Systems thinking approaches to address complex issues in project management
Dr. Tim Haslett, Visiting Lecturer, University of Technology Sydney, Australia
Dr. Jim Sheffield, Senior Lecturer, Victoria University, Wellington, New Zealand
Project management and systems thinking overlap. Surprisingly, project managers do not seem to use simple systems thinking tools even though these provide unique benefits in framing and solving problems that arise from multiple perspectives and relationships. The purpose of this paper is to introduce selected systems thinking concepts and tools and describe their application to the management of complex projects. The benefits of the application of “hard” and “soft” systems thinking tools at particular phases of the project lifecycle is also discussed.
Project management has had a close relationship with systems approaches, in particular, systems engineering (SE), when operations research, systems engineering and project management were used to manage large research and development projects by the defence in the mid 90s. According to Johnson (1997), project-based working evolved during this time to assist scientists and engineers who employed technologies that were complex, innovative, and uncertain in the development of new weapon systems. For example, during this period, the matrix organization was developed to help overcome the limitations of conventional line and staff organizations in managing projects effectively. Soon after World War II project management and SE began to evolve as separate academic disciplines. This was followed by the development of separate bodies of knowledge for these professions, such as the PMBOK® and SEBOK®, respectively, to promote best practice. Although project management and SE have taken separate paths in terms of professional practice, they share many concepts, such as the life-cycle approach to conceptualise, design, and implement systems and projects. The Defence sector and its contractors continued to develop project management techniques, the use of which soon spread to large engineering, construction, information systems, and organizational change projects. As project management spread its wings, “softer” or more people-centric issues, such as personnel management, motivation, team performance, team structure, stakeholder management, negotiation, communications management, and leadership, were added to the list of Best Practices (McConnell, 1996). Systems thinking approaches also started to move away from “hard systems” (product and technology-centric) to “soft systems” (people and process) approaches (McConnell, 1996, p. 11). Jackson (2003, p. 22) states that “Soft systems thinkers abandoned the notion that it was possible to assume easily identifiable, agreed-on goals that could be used to provide an objective account of the systems and its purposes. This was seen to be both impossible and undesirable given multiple values, beliefs and interests.” Through the work done by Peter Checkland and his associates at Lancaster University (Checkland, 1981), soft systems approaches found their way into information systems projects where success is more dependent on people and process than on intendedly objective accounts of the product and its constituent technologies. The idea that different methodologies were needed to manage projects also became a concern for project management scholars and practitioners (Turner & Cochrane, 1993), and research was carried out by PMI to categorise projects and recommend ways to manage projects based on their characteristics (Crawford, Hobbs, & Turner, 2004). In the general management field, the publication of The Fifth Discipline (Senge, 1990) in the 1990s by Peter Senge and his associates helped managers to appreciate how their actions were influenced by feedback and delays. Several papers are now being published in project management journals urging project managers to use system thinking approaches, especially in efforts to “tame” complexity (Pollack, 2007; Winter & Checkland, 2003; Yeo, 1993). The current focus on managing complex projects (Cicmil, Cooke-Davies, Crawford, & Richardson 2009; Remington & Pollack, 2008; Williams, 2002) suggests that project managers would benefit by understanding the application of systems thinking approaches to deal with the complexities arising in their projects. The purpose of this paper is to introduce selected systems thinking concepts and tools and describe their application to the management of complex projects. The benefits of the application of “hard” and “soft” systems thinking tools at particular phases of the project lifecycle is also discussed. The paper is based on the experience of two of the authors in teaching systems thinking to practising project managers in a masters course in project management at the University of Technology Sydney in Australia accredited by the PMI Global Accreditation Centre (GAC).
Why Systems Thinking?
