A Decision-making Process Incorporating Cognitive Biases

 

This document is taken from Chapter 9 of ‘People Risk Management: A Practical Approach to Managing the Human Factors That Could Harm Your Business’ by Keith Blacker, Patrick McConnell published by Kogan Page© (April 28, 2015), hereafter referred to as [PRM}.

References are identified in the book’s bibliography.

 

9.1        From [PRM]Chapter 9 – People Risk Management

 

9.2        Improving Individual Decision-making

9.2.1        Human decision-making

Good decision-making is a skill.  But humans are not born with the full set of skills needed to make good decisions, other than the natural ‘flight or fight’ choices that have meant, for example, that our personal ancestors have chosen to flee more often than fight predators, which is why we happen to be  here today. Humans have developed over millennia the capability, which appears to be unique in the animal world, to make complex, considered decisions in our Systems 2 brain as described by Daniel Kahneman[1].   But, as described in [PRM]Chapter 3, we also have an impetuous Systems 1 brain that means that, when making decisions, people are a curious mix of the thoughtful and thoughtless, responsible and reckless, and open and secretive.

People did not evolve to make business, as opposed to day-to-day life, decisions and until recently such decision-making was the purview of only a select few in society[2].  Nor are we generally taught ‘decision-making’ in school. We tend to learn decision-making ‘on the job’ and thus we are subjected to biases already in the existing environment.  Outside of our jobs, we learn by games (called scenarios in business schools) or through case studies where decisions, usually ones that have gone wrong, are analysed.  Our natural optimism and confirmation biases will convince us that we would never have made the mistakes that the people in those case studies made.  Our alternative decisions will, of course, always remain perfect because, unlike real-world decisions, they will never get tested[3].  The thorny question of how business decision-makers should be educated has been raised by the respected business academic, Professor Henry Mintzberg[4] in his book, Managers not MBA.   In short, Mintzberg is strongly critical of traditional MBA programs as, to paraphrase, they teach neither business nor administration and reinforce arrogance rather than teach humility in the face of uncertain realities[5].

Our brains are not equipped to make perfect business decisions every time, so making that a goal for decision-making would be naive.  However, managers can take practical steps to improve their own and others’ decision-making and to detect bad decisions.  But why would we want to be better decision-makers ourselves and why would we try to persuade others to do likewise?  It is the premise of this book that bad decisions can cause sometimes disastrous losses.  [PRM]Chapter 4 gives only some of the many examples available of bad decision-making and other [PRM]Chapters give some reasons why some of these bad decisions were made.  We do not claim that, in the same situations, we personally would have made better decisions but do contend that if the decision-makers in those cases had been better at making decisions, in particular countering some of their own biases at least some of the losses and disasters would have been lessened, if not completely averted.

9.2.2        Individual versus Group decision-making

This [PRM]Chapter is about decision-making but it is focused on individuals making (important) decisions in a business context.  It is not about how people should make decisions in their private lives, which is a different set of discussions around ethics and personal freedoms. We do, however, believe that it is highly unlikely (but not impossible) that someone who is unethical in their private life will behave completely differently in business situations.  In their excellent book, Obstacles to Ethical Decision-Making[6], Patricia Werhane and her colleagues, tackle the thorny issue of personal ethics, which has engaged the attention of philosophers for centuries.    They conclude that although ethical decision-making is very hard (not least because of the human tendency towards self-deception[7]), there is nevertheless hope that decision-making in business can be viewed not as a contest between ‘ethics’ and ‘profits’, in other words, ‘what should be done’ as against ‘what can be done’.  They illustrate this false dichotomy with a case study of a subsidiary of Bayer, the large German chemical company, which, when faced with an entrenched mind-set of using child labour in India, rather than continue the practice, developed innovative solutions that created a situation where, through incentives and education, peasant farmers embraced the notion that sending their children to school would be beneficial for everyone in the longer term.  For Bayer, it was not a question or ‘either/or’ but instead ‘both’.   Werhane calls this “moral imagination” and argues that what is needed is:

A sound decision model, a strong dose of self-awareness, consciousness of our fallibilities, and a well-developed moral imagination

We agree with these arguments and suggestions and will address them below in the narrower context of business decision-making.  But we do not address the issue from a philosophical perspective that focuses on ‘what an individual should do’ but one that considers the more negative viewpoint of ‘how can we stop individuals making (particularly) bad decisions’. In other words, we are not making value judgments other than on the quality, or otherwise, of the decision-making process.

