See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/222309792

Mechanisms of governance in the project-based organization: Roles of the

broker and steward

Article  in  European Management Journal · June 2001

DOI: 10.1016/S0263-2373(01)00022-6

CITATIONS READS

280 3,287

2 authors:

Rodney Turner Anne E Keegan

Kingston University London University College Dublin
63 PUBLICATIONS   8,780 CITATIONS    67 PUBLICATIONS   3,544 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

HRM and Paradox View project

HRM in Project Based Organizations View project

All content following this page was uploaded by Anne E Keegan on 24 October 2017.

The user has requested enhancement of the downloaded file.


Pergamon
PII: S0263-2373(01)00022-6
Mechanisms of
Governance in the
Project-based
Organization:
Roles of the Broker and
Steward
J. RODNEY TURNER, Erasmus University, Rotterdam
ANNE KEEGAN, Erasmus University, Rotterdam
The last 50 years has seen a shift in the nature of
work, from mass production, with stable customer
requirements and slowly changing technology, to
the current situation where every product or service
may be supplied against a bespoke design, and
technology changes continuously and rapidly. This
modern environment is a more project-based econ-
omy. The management of the former situation was
well understood, based on classical management
theory, developed over the previous 100 years.
Classical management offers the traditional organi-
zation many strengths derived from the functional
hierarchy at its core. These include strong central
planning, governance and control, knowledge man-
agement and human resource development. The
project-based organization requires a new approach
to its management, which addresses the unique,
novel and transient nature of its work, but retains
the strengths of classical management. This paper
is one of a series aimed at deriving such a manage-
ment paradigm for the project-based organization.
In this paper, we describe governance structures
adopted by successful project-based organizations,
and how they use them to manage the interface
between projects and their clients. We describe two
roles observed at this interface, labelled the broker
and steward. We provide a Transaction Cost per-
spective of the governance mechanisms observed
254
European Management Journal Vol. 19, No. 3, pp. 254–267, 2001
 2001 Elsevier Science Ltd. All rights reserved
Printed in Great Britain
0263-2373/01 $20.00 + 0.00
and the two roles. We note that the same govern-
ance mechanisms are adopted whether the project
is managed in the market or the hierarchy. This is
in stark contrast to the classically managed organi-
zation, and suggests different pressures act on the
project-based organization requiring hybrid
governance structures to be adopted for all projects.
We outline the roles of the broker and steward in
the different project governance structures we have
identified. We consider why it is necessary to have
two roles, a broker and a steward, and not one per-
son fulfilling both.  2001 Elsevier Science Ltd. All
rights reserved.
Keywords: Governance, Transaction cost econom-
ics, Project-based organizations, Projects
Introduction
In this paper we describe governance mechanisms
adopted by project-based organizations to manage
the interface between project teams and their cus-
tomers. We describe two key roles adopted at this
interface, which we have labelled the broker and the
steward, and provide an explanation of the mech-
anisms adopted from a transaction cost perspective.
European Management Journal Vol. 19, No. 3, pp. 254–267, 2001
 MECHANISMS OF GOVERNANCE IN THE PROJECT-BASED ORGANIZATION
Turner and Keegan (1999, 2000) described oper-
ational control processes in project-based organiza-
tions, observed that sometimes the project organiza-
tion creates an interface between projects and their
clients, and noted the two roles now labelled the
broker and the steward. The second paper introduced
a schema for categorizing project-based firms,
depending on whether:
❖ they undertake a few, large projects or many,
small projects
❖ those projects are undertaken for a few, large,
dominant clients or many, small clients
The interface between projects and their clients was
initially observed in the case of project-based organi-
zations undertaking many small projects for many
small clients, and it was suggested that it may be a
transaction cost issue. Closer inspection of the data
showed that the interface and the roles of broker and
steward exist in almost all cases. It therefore appears
to be a governance structure adopted by successful
project-based organizations, and so is, according to
Williamson (1996), by definition a transaction cost
issue.
We start by describing the background to our
research, where we compare the project-based
organization to the classically managed (functional
hierarchical line management) organization. We
briefly summarise four operational control processes
in each of the four cases implied by the schema
above, as described in Turner and Keegan (2000). We
then describe the governance mechanisms we have
observed in each of the four scenarios. We show that
substantially similar mechanisms are observed
regardless of whether the project is undertaken in the
market (it is done for an external client), or in the
hierarchy (it is done for an internal client). This is in
stark contrast to the classically managed organiza-
tion, where different structures are used for (routine)
work undertaken in the market and the hierarchy.
We then review the concepts of Transaction Cost Eco-
nomics (TCE) and what it says about governance
structures of the classically managed organization.
We also conjecture how it might apply to project-
based organizations. We suggest that asset specificity
is not the main pressure in the relationship between
project teams and their clients. Instead it is the risk
and uncertainty in the product of the project, and the
process of its delivery, and the need to use configur-
ation management to manage the refinement of both
(reduction of uncertainty), as the project progresses.
There is a need to maintain a constant dialogue
between project teams and their clients, and hence it
is necessary to maintain hybrid governance mech-
anisms. From this we conclude that the idea of man-
aging the entire organization as a project is fatally
flawed since the governance mechanisms required to
manage a project have higher transaction costs than
European Management Journal Vol. 19, No. 3, pp. 254–267, 2001
the markets and hierarchies described by Williamson
(1996) for the classically managed organization.
We conclude by outlining the roles of the broker and
steward in the different project governance structures
we have identified. We consider why two roles are
required, and suggest that it is because the broker is
essentially and extrovert, entrepreneurial role and
the steward an introvert, intrapreneurial role. We
briefly consider what Agency Theory may have to
say about the two roles.
