1
Week 2
Montibeller, G., & Franco, L.A. (2010). Multi-Criteria Decision Analysis for Strategic Decision
Making.
Introduction
A strategic decision has been defined as one that is important in terms of the action taken, the
resources committed, or the precedents set. Strategic decisions are infrequent decisions made by the
top leaders of an organization that critically affect organizational health and survival.
- The process of creating, evaluating and implementing strategic decisions is typically
characterized by the consideration of high levels of uncertainty, potential synergies between
different options, long term consequences, and the need of key stakeholders to engage in
significant psychological and social negotiation about the strategic decision under
consideration.
Strategic Decisions and Strategic Decision Making. The popular view of strategic decisions is that
they typically involve a high degree of uncertainty, high stakes, major resource implications, and
long-term consequences. This view is associated with the traditional conceptualization of strategic
decisions as the product of intentional attempts at rational choice, and context setters for subsequent
strategic action.
- It conceptualizes the strategic decision-making process in a way that is consistent with the
reality faced by practising managers.
- A rational and linear relationship between decisions and actions has been assumed. However,
organizational decisions are not always decisive, in the sense that they not always imply the
presence of ‘commitment to act’.
In our experience, managers act in accordance to the belief that strategic decisions must be intentional
acts and the result of a well-designed rational process.
Technical Complexity. From a decision analytical perspective, the two most troublesome challenges
in dealing with strategic decision are the inescapable presence of high levels of uncertainty and
decision complexity. There are different sources of uncertainty:
- Epistemic uncertainty refers to a lack of complete knowledge about an organization’s external
environment and its impact on the performances of potential strategies.
- Uncertainty associated with organizational values, which happens when there is doubt about
what strategic objectives, or policy values, should guide the decision or choice of action.
- The inter-relationship among choices.
Social Complexity. We posit that strategic decisions are socially produced and reproduced mental
frameworks through which managers make sense of their strategic concerns and so are able to act
upon them.
- The strategic decision making thus provides the cognitive structure within which strategic
change takes place in organizations.
- The reality of strategy and change is ‘socially constructed’ in the form of strategic decisions.
Strategic decision-making takes place through different communication channels, but also through
other communication channels, such as strategy workshops. Strategy workshops typically involve a
group of managers representing key organizational stakeholder groups that come together to impose a
structure to a decision problem which they perceive as ‘strategic’.
,2
- Participants bring to the workshop their individual mental frameworks of the issues
constituting the problem, how these interrelate and their perceived implication in relation to
the strategic choices open to them.
- Differences in interpretation of the issues are possible, which in turns creates cognitive
conflict.
- To resolve the conflict, participants engage in negotiation in order to produce a shared mental
framework of the strategic decision. However, the negotiation process can be hindered by
cognitive limitations (biases).
- Managers will compete to instil their own mental frameworks.
It has been argued that successful agreements regarding strategic decisions may depend on the
willingness of managers to engage in open dialogue if the choices they face are not to be dictated by
means other than an overt exercise of power.
Episodes of strategic decision and political internal negotiation, where issues of decision structuring,
group dynamics and power become critical in building up momentum for strategic action. Conducting
strategy workshops requires a shift in emphasis by decision analysts from the individual to the group.
MCDA for Strategic Decision Making: Modelling Content
Tackling Uncertainty with Future Scenarios. In traditional decision analysis, the standard way of
analysing decisions under uncertainty is to represent options and uncertainties as a decision tree and
then select the option with the highest expected value. If multiple criteria are considered in the
evaluation, usually a multi-attribute utility function is employed to aggregate the partial performances;
the option selected is the one with the maximum expected utility
There are three main assumptions in this type of analysis:
- The outcomes from a chance node should be mutually exclusive (i.e., only one of them will
happen) and collectively exhaustive (i.e., they cover all possible outcomes that may happen in
the future).
- It is possible to obtain, in a reliable way, accurate probabilities of outcomes.
- The use of the expected value rule as a way of selecting the best alternative.
Since 1980s, scenario planning has been suggested as an alternative way of considering uncertainty in
strategic decisions, instead of traditional forecasting. Once scenarios are developed and suitable
strategies are devised, a table can be built – which describes qualitatively the outcomes of each
strategy under each scenario.
- Scenario planning has been widely used in practice, and seems to be a tool which managers
are comfortable to work with.
What distinguishes scenarios from outcomes of events in a decision tree is:
- Considering an extreme scenario also leads managers to think about how robust the strategy
they are assessing is. This may support the creation of new, more robust, strategies and make
managers think about the future without merely extrapolating present trends.
- Scenarios are not, necessarily, mutually exclusive and exhaustive.
Considering Multiple Objectives.
The popularity and advantages of scenario planning, combined with the power of evaluation of
MCDA, provides a potent set of decision-support tools for strategic decisions.
