1 | P a g e R e v i s e d F I N A L N o v 2 0 1 6
Understand, reduce, respond.
Project complexity management theory
and practice.
Purpose:
This paper contributes to the understanding of complexity and its management from an OM
perspective, building on and extending the systematic literature review published in this journal in
2011, and provides a foundation for exploring the interactions between complexities and responses.
Design methodology / approach:
The paper takes a subjective view of complexity, focusing on the ‘lived experience’ of managers. It
takes an updated systematic literature review, and demonstrates the comprehensiveness of a
framework to classify complexities of projects. It reports the findings from 43 workshops with over
1100 managers.
Findings:
Firstly, the complexity framework is effective in aiding understanding. Secondly, and somewhat
unexpectedly, managers were able to identify strategies to reduce the majority of complexities that
they faced. Thirdly, the workshops identified a typology of responses to residual complexities.
Research limitations / implications:
The framework has demonstrated its utility and a gap in understanding emergent complexities is
identified. The framework further presents the opportunity to explore the recursive nature of
complexity and response.
Practical implications:
This paper provides a framework that is both comprehensive and comprehensible. We demonstrate
that complexities can be reduced and provide a means to assess responses to residual complexities,
including potentially matching managers to projects.
Originality / value:
This work extends the previous systematic review combined with extensive empirical data to
generate findings that are having impact in practice, and have the potential to strengthen a
relatively neglected area within OM. A research agenda is suggested to support this.
International Journal of Operations and Production Management, Volume 37, Issue 8, 2017, pp. 1076-1093
DOI:10.1108/IJOPM-05-2016-0263
Published by Emerald. This is the Author Accepted Manuscript issued with: Creative Commons Attribution Non-Commercial License (CC:BY:NC 3.0).
The final published version (version of record) is available online at DOI:10.1108/IJOPM-05-2016-0263. Please refer to any applicable publisher terms of use.
2 | P a g e R e v i s e d F I N A L N o v 2 0 1 6
1 Introduction: a journey in complexity
The point of departure for this journey was the work of Stacey (1993) in organisational strategy, and
subsequently Williams (2005). The former was highly vocal in the discussion of complexity in
business and management, and how organisations displayed many of the characteristics of complex
systems. The latter considered how organisations were responding to complexities in the context of
projects. This is particularly relevant to project-based operations as they are noted to be, in many
cases, highly dependent on attempts at standardised processes (Geraldi, Maylor and Williams, 2011)
despite being defined in OM terms by high variety of process and low volume of throughput (Slack
and Brandon-Jones, 2015).
‘Complex responsive processes’ (Stacey, 2001) subsequently became popular topics in strategy.
These went beyond mechanistic notions of strategy, with linear conceptions of cause and effect in
organisations supplemented by a more evolutionary, organic conceptualisation. The approach
provided better explanation of observed behaviours of organisational systems than had been
achieved previously. Organisations were shown to be increasingly complex systems as a result of
ever more dynamic environments, extended control loops and ongoing addition of system
constraints such as legislation and corporate social responsibility requirements.
Related to this development was the application of analogies for social systems with other complex
systems that proved insightful, including weather systems, mathematical systems (e.g. Mandelbrot
sets) and ant colonies (Doerner et al, 2006). These insights included sensitivity to initial conditions,
the role of ‘weak signals’, unpredictability of cause and effect and a boundary region, the edge of
chaos. On this last insight, Pascale et al. (2001, pp. 61) note that, “The edge [of chaos] is not the
abyss. It’s the sweet spot for productive change.” Kaufmann’s NKC (Kaufmann, 1993; Vidgen and
Wang, 2006) also provided some interesting discussions and a language around complexity
landscapes and co-evolution of systems.
Williams (2005) subsequently identifed the challenge to “understand what makes projects complex
to manage and to provide a common understanding of the ‘lived experience’ of managing in a
project context. This will provide both academics and practitioners with a shared language to name
and make sense of …how to both shape and respond to this complexity. Such a common language
will enable us to connect findings experiences and knowledge accumulated in different
environments… as well as in different parts and phases of projects.” (paraphrased in Geraldi et al,
2011, pp. 968).