Background
For two centuries, from the late eighteenth century
to the late twentieth century, functional hierarchical
line management was the main paradigm for the
management of organizations. Classical management
(Morgan, 1997; Huczynski, 1996) is based on the
work of:
1. Smith (1776), who from the start introduced the
linear-rational model of economics;
2. Fayol (1949), who designed organizations with
administrative routines and command structures
through which work could be planned, organized,
and controlled;
3. Taylor (1913), who developed concepts of job
design, to design the work of specialists within the
functions in Fayol’s command and control hier-
archy
Under orthodox economic thinking, the firm is
viewed as a production function (Williamson, 1996)
(Figure 1). Following the ideas of Fayol, a single
model, or structure, is created for the firm. This struc-
ture fulfils several roles, including:
❖ the setting of policy
❖ the creation of governance mechanisms to
implement that policy
❖ control of the operations
❖ selection, recruitment, and development of
human resources
❖ retention, maintenance and development of
organizational learning
It is because the functionally managed organization
fulfils these roles so well, that classical management
was so successful for so long. The classically man-
aged firm works well if markets, products and tech-
nologies are slow to change. The organization works
like a machine (Morgan, 1997; Williamson, 1996).
Raw materials and intermediate products procured
in the market, and intermediate products made in the
hierarchy will be stable, effectively decoupling the
various production functions (illustrated as being
separated by brick walls in Figure 1). This enables
the ideas of Taylor (1913) to be applied to the precise
255
MECHANISMS OF GOVERNANCE IN THE PROJECT-BASED ORGANIZATION
Figure 1 The Classically Managed Organization
definition of the workings of each function
(separately).
Since the 1950s, project-based ways of working have
become increasingly necessary to deal with oper-
ations that are substantially unique, novel and transi-
ent (Turner, 1999), especially with the recent explos-
ive development of markets, products and
technologies (Kanter, 1983; Handy, 1995). We define
the project-based organization as one in which:
the majority of products made or services supplied are
against bespoke designs for customers
If the end product is bespoke, so too will be the inter-
mediate products, and thus the processes to produce
them will be novel on every project. The consecutive
production functions will also be linked (bilaterally
dependent), since uncertainty in the process and
intermediate products consumed and produced in
one step will feed into uncertainty in the processes
upstream and downstream (Figure 2). Thus, classical
management cannot apply (Turner and Keegan,
1999), because:
❖ the functions are linked (bilaterally dependent)
Figure 2 The Project-based Organization
256
❖ the work cannot be predicted with absolute cer-
tainty so classical job design cannot apply
❖ operational control is aligned with the project and
not with the functional hierarchy
New management models are required, but ones that
retain the strengths of classical management. We aim
to develop such models. Turner and Keegan (1999)
showed it is essential to retain functions (competence
pools). The trick is to configure those functions into
versatile production networks. Turner and Keegan
(2000) described operational control processes
enabling that to happen. Keegan et al. (1999) and Kee-
gan and Turner (2000) and Keegan and Turner (2000)
described human resource management practices in
project-based organizations, including individual
selection and development. Turner et al. (2000)
described experiential learning practices. Here we
describe governance mechanisms adopted to support
the operational control processes, and to manage the
interface between project teams and their clients. For
a detailed description of our research project, and the
organizations interviewed, please refer to Turner and
Keegan (1999 or 2000).
European Management Journal Vol. 19, No. 3, pp. 254–267, 2001
 MECHANISMS OF GOVERNANCE IN THE PROJECT-BASED ORGANIZATION
Operational Control Processes in the
Project-Based Firm
Turner and Keegan (2000) described operational con-
trol processes in the project-based firm. A schema is
introduced (Table 1), that classifies firms according
to whether they undertake:
❖ a few, large projects or many, small projects, for
❖ a few, large clients or many, small clients
We continue to question this classification. First it
should be said that many organizations operate in
more than one of the four quadrants implied. But
where they do, successful organizations adopt the
appropriate operation control processes and govern-
ance mechanisms for each scenario (see below). We
are also embarrassed at the cliché of a two-by-two
matrix. We tried expanding to three parameters on
each axis, but found two appropriate. Four scenarios
are implied:
1. Large projects for large clients are managed as tra-
ditional large projects. A large, dedicated team is
created, effectively a firm within a firm, and the
project effectively has its own command and con-
trol structure. Team members often change as the
project progresses, and isomorphic team structures
are adopted, where the nature of the team is
adapted to suit the stage of the project life-cycle
(Frame, 1995).
2. Many small projects for large clients are managed
as programmes of projects. The projects will con-
tribute together to a larger end. Homogeneous
teams are created to be responsible for each project
from concept to completion. The teams work fairly
independently under the overall guidance of a
programme manager, who will be responsible for
the larger concept. Within the programme, the
Table 1 A Selection of our Sample Classified According to the Schemaa
FEW LARGE CLIENTS Isomorphic projects
A firm within a firm
ABB Lummus
Fluor Daniel
Washington
MANY SMALL CLIENTS Start-up company
Entrepreneur/Intrapreneur
SDS Systems
FEW, LARGE PROJECTS
a
Normal script: Archibald Type I Firms, external client; Italic script: Archibald Type II Firms, internal client
European Management Journal Vol. 19, No. 3, pp. 254–267, 2001
projects are managed as a fish-tail (Figure 3). The
nature of projects is substantially different than in
1 above, and so we have labelled them projettes
to differentiate.
3. Many small projects for many small clients are
managed as a portfolio of multi-projects (multi-
projettes); they share a common resource pool, but
have independent outputs. The projettes are man-
aged substantially in the same way as in 2, with a
homogeneous team responsible for the projette
from concept to completion. But, we noticed here
the creation of the broker/steward interface to
manage the relationship between the client and the
project teams (Figure 4).
The broker is responsible for the relationship with
the client, and the steward with the project team.