Kim (1999, p. 2) defined a system as “any group of interacting, interrelated, or interdependent parts that form a complex and unified whole that has a specific purpose.” Based on this definition, a project itself can be considered a system. Some terms commonly used by systems thinkers are introduced in this section, before systems thinking concepts are explored further. The boundary of a system is the scope of interest or concern and can change as the scope of interest changes. The parts of a system inside the boundary interact with each other but also with the environment that exists outside the boundary. A project may have a boundary based on its initial scope but this boundary could change as the scope changes. The project teams inside the project boundary could have interactions with external stakeholders if they are considered to be part of the environment. A system takes inputs from its environment and transforms them into desired outputs. In a project the requirements could be considered as inputs that are transformed by the project team into products or services as outputs. A system has structure that defines its parts and their relationships and uses processes or a sequence of activities to perform a function. Project implementation employs structures, processes, and activities. Systems are generally open, i.e., they interact with the environment. They are organized by a hierarchy and exhibit emergence. The notion of hierarchy in systems thinking has to do more with vitality, survivability, and purpose rather than the notion of command and control usually depicted through organization charts (Boardman & Sauser, 2008). The idea of emergence is often described as “a system is more than a sum of its parts” and is a result of the dynamic interactions between the parts. Such interactions may lead the system into a chaotic state that settles down to a new state after a while. Project managers working on complex projects often state that they have experienced situations when everything seems to be going out of their control but the project finally settles down into a new state of equilibrium (Syed & Sankaran, 2009).
Causal Loops and System Archetypes
Levels of Thinking
Senge (1990) helped to introduce the idea that cause and effect are often linked (perhaps via intervening events) in a circular arrangement called a “causal loop.” An appreciation of the working of a causal loop, and its effect on other loops with which it may be entangled, promotes an understanding of the more complex issues that bedevil project management. For example, the study of causal loops reveals recurring patterns called system archetypes that have proven most useful in identifying leverage points and the root causes that underlie difficult problems.
Exhibit 1 shows this relationship. For example, the mental model of a large organization steeped in project management methodologies may create a bureaucratic structure that has to be used in every project that may lead to a pattern of cost overruns in small projects that may not need all the processes used for larger projects.
Exhibit 1. Four levels of thinking (Maani & Cavana, 2007, p. 15)
A causal loop diagram (CLD) provides a visual representation useful in identifying the cause-and-effect relationship among a set of factors that operate together as a dynamic system. For example, stress of overtime in a project could result in burnout. Such relationships are accentuated by feedback loops. Exhibit 2 shows a causal loop that could be found in projects on the left side of the exhibit.
As project completion times blow out, the pressure on staff, particularly the project manager, increases, resulting in an increase in the number of staff leaving, or turnover. Staff turnover erodes the skill base, which then contributes to increasing rework. There are delays in the system as the increasing pressure is often not immediately obvious as it is absorbed until a crisis develops. There are also further delays in hiring staff. Both these factors can exacerbate the downward spiral. The idea of drawing a CLD is to look for leverage points.
The diagram on the right indicates a possible leverage point where the effect of the emergence of rework is mitigated by hiring as the rework begins to surface. This policy intervention may short-circuit the crisis generated through staff turnover. The purpose in designing CLDs is to understand the fundamental dynamics of the system and to develop policy levers to control variation in the system.
Exhibit 2. Two CLDs with a policy lever included on the right side
It is also necessary to understand the concept of Behaviour Over Time (BOT) when analysing causal loops. A CLD is usually analysed using a series of BOT graphs. The left side of graph in Exhibit 3 shows the freehand BOT with the planned staffing for the project, the one that is feared if the project deadlines blow out and the actual behaviour. Lyneis, Cooper, and Els (2002) made the point that projects need to have a learning structure built into the project so that predicted and actual behaviour can be understood, and the learning from the past built into future projects. Lyneis and colleagues employed the dynamic relationships represented in CLDs to identify the root causes of the decline in staff productivity during the first part of a project, the so-called “Rookies and Pros” effect shown in Exhibit 3.
Exhibit 3. Behaviour over time graphs for staffing levels and productivity after Lyneis et al. (2002)
In the example in Exhibit 3 it would be expected that productivity would decline slightly when new staff are taken on. This enables the project manager to resist the temptation to hire additional staff as a decrease in productivity is detected, and to find ways to train or mentor new staff to come up to speed faster.
Closely associated with causal loops is the concept of system archetypes. Kim (1992) explained that “system archetypes are one class of tools that capture ‘common stories’ in systems thinking—dynamic phenomena that occur repeatedly in diverse settings.” Typical system archetypes found in projects are: fixes that fail, shifting the burden, and tragedy of the commons. The structure of tragedy of the commons is shown in Exhibit 4 and briefly explained using an example that can occur in a project organization.
