;
s.d. = 9.16) , and the percentage of their work time v^hich was devoted
to the project (x = 59.62 percent; s.d. = 37.98). In addition,
respondents indicated whether or not they had experience in R & D,
marketing, and manufacturing (0 = no, 1 = yes) .

Results

Patterns of Boundaty Activities

The first general research question addressed the existence of
multiple independent dimensions of boundary activity to deal with the
coirplex set of dependencies that the new product teams faces with other
parts of the organization. To investigate this, a principal conponent
analysis with a varimax rotation was performed on the 24 boi.'ndary
activity items included in the questionnaire. Table 1 summarizes this
analysis and shows the item loadings greater than .35.

Insert Table 1 About Here

Although some cross-loading exists, four interpretable factors
emerged. The first factor is defined by 12 items with loadings greater
than .50. The items appear to reflect primarily buffering and repre-
sentational activities. Exanples of buffering included such things as
absorbing pressures and protecting the team. Representational ac-
tivities included persuading others to support the team, keeping higher
levels informed of team activities, informing the team of coirpany

14

strategy and of potential threats or opposition the team roay face, and
lobbying for resoitfces. Since these activities contain both protective
and persuasive goals, we label the set of behaviors identified by this
factor as ambassador activities.

Five items with loadings greater than .50 primarily define the
second factor. These items represent interactions aimed at coordinat-
ing technical or design issues. We term this activity set, task
coordinator. Examples of behaviors in this set include discussing
design problems with others, obtaining feedback on the product design,
coordinating and negotiating with outsiders.

The third factor is made up of four items with loadings greater
than .50. This factor is labelled scout because it describes ac-
tivities that involve general scanning for ideas and information about
the competition, the market, or the technology'. While task coordinator
activities appear to be aimed at handling specific technical and
coordination issues, scout activities appear to incorporate more
general scanning.

The fourth set of activities are guard activities. The three
items that comprise this factor all involve controlling the team's
release of infonnation. Activities here are aimed at keeping informa-
tion within the team's boundaries in orTder to protect the team or
present a specific image of the team to outsiders.
Further analysis of Boundary Activity

To ensure orthogonality among the four boundary dimensions, factor
scores were computed and used in subsequent analyses. As a partial
test of the validity of the structure of the boundary activities, and

15
to learn more about the destinations of the four types of boundary
activity, correlations betu'een the individual boundary factor scores
and the individual communication variables were calculated. Although
not shown as a table, the ambassador factor scores are correlated with
frequency of communication with manufacturing (r = .12, p < .05),
marketing (r = .25, p < .001), top R&D management (r = .15, p < .01),
top division management (r = .20, p < .001) , and top corporate
management (r = .15, p < .01). In contrast, the task coordinator
factor was associated with higher levels of communication with
manufacturing (r = .34, p < .001) , and R&D below the top management
level (r = .10, p < .05) . Individuals with high scores on the scout
factor engage in more frequent communication with R&D below the top
management level (r = .10, p < .05), marketing (r = .29, p < .001) and
sales and service (r = .16, p < .01) . The guard factor was unrelated
to frequency of communication.

The relationships noted above lend support for the validity of
differentiated boundary activity sets. Individuals who believe
themselves responsible for protecting the team and representing the
team to outsiders have the broadest set of communication links. Those
individuals reporting responsibility for technical problem solving and
coordination interact with lateral groups particularly manufacturing
and R&D, while those who are scanning for competitive, market, and
technical ideas, have more frequent interaction with R&D, marketing
and sales.