They work together to manage the interface. A cli-
ent with a requirement approaches an appropriate
broker, who identifies the steward with access to
the resources able to meet the requirement. The
steward compiles the team, including project man-
ager, who deliver the requirement. The project
manager, steward and broker work together with
the client to ensure the project’s deliverables
indeed meet the client’s requirements. It is neces-
sary to create the broker/steward interface for
reasons of efficiency and effectiveness. It becomes
inefficient for all the project teams to deal with all
Figure 3 Fish-tail Programme of Projettes
Programme management
Fish-tail programmes
Ericsson
Reuters Networks
Multi-projects
Broker
Virtual Factory
EDS (NZ)
Reuters PD
BT Back Office
MANY, SMALL PROJECTS
257
MECHANISMS OF GOVERNANCE IN THE PROJECT-BASED ORGANIZATION
Figure 4 Multi-projettes with the Broker/steward
Interface
the customers. It is better for each team to deal
with one steward and each customer with one
broker, and the brokers and stewards manage the
interrelationships. This can be posed as a Trans-
action Cost issue. If there are m competence pools
and n potential customers, if the cost of managing
a relationship is c, and the cost of maintaining the
broker/steward interface C, then the broker/
steward interface economizes the relationships
when:
(m + n) × c + C ⬍ (m × n) × c Equation (1)
4. The one case from our sample of a large project
undertaken for many clients was a start-up com-
pany, where the company and its new product are
the project. The managing director (the original
entrepreneur whose idea the product was), man-
aged the teams producing different versions of the
product. The marketing director worked with
potential clients to try to identify their require-
ments, and match the versions of the product to
them.
[We differentiate between programmes of projects
Figure 5 Governance Structure Between BT’s Front Office and Projects in the Back Office
258
and portfolios of multi-projects. Multi-projects are
many small projects sharing a common resource
pool, but with no relationship between the products
of the projects; programmes are multi-projects where
the products are combined into a larger facility
(Murray-Webster and Thiery, 2000; Turner, 1999)].
Turner and Keegan (2000) describe the four oper-
ational control processes in greater detail.
Table 1 lists a few companies from our sample
against the above categorization. It also classifies
them according to another schema introduced by
Archibald (1992).
1. Type I organizations undertake projects as their
main or even exclusive business, supplying
bespoke products or services to customers. These
firms undertake projects for clients who have
placed their project work in the market.
2. Type II organizations have projects supporting
their mainline business, which is routine in nature.
Projects are undertaken in the support function to
deliver new markets, products or technologies.
These firms retain project work in the hierarchy,
often because they work in a rapidly changing
business. One of our sample companies told us
that they need to change their business radically
every two years to keep pace with market changes.
Essentially clients of Type I firms are placing their
projects in the market, whereas clients of Type II
firms are managing their projects in the hierarchy.
Governance Mechanisms Adopted by
Project-Based Firms
The broker/steward interface was identified in the
case of many small projects undertaken for many
small clients. However, when we revisited our data,
we also noticed it in cases 2 and 4 above. Figures 5,
European Management Journal Vol. 19, No. 3, pp. 254–267, 2001
 MECHANISMS OF GOVERNANCE IN THE PROJECT-BASED ORGANIZATION
Figure 6 Governance Structure for a Programme of Projects
6 and 7 illustrate the operation of the broker/steward
interface as a governance mechanisms in scenarios 3,
2 and 4 above (respectively).
Governance Structures for Multi-Projects to Many
Small Clients
Figure 5 illustrates a governance structure for a pro-
ject-based organization delivering a portfolio of
multi-projects to many small clients, as described
above. Figure 5 actually illustrates the structure
adopted by BT’s back office to deliver projects to the
front office, but we observed similar structures
adopted for new product development in Reuters, by
EDS(NZ) in servicing New Zealand Telecom and the
University of St Galen’s virtual factory in Eastern
Switzerland. The interesting point to note is that the
same structure is adopted whether the projects are
undertaken internally (within the hierarchy, BT),
externally (in the market, the Virtual Factory), or in
an intermediate, hybrid structure (EDS and NZ
Telecom). Effectively a hybrid structure is adopted in
all cases; in the first case, the project is pushed out
of the hierarchy into the hybrid structure; and in the
second it is pulled out of the market into it.
Figure 7 Governance Structure for an Entrepreneurial
Start-up Company
European Management Journal Vol. 19, No. 3, pp. 254–267, 2001
BT explained their situation to us as follows. In 1990,
BT radically changed its structure through a change
project called Project Sovereign (MPA, 1992). That
created the back office, and then three operating
divisions: domestic, business, and international.
However, subsequent market segmentation had led
to further division of the operating divisions, so at
the time of the interview there were 16, shortly to
rise to 19. The back office could not cope with this
constantly changing interface. So they created the
governance structure. At the interface, each division
had an Account Manager (broker), representing it. In
each project, one of several disciplines in the back
office will lead. Hence there are a series of Solutions
Managers (stewards), who will create project teams
consisting of the lead and other disciplines in
response to a need brought in by a broker. If a new
operating division is created, a new broker is created.
But the essential point is the competence pool con-
tinues to deal with its steward, with whom it is fam-
iliar.
Names given to brokers and stewards in other cases
are listed in Table 2.
Governance Structures for Programmes of Projects
for Large Clients
Figure 6 illustrates the governance structure adopted
by Ericsson for the management of its work with its
clients to deliver new networks for national tele-
phone operators. Above, we describe the programme
as a fish-bone of small projects, small, clearly defined
projects, building towards the programme objective
(which may be less clearly defined, and indeed never
achieved as it is constantly advancing). In this case
the broker is called the Account Manager, and the
steward the Solutions Manager. Because of the need
for client confidentiality, Chinese walls are main-
tained between the programmes. A client deals with
one Account Manager and that Account Manager
with just one Solutions Managers. Project team mem-
259
MECHANISMS OF GOVERNANCE IN THE PROJECT-BASED ORGANIZATION

