Complex systems lack explicit boundaries. Any boundaries which exist are imposed by people who are attempting to constrain and simplify the system for a particular purpose and therefore these boundaries are artefacts affected by the designer’s biases, interests and vision. This is not bad situation, we just need to remain aware of how boundaries were set. Boundaries are essential because without them we are forced to consider an infinite number of connections—everything is connected to everything else—which is hopeless. Designers, therefore, need ways to define system boundaries which delineates the system in ways which are both relevant and manageable.

Typically, boundary definitions are set without thought. Designers rely on their intuition and make decisions like: “the culture change programme will focus on the call centre and we will concentrate on the managers’ viewpoints.” This decision leaves out other stakeholders such as the call centre operators and the human resources department in headquarters. Whenever boundaries are set people are left out. The question for designers is simply: “are we leaving people out for the right reason?” Boundary setting is the first important step in designing interventions for a complex system. A practical approach to boundary setting is a fundamental tool for complexity-base designers.

Weiner Ulrich (1983; 1996) provides a practical boundary-setting approach based on considering four types of stakeholder and asking three questions from the perspective of each. The four stakeholders are:

• clients—the people or groups who benefit from the interventions;
• decision-takers—those people who allocate the budget to implement the interventions; and
• planners—the people responsible for designing the interventions;
• bystanders—people affected by the decisions but not involved in the process (Ulrich called this type, the witness).

Ulrich suggests the questions be asked from two perspectives: what ought to be the answer and what is the answer. These questions can be summarised as having the following dimensions:

• client—1) sense of purpose; 2) clash of purposes;
• decision taker—3) control of resources; 4) lack of control (environmental conditions);
• planners—5) types of expertise; 6) likelihood of success; and
• bystanders—7) voice of the affected; 8) clash of worldviews.

Designers can run a simple workshop format with representatives from each stakeholder type. I have used a challenge-and-respond format combined with mixing people throughout the workshop to ensure we expose the maximum variety of viewpoints.

Setting boundaries in this way ensures the system is defined so that it is both relevant and manageable. Perhaps more importantly, it specifically includes a broad set of stakeholders in the improvement process. Complex problems never have a right or wrong answer. In fact it’s impossible to objectively measure whether a complex problem has been ‘fixed’. Rather, the stakeholder must believe that improvements are being made and this requires their active involvement in the intervention design and monitoring.

Ulrich, W. 1983. Critical Heuristics of Social Planning: A New Approach to Practical Philosophy. Bern: Haupt.
Ulrich, W. 1996. A Primer to Critical Systems Heuristics for Action Researchers. Hull: Centre for Systems Studies, University of Hull.

Thanks to Bruce McKenzie for putting me on to Ulrich’s work.