In business few decisions are made by a lone individual, although the myth persists of high-powered CEOs doing just that. In business, decisions, and certainly important decisions, are generally made collectively if only for the reason that the resources needed to implement any meaningful decision will come from multiple business units/divisions within a firm. In other words, there must be some sort of consensus about how a decision will impact various ‘stakeholders’, otherwise a decision may fail because of lack of commitment.  Before proceeding it should be noted that very many, probably the vast majority of, day-to-day decisions appear to be made by line managers and staff, working alone.  However, these decisions are generally pre-programmed and repetitive in that they are taken within a set of policies, procedures and controls that have been developed collectively beforehand.  In practice, a certain amount of licence to ‘change the program’ is afforded to managers and staff in order to handle situations where the program doesn’t work.  It is where such reactive changes grow and are adopted piecemeal that ‘rules’ become obsolete and dangers arise.

So what is the role of individual responsibility in collective decision-making?  It is all too easy to hide within, and behind, collective terms such as the team, group, division or firm.  In practice, it is very difficult to go against the general consensus of such collectives, because as Janis[8] points out, one of the defining characteristics of Groupthink is ‘direct pressure on dissenters’, ie groups freeze out nonconformists[9].   While such pressure may be seen as ‘bullying’ and in some cases, such as Enron and Washington Mutual described in [PRM]Chapter 4, it certainly was, the pressure is often subtle and invisible. In recent research into pressures on individuals to behave unethically, Veronica Bohns[10], and her colleagues at the University of Waterloo, found that people generally underestimate how much influence they have over others, even when encouraging decisions that are clearly unethical. The researchers found that even the mildest of pressure can influence people to do the wrong thing.  These findings mean that individuals can be influenced by subtle factors, such as the body language and tone of voice of someone they respect, as much as the arguments being advanced.   The desire to conform is part of the human make-up.  Many thinkers, including Janis and Werhane, suggest that the role of the ‘leader’ is to promote diversity of thinking, like a Philosopher King sitting above, and remote from, discussions.  While such objectivity is indeed desirable, even with the best will in the world such fine intentions can be undermined not only by the leader’s behaviour but also by the way in which questions and decisions are framed.

We argue that the role of the individual decision-maker in the real world of collective decision-making is to take each decision seriously and to think critically.  This means, as far as possible, ensuring that each decision is made: (a) following a sound process; (b) with an awareness of one’s own biases and fallibility; (c) with respect for others; and (d) with a commitment to raising issues that are producing personal concerns.  We recognize that in the real world this is much easier to say than do and that in order to promote critical thinking there is a need for disciplined decision-making processes, such as that described in [PRM}section 9.3.

It should be noted that this process (or the decision model as described by Patricia Werhane) is NOT a replacement for the existing decision-making processes used in a firm (as each firm will almost certainly have one or more processes for making decisions) but as an addition to these models.  Existing decision-making processes in firms deal reasonably well with visible factors, such as the financial inputs and outputs and the plans produced.  What they do not do so well is deal with the invisible, especially the non-financial personal and group biases.

9.3        Decision-making Process

Figure 9.1 shows three phases of a suggested decision-making model:

  • Pre-Decision: the activities involved in preparing for a decision;
  • Decision: the activities involved in making a decision; and
  • Post-Decision: the activities undertaken after a decision is made. Note this is not considering the plan arising from the decision but a post-hoc analysis of the quality of the decision that was made.

Linked 1 Figure 9-1

Figure 9.1 – Three Phases of Decision-making

The three phases shown in the diagram reinforce one another in that information considered, and importantly documented, in the first phase (Pre-Decision) is used in the second (Decision) to drive the analysis of the decision and in the third (Post-Decision) to review the actual decision-making process.  In particular, the Post-Decision phase is best seen as a learning opportunity rather than, for example, a formal audit of the decision.   As shown in the diagram, however, there will be some decisions that are found to be unsatisfactory (for example, incomplete process or unacceptable pressure on participants) and should be returned for review by the original decision-maker(s) before they can be actioned[11].  If some measure of ‘quality in decision-making’ is also used in performance reviews, then there will be positive/negative feedback for participants in the decision-making process.

In addition to the decision-maker (shown singly in the diagram but in most cases of important decisions usually a part of a group of ‘decision participants’) two other key roles are shown:

  • Decision Support: analysts trained in providing technical support and advice for making decisions.  For example, in addition to providing decision tools, such as ‘decision trees’, these analysts would also collate checklists and maintain an ‘issues/concerns register’ and a ‘risk register’ during the decision-making process;
  • Decision Process Reviewer: an analyst trained in analysing and reviewing decisions.  Such an individual could be a senior Decision Support analyst and/or a senior individual in the firm who is trained in decision analysis.  A reviewer may also, when required, call upon other expertise, such as Project Management experts to evaluate not the ‘plan’ but rather whether the ‘plan’ was constructed within a considered and comprehensive process.

