Decision Making / 5 / 5.1 / 5.1.1

5.1.1 Reliance On Stakeholders’ Views

It is an unrealistic expectation, even given what we now know, to walk straight in and solve wicked problems that have appeared in risky situations. Our starting position frequently involves relying heavily on stakeholders’ views of a particular problem or issue.

Recently I witnessed (and photographed for my grandchildren) the magical transformation of a cicada (hemiptera: cicadidae) from its earth burrowing beetle-like form into a flying insect with gossamer wings. This insect spends some seven years in the ground and in a single night burrows its way out of the ground to climb the nearest tree where it attaches itself by the claws at the end of each of its legs. See Figure 5-1.

In the following few hours it breaks out of its hard shell which splits open to reveal the pulsating soft body of the emerging winged insect. Gossamer wings magically drawn from small pockets no bigger than a few millimetres to uncurl to be several centimetres long. During this period when the winged insect is emerging, and indeed for the whole night, it is totally defenceless. It has no strength. It is unable to crawl until it is completely free of its shell. Even then, it is unable to fly. Its wings fully unfurl over the next few hours but do not become strong enough to sustain flight until exposed to sunlight. When risk managers find themselves highly reliant on a single gatekeeper of dubious competence, such as the shot-firer in the Royal Canberra Hospital implosion they become as vulnerable as the emerging cicada.

Doyle and Ford (1998: 3-29) argue that stakeholder views are almost always incomplete, fuzzy, linked to ingrained assumptions or involve imperfect knowledge, yet those views frequently form the basis for problem conceptualisation and strategy development. In preceding chapters, a strong case was put for maximum stakeholder involvement. Our reliance on the client’s view may be heaviest when dealing with unfamiliar or unusual problems, or during the conceptualisation phase of a project. In the conceptualisation phase we rely heavily on the views of one or more key stakeholders that are considered to be subject matter experts, as correct or incorrect as their views may prove to be, until we can formulate and validate our own views. As we proceed through conceptualisation to strategy development, we make many choices:

• Where do the boundaries of the problem space lie? What do we include in our consideration of the problem? What do we exclude?
• Where should we draw the boundaries to the solution space? What are the limits to what we include or exclude from our analysis?
• What are the main drivers that influence the observed behaviour, and to what extent?
• What is the most appropriate way to conceptualise the combining of influences at various points or nodes in the problem space?
• Where a model is built, is the model a sufficient replication of observed real-world behaviour? At what point do we know that the model we have built is sufficient?

In risk management, team members are involved in ongoing rounds of choice and decision making. It is helpful to view risk management as participatory management where the team is involved in group-model-building activities. If an outside agent, such as a consultant, is involved we should view the activities as action research.

Applying the risk-management process is an iterative modelling process. It involves building, repeatedly stepping through, mentally simulating and making adjustments to the model; iterations of choice and decision making continue through every stage¹. Just how good those choices and decisions are dictates the suitability of the risk-management strategies that are ultimately produced.

The formative stages of decision making cannot be trivialised without risk of basing decisions on invalid, biased assumptions or inappropriate problem conceptualisation. This conceptualisation process, and subsequent analysis must recognise and as far as possible compensate for the human cognitive limitations, whilst building on strengths. Further, the model building and mental simulation processes must be cycled through a sufficient number of times and sufficiently quickly that they actually become processes of dynamic review and provide the basis for monitoring. What we are trying to achieve is akin to rapidly flicking through a sequential set of picture cards sufficiently fast that we see, in our recent memory at least, a continuous ‘movie’ in which we can relate what is occurring now with what went before.

Kleinmuntz (1993) recognised when it comes to decision making in dynamic environments, such decision making can only be considered reliable when there is juxtaposition of previous decisions and observed feedback effects:

‘… Diehl (1992) examined decision rules in a very simple inventory management task. Diehl varied both the length of the delay and the complexity of feedback structure. Performance was most effective with simple, undelayed feedback structure. As both delays and complexity increased, performance deteriorated… (Kleinmuntz, 1993: 226).’