Table 2 Names Given to the Broker and the Steward

Case Broker Steward

Multi-projects
쐌 Virtual factory Broker Network coach
쐌 BT Back office Account manager Solutions manager
Reuters new product development Product manager Project manager

Programme
쐌 Ericsson telephone network development Account manager Solutions manager
쐌 Reuters infrastructure development Champion Project manager

Large projects
쐌 ECI (Contracts manager) Programme director
쐌 STS Marketing director Research director
(Managing director)

bers should not be simultaneously working for two
clients. Ericsson talk about the relationship between
the Account Manager, Solutions Manager and a pro-
ject manager as ‘core three’, shown as a triangle by
Turner and Keegan (1999) (Figure 6). Here, we show
it as a governance chain, with the project manager
essentially shown as being on a different level to the
other two. Essentially the same structure is adopted
in Reuters for the delivery of internal infrastructure
(the networks over which they deliver their data
products), except there the broker (Project
Champion), will be attempting to sell the project to
several clients, regional and national operating
divisions. Thus, again, the same hybrid structure is
adopted regardless of whether the project is under-
taken internally or externally; it is pulled out of the
market or pushed out of the hierarchy into a
hybrid structure.
You might say that infrastructure development and
new product development have the same customers
in Reuters, and so question why we labelled them
large or small. They are not the same customers. Cus-
tomers for NPD are operating units (Front Line Busi-
ness Units), whereas customers for infrastructure
development are country or even continental
divisions. Thus the sizes are substantially different,
as are the power relationships.
director employed to sell the concept to the various
potential users of the new product. He was brought
in for his selling skills, something the managing
director did not feel confident with; the managing
director enjoyed the science, the marketing director
the sales. However, the marketing director was also
fulfilling one of the steward’s roles. The company
was beginning to think about gearing up to produce
product, and the marketing director was looking to
find production capacity.
Governance Structure for Large Projects from the
Engineering Construction Industry
The last scenario is for large projects from the Engin-
eering Construction Industry, ECI (Figure 8). In this
case we do not see the broker and steward roles so
clearly. The Project Director maintains the relation-
ship with the client directly. The Marketing Depart-
ment fulfils one of the roles of the broker, in that it is
they who establish the initial contact with the client.
Similarly, the Project Director fulfils some of the roles
of the steward, in that it is he or she who sources the
resources required to undertake the project. How-
ever, throughout the project delivery, the Programme

Governance Structure for an Entrepreneurial Start-

Up Company
The third case is the entrepreneurial start-up com-
pany (Figure 7). Here there is only one project effec-
tively, to develop the company’s new product. How-
ever, there are several potential applications, with
sub-projects delivering the different versions. The
sub-project to deliver the main one is managed by the
managing director, the original entrepreneur, who is
also the steward assigning resources to all the sub-
projects. (Team leaders manage the others. There
were only 25 people in the company at the time of Figure 8 Governance Structure for Large Project in
the interview). There is only one broker, a marketing the ECI