,3
One important change that organisations may experience, when using MCDA for strategic decisions,
is the use of a value-focused framework [39] to guide the decision-making process. In this case,
strategies are seen as means to the achievement of the organisation’s strategic objectives. This may
help both in aligning the strategic vision of the organisation with its strategic objectives, and in better
scoping the strategic choices it is considering.
- A key aspect in supporting strategic decisions using value-focused thinking is the need to help
the definition and structuring of these strategic objectives.
Identifying Robust Options. The early focus of traditional decision analysis was in providing a
single solution, the one that maximises the expected value/utility. As we discussed before, this is a
feasible aim, as long as the conditions required hold. In conditions of deep epistemic uncertainty,
however, decision makers may be unable to define a set of exhaustive outcomes to events and/or
attach realistic probabilities to them. Furthermore, if one is using scenarios and MCDA to assess the
value that each of these strategies generates for the organisation under each scenario, it is not feasible
to calculate an expected overall value for each strategy as probabilities should not be attached to
scenarios. Instead of maximum expected utility, scenario planning proposers have stressed the need
for finding robust strategies, those that perform relatively well across the scenarios.
If the analyst is using MCDA with multiple scenarios, two aspects should be of concern:
- The robustness of performances of a strategy across scenarios (called inter-scenario
robustness).
- The spread of performances across scenarios (called inter-scenario risk).
In practice, we have found that the most helpful way for supporting decision-makers’ choices of
strategies has been a visual inspection of the performances and spreads, with a focus on inter-scenario
robustness and inter-scenario risk.
Designing Robust Options. Much of the focus of the MCDA literature has been on evaluating
options, given a predefined set of alternatives. While this is an important aspect of many decisions,
our experience is that most decisions, particularly at the strategic level, do not start with a well-
defined set of options.
In this regard, the decision analyst can help decision makers in:
- Identifying strategic options.
- Designing better ones.
Designing and Appraising Complex Strategic Options. At a strategic level, many times decision-
makers are faced with very complex strategic choices or policies that are composed by a large set of
sub-options. A challenge in this type of problem is the cognitive burden involved in appraising
holistically the performance of each policy and the time burden that may be required to evaluate a
large set of options.
There are some methods that deal with this situation:
- The analysis of interconnected decision areas (AIDA) technique, where the links between
several decision areas are represented, each one with several options, and whose compatibility
is explored in order to generate a list of possible option portfolios.
- To represent the problem as a portfolio problem, with decision areas which group options. An
MCDA model can then be built to assess the multiple benefits that each option generates, and
, 4
software can then calculate the best portfolio of options (i.e., the one with the highest
marginal benefit per unit of cost, given the organization’s budget).
Considering Long Term Consequences. Most of the MCDA applications reported in the literature
assess single-point out- comes, which try to represent the performance of an option if it were
implemented. Particularly in strategic decision-making, however, considering long-term consequences
is relevant and, many times, crucial. Ways of considering long-term consequences in these cases is
by:
- Applying time discounting, as in net present value (NPV) analysis.
- The use of several MCDA models, each one concerning a particular time frame.
MCDA for Strategic Decision Making: Facilitating Process
This section will focus on designing decision support processes to tackle the social aspects associated
with strategic decision making, and propose facilitated decision modelling as an effective means to
provide that support. The focus of support will be at the group rather than the individual level, which
is consistent with an increasing recognition of the importance of strategy workshops. Such strategy
workshops, by definition, involve working with groups of diverse composition which are likely to
include key organisational stakeholders.
Facilitated Decision Modelling. This describes a process by which formal decision models are
jointly developed with a strategy team, in real time, and with or without the assistance of computer
support. We consider a decision model as formal if it represents a strategic decision problem either in
terms of cause-and-effect relationships; or of relationships between decision choices and their
(deterministic or uncertain) consequences. A formal decision model is amenable to analysis and
manipulation, but not necessarily fully quantifiable.
A decision model produced in a facilitated manner is used by the strategy team members as a
‘transitional object’. It allows them to share their strategic concerns and increase their individual
understandings of the strategic issues, appreciate the potential impact of different strategic choices,
and negotiate strategic action that is politically feasible. When members of a strategy teams
participate in a facilitated modelling process, they engage in ‘conversations’ to exchange their
understandings and views about the strategic decision that is being analysed. Because interaction
between the participants in the decision modelling process, and of the participants with the decision
analyst, is needed to jointly build a decision model of the strategic situation, facilitated decision
modelling is also an interactive process.
Facilitated modelling is typically organized into group work stages, which roughly correspond to:
- Structuring the decision problem and agreeing a decision focus.
- Developing a model of organizational objectives.
- Creating, refining and evaluating options.
- Developing action plans.
The preceding discussion makes it clear that facilitated decision modelling is different from
standard decision analysis modelling. It requires a decision analyst able to support a group model-
building process that must be participatory, interactive, staged, non-linear, adaptable, and supported
by appropriate technology. At the same time, the facilitative decision modeller and his/her chosen
modelling approach must be responsive to the dynamics of group work and the particularities of the
situation at hand.