Hence it is necessary to cycle repeatedly through the risk-management process from the very beginning where we establish the context, through risk identification, risk analysis and evaluation and development of risk-management treatments.

Without this continuous and rapid cycling through the process, we become susceptible to the threat of failing to see the relationship between cause-and-effect and how feedback loops operate. This raises the likelihood that we will suffer the ‘misperception of feedback’ described by Sterman (1989a; 1989b; and 1989c.) and by Kleinmuntz (1993) as the ‘misperception of the implications of feedback’.

There are many things such as being presented with a mass of information that serve to mask the systemic structures of the risk context; it is these systemic structures that underlie the dynamic behaviour we face as risk managers.

Further, systemic response to risk managers’ remedial strategies is frequently counter-intuitive. This counter-intuitive behaviour is created by the systemic structure mentioned above.

In the field of system dynamics modelling², it is becoming commonplace to build management flight simulators with the aim of exposing executives to the dynamic response of the systems they are responsible for managing. Executive managers have the opportunity to interact with management flight simulators and test their strategies. They get to play in these micro-worlds as frequently as they choose. Despite being able to play the game over and over, research has shown that they have little chance of uncovering the complex set of interrelationships and underlying operating mechanisms that conspire to produce the observed dynamic responses. This is also the case in real-world management.

When a manager is forced to rely on intuition, that is the application of decision heuristics, the outcomes in this dynamic world are strongly dependent on the ecological validity of the heuristics used by the player/manager. As suggested at Chapter 1, we all carry a number of heuristics in what Gigerenzer calls our adaptive toolbox. Ecological validity of items in our adaptive toolbox determines the extent to which they are effective in specific environments.

Adapting to new cues in the contrived environment, such as that replicated by the management flight simulator or micro-world is difficult because much of the behaviour being observed at any time, in part at least, is the product of earlier attempts at corrective action. This is exacerbated when the player cannot possibly discover the underlying systemic structures, which produce the dynamic behaviour. This situation of ‘flying blind’ changes only when we as managers become involved in fundamental and incisive investigations which lead to identification and analysis of the systemic structures which actually underlie and create the behaviour and response of the simulation. If the systemic structures which create the observed modes of behaviour are masked, deliberately or inadvertently, from the player the player will continue to flounder endlessly in a complex, dynamic maze.

To be effective the player must understand the business rules that dictate how the model behaves. For example, the business rules that dictate how a card game can be played are few in number and structure is relatively simple. The systemic structure of a card game is a product of the basic game business rules, the suits, and hierarchy of the cards in the deck. Before playing, the card player knows exactly what the business rules are and composition of the deck. If the composition of the deck were to be changed without the players being told, this would completely, and unexpectedly, change the game. We could only imagine how a game of poker could play out if there were nine aces, five kings and seven queens instead of the usual expected four, but nobody knew beforehand. Such a situation, as perplexing as it might appear to the player, is trivial by comparison to many risk-management situations.

Unexpected situations could be avoided by, before the game started, explicitly stating the game’s rules and ensuring that the pack of cards was free from tampering. In effect this is what we are attempting to do in risk management, though it would appear many managers are content to play games such as our corrupted game of poker. It should come as no surprise that in the life of the risk manager, many players believe they can sit at a control panel of our example flight simulator, monitoring the changing patterns displayed by the dials, gauges and lights, pulling the levers and flicking the switches as they feel necessary to make adjustments ranging from minor trim to major changes in direction.

The pilot of a Boeing 747 might be trained on a cockpit simulator, but only commences such training after studying aerodynamics and developing a detailed understanding of how an aircraft behaves under a wide range of dynamic conditions. A teenager might learn to fly the same 747 simulator drawing heavily on coordination skills and heuristics developed on a home game machine, or in an amusement arcade. There is risk that our juvenile ‘pilot’, having successfully flown the simulator for a period of time, might assume he has developed a sufficient understanding of dynamic behaviour. Without a real understanding of aerodynamics, it would only be a matter of time before the teenager, placed at the controls of a real 747, crashed. Considerations of risks, particularly the risk of not really understanding what produces dynamic behaviour must be included in our principles of method.