260 European Management Journal Vol. 19, No. 3, pp. 254–267, 2001
 MECHANISMS OF GOVERNANCE IN THE PROJECT-BASED ORGANIZATION
Director maintains direct contact with the client. This,
though, is as we would expect. With a small number
of large projects undertaken for a small number of
large clients the relationships are economized with a
direct relationship between the client and the project
teams (equation 1).
So what is it that creates cost in the relationship
between the project teams and their clients? And why
is it that a hybrid relationship is appropriate in all
cases, for both Archibald’s Type I and Type II organi-
zations. To answer these questions we turn to Trans-
action Cost Economics to see what perspectives it
can bring.
Mechanisms of Governance: Transaction
Cost Economics Perspective
The core hypothesis of Transaction Cost Economics,
TCE, is that one should:
Align transactions (which differ in their attributes) with
governance structures (which differ in their costs and
competencies) in a discriminating (mainly transaction cost
economizing) way. (Williamson, 1996, p. 46)
Williamson says that TCE makes the transaction the
unit of analysis, but he also says:
Transaction cost analysis [is] an examination of the com-
parative costs of planning, adapting, and monitoring task
completion under alternative governance structures. (p. 58)
In the classically managed organization making rou-
tine products, the fundamental choice is a buy-or-
make decision, between:
❖ buying intermediate products in the open market,
where economies of scale and price competition
lead to lower price
❖ making them internally in the firm, where high
management costs are offset by a better ability to
manage maladaptation costs
There is a third, intermediate option, the hybrid
structure, where the client imposes some controls on
the supplier, as in the case of franchising or inte-
grated supply chain management. The decision
between market, hybrid or hierarchy is dependent on
three variables:
❖ the frequency at which the transactions occur
❖ the degree or type of uncertainty to which they
are exposed
❖ the condition of asset specificity
Asset specificity is the amount to which assets used
in the production function are specific to the task or
are redeployable to alternative uses without loss of
value. Williamson (1996) identifies six types of asset:
European Management Journal Vol. 19, No. 3, pp. 254–267, 2001
❖ physical location
❖ dies, moulds and jigs
❖ tacit knowledge
❖ production equipment
❖ brand name
❖ time
Uncertainty arises from three sources:
❖ risk
❖ unintentional miscommunication
❖ strategic non-disclosure by one party, particularly
the supplier
Williamson clearly states that he is dealing only with
repetitive production (transactions), and so the fre-
quency of transaction is high. He shows then that
what determines the need for control of the trans-
action is the frequency at which disturbances to the
transaction arise.
It is assumed production costs are always cheaper in
the market due to incentives, economies of scale and
avoidance of administrative overhead associated
with the bureaucracy of managing the hierarchy.
Hence at low levels of asset specificity and disturb-
ance, intermediate products should be bought in the
market to minimize cost. However, at higher levels
of asset specificity, assets will become maladapted to
the task, or the products produced will be malad-
apted to the requirement, or the assets will achieve
low utilization. This increases the need for control of
the transaction. At intermediate levels of asset speci-
ficity, control is most cheaply achieved through
hybrid forms of governance, and at high levels it is
best achieved through the hierarchy. As frequency of
disturbance increases, the hierarchy becomes increas-
ingly the preferred form of governance at even low
levels of asset specificity.
Williamson (1996) also covers the concepts of
bounded rationality, incomplete contracting, auton-
omous and cooperative adaptation, opportunism,
bilateral dependency and the Fundamental Trans-
formation:
Bounded Rationality: Orthodox economics assumes
people behave rationally. Williamson says that
because of human frailty this is impossible. People
would like to behave completely rationally, but that
requires people to be omniscient and clairvoyant.
Behaviour is intendedly rational, but only limitedly
so, limited by our ability to receive, store, retrieve
and process information. Bounded rationality adds to
the uncertainty of the relationship.
Incomplete Contracting: All complex contracts are
unavoidably incomplete, again because of human
frailty. To foresee all contingencies people again need
to be omniscient and clairvoyant. However, there is
no need to be myopic. Contracting parties can put in
261
MECHANISMS OF GOVERNANCE IN THE PROJECT-BASED ORGANIZATION
place control and governance structures to deal with
the risks as they arise.
Adaptation: Bounded rationality and incomplete con-
tracting require adaptation. Williamson identifies
two types of adaptation, autonomous adaptation and
controlled adaptation. The former arises best in the
market without controls, and the latter in the hier-
archy (or hybrid governance structures). The greater
asset specificity or the frequency of disturbances, the
greater the need for controlled adaptation (see Note).
Opportunism: Some people behave opportunistically
some of the time. However, Williamson discounts the
notion of trust. He says calculativeness, which he
views as part of risk, causes most people, most of the
time, to see that it is in their interest to cooperate.
They will take good bets, and do not rely on mani-
festly bad games such as the Prisoners Dilemma.
Bilateral Dependency: This is a state of mutual depen-
dency between supplier and buyer. It evolves during
an ongoing contractual relationship. Incomplete con-
tracting and opportunism will increase bilateral
dependency, and it will require controlled adaptation
to manage.
The Fundamental Transformation: A buyer and supplier
may initially have no mutual dependency. If there
are many potential suppliers of a product, then in the
initial bid process, the buyer will have many sup-
pliers to chose from. However, once the supplier has
chosen a preferred supplier, and especially if the sup-
plier then needs to invest in specific assets, the Fun-
damental Transformation will occur, they will
become bilaterally dependent.
The Project-Based Organization
As we said above, Williamson (1996) clearly states he
is dealing with repetitive, routine transactions,
undertaken by the classically managed (functional,
hierarchical, line management) organization. In the
book he invokes the machine metaphor (Morgan,
1997) to illuminate his points. However, he says the
ideas should apply to all transactions, including pro-
ject-based ones. We conjecture here how the concepts
may apply in the case of projects, where the trans-
action is unique, novel and transient, with high
degrees of uncertainty. We consider first the buyer,
and then the supplier and finally the project relation-
ship.
The Buyer
The buyer has a problem. Even if an asset has high
specificity, it may only be used once. The client can-
not afford to spend a lot of money to invest in the
asset to only use it once. (Here we are talking about
the asset required to deliver the project, not the asset
262
produced by it. One of the most significant assets
here is tacit knowledge of successful project manage-
ment processes, Turner et al., 2000). This often
requires client organizations to take projects out of
the hierarchy (where other conditions suggest they
ought to be), and place them in the market. One of
the firms in our sample is the supplier of information
systems services to a national telephone operator in
an outsourcing relationship. During the late 1980s
and early 1990s, companies in the Engineering Con-
struction Industry outsourced their project capabili-
ties under the concept of partnering (Bower and
Skuntzos, 2000). This places clients and suppliers in
a state of bilateral dependency even outside the dur-
ation of the contract.
The Supplier
The supplier’s problem is to deliver a product that
meets the client’s requirements. Because of bounded
rationality and incomplete contracting, the project’s
product and delivery process will be imperfectly
defined at the outset. The goals and the methods will
not be perfectly defined (Turner and Cochrane, 1993).
This will require the use of configuration manage-
ment to refine their definition as the project pro-
gresses, to ensure what is delivered at the end does
indeed meet the client’s requirements. Thus the client
needs to be involved in the client process, creating a