Many projects in an organization use a Project Management Office (PMO) to provide essential services at competitive costs to projects based on a priority system. In order to get the attention of the PMO each project manager may increase the priority of his or her project's requests in an unsuccessful attempt to gain advantage over competing projects. This leads to a PMO that appears to be overloaded and this, in turn, may result in outsourcing of services and an increase in costs defeating the purpose of providing a PMO as a common resource to reduce cost. An appreciation of the dynamics of the situation explains why the intendedly rational actions of individuals produced a collective behavioural system that is dysfunctional. When such a pattern is detected it is difficult to resolve the problem at the individual project manager level. Asking questions about the dynamics of the system such as “How can the long-term effects of this misuse of PMO be made transparent to the individual project managers?” can assist in finding a mutually acceptable solution.
Exhibit 4. Typical structure of the “tragedy of the commons” archetype (http://en.wikipedia.org/)
Stock and Flow Diagrams
Although CLDs are useful in understanding many situations, they become unwieldy as the number of loops increase. The dynamics associated with the loops are also difficult to simulate using computers. Stock and flow diagrams can be used to overcome this limitation
Lyneis et al. (2001) cited a review by Morris and Hough that found that in 3,500 projects overruns were typically between 40% and 200%. They also note that recent developments in project management techniques do not seem to have improved the situation significantly. They provide an example of the typical BOT graphs that indicate staffing build-up through a project. Ideally, staffing builds up and falls away as work is completed. Lyneis et al. (2001, p. 274) observed that project staffing is “Often slower to build up than planned and exceeds planned levels for an extended period. Often there is a second peak once rework is discovered and needs to be corrected.”
Lyneis et al. (2001) also pointed out that productivity does not remain constant for the duration of the project. In reality, productivity typically falls from the beginning through the middle of the project before rising at the end. “Projects continue to perform poorly because projects are fundamentally complex dynamic systems and most project management concepts and tools either (1) view the project statically or (2) take a partial, narrow view in order to allow managers to cope mentally with complexity. These tools lead each manager to believe that each project is unique, which makes systemic learning across projects difficult” (Lyneis et al., 2001, p. 239).
The use of SD at the beginning of the implementation phase lays out the workflow for the project. There are two main advantages in doing this. First, the resource requirements for the project can be identified; and, second, the effects of feedback systems, such as those created by lags and delays rework, can be modelled. The inclusion of co-flows of staff can help determine the time required for the project and an estimation of likely delays as a result of resource constraints. Exhibit 5 is a simple example of such a structure. With very little modification, this model can be adapted to change the staffing co-flows to finance or material co-flows.
Exhibit 5. Work and staffing co-flows
The main flow is the work that comes to the design department and work that leaves the design department and goes to the construction department. There is a re-work loop for work that must be redesigned. The co-flows are two stocks of staff for each department. The work progresses through the main chain at a rate at which that number of staff can complete their respective tasks. The numbers of staff, in turn, are dependent on the respective hiring and quitting rates. This model enables managers to understand workflow efficiencies and potential delays occasioned by the number of staff available and the amount of rework generated.
Exhibit 6 is a stock and flow version of the staffing model known as the “Rookies and Pros” model. This model allows managers to understand the negative effects caused by experienced staff leaving the project, particularly when the workforce is being expanded either a result of increased work or increased rework.
Exhibit 6. The Rookies and Pros model
The critical dynamic here is the pros’ workload, made up not only of their normal workload, but also by the number of rookies they need to train. As the number of rookies increases, the pros’ workload increases, with the lagged effect that pros begin leaving. This increases the number of rookies, as a result of new hires, and further increases the average workload for the remaining pros. This situation can be exacerbated when there are extra hires as result of project changes or rework.
Simple models, such as those just illustrated, may include 20 or more simultaneous difference equations that must be calculated for each day of the project, a task well beyond the human brain even when using a spreadsheet. The impact of more complex interactions and feedback systems can only be understood through simulation modelling. Building dynamic models before the planning and contract negotiation stages of the project provides insights into resource requirements and the impact of delays caused by rework cycles.