Not only do identified boundary factors relate to the type and
destination of their communications, they also relate to how much time

16
individuals spend working alone on the project, working with other team
members on the project, and working with non-members on the project.
Ambassador factor scores are negatively related to percentage of time
spent alone (r = -.18, p < .001), but positively related to time spent
with team members (r - .11, p < .05), and outsiders (r = .14, p < .01).
Similarly, task coordinator factor scores are negatively correlated
with time spent alone (r = -.13, p < .01) , but positively correlated
with time spent with outsiders (r = .29. p < .001) . Scout and cfuard
factor scores are unrelated the amount of time team members spend
working with others.
Factors Influencing Boundary Activity

The second goal of this study was to identify the environmental
and individual variables that influence the extent to which team
members engage in boundary activities. To identify the relationships
between boundary activities and the other variables, separate regres-
sions were computed for each of the four, independent boundary activity
factors. For each factor, a hierarchical regression model was
developed. First, dummy variables representing the companies in the
sample were entered into the equation to control for company differen-
ces. Second, the five task environment variables were entered: degree
to v^ich product is revolutionary (REVOL) ; experience of the conpany
(EXPER) ; degree of competition ((XMPET) ; availability of resources
(RESAVL) ; and market stability (MSTABIL) . Third, dummy variables
representing the phase of the product development process were entered.
Finally, the individual variables were entered. These were: whether or
not the individual was a team leader (LEADER) ; the number of teams on

17
which the individual has served (NUMTEAMS) ; years of experience in the
ccmpany (YEARSCD) ; the percentage of time the individual spends working
on the project (PCTPRQJ) and v^ether or not the individual has
experience in R & D (EXRD) ; manufacturing (EXMFX3) ; or marketing
(EXMKT) . Following the entry of each set of variables, the incremental
increase in the variance explained was examined.

Table 2 shows the correlation matrix for all variables in the
regressions and Table 3 shows the regression results. As shown, the
conplete set of variables explain significant amounts of the variance
in all four of the boundary activity factors. Looking at each set of
variables, a number of findings are worthy of note.

Insert Tables 2 and 3 About Here

First, the dummy variables indexing the company explained sig-
nificant variance on both the ambassador and guard factor. This
suggests that particular corporate environments may encourage certain
external activities more than others.

Second, task environment variables do not seem to have a large
irtpact on boundairy activity. The exception here is with regard to
resource availability. In lean environments (low resource avail-
ability) individuals display higher levels of scout activity and lower
levels of cfuard activity, than in rich environments.

Third, the phase of the project is related to the level of
activity on the ambassador dimension. Although not shown here, an
inspection of ambassador activity across the three stages of the

18

product development process suggests that such activity is highest
during the creation and diffusion stage, and lower during the develop-
ment stage.

Finally, the individual variables have a significant inpact on all
four factors. Whether or not an individual is a team leader is
strongly related the extent to which the individual assumes respon-
sibility for both ambassador and task coordinator activities. Longer
tenure in the organization is related to reported responsibility for
task coordinator activities, while the percentage of work time com-
mitted to the team is positively related to responsibility for guard
activities. Finally, there appears to be support for the hypothesis
that similarity in background and experience is related to frequency
and type of communication. Experience in manufacturing suggests high
levels of task coordinator but low level of scout activity while ex-
perience in marketing suggests high levels of ambassador and scout
activity, but low levels of task coordinator activity.

Discussion

Despite the increasing emphasis on the use of teams in facilitat-
ing product development, there is little concrete evidence regarding
how particular team activities or structures can enhance the product
development process. In this study we have concentrated on identifying
the pattern of cross-boundary activities a team uses to manage its
dependence on other parts of the organization. Our goal was to move
beyond examining the frequency of external communication in order to

19
describe the varied types of cornmunication needed to handle complex
interdependence. Several findings from this research may be useful in
\jnderstanding and inproving new product team performance.

First, the data suggest that the boundary activities of new
product teams are multi-faceted and complex. Boundary spanning goes
beyond the importation of technical information and includes handling a
broad range of inputs and outputs, as well as functions such as
protecting the team and representing the team to others. This model of
boundary spanning is much closer to that taken by organization theor-
ists than group researchers. More specifically, teams attempt not only
to exchange technical information but also to model the environment of
the organization, facilitate coordination with other organizational
units, and to mold the views, expectations and behaviors of those who
control critical resources.