Obviously the relative time and effort allocated to each of these three phases, will depend on: (a) the importance of the decision; (b) the ‘routine-ness’ of the decision; and (c) the breadth of the impact, in that the greater the number of ‘stakeholders’ impacted the more time and effort is required to make a decision.  Not all decisions will require a full-blown process. Some decisions, such as for example hiring temporary staff, which are routine, relatively unimportant and with minimal impact, will not require a complete process (although such decisions should be documented thoroughly for later review if required).  However, some decisions, such as major projects or new corporate strategies, would, and should, require a detailed and comprehensive process.  The level of detail for any particular decision should depend on the pre-determined ‘Risk Appetite’ of the Board and management as described in [PRM]Chapter 5.

Some decisions are routine but important, such as, for example, the take-off of a jumbo jet. Pilots, flight engineers and air-traffic controllers spend much more time in pre-flight checks (Pre-Decision), than they do in taking off (Decision) because by the time pilots throttle the engines to accelerate down the runway they need to be pretty sure that everything has been checked out.  And when they are in the air, climbing to their flying altitude, there is a further set of post take-off checks (Post-Decision) to ensure that the take-off has been successful. If the take-off has experienced a problem, such as an undercarriage failing to retract, the plane will typically circle while a diagnosis takes place, often requiring that pre-flight checks be re-run or re-evaluated, before determining the next action.  Much of the checking for a jumbo jet take-off uses pre-prepared checklists which are run through by the pilots/flight engineers with each step being checked by at least two individuals. These, of course, are the visible factors being checked.

Less obvious are the factors that are invisible, such as the mental state of the pilots and engineers.  Before entering the cockpit, pilots are required to assess their own health for each flight using a suggested checklist for invisible factors with the mnemonic IMSAFE[12]: Illness; Medication; Stress; Alcohol; Fatigue; and Eating.  Of course, as humans, pilots may deceive themselves into thinking that they are ‘fit to fly’ when they are not, but such a checklist follows the principles of counteracting behavioural biases by reminding/nudging pilots of their individual responsibilities, before they fly. What if decision-makers were to give a moment’s consideration as to whether they are mentally and physically equipped to make a ‘good’ decision before they start?  For example, have they read the material fully in preparation for a decision? While there may a loss of face, delaying a decision until people are ‘fit to decide’ may even avert a disaster.

Some decisions, such as strategy and product development are non-routine, important and will likely have a serious impact on a range of stakeholders. In such decisions, the importance of invisible factors is amplified, as is the importance of the Pre-Decision process.

9.3.1        Pre-Decision Phase

A business decision must have a purpose for being taken, such as a problem that has to be solved or a transaction that has to be completed. Otherwise, the decision is merely capricious.   But surprisingly the real purpose for a particular decision is often not stated clearly, being assumed in the decision that is actually taken, giving rise to confirmation bias.  The case of the take-over of ABN-AMRO by RBS, described in [PRM]Chapter 4, is one where the purpose was not only unclear at the outset but also changed as the decision was being taken[13].  At the beginning there was no problem to be solved, as the Board had agreed to go forward with an organic rather than an acquisition strategy.  However, an opportunity emerged to acquire the Dutch bank.  During the prolonged take-over battle, however, the ‘purpose’ changed to one where ‘Beating Barclays’ became the main driver of the Board’s thinking.   The lack of clarity of its purpose and its detailed objectives was one of the reasons why the take-over was a financial disaster, with RBS paying far too much for the bank and consequently having to be acquired itself by the UK taxpayer.

It may sound obvious, but the first step in any Pre-Decision checklist must be to ask the question ‘What is the purpose for this decision?’ and then to gain general agreement that the answer arrived at makes sense for the firm.  At this point, when personal investment in any particular answer is not yet strong, constructive criticism may not only head off a potential disaster but also help to hone the rationale for taking a particular decision, making it more likely to be successful.  Table 9.1 shows an example of a Pre-Decision Checklist for Invisible Factors, such as Possible Biases.  Documenting the answers for such a checklist will, in fact, create a semi-formal ‘terms of reference’, against which the final decision can be checked.  It should also be noted that the debates and answers to the challenges raised in the checklist would be documented in a fashion that would support a Post-Decision Review, so that those involved will be aware that the quality of their decision-making will be under scrutiny.