Soft Systems Methodology
The concept and design phases involve the initial conception of a project. Soft Systems Methodology (SSM) may be used at this stage of the project to clarify the benefits that each stakeholder may expect to accrue from the project. SSM was developed by Peter Checkland and his associates at Lancaster University (Checkland, 1981) to deal with ill-structured problems that were not handled well by the hard systems approach associated with product and technology-driven approaches to systems engineering. According to Jackson (2003, p. 185), a soft systems approach “seeks to work with different perceptions of reality, facilitating a systemic process of learning in which different viewpoints are examined and discussed in a manner that can lead to purposeful action in pursuit of improvement.” SSM provides some tools and concepts, such as the use of rich pictures and root definitions that can be used by project managers to clarify the purpose of a project.
Rich pictures are informal drawings that express how an individual feels about a situation. The goal is richness of personal expression, unrestrained by social conventions and unconstrained by predetermined frameworks. The use of drawing conventions and tools for more objective representation of a technical problem is expressly discouraged. Rich pictures are often constructed by stakeholders at the start of a SSM cycle as part of an interview or small group interaction. Monk and Howard (1998, p. 22) stated that a “rich picture is intended to be a broad, high-grained view of the problem situation.” Rich pictures are usually drawn to identify structures, processes, and concerns (Monk & Howard, 1998). A rich picture expresses how people perceive the problem, not how it is analysed.
Exhibit 7 is a rich picture developed by students in the project management course at the University of Technology Sydney (UTS). The students were asked to develop a rich picture for an organization that wished to shift the Cricket World Cup from India to another venue immediately after the bomb blasts in Mumbai in 2008.
Exhibit 7. UTS students’ rich picture of the Cricket World Cup organization
The rich picture identifies the issues and related processes in a problem situation. Rich pictures are also useful in surfacing the mental models and metaphors associated with the situation. The metaphors are often indicative of mental models, values, and attitudes that are unstated but extremely influential in the governance and management of the project.
Rich pictures provide an excellent method of surfacing the true diversity of stakeholders’ goals. Understanding the diversity of the goals that motivate each stakeholder at the beginning of the project can be extremely useful.
The root definition helps to prepare a concise statement of what a system is expected to achieve in its most fundamental form (Jackson, 2003, p. 192). A root definition of a project serves as a high-level definition of overall purpose. A “CATWOE analysis” is usually performed to develop the root definition that clarifies a problem situation which is surfaced using rich pictures.
CATWOE is an acronym that includes:
- Clients—Those who benefit (or suffer) from the operations of the organization, e.g., customers.
- Actors—The individuals, groups, institutions and agencies who perform the functions of the organization.
- Transformations—The processes that transform inputs to outputs.
- Weltanschauung—The world-view or bigger picture that encapsulates the problem situation and expresses the way the organization views the world.
- Owners—The people, who have the ultimate say over the project, provide the resources and can pull the plug on the project.
- Environment—The broader constraints that act on the situation. These may be ethical limits, the law, financial constraints or limited resources.
CATWOE is a simple descriptive framework that requires little explanation apart from the fact that it normally helps to do the Weltanschauung or world-view first. This acts as a framing exercise and provides a useful perspective for the rest of the analysis. The CATWOE analysis helps in the formulation of the root definition. Checkland and Poulter (2006, pp. 92-93) provided an example of the root definition of a system to upgrade the skills of a department in a chemical firm.
“A system owned by XXXX which together with users, identifies those scientific, technical and commercial staff in research, technical services, development, production and business functions who require professionally provided information to do their jobs effectively and key users in particular—which identifies the level and nature of those requirements, i.e., breadth and depth of subject matter, detail, and precision output.” Note that the root definition should identify specific actions and associated measures of efficacy, efficiency and effectiveness.
In this paper, we have tried to provide a brief overview of the concepts and tools used by systems thinkers that project managers have found useful in taming complex projects. These tools supplement rather than replace the tools recommended as Best Practices by professional project management institutes such as PMI. They can be used in addition to conventional tools to deal with some of the “wicked problems” that arise in projects that cannot be effectively handled by standard tools and techniques. Wicked problems are those that are not amenable to linear approaches that assume a system composed of relatively static objects with simple well-understood interconnections.