state of permanent bilateral dependency between cli-
ent and contractor throughout the project.
The Project Transaction
We identified above that, whereas in the classically
managed organization the steps in the process chain
are decoupled because the intermediate products are
stable and unchanging, in the project-based organiza-
tion the steps of the production process are inextri-
cably linked. They are linked by the uncertainty of
the project process, they are linked by the lack of clar-
ity in the definition of the final product, and they are
linked by the need to take a holistic view of the pro-
ject process and the need to manage the risk through-
out. This creates a state of bilateral dependency
between all the stages of a project, and the functions
working on it. The PRINCE 2 process (CCTA, 1996),
suggests that:
❖ at the start of the project the firm’s standard pro-
ject process should be tailored to the needs of the
project in hand
❖ at the hand-over from one stage of the project to
the next, the intermediate product needs to be
checked against the current business case, the
needs of the final customer, and the needs of the
subsequent steps of the process
❖ a governance mechanism needs to be put in place
across the entire project process (Figure 9).
European Management Journal Vol. 19, No. 3, pp. 254–267, 2001
 MECHANISMS OF GOVERNANCE IN THE PROJECT-BASED ORGANIZATION
Figure 9 The Project Process and Governance Structures According to PRINCE 2, (CCTA, 1996)
This interrelationship between the client, supplier
and project has several consequences:
1. It creates a state of bilateral (or multi-lateral)
dependency between the client and the suppliers
(and between suppliers and suppliers) for the dur-
ation of the project (and even outside it), creating
the need for a governance structure.
2. It means the contract is unavoidably incomplete,
increasing the mutual dependency of the parties
to the contract, and the need for the governance
structure.
3. It creates an environment for opportunistic behav-
iour, although calculativeness should illustrate the
benefit of cooperating. However, it depends on
whether the client and contractor expect to work
together again. In the ECI there is only a small
number of clients and contractors, enhancing
dependency and the need to cooperate. There is
no Fundamental Transformation because client
and contractor are in a state of dependency from
before the project to after and on to the next pro-
ject. In the building industry this is not the case.
Here the Fundamental Transformation does take
place. One of our respondents from an Infor-
mation Systems supplier told us that their clients
from the building industry did not seem to
appreciate that they were in a state of bilateral
dependency for the duration of the project, and
behaved as if the relationship could be broken at
any time. He said the client may have no depen-
dency on a bricklayer, because other bricklayers
are available in the market with no asset speci-
ficity, but that is not the case with an Information
Systems supplier.
4. All of these create a need for controlled adaptation
(using configuration management), and a govern-
ance structure for the duration of the project that
involves the client and supplier(s) in the process.
It is this governance structure we have observed, and
European Management Journal Vol. 19, No. 3, pp. 254–267, 2001
the broker and the steward are two key roles in the
structure.
Economizing
Williamson (1996) devotes some pages to economiz-
ing. He says that economizing is one of the two key
elements of strategy (the other being strategizing —
power and policy).
In the classically managed organization, this entails
choosing the appropriate governance structure
(market, hybrid or hierarchy), and improving the
efficiency of those transactions or production units
managed by the hierarchy. Because the successive
steps in the production process can be decoupled
through the stable intermediate products, this last
involves increasing the efficiency of resource utiliz-
ation at each step. (The cost of the raw materials or
intermediate products will have been economized by
the choice of appropriate governance structure.) This
drive for increased resource efficiency we have called
an economic orthodoxy (Turner and Keegan, 1999)
and is almost equally well explained by the other
economic orthodoxy of viewing the firm as a pro-
duction function (Williamson, 1996).
In project-based firms the situation is different. Steps
of the process are linked, so the whole chain needs
to be economized. As is well known, where steps are
linked, the optimum for the whole is not necessarily
obtained by optimizing each part (Wagner, 1969). It
may be necessary to maintain resources at a given
step more than are required for that step to cope with
maladaptation (manage the risk). There are several
examples from our interviews of companies not
understanding this. In an information systems sup-
plier, the new managing director changed the cri-
terion against which project managers were judged
from profit to resource utilization. In another, a pro-
263
MECHANISMS OF GOVERNANCE IN THE PROJECT-BASED ORGANIZATION
ject director from a firm in Engineering Construction
with a parent in defence electronics complained that
he wanted to be judged on the overall profitability
of his projects, whereas the parent judged him work-
package by work-package.
What is Going on in the Project-Based
Organization?
We now consider how this may explain the govern-
ance structures we have observed in project-based
firms. We recall governance structures will be chosen
by the need to manage:
❖ the bilateral dependency that exists between cli-
ents and the project team, between stages of the
project or programme, and between different com-
petence pools or suppliers working on the project
❖ the configuration, (uncertainty in the project’s pro-
duct and process). It is configuration management
that creates the cost in the relationship between
the project team and the client, both through the
cost of undertaking the process, and in the chance
of making an error and the cost of that error.
The Four Cases
Case 1 — Large Projects for Large Clients: Because both
project team and client are involved in a small num-
ber of relationships, the need to economize is best
served by direct contact. Small projects are thought
to be more risky than large ones, because there, an
error has a larger percentage impact, (Turner, 1999,
Chapter 14), so on large projects the potential cost of
errors is less. Finally, in the Engineering Construction
Industry, assets tend to be less specific. Thus, both
the number and cost of relationships is small, which
means there can be a direct, and perhaps more
remote, relationship between project director and cli-
ent, more akin to Williamson’s market. There is a
strong, bilateral dependency between different stages
of the project, and between different functions work-
ing on it, but this can be handled internally by the
project director. It does explain, however, why the
project director will try to optimize the whole project,
not individual packages of work, as the project direc-
tor from the Engineering Construction Industry
described above was trying to do.
Case 2 — Large Programme of Small Projects for Large
Clients: Now the need is to manage the configuration
of the product and process of the whole programme.
Indeed, as we said above, the vision of the final pro-
duct of the programme may evolve as the pro-
gramme progresses. This requires strong communi-
cation between the client and the programme team
throughout. This communication will be the role of
the broker and the steward, (Account Manager and
264
Solutions Manager in the case of Ericsson’s telephone
networks business). Indeed, the individual project
teams can be made relatively independent of each
other, (decoupled), and so be optimized individually,
explaining why the manager from the defence elec-
tronic industry described above focused on optimiz-
ing individual work packages.
Case 3 — Many Projects for Many Clients: Now the
need is to maintain the communication between the
many projects and their many clients, and between
the many clients and the projects working for them.
The need to economize is not best served by a direct
relationship between clients and project teams. That
would create confusion, lead to potential maladapt-
ation of the projects’ products, and high risk of error,
with a heavy impact of seemingly small errors (see
above). This leads to the creation of the
broker/steward interface. The broker represents the
client in the communications on all projects being
undertaken for that client. Here, because the projects