Questions Possible Biases Possible Challenges
What is the Purpose of this decision?

Has the Purpose been ‘framed’ properly?

Who will sign off on the Purpose?

Overconfidence, Groupthink, Framing,

Anchoring,

Loss Aversion, Confirmation bias, Sunk Cost Fallacy,

Status Quo bias,

Planning Fallacy

Why is this purpose important?

What if the decision was not made?

What if the purpose was re-stated in a different way?

What if we did the opposite?

What if we did nothing?

Are we doing this because others are doing it?

Why do we believe that we will do better than others in achieving this purpose?

Why is this decision being taken at this time? Availability, Groupthink,

Action bias

Why the hurry?

Why should we be ‘first movers’ in this decision?

Are ‘deadlines’ real or manufactured?

Who is proposing the decision?

Who is taking the decision?

Who should be taking the decision?

Who will be impacted by the decision?

Who will sign-off on any decision?

Who will review the decision process for quality?

Who has been assigned the role of ‘devil’s advocate’?

Groupthink, Conflicts of Interest,

‘Halo’ effect,

Attentional, Availability

Are the owners of the decision clearly identified?

Have all ‘stakeholders’ been identified?
Have any stakeholders been excluded, deliberately or otherwise?

Who is likely to support/ oppose the decision?

Who are the ’squeaky wheels’?

 

How much time, effort and resources have been allocated to the Decision? Action Bias,

Availability,

Planning Fallacy

Is there a project plan for the decision process?

Is the time, effort and resources adequate for making the decision?

What are the Conflicts of Interest? Self-serving Bias,

Blind Spots

Who benefits if the purpose is achieved?

Who benefits if the purpose is NOT achieved?

What are the Assumptions being made? Anchoring,

Availability,

Confirmation

Have all assumptions been recognized?

Have hidden assumptions been articulated?

What is the Preferred decision? Confirmation bias,

Status Quo,

Blind Spot,

Availability,

Sunk Costs

Has a decision already been made?

Is an alternative ‘too hard work’?

Are there good reasons why the decision has not already been taken?

What is specifically Included in the decision?

What is specifically Excluded from the decision?

Availability,

Attentional,

Groupthink

Why has something been included/excluded?

If something was included/ excluded, would the purpose change?

What information is essential to making this decision?

What information is available to make this decision?

What information is not available to make this decision?

Availability, Attentional,

Confirmation bias

What do we need to know to make this decision?

How much will it cost to acquire the necessary information?

What are the risks in not acquiring the information?

What Internal Expertise is needed to make this decision?

What External Expertise is needed to make this decision?

Groupthink,

Availability,

Halo Effect

Where is the best expertise for analysing this decision?

What are the reasons (eg costs/ opposition) for not involving the necessary expertise?

What Constraints (eg timing, costs) are there in taking the decision? Availability,

Groupthink,

Blind Spots,

 

Are constraints real or manufactured?

What if we had unlimited resources, eg time and money?

What are the potential impacts of any decision?

What are the acceptable impacts of any decision?

What are the unacceptable impacts of any decision?

Confirmation bias,

Sunk Cost Fallacy,

Planning Fallacy

What is ‘worst case’ scenario?

Have we asked stakeholders how they would be impacted, or made assumptions instead?

What if we ended up on the front page of the newspapers?

What Risks are there in the decision?

What is the ‘Appetite’ for Risk?

Availability,

Groupthink,

Blind Spots,

Planning Fallacy,

Illusion of Control

Have we considered possible risks?

Have we clearly articulated a ‘risk appetite’ for this decision?

Who analysed the risks, were they independent?

Have we looked outside for similar ventures?

What are the criteria for a ‘good’ decision in this case How have the criteria been established/prioritised? Availability,

Blind Spots,

Planning Fallacy

What would trigger a re-think?

What are the desired outputs?

What is the range of desired results?

What if the decision was ‘bad’, see Pre-Mortem?

     
Decision Ranking (1-9)

Routineness (1-9)

Importance (1-9)

Impact (1-9)

Each individual would assess the decision checklist according to pre-defined criteria Average and range of responses

Plus anonymous reasons

 

Table 9.1 – Pre-Decision Checklist –Possible Biases and Challenges

At first glance it may seem that there are a lot of questions to be answered before a decision can even start to be considered, but there is also a lot of very important questions asked in a pilot’s take-off checklist.  In fact, all of these questions are, or should be, addressed in any reasonably structured decision process and are usually answered either explicitly (visible) or implicitly (invisible).  The danger of not addressing key questions up-front is that they will be addressed further into the decision process and possibly incorrect assumptions will be made, resulting in a bad decision.