From the resource dependence perspective, the most critical
determinant of an entity's viability is the ability to obtain critical
resources (Astley & Van de Ven, 1983; Pfeffer & Salancik, 1978). Our
expanded view of boundary spanning suggests that technical know-how is
only one critical resource for new product teams. In addition to
existing in a technological environment, teams belong to an organiza-
tional environment in which there is competition for time, money,
personnel, influence, support and information. Team members in this
study worked to obtain these organizational resources, v^ile simul-
taneously protecting other team members who could then concentrate on
the core technical work. Clearly, the management of this broader
organizational dependence needs to be added to that of technological

20

dependence in order to understand the requirements of boundary span-
ning. In fact, as 'global corpetition grov/s and authority within the
organization gets pushed down the hierarchy, functional groups must
work more closely together (Galbraith, 1982; Malone, 1987). This means
that the management of this organizational dependence may well become
increasingly inportant.

Second, the data indicate that different types of boundary
activities are needed for different types of dependence. Vertical or
hierarchical dependence (Gresov, 1988) , whereby top levels of the
organization distribute personnel, funding, equipment, legitimacy, and
priority, is handled through ambassador activity. Ambassador activity
appears to be somewhat political; identifying threat and opposition in
top levels of the organization and working to build support from these
powerful outsiders. Horizontal or lateral dependence, whereby other
functions have critical information, expertise, and creative ideas, is
harnessed through task coordinator and scout activity. Task coor-
dinator activity is the most time-consuming, requiring the largest
aiTKDunt of time devoted to external, as opposed to internal, activities.
Coordination and synchronization with other technical groups appears to
require more focused communication than the more general scanning done
through scout activities. Scout activities aimed at obtaining competi-
tive, market, and technical ideas illustrate yet a third set of
behaviors, aimed at a different part of the organization (those with
access to information about external competitive trends) , to handle a
second form of lateral dependence. In contrast, guard activities
represent a means of decreasing, rather than meeting, dependence

21
through the control of information flow out of the group.

Third, the findings indicate that team members do not vary their
boundary activities based on characteristics of the task environment.
One clear exception to this is the response to resource scarcity — yet
even here teams respond through an increased openness in boundary
permeability (scouting allows for information to flow into the group
while low guard activity means that information also flows out of the
group) and not the predicted ambassador activity aimed at obtaining
resources. Future research will need to determine whether this lack of
adaptation of boundary activity to environmental conditions is detri-
mental to team performance, or whether the teams in this sanple are
already exhibiting high levels of boundary activity to deal with
organizational dependence and thus task environment demands are also
met.

Finally, the data suggest that individual characteristics are
inportant in understanding who assumes boundary activity. Not surpris-
ingly, leaders display higher levels of some inportant activities —
ambassador and task coordinator — than non-leaders. In addition,
functional experiences facilitate particular boundary activities and
inhibit others. Those with current or previous experience in marketing
are prone to take on scout activities, scanning and obtaining ideas
from sales, marketing, and lower levels of R & D, yet shy away from
task coordinator activities with R&D and manufacturing. The opposite
holds for those with current or previous experience in manufacturing
and to some extent R&D. Clearly, experience, knowledge, and comfort
with the language, values, and priorities of another function or

22

"thought-world" (Dougherty, 1987) facilitates communication with those
similar others. These findings suggest that training for new product
team members and leaders should focus on gaining cross-functional
experience and obtaining skills in a wide range of external boundary
activities.

In short, we have identified a set of activities beyond tech-
nological boundary spanning: ambassador, task coordinator, scout, and
guard activities. This broad range of activities suggests that new
product team members, and members of other externally-dependent groups,
must learn to manage a complex set of organizational, as well as,
technological dependencies. We look to future researchers to test the
validity and generalizability of our findings, and the relationship
between boundary spanning and key outcome variables such as perfor-
mance, development, and innovation.

23

REFERENCES

Abemathy, W.J, & Clark, K. (1985) . Innovation: Mapping the winds of
creative destruction. Research Policy, 14, 3-22.