are fairly small, with homogeneous teams, the main
issue is configuration management of the product
and process, ensuring the right product is delivered
to the client. The broker and steward work together
to ensure that the right resources are used to deliver
the project, that the best overall process is followed,
and that the product is correctly configured to meet
the clients needs.
Case 4 — Start-up Company: The issue in this case was
mainly that the managing director wanted to focus
on the science, that was his skill. He therefore
appointed a marketing director to liaise with the
potential clients. However, if the managing director
had liaised with clients, then his scientific skill would
have been wasted, there would have been maladapt-
ation in the use of his time. Also, if he were not
skilled in commercial issues, then if he were to do
that it may not lead to the best outcome. Thus the
need to economize is best served by the managing
director fulfilling the role of the steward, where his
skills lie, and employing a marketing director to
negotiate with clients and achieve the optimum out-
come there.
Why is the Project Pulled Out of the Market or
Pushed Out of the Hierarchy?
We also wish to address the question of why Archib-
alds’s Type I firms pull the project out of the market
and manage it in a project-based hybrid, and why
Type II firms push the project out of the hierarchy,
and also manage it in the same project-based hier-
archy.
Type I Firms: The answer in this instance is easy. With
strong bilateral dependency between client and pro-
ject team, direct relationships need to be maintained.
This cannot be achieved in pure market-based
governance structure. It is necessary to create a
European Management Journal Vol. 19, No. 3, pp. 254–267, 2001
 MECHANISMS OF GOVERNANCE IN THE PROJECT-BASED ORGANIZATION
hybrid, and this will be necessarily focused on the
project.
Type II Firms: Now bilateral dependency between the
functions working on the project means that they
cannot be decoupled, (as Taylor and Fayol would
have us do). Thus project centric operational control
and governance structures need to be adopted, separ-
ate from the functional hierarchy. This has been our
premise from the start, and the basis on which our
research project is based, (Turner and Keegan, 1999),
and is now confirmed.
The Same Governance Structure: There are different
reasons why the project is pulled out of the market
or pushed out of the hierarchy, but they do not
answer the question why similar governance struc-
tures are adopted. The answer must be that project-
centric operational control and governance structures
are required in both instances, and so once the project
has been made the unit of analysis, then the same
solution will be achieved in each case.
It is worth remarking that risk management is best
achieved by making the risk management process
internal to the project team, (Turner, 1999). If the risk
of error is part of the Transaction Cost, then the need
to economize again will be best served by making
the project the unit of analysis.
The Purely Project-based Organization
It has been suggested (Levett, 1998) that organiza-
tions should adopt a totally project-based approach
to their operations. Everything down to accounting,
marketing and human resource functions should be
managed as projects. It has been one author’s opinion
since the late 1980s that some things are essentially
routine, and are best managed as routine operations.
Now we find support for that conjecture. Managing
operations as projects requires the creation of govern-
ance structures over and above the normal hierarchy,
adding to the transaction cost. Bilateral dependency
in projects requires projects to be taken out of the
hierarchy and placed in separate hybrid structures,
creating additional cost. Functional, hierarchical, line
management may have its flaws, but it is the best
option to manage essentially routine operations. In
choosing a governance structure, Coase (1964) said:
The relevant comparisons are with the feasible alternatives,
all of which are flawed.
and Williamson (1996) says:
Many hypothetical forms of organization never arise, or
quickly die out, because they combine inconsistent fea-
tures. (p. 95)
Classical management may not be perfect, but it is
European Management Journal Vol. 19, No. 3, pp. 254–267, 2001
the best for what it was designed, to manage routine
operations in the context of stable markets, products
and technologies. Project-based ways of working are
essential to manage essentially unique, novel and
transient operations and activities, but they need a
governance interface between them and routine oper-
ations, outside the normal hierarchy. An organization
that tries to manage the essentially routine as a pro-
ject will not flourish or quickly die out.
The Roles of the Broker and Steward
So, what are the roles of the broker and steward?
The Italian Connection
Boissevain (1974) in studying the Sicilian Mafioso
identified the roles of broker and patron. Brokers are
social fixers, who use secondary resources such as
information and contacts, to achieve their objectives.
Belbin (1993) might call them resource-investigators.
They build their reputation through successful deals
with patrons. They also act as gatekeeper, controlling
access to the patron. The patron has control over pri-
mary resources, funds and people, which can be
exchanged for information or the resolution of a
problem. The broker and patron cannot function
without each other. Here the balance of power is dif-
ferent than in the cases we observed. People want
access to the patron’s primary resources, which he or
she releases guardedly, and the broker acts as gate-
keeper. In the Type I firms in our sample, the broker
acts as sales and marketing, as much seeking cus-
tomers as controlling access. (Belbin’s resource-inves-
tigator role includes the skill of ambassador). The
steward controls the primary resources (people and
money), in the supplier, but is marshalling them to
meet the customer’s requirement rather than handing
them out reluctantly. In Type II organizations, the
situation is different, resources are scarce, so the
broker has a prioritization role. Similar roles also
exist in the Italian textile industry (Mariotti and Cain-
ara, 1986).
The Broker
The role of the broker is in maintaining the relation-
ship with the client. Turner and Keegan (1999) sug-
gested that this entails:
❖ identify and attract new clients
❖ bid for and win work
❖ liaise with the client during the work and delivery
of the product
❖ ensure the client is satisfied
❖ win follow-on business
265
MECHANISMS OF GOVERNANCE IN THE PROJECT-BASED ORGANIZATION
Their role combines ambassador for the firm and
resource investigator for the client.
The Steward
The steward puts together the network of resources
to deliver the project, ensuring that the right people
are in the right place at the right time to ensure it
happens. It is the project manager’s role to manage
the process. We called the role ‘steward’ after the
stewards at a sporting function (Henley, Wimbledon,
St Andrews). Their role is to ensure that the resources
are available on the day, but not to manage the pro-
cess — that is the job of the referee. A similar analogy
is with the producer of a film. The producer marshals
the resources, the director manages the process.
Turner and Keegan (1999) did not identify this role.
Resource managers perform the input management
as described there, and the project manager or direc-
tor perform the process and output management. The
role of the steward is almost abstract, but an essential
one, complementing that of broker and project man-
ager in the ‘core three’ of Ericsson (Turner and Kee-
gan, 1999, Figure 6).
Two Roles?
Why are there two roles? Why cannot just one person
be both broker and steward? This requires further
investigation. Agency Theory may provide some
insight, (see below). However, first we present some
anecdotal evidence. When discussing this issue with
an Irish psychiatrist working in Malta, and a psychol-
ogist in Australia, both gave similar responses. The
former said that the broker must align with the cul-
ture of the client, whereas the steward must align
with the culture of the competence pools. The latter
said that the broker is essentially an extrovert,
entrepreneurial role, whereas the steward is essen-
tially an introvert, intrapreneurial role. We have no
empirical evidence yet to support these conjectures,
except possibly the marketing and managing direc-
tors of the start-up company, who both match the
psychological profiles of them.
With the repeat orders, however, we find the roles