An example of this is the case of the JPMorgan Whale[14], discussed briefly in [PRM]Chapter 4.  In that case, the purpose of a decision to reduce overall risk was very clearly articulated by the Board and senior management.  However, what was not articulated were the constraints, such as acceptable costs, nor were possible risks addressed.  This resulted in decision-makers making assumptions about costs and risks that were at odds with the original purpose of the Board and in doing so they made a disastrous decision.

In developing a Pre-Decision Checklist for a particular decision, the importance of the ‘Possible Challenges’, in Table 9.1, should not be underestimated as these will prompt or ‘nudge’ individuals to consider questions that they would never have thought of and to raise any concerns that they may have.  It is in the interest of everyone, except potentially those who have Conflicts of Interest, to encourage openness at this stage as the costs and risks of dissent later in the process will be lessened. Do it Once, and Do it Right! And, of course, the situation may arise that, after due consideration of the Pre-Decision checklists, it may be decided not to proceed because, for example, the Purpose is not well defined.

Note that one of the questions in this checklist is – who has been assigned the role of devil’s advocate?  In order to minimize Groupthink, Janis recommends[15] that for every major decision someone is designated to ask ‘hard questions’ and to make a nuisance of themselves.  In considering the obstacles to making ethical decisions, Patricia Werhane and her colleagues describe the situation where Colin Marshal, a former CEO of  British Airways, created the official role of ‘Corporate Fool[16]’ and put a senior manager into the role.  The purpose of this self-perceived ‘court jester’ was not only to “draw attention to things that are going wrong” but to become a means, or lightning rod, to encourage critical thinking and dissent by others

When things go wrong employees usually have a good idea of how to fix them. You need to create a state in which they’ve got the courage to do something. You want to build organizations where everyone sees provocation as one of their essential roles

In management mythology, Chairs of Boards and CEOs are supposed to somehow sit as  objective ‘judges’ analysing competing views on a particular decision/ strategy before leading ‘their people’ to a common decision.   This, of course, ignores human nature as Chairpersons and CEOs are as prone to behavioural biases as the rest of us and often have an interest in a particular course of action, if only because it vindicates prior decisions, ie there is a confirmation bias. There is nothing wrong with a leader having a particular viewpoint, provided that he/she is prepared to let other competing perspectives be heard.   Ideally, there should be no need for a devil’s advocate in a decision process, because, in theory, everyone would have the right, and responsibility, to speak-up.  Human nature being what it is, however, individuals will be reluctant to do so, and, as in flight cockpits, well-designed checklists provide a platform for junior staff to raise concerns and ask uncomfortable questions.  Likewise, checklists provide a mechanism for proposers of a particular decision to pre-empt concerns by addressing possible challenges ahead of time, rather than being in the invidious position of having to ‘defend the indefensible’ later on.  The advantages of using checklists as decision-making tools are explained later after the Decision Phase itself is described.

9.3.2        Decision Phase

The classic process for making a decision, developed by Herbert Simon and others in the1960s, is a sequence of:

  • Defining a ‘problem’; note this is already covered by the purpose described in the Pre-Decision checklist;
  • Developing the options (possible solutions) for achieving the purpose of the decision (ie solving the problem);
  • Collecting the information required to analyse the options;
  • Analysing each option by comparing it against the predefined purpose;
  • Selecting between options, ie taking a decision, which may in fact involve a combination of several options. At this stage an outline Plan will usually be developed.

In real life, however, the decision-making process is rarely that simple and, as shown by Henry Mintzberg[17], most decision-making in business is messy in that the process of analysing options may turn up new options that have to be analysed or may even rule out all options necessitating a rethink of the original purpose.  It is this iterative nature of real-life decision-making that creates risk, because individuals, however diligent, take short-cuts and make unwarranted assumptions in their haste to come to a decision and to be seen to be decisive (ie Action bias). Table 9.2 shows an example of a Decision Checklist, using Simon’s decision phases, which describes Possible Challenges that may be raised during the decision-making process to counteract possible biases. Again it should be noted that the debates and answers to the questions raised in this checklist would be considered and scored by decision participants and documented in a fashion that would support Post-Decision Review.

Decision Phase Possible Challenges
DEVELOPING OPTIONS  
What options are already on the table? Why have some options been selected?

Who benefits from the options already selected?

Why have these pre-selected options not been accepted in the past?

Who is proposing each option? Who benefits from a particular option?
What options have already been excluded? Why have some options already been rejected?

Who benefits from the options already rejected?