This paper seeks to convey an appreciation of the dynamic aspects of systems, tools and techniques for managing projects that involve circular patterns of cause and effect, and the need to surface and stabilize stakeholders’ multiple and conflicting interpretations of system requirements. These are issues that require paying attention to the emergent understanding of people, and the coordination of their thinking and behavioural processes. The project management literature is replete with examples of systems that failed because they were developed in the face of unresolved people-centric or “soft system” issues. Such projects must be managed with appropriate attention given to both “soft systems” and “hard systems” at various phases of the system development lifecycle.
To illustrate where and when “soft system” techniques and tools can be used, we have divided the project lifecycle into three broad phases (Haslett & Sankaran, 2009):
- 1 The concept and design phase
- 2 The implementation phase
- 3 The evaluation phase
Exhibit 8 summarises the phases of a project where some of the tools and techniques discussed can be applied.
Exhibit 8. Recommended systems methodologies to be used at various phases of a project
Boardman, J., & Sauser, B. (2008). Systems thinking: Coping with 21st century problems. Boca Raton, FL: CRC Press.
Checkland, P. (1981). Systems thinking, systems practice. Chichester, NY: Wiley.
Checkland, P., & Poulter, J. (2006). Learning for action: A short definitive account of soft systems methodology and its use for practitioners, teachers and students. Chichester, NY: John Wiley.
Cicmil, S., Cooke-Davies, T, Crawford, L., & Richardson, K. (2009). Exploring the complexity of projects: Implications of complexity theory for project management practice. Newtown Square, PA: Project Management Institute.
Crawford, L., Hobbs, B., & Turner, J. R. (2004). Project categorization systems and their use in organizations PMI research on categorization of projects: An empirical study. In Slevin, D. P., David, C., & Pinto, J. K. (eds.) Innovations: Project management research. Newtown Square, PA: Project Management Institute.
Jackson, M. C. J. (2003). Systems thinking: Creative holism for managers. Chichester, NY: Wiley.
Johnson, S. B. (1997). Three approaches to big technology projects: Operations research, systems engineering and project management,. The Society for Higher Technology, pp. 891- 919.
Kim, D. H. (1993). System archetypes I: Diagnosing systemic issues and designing high-leverage interventions. Waltham, MA: Pegasus Communications.
Kim, D. H. (1999). Introduction to systems thinking. Innovations in Management Series. Waltham, MA: Pegasus Communications.
Lyneis, J. M., Cooper, K. G., & Els, S. A. (2001). Strategic management of complex projects: A case study using system dynamics. System Dynamics Review, 17(3), 237 – 261.
Maani, K. E., & Cavana, R. Y. (2006). Systems thinking, system dynamics: Managing change and complexity. North Shore: Pearson Education New Zealand.
McConell, S. (1996). Rapid development: Taming wild software schedules. Redmond: Microsoft Press.
Monk, A., & Howard, S. (1998). The rich picture: Reasoning about the work context. Interactions, March-April, 21-30.
Morris, P., & Hough, G. (1987). The Anatomy of Major Projects. New York: Wiley.
Remington, K. & Pollack, J. (2008). Tools for complex projects. Aldershot: Gower.
Pollack, J. (2007). The changing paradigms of project management. International Journal of Project Management, 25(3), 266-274.
Haslett, T., & Sankaran, S. (2009). Applying multi-methodological system theory to project management. Proceedings of the 53rd Meeting of the International Society of Systems Sciences- Making Living Systems Unremarkable. Brisbane, Australia, July 12-17.
Senge, P. (1990). The fifth discipline: The art and practice of the learning organizations. Doubleday.
Syed, G., & Sankaran, S. (2009). Investigating an interpretive framework to manage complex information technology projects, IRNOP IX Conference, Berlin, Oct. 11-13.
Turner, J. R., & Cochrane, R. A. (1993). Goals-and-methods: Coping with projects with ill-defined goals and/or methods of achieving them. International Journal of Project Management, 11(2), 93-102.
Williams, T. (2002). Modelling complex projects. London: Wiley.
Winter, & Checkland, P. (2003). Soft systems: A fresh perspective on project management. Civil Engineering, 15(4), 187-192.
Yeo, K. T. (1993). Systems thinking and project management: Time to unite. International Journal of Project Management, 11 (2), 111- 117.
©2010 Sankaran, Haslett and Sheffield
Originally published as a part of 2010 PMI Global Congress Proceedings – Melbourne, Australia