Pursuant to Williams’ challenge were our own investigations into the complexity of projects. We
observed project practitioners struggling with the complexity of their work, and noted how
organisations appeared to be able to create complexities for project delivery, rather than manage
them effectively. Indeed, managers we worked with said that they perceived complexity growing in
projects at a faster rate than the capability to work with this complexity was evolving. Although
subjective and therefore difficult to quantify, this represents a potential complexity crisis; delivery of
vital projects was (and still is) hampered by the sheer complexity of the challenge placed before
managers.
In attempting to apply the insights from complexity science to help with this challenge of managing
complexity more effectively, there were some particular issues. Notably, the notion of complexity
itself was not well understood in the project context (Vidal and Marle, 2008), with arguments
prevailing about whether something is ‘complex’ or ‘complicated’. Typically, these arguments
generate much heat but little light; one person’s complex was another’s complicated. In addition,
3 | P a g e R e v i s e d F I N A L N o v 2 0 1 6
whilst the analogies initially yielded insight, they became increasingly distant from the problems of
practice. Social systems are not equivalent to weather systems, mathematical systems or biological
systems, even if some of the characteristics are comparable.
We were assisted in our journey by Cicmil et al. (2009) who used the classification of complexity in
projects and the complexity of projects to describe two approaches to complexity in the PM
literature. The first denoted a rationalist approach to the phenomenon, objectifying complexity and
the application of the body of work of complexity science. The second, ‘complexity of’, is a subjective
approach, representing the lived experience of managers of what they termed ‘complexity.’ Our
work has developed in this latter vein. This trajectory has echoes of the development of quality
management in OM, where initial work focused on quality as an objective construct, defined by
conformance to standard. The result was a concept of quality management that focused on ensuring
such compliance through objective measurement. The assumed goodness of the standard was never
questioned. The work of Garvin (1984) broadened the concepts to initially include perceived quality
– moving beyond the objective and measurable into subjective and perceptual. Parasuraman et al
(1985) subsequently expanded this into a whole range of measures for service quality. Similarly,
these rely on a subjective approach to the topic of interest being taken.
Our initial conceptual model is shown in Figure 1 and places ‘complexity of project’ as the
independent variable, with the managerial response being a function of that complexity. We
reasoned (Geraldi et al., 2011) that if we understood this independent variable, then we could make
an input to the call for greater understanding of contingency in practice, for OM (Sousa and Voss,
2008), project management (Thomas and Mengel, 2008) and more widely in organisational studies
(Brown and Eisenhardt, 1997).
Figure 1: Initial conceptual model
The first stage in the complexity journey is understand: understanding the independent variable,
Williams’ ‘lived experience’ or Cicmil et al.’s complexity of projects. The framework used here has
been derived from both a systematic literature review (Geraldi et al, 2011, updated here) and
extensive empirical work (Maylor, Turner and Murray-Webster, 2013). It comprises the dimensions
of structural, socio-political and emergent complexities.
The second stage in the complexity journey is an intervening step in our original conceptual model –
reduce. The nature of our work is to be highly engaged with practice and practitioners, and during
this work the discussion emerged concerning how complexity itself is ‘managed’ (or not), or has
potential for being managed. Subsequently, workshops with over 1100 managers demonstrated that
the majority of complexities they faced could either be removed or reduced. We were surprised by
the strength of this finding, but encouraged that it demonstrated the value for practitioners of
having a nuanced understanding of complexities, beyond the debates of complex vs complicated.
The third stage in the complexity journey was again from the initial challenge – respond. This was to
explore how the residual complexities can be conceptualised as providing the basis for the
4 | P a g e R e v i s e d F I N A L N o v 2 0 1 6
managerial ‘response’ and thus to understand better one aspect of contingency in this domain. We
report how this has been explored in its impact on project-level practices. Finally, we reflect on how
this response can be framed from a complexity perspective, and provide interesting avenues
pedagogically and for further research.