Adams, J.S. (1976) . The structure and dynamics of behavior in organiza-
tional boundary roles. In M. Dunnette (ed.) Handbook of Industrial
and Organizational Psychology. Chicago: Rand-McNally

Adams, J.S. (1980) . Interorganizational processes and organization
boundary roles. In Research in Organization Behavior, (vol. 2,
pp. 321-355) . Greenwich, CT: JAI Press.

Aldrich, H.E. (1979). Organizations and Environirtents. Englewood
Cliffs, NJ: Prentice Hall.

Allen, T. (1970) . Communication in R&D laboratories. R&D Management,
1, 14-21.

Allen, T. (1984) . Managing the Flow of Technolocr/: Technology Transfer
and the Dissemination of Technological Information within the R&D
Organization. Cambridge, MA: M.I.T. Press.

Ancona, D. & Caldwell, D. (1987). Management issues facing new-product
teams in high technology companies. Advances in Industrial and
Labor Relations, 4, Greenwich, CT. JAI Press

Argote, L. (1982) . Input uncertainty and organizational coordination
in hospital emergency units. Administrative Science Quarterly,

27, 420-434.

Astley, W.G., & Van de Ven, A.H. (1983). Central perspectives and

debates in organization theory. Administrative Science Quarterly,

28, 245-273.

Bagozzi, R. (1975) . Marketing as exchange. Journal of Marketing, 39,
32-39.

Brown, J. & Utterback, J. (1985) . Uncertainty and technical communica-
tion patterns. Management Science, 31, 301-311.

Burgelman, R. (1983) . A process model of internal corporate venturing
in the diversified major firm. Administrative Science Quarterly,
28, 223-244.

Crawford, J. (1974) . Task uncertainty, decision importance, and group
reinforcement as determinants of communication processes in
groups. Journal of Personality and Social Psychology, 29, 619-
627.

24

Daft, R. & Macintosh, R. (1981) . A tentative exploration into the

amount and equivocality of information processing in organization-
al work units. Administrative Science Quarterly, 26, 207-224.

David, E. (1984). Trends in R&D. Research and Development, 8, 56-67.

Dewer, R. & Dutton J. (1986) . The adoption of radical and incremental
innovations: An empirical analysis. Management Science, 32, 1422-
1433.

Dou(^erty, D. (1987) . New products in old organizations: The myth of
the better mousetrap in search of the beaten path. Hi.D.
Dissertation. Sloan School of Management, MIT.

IXincan, R. (1972) . Characteristics of organizational environments and
perceived environmental uncertainty. Administrative Science
Quarterly, 17, 313-327.

Ebadi, Y. & Dilts, D. (1986). The relation between research and

development project performance and technical communication in a
developing country. Management Science, 32, 822-830.

Ebadi, Y. & Utterback, J. (1984) . The effects of communication on
technological innovation. Management Science, 30, 572-585.

Galbraith, J. R. (1977) . Organization Design. Reading, MA: Addison-
Wesley.

Galbraith, J. R. (1982) . Designing the innovating organization.
Organizational Dynamics, 10, 5-26.

Good, L. & Nelson, D. (1971) . Effects of person-group and intra-group
attitude similarity on perceived group attractiveness and
cohesiveness. Psychonomic Science, 25, 215-227.

Gresov, C. (1988) . Exploring fit and misfit with multiple contingen-
cies. Academy of Management Proceedings, New Orleans.

Hackman, J. & Walton, R. (1986) . Leading groups in organizations. In
P. Goodman (ed.) Designing Effective Work Groups. San Francisco:
Jossey-Bass.

Houston, B. & Holmes, D. (1974) . Effect of avoidant thinking and

reappraisal for coping with threat involving temporal uncertainty.
Journal of Personality and Social Psychology, 30, 382-388.

Kanter, R. (1983) . The Changemasters . New York: Simon & Schuster.

25

Kazanjian, R.K. & Drazin, R. (1986) . Inplementing manufacturing
innovations: Critical choices of structure and staffing roles.
Human Resource Management, 25; 385-403.