What is the process of developing options? Who will be involved in developing options?

 

What are the criteria for selecting an option for analysis? Are criteria clearly identified?

Who will sign-off on the options to be analysed?

DOCUMENTING OPTIONS  
What is the Purpose of each option? Has the purpose of each option been clearly identified and agreed?

Have the options been framed properly?

COLLECTING INFORMATION  
For each option, what information is necessary to perform analysis for this decision? Have information sources been identified?

Has the information been acquired or must it be purchase/created?

 

What information will be used? Has each source of information been verified for relevance, completeness and accuracy?

Have data quality checks been carried out?

Have data capture processes been verified?

If existing information sources are used, has their data quality been checked?

Has the information used been extracted and stored securely for later analysis and review?

Have standard statistics been produced for sense-checking?

ANALYSING OPTIONS  
For each option, what is the plan for analysis? Does a plan for analysis of each option exist?

Are appropriate tools being used to address the agreed criteria for a good decision

For each option, what assumptions and constraints will be used Has each assumption been articulated, justified and documented?

Are the constraints documented and credible?

How will each option be analysed? What are the results of each option analysis?

Has sensitivity analysis been performed?

For each option, how will risks be identified and analysed? Has independent risk analysis been performed?

Is there a risk register?

How will unknowns and uncertainties be identified and analysed? Are the uncertainties recorded, eg in a risk register.

Has a sensitivity analysis of uncertainties been undertaken?

SELECTING OPTIONS  
What is the process of selecting between options that have been analysed? Is there a description of the selection process and has it been followed?

Has each option been analysed to the same standard?

Have the same assumptions been used in analysing each option?

Have the same uncertainties been analysed in all options?

What is the preferred option(s)? Have options been compared objectively and to the same criteria?

Have challenges (for and against) been given equal weight?

What is the Decision? Does the Decision follow the logic of the selection?

Is there a formal Decision with conclusions clearly specified?

Have the Impacts been clearly identified?

What is the agreed Plan? Has a comprehensive Plan been developed, reviewed and internally agreed?

Has a realistic budget been identified, reviewed and internally agreed?

Have the risks of implementation been identified, documented and addressed?

What is the Risk Mitigation Plan? Has the mitigation for each important risk been identified, agreed and action planned, including integration into budget?

Have Key Risk Indicators (KRIs) been identified?

What are the unknowns and uncertainties? Have the unknowns and uncertainties been documented and monitoring actions agreed?

Has the need for further information been identified?

What is the sign-off process? Has the Decision process been documented fully?

Who will sign-off the Decision?

Who will review the Decision?

   
Assessment of Decision Quality (1-9) Average and range of assessments. Plus anonymous comments

Table 9.2 – Decision Checklist –Questions and Possible Challenges

9.3.3        Decision Phase – Tools and Techniques

There are many books on how decisions should be made (known as ‘normative decision-making’) with descriptions of many tools and techniques for considering decisions.  In fact, there is an emerging international standard (ISO 31010[18]) which describes such tools, including (as a sample of many):

  • Decision Tree Analysis: which is a mechanism for ordering and ranking options (as a tree) and assigning weights to each branch to arrive at the ‘best option’;
  • Multi-Criteria Decision Analysis (MCDA): a more sophisticated mechanism for ordering and ranking options against multiple rather than a single criteria;
  • Brainstorming: a mechanism for encouraging a group of individuals to ‘envisage’ possible outcomes and risks;
  • Discounted Cash Flow (DCF): a mechanism for comparing the financial impacts of various options by using standardized assumptions about prevailing interest rates;
  • HAZOP (Hazard and Operability study): a structured technique for identifying risks; and so on.

Such tools are invaluable in trying to come to an objective decision because their purpose is to provide a template into which information can be loaded and (at least semi-) objective results can be determined.  But in general such tools deal with Visible factors, such as projected cash flows and losses.  It is not the purpose of this book to suggest which tools should be used for analysis of visible factors, only that there should be a conscious choice of tools for a particular decision within a particular industry.  We are interested here in tools that will bring to the fore Invisible factors, which can then be used in conjunction with more conventional decision tools.