This work is important because projects continue to exhibit Flyvbjerg’s performance paradox
(Flyvbjerg et al, 2003) with the level of performance being in contrast to the levels of importance of
projects to organisations. Their complexity is rated as one of the reasons for this performance,
indeed reinforcing the notion of a complexity crisis.
This paper is structured as follows. The following section presents the results of the investigation
into the construct: complexity of projects. A summary of and an update to the systematic literature
review (Geraldi et al., 2011) published in this journal on complexity of projects is presented. It
describes the framework derived for understanding complexities and demonstrates that there have
been no additional concepts added in work published in the intervening years. The workshop
approach for the study in the management of complexity is described and the results presented. The
discussion then considers the response to the complexities and we conclude with practical and
research implications.
2. Literature review
The purpose of this review is not to repeat the systematic review of Geraldi et al (2011), which is
summarised here. Instead, our question to the literature was highly specific: have there been any
additions to the conceptual basis of ‘the complexity of projects’ since 2011? We eliminated the
complexity science approaches and mathematical modelling of specific contexts (including transport
systems and shipbuilding) as these did not meet the criteria of being relevant to complexity of projects.
The new search was based on peer-reviewed English-language journal articles since 2011. The first
stage of the search, using the EBSCO database was for “complex*” AND “project management” in the
subject (40 responses). We subsequently searched for the keyword ‘complex*’ in International Journal
of Operations and Production Management (11 responses) the International Journal of Project
Management (39) and Project Management Journal (10). We eliminated any non-peer-reviewed work
(notably book reviews) and the overlaps between the second stage of the search and the first.
Main findings from 2011 review
By conducting a temporal analysis of the literature on complexity of projects, the findings were
represented as shown in Figure 2. The build-up of the concepts into five main categories was
substantiated by both literature and empirical research (e.g. Shenhar and Dvir, 1996; Maylor, et al,
2008). This provided the basis for the claim that complexity of projects was sufficiently well
understood to move forward – conceptually, it had reached saturation as no new concepts were being
added.
5 | P a g e R e v i s e d F I N A L N o v 2 0 1 6
Figure 2: Historical Development of Complexity Frameworks (Geraldi et al, 2011)
First
appear
ance
(Baccarini,
1996)
(Dvir and
Shenhar,
1998; Shenhar
and Dvir,
1996;
Williams,
1997, 1999)
(Ribbers and
Schoo, 2002;
Xia and Lee,
2005)
(Dvir, et al.,
2006;
Shenhar and
Dvir, 2007;
Williams,
2005)
(Geraldi and
Adlbrecht,
2007;
Remington
and Pollack,
2007)
Later
appear
ance
(Austin, et
al., 2002;
Clift and
Vandenbos
ch, 1999;
Jaafari,
2003)
(Little, 2005;
Tatikonda and
Rosenthal,
2000)
(Benbya and
McKelvey,
2006;
Maylor, et
al., 2008;
Xia and Lee,
2004) (Haas, 2009) (Howell, et al., 2010)
Since 2011 – Understanding complexity of projects
Our own follow-up of this work included testing both the comprehensiveness and comprehensibility
of the framework. As reported in Maylor et al. (2013), the 5-category model met the requirements
for the first, but less so for the second. The solution was to include ‘pace’ into structural complexity.
The rationale for this was that as we saw from our empirical work, ‘pace’ is a measure of relative
resource intensity, and was consistent with other concepts in that dimension. The second was to
combine ‘uncertainty’ and ‘dynamics’ into a dimension called ‘emergent complexity’. The rationale
for this was that they were linked in practice with an uncertainty at one point in time often leading
to a dynamic situation at a later point in time. The three dimensions that have been demonstrated
to be of value in nuancing the description of complexity of projects, and typical concepts within each
dimension are shown in Table 1.
Uncertainty
Structural
Complexity
Dynamic
Pace
Structural
Complexity
Structural
Complexity
Structural
Complexity
Uncertainty
Structural
Complexity
Uncertainty
Pace
Dynamic
1996
Time
1997/99 2002/04 2005/06 2007/08
Socio-
Political
Structural
Complexity
Uncertainty
Pace
Dynamic
2009/10
Socio-
Political