disappearing. With the second or third orders the
broker disappears first, and the steward maintains
direct contact with the client. The risk of error is
reduced with experience, and the need to manage
configuration reduces as the solution becomes better
understood. Eventually the supply becomes routine,
the contract is placed in the market or the hierarchy,
and the role of the steward also disappears.
Agency Theory
Agency Theory (AT) is a theory of governance of the
firm with similarities and differences to TCE
266
(Williamson, 1996). AT is concerned with incentive
alignment, TCE with dispute avoidance. TCE
assumes bounded rationality and incomplete con-
tracting and manages transactions to control adap-
tation to avoid disputes. AT creates incentives for
people to behave rationally and avoid opportunism.
AT says that, in firms with high asset specificity, it
is necessary to have a Board of Directors, inde-
pendent of workers and managers, to control the
activities of the managers and look after share-
holders’ interests. Firms from services industries such
as advertizing, accounting or law, with low asset
specificity, can be financed by debt and managed by
worker/managers. Firms with high asset specificity,
must be financed by equity, and have independent
directors to represent shareholders interests. The
Board performs decision control while managers per-
form decision management.
In a project, which operates as a firm within a firm,
with bilateral dependency, someone needs to rep-
resent the interests of the client to ensure ex ante
incentive alignment. That is the role of the broker.
The steward and the project manager manage the
resources and process to overcome the risks and
uncertainty, to iron out imperfections in the contract,
to achieve ex post dispute avoidance. This is why the
broker and the steward are two separate roles. They
cannot be the same person, and they cannot do with-
out each other.
Conclusion
We have described governance structures adopted by
project-based firms, and the roles of the broker and
steward observed in the management of projects. We
observed similar governance structures adopted by
firms undertaking projects for clients and by firms
undertaking projects internally. This is in contrast to
classically managed firms, where different
approaches are used to manage contracts in the mar-
ket or the hierarchy. We have explored the perspec-
tives the theory of Transaction Cost Economics brings
to the governance of projects, and have shown that
it can be used to explain the project-focused, hybrid,
governance structures we have observed in project-
based firms. We have also shown that it explains the
roles of broker and steward. We have outlined the
roles performed by these two parties to a project.
Note
1. Turner and Keegan (1999) suggested that central planning,
as a model for the governance of nations, has been a disas-
ter. The concepts of autonomous and controlled adaptation
confirm this. UK politicians talk about the European Union
becoming a ‘federal superstate’. But ‘federation’ and
‘superstate’ are different things. Federation is what they
have/had in the USA, Australia, or the Holy Roman
Empire for its first nine centuries. In the USA, each state
European Management Journal Vol. 19, No. 3, pp. 254–267, 2001
 MECHANISMS OF GOVERNANCE IN THE PROJECT-BASED ORGANIZATION
sets its own tax rates. Autonomous adaptation (the market)
ensures tax harmonization as each state cannot set taxes
significantly higher than others. Superstate is what they
had in the Holy Roman Empire for its last century, in the
USSR, and what we are getting in the EU. Central planning
brings controlled adaptation. Williamson (1996) says this
is an unsuccessful strategy. The hierarchy, with controlled
adaptation, should be the governance structure of last
resort, and the USSR was brought down by the high cost
of its bureaucracy, not economic inefficiency. Williamson
quotes Alfred Whitehead who said that civilization
advances by reducing the number of activities that need to
be subjected to conscious control.
References
Archibald, R.D. (1992). Managing High-Technology Programs
and Projects (2nd edn). Wiley, New York.
Belbin, R.M. (1993). Team Roles at Work. Butterworth-Heinem-
ann, London.
Boissevain, J. (1974). Friends of Friends. Blackwell, Oxford.
Bower, D. and Skuntzos, F. (2000) Partnering, benchmarking
and alliancing. In The Gower Handbook of Project Manage-
ment (3rd ed), eds J.R. Turner and S.J. Simister. Gower,
Aldershot.
Coase, R.H. (1964) The regulated industries: discussion. Amer-
ican Economic Review 54(May).
CCTA (1996). PRINCE 2: Project Management for Business. The
Stationery Office, London.
Fayol, H. (1949). General and Industrial Management. Pitman,
London.
Frame, J.D. (1995). Managing Projects in Organizations (revised
ed). Jossey-Bass, San Francisco.
Handy, C.B. (1995). The Age of Unreason. Arrow Business
Books, London.
Huczynski, A.A. (1996). Management Gurus: What Makes Them
and How to Become One. International Thomson Business
Press, London.
Kanter, R.M. (1983). The Change Masters: Corporate
Entrepreneurs at Work. Allen and Unwin, London.
Keegan, A.E., Turner, J.R. and Paauwe, J. (1999) The people
side of project-based organizations. RIBES Research
J. RODNEY TURNER,
Erasmus University, Fac-
ulty of Economics, Depart-
ment of Marketing and
Organization, Room H15-3,
Burgemeester Oudlaan 50,
3062 PA, Rotterdam, The
Netherlands. E-mail: turner
@few.eur.nl/rodney.turner@
supanet.com
J. Rodney Turner is Pro-
fessor of Project Management at Erasmus University,
Rotterdam, Operations Director for the European Con-
struction Institute, Benelux Region, and Director of
EuroProjex, the European Centre for Project Excel-
lence. His many books include the best-selling Hand-
book of Project-based Management, and he is Editor
of The International Journal of Project Manage-
ment.
European Management Journal Vol. 19, No. 3, pp. 254–267, 2001
View publication stats
Paper 9924. Rotterdam Institute of Business Economic
Studies, Erasmus University Rotterdam.
Keegan, A.E. and Turner, J.R. (2000) Managing human
resources in the project-based organization. In The Gower
Handbook of Project Management (3rd edn), eds J.R. Turner
and S.J. Simister. Gower, Aldershot.
Levett, C. (1998) The organization design of the project-ori-
ented company. In Projects and Processes, Proceedings of
the Austrian PM Days 1998, eds R. Gareis and D. Zuchi.
Project Management Group, University of Economics
and Business Adminstration, Vienna.
Mariotti, S. And Cainara, G. (1986) The evolution of trans-
action governance in the textile-clothing industry. Journal
of Economic Behaviour and Organization 7(September).
Morgan, G. (1997). Images of Organization (2nd ed). Sage Publi-
cations, Thousand Oaks, CA.
MPA (1992) Project Sovereign. Proceedings of Seminar 48. Major
Projects Association, Oxford.
Murray-Webster, R. and Thiery, M. (2000) Managing pro-
grammes of projects. In The Gower Handbook of Project
Management (3rd ed), eds J.R. Turner and S.J. Simister.
Gower, Aldershot.
Smith, A. (1776). The Wealth of Nations. Stratton and Cadell,
London.
Taylor, F.W. (1913). The Principles of Scientific Management.
Harper and Row, New York.
Turner, J.R. (1999). The Handbook of Project-based Management
(2nd ed). McGraw-Hill, London.
Turner, J.R. and Cochrane, R.A. (1993) The goals and methods
matrix: coping with projects with ill-defined goals
and/or methods of achieving them. International Journal
of Project Management 11(2).
Turner, J.R. and Keegan, A.E. (1999) The versatile project-
based organization: governance and operational control.
The European Management Journal 17(3).
Turner, J.R. and Keegan, A.E. (2000) The management of oper-
ations in the project-based organization. Journal of Change
Management 1(2).
Turner, J.R., Keegan, A.E. and Crawford, L. (2000) Learning
by experience in the project-based organization. In Pro-
ceedings of PMI Research Conference 2000, Paris, ed. L. Ged-
ansky. Project Management Institute, Sylva, NC.
Wagner, H.M. (1969). Principles in Operations Research. Pren-
tice-Hall, New York.
Williamson, O.E. (1996). The Mechanisms of Governance. Oxford
University Press, New York.
ANNE KEEGAN, Erasmus
University, Faculty of Eco-
nomics, Department of Mar-
keting and Organization,
Burgemeester Oudlaan 50,
3062 PA, Rotterdam, The
Netherlands. E-mail: kee-
gan@few.eur.nl
Dr Anne Keegan is Univer-
sity Lecturer at the Rotter-
dam School of Economics
and, jointly, at the Erasmus University. Her research
centres on project-based organizations, HRM in
knowledge-intensive firms, and new forms of organiz-
ing. She is a Partner in a major European-wide study
into the Versatile Project-based Organization.
267