One such tool is the Pre-Mortem.  As noted throughout this book, people tend to be optimistic, preferring to see the upside of any decision rather than its risks.  The concept of a pre-mortem has been developed to help counteract over-optimistic assessments[19].  A pre-mortem is a post-mortem that takes place before death rather than afterwards, but does not answer the question ‘what did cause death’, rather ‘what could have caused death’?  In other words, it is a risk analysis tool.  A pre-mortem is usually conducted as a brainstorming exercise after a decision has been made in which a facilitator develops a scenario where the decision being considered has not just gone wrong but has gone horribly wrong.  Each individual in the brainstorming group is then given time to come up with a long list of reasons for why the decision could have gone so wrong. These reasons are then revealed, discussed, consolidated and then analysed. While noting that a pre-mortem is not a panacea, Daniel Kahneman considers its main virtue is that, like a devil’s advocate, it legitimizes doubts and “encourages even supporters of the decision to search for possible threats that they had not considered earlier”. As a pre-mortem is specifically designed to ‘think the unthinkable’ it is more likely to make unpalatable risks more visible.

Daniel Kahneman and his late colleague, Amos Tversky, identified a type of “pervasive optimistic bias”, which they termed the Planning Fallacy[20], in which decision-makers develop forecasts and plans that are close to a ‘best case scenario’ but are at odds with similar situations elsewhere.  This is a form of arrogance that results from a conceit that the decision-makers know better than their counterparts in other firms. Bent Flyvbjerg[21], a Danish transportation planning expert, argues that the major source of error in forecasting is the “prevalent tendency to underweight or ignore distributional information” and his solution, called “reference class forecasting” is to develop databases of the outcomes of similar projects.  By looking at the ‘distribution’ of the outcomes of similar decisions, the optimism of decision-makers will be somewhat tempered.    In writing on ‘strategic risks’, Adrian Slywotzky[22] makes the same point by providing statistics on failure rates on ‘typical’ business projects, such as the well-known 70% failure rate for large IT projects. Slywotzky points out that “when you overestimate the odds [of success], you underestimate the investment needed to win”.  In other words, you almost guarantee failure.  Even when such databases are not available, the basic question that individual decision-makers should ask is “why do we think that we are better than [others] who have failed at the same thing”?  The objective is to force decision-makers to search for disconfirming information at the outset. The answers, like a pre-mortem, will help to bring invisible overconfidence to the fore.

Simulation is a well-developed technique for unveiling the range of outcomes of a particular decision option being analysed[23].  The technique uses the power of computers to run millions of possible scenarios varying the underlying assumptions in each scenario. For example, in considering the introduction of a new product the range of profits from a varying mix of assumptions about costs, pricing and market share can be determined.  Simulation is already widely used in industries such as energy exploration as part of their toolkit for analysing visible factors especially where there is a great deal of uncertainty involved in estimating various factors. It can also be used for unearthing invisible factors using so-called ‘sensitivity analysis’.  As part of their standard set of outputs, simulation software packages not only provide the average and range of results but also importantly the sensitivity of the results to underlying assumptions.  For example, the profitability of a new product will be highly sensitive to factors such as market share and price.  By discussing how assumptions about such factors have been arrived at, then biases such as overconfidence and confirmation, can be unearthed.

What is the role and responsibilities of the individual in such analysis?  The checklists, tools and techniques described here are designed to bring to the fore the biases and (sometimes unwarranted) assumptions that everyone brings with them to considering every problem.  The responsibility of the individual in making decisions then is not to be defensive but to admit that they may have biases and to be open to, and to look for, opportunities to question one’s own understanding of a particular situation.  It’s not wrong to be wrong, but it is wrong to believe that you can never be wrong.

9.3.4        Post Decision Phase

The purpose of the Post Decision Phase is NOT to second guess or re-make a decision but to assess a particular decision for quality.  In particular, did the decision-makers follow a good process and as far as possible identify and try to counteract any individual and group biases?   Nor is such a review considering the outcomes of a decision but ideally would be conducted immediately after a decision was made and before any plans are signed-off for implementation.  The review can be seen as a last chance to catch any problems with the decision before resources are expended and to identify any unresolved concerns.

Table 9.3 shows an example of a Post-Decision checklist that describes ‘Possible Tests’ of good decision-making with Assessments and an overall assessment as to whether a formal Decision Review is required. Note the assessments here are examples from a range of 1 to 9, but other measures of decision-quality could be used.

Questions Possible Tests (Examples) Assessment
Did the decision-makers follow a good decision-making process? Was the documentation adequate?

Was sufficient time given to raise and address concerns?

Was there an Action bias?

1= Minimal process

…..

9= Comprehensive well-documented process

Has the Purpose of the decision been identified properly?

 

Did the Purpose change during the decision process? 1= Purpose unclear

…..

9= Clear well-documented Purpose

Was participation in decision-making adequate? Was there evidence of bullying or groupthink?
Was there evidence of ‘passivity’?

Was there evidence of ‘denial’ among participants?