Katz, R. (1982) . The effects of group longevity on project communica-
tion and performance. Administrative Science Quarterly, 27, 81-
104.

Leblebici, H. & Salancik, G. (1981) . Effects of environmental
uncertainty on information and decision processes in banks.
Administrative Science Quarterly, 26, 578-596.

Malone, T. W. (1987) . Modeling conditions in organizations and
markets. Management Science, 23, 1317-1332.

Newcomb, T. (1961) . The Accpjaintance Process. New York: Holt, Rinehart
& Winston.

Pennings, J.M. (1980) . Interlocking directorates. San Francisco, CA:
Jossey-Bass .

Pfeffer, J. (1972) . Merger as a response to organizational inter-
dependence. Administrative Science Quarterly, 17, 382-394.

Pfeffer, J. & Salancik, G. R. (1978) . The external control of
organizations: A resource dependence perspective. New York:
Harper & Row.

Quinn, J.B. & Mueller, J. A. (1963). Transferring research results to
operations. Harvard Business Review, 41 (January-February) , 44-
87.

Taylor, R. (1972) . An analysis of the two-step process of scientific
and technical communication in a military research and development
laboratory. Unpublished dissertation, Indiana University.

Thonpson, J.D. & Tuden, T. (1959). Strategies, Structures and

Processes of Organizational Design. In J.D. Thoirpson, et al.
(Eds.). Comparative Studies in Administration. Pittsburgh, PA:
The University of Pittsburgh Press.

Tushman, M. (1977) . Special boundary roles in the innovation process.
Administrative Science Quarterly, 22, 587-605.

Tushman, M. (1979) . Work characteristics and solo unit communication
structure: A contingency analysis. Administrative Science
Quarterly, 24, 82-98.

Tushman, M.L. , & Nadler, D.A. (1978). Information processing: An

integrating concept in organization design. Academy of Management
Review, 3, 613-624.

26

Van de Ven, A. (1986) . Central problems in the management of innova-
tion. Management Science, 32, 590-607.

Van de Ven, A.J. , & Walker, G. (1984). The dynamics of interorganiza-
tional coordination. Administrative Science Quarterly, 29, 598-
621.

27

TABLE 1

VARIMAX FACTOR LOADINGS FOR BOUTroARY MAJJAGEMQIT DIMENSIONS

n=409

Absorb outside pressures for the team so it

can work free of interference. .785

Protect the team from outside interference. .740

Prevent outsiders from "overloading" the

team with too much information or too many

requests. .719

Persuade other individuals that the team's

activities are inportant. .654

Scan the environment inside your organization

for threats to the product team. .636 .417

"Talk up" the team to outsiders. .602

Persuade others to support the team's decisions .592 .416

Acquire resources (e.g. money, new members,

equipment) for the team. .587 .417

Report the progress of the team to a higher

organizational level. .553 .403

Find out v^ether others in the company support

or oppose your team's activities. .551 .449

Find out information on your company's

strategy or political situation that may

affect the project. .549 .430

Keep other groups in the company informed

of your team's activities. .519 .421

Resolve design problems with external groups. .776

Coordinate activities with external groups. .660

Procure things which the team needs from other

groups or individuals in the company. .657

Negotiate with others for delivery deadlines. .618

28

Review product design with outsiders.

Find out what conpeting finns or groups are
doing on similar projects.

Scan the environment, inside or outside the
organization for marketing ideas/expertise.

Collect technical information/ideas from
individuals outside of the team.

Scan the environment inside or outside the
organization for technical ideas/expertise.

Keep news about the team secret from others
in the conpany until the appropriate time.

Avoid releasing information to others in the
company to protect the team's image or
product it is working on.

Control the release of information from
the team in an effort to present the
profile we want to show.

.515

.424

.491

.404

.791

.719

,645

.587

.823

.817

.592

in

o *

o

rH -X r-- -K O

n -tc rsj -x rH
I I

r~ -x '^

(M -x o

• -x

I I

i-l -x o
■ -x

in -x tH -x r

OO -X CM -X c

. -x • -x

I I

(H -X

(N -X

• -X

CO
o

o

O

rsj

-X
■X

■X o

■X

0^

o

CO

o

o

O

in
O

o

CO

iH -x o r^

o-o -x ^ o

• -X • -x

1

r~ -x (M -X ■^

iH -x ro -X O

. -X • -X

I

rsj

o

CS)

0^
O

I

ro -x m

n -x o

• -x •

I

o

in
o

o o
• *
I

n
o

IX) -X (Ji *

rH -X rH *

. -x • *

I I

IX)

in -x <* -x (-0 rs)

n -x ■^ -x iH -x o

• -x • -x • -x

I I

n CO -x in -x CO <m

o rH -x rv] -x o o

• -x • -x •

I I I

(M

o

in

^■Xvo-Xro-XrH ^

in-x n-x in-x .h-x o

. -X . -X • -X • -X

I I I I

o -x in
n -x o
■X • -x

o

(M

I I

I I

I I

I I

in t^ rH o-xr\) rH CO og co-x^X) in t^

O O O rH-XO <-{ O O rH-XO o o

.-);■ .-X. • .-X-

I III II

cocnconoocotNjr^oinr^*

O O O rH-XO O O O O O rH-XrH-X

.-x.-x.-x -x.-x

II I II III

M o in rs] n r--xo (m o r~ n n

O O O O O ^ * r-i rH-XO O rH-XrH-*:
-JC..X--X. • .-X.-K

11 II

v£) -St in m o o-xo^ in r-t t-t ■^ co-»<

O O O O O -^-XO O o o o ^ *

. -x . * . . • • • *

III I I

B

CM M

p

in vo

I i

in ^

CO 0^

rH
O
O

0-
*
*

-jc

TABLE 3

REGRESSION RESULTS"*
n=409

Boundary Factors
2 3

30

Task
Ambassador Coordinator

Scout

Guard

Compriny

-.20**

Dummy 1

-.01

.04

.14

EXimmy 2

-.16*

-.12

-.07

-.07

Dummy 3

.23**

.00

.01

-.04

Dummy 4

.03

.13

.10

.12

AdjustPd R Square
F

.04
5.03***

.00
1.31

.00
.59

.02
2.97

Task Eiivironment

REVOL

-.02

.06

-.02

.07

EXJ^KR

-.06

-.01

-.02

-.06

CO^PhT

-.12*

.06

.05

.02

RESAVL

.08

.02

.13*

-.22

MSTABIL

.05

.05

-.04

-.02

Adjusted R Square

.05

.00

.01

.05

F

3.06**

.79

1.40

2.92

E-R Square Change

1.46

.38

2.04*

2.82

Riase

EXmimy 5

-.05

.02

.01

-.05

IXimmy 6

.19**

.06

-.05

.06

Adjusted R Square
F

.06
3.23***

.00
.81

.01
1.19

.04
2.43'

F-R Square Change

3.79*

.89

.28

.29

Individual

LEADER

.43***

.20***

-.04

-.10

NIMIEAMS

.06

.02

-.07

.04

YEARSCO

-.01
.00

.12*
.00

-.01
-.03

.03

PEi'PRQJ

.18'

EXRD

EXMFG

EXMKT

-.08
-.07
.14*

.09

.22***
-.24**

-.03
-.21***
.25***

.01
.02

-.01

Adjusted R Square
F

.26
7.86***

.10
3.03***

.07
2.42***

.06
2.35'

F-R Square Change

13.00***

5.95***

4.23***

2.15'

+ Entries are standardized regression coefficients
* p<.05 ** p<.01 *** pK.OOl

**

**

**

**

***

€>-i<^-&^

Date Due

'■",

It

^

'HflJt ? 9 1S9f ,

DEC 0 8 mr

Lib-26-67

MIT LIBRARIES DUPl 1

hill till !lun|

3 TOAD DOSblSlS S