1= Lack of challenge

…..

9= Decision challenged and concerns resolved satisfactorily

Were decision-makers and stakeholders involved? Was there sufficient challenge or was the decision a ‘fait accompli’?

Were there any attempts to close down discussion prematurely?

1= Lack of involvement

…..

9= All decision-makers and stakeholders involved

Were Conflicts of Interest Identified and addressed? Did individual participants identify their own and others’ Conflicts of Interest? 1= No Conflicts identified

…..

9= Conflicts clearly identified and addressed

Were Assumptions, and Constraints identified and addressed? Were Assumptions, and Constraints fully documented?

 

1= No Assumptions, and Constraints identified

…..

9= Assumptions and Constraints clearly identified and addressed

Was Information collected and analysed properly as to relevance, completeness and accuracy? Were sufficient reasons given for each source of information used?

Was each source of information analysed for relevance, completeness and accuracy?

Did the decision-makers search for disconfirming information?

1= No rigorous analysis of Information quality

…..

9= All Information collected and analysed

Was Internal and External Expertise sought and listened to? Was the rationale documented for seeking the expertise used?

Did the decision-makers incorporate the advice of external experts?

1= No Expertise sought

…..

9= Relevant expertise sought and opinions used

Were the interests of non-participant stakeholders considered? Did the decision-makers seek to engage non-participants?

Were the concerns of non-participants taken into account?

1= Non participants ignored

…..

9= Concerns of non-participants sought and considered

Were the full range of Options considered in the decision? Was there comprehensive documentation on each option?

Were the arguments for discarding certain options credible?

Was there evidence of ‘Confirmation bias’?

1= Options not considered

…..

9= All Options analysed methodically and objectively

Were the full range of Risks considered in the decision? Was there a formal Risk Management process and did it complete properly?

Were Risks assessed properly?

Have risks been reviewed by Risk Management professionals?

1= Risks not considered

…..

9= Risks identified and assessed methodically

Were unknowns and uncertainties adequately documented? Have uncertainties and unknowns been identified?

Has the rationale for not resolving uncertainties been given and is it credible?

Have processes for monitoring uncertainties been proposed?

1= No Analysis of Uncertainties

…..

9= All Uncertainties identified and monitoring initiated

Was there a formal Decision? Was a formal decision agreed with outstanding issues recorded? 1= No formal Decision

…..

9= Formal Decision documented and agreed with outstanding issues highlighted

Have plans been developed and agreed? Have the plans been documented in sufficient detail?

Have the plans been reviewed by external experts?

1= No agreed Plans

…..

9= Detailed plans produced and reviewed

Have concerns been addressed? Have documented concerns been addressed to the satisfaction of those who raised them? 1= Concerns ignored or overridden

…..

9= Concerns addressed

     
Assessment of Decision Quality (1-9) Decision Reviewer (s) and participants would review decision. Average and range of assessments

Plus anonymous assessments

Review Required? Yes/No Reasons for decision-maker review

Table 9.3 – Post-Decision Checklist –Decision Quality

In this model of decision-making, Post-Decision analysis would be conducted by an independent, skilled Decision Process Reviewer, whose role would be to consider the documentation provided by the decision-makers to support the decision.  The documentation provided to the reviewer would not only include formal reports and analyses but also meeting notes (including in the case of Board decisions, official Board minutes) and detailed reports/artefacts used to support analysis (such as spreadsheets).  Individuals involved in the decision-making process would also be asked to complete the Post Decision checklist(s), anonymously if preferred, with annotation as to their reasons for a particular viewpoint.  The Decision Process Reviewer would take the reasons proffered by decision-makers and, in particular dissenters, into account when assessing the quality of a decision that was made and requesting a Decision Review by decision-makers, if required.

It is likely that a Post-Decision Review would not be a one-off exercise but somewhat iterative as it is possible, even probable, that further documentation would be requested to back up parts of the decision process that were not adequately documented. In such cases, the overall assessment of the decision ‘quality’ would reflect the lack of comprehensive documentation. Note here we do not discuss the issue of measuring ‘decision quality’ as we believe this should be determined by management as part of setting up a comprehensive decision process.   As noted elsewhere, the action of developing such assessment (and subsequent reward) criteria would provide a valuable learning opportunity for management and HR experts to tailor the concept to the unique situation of their firms.  The danger, as with any assessment scheme of course, is that it may be ‘gamed’ by individuals to their benefit.

9.3.5        Decision Process – Output

The output from any decision-making process would normally be a formal Decision that is ready for sign-off/approval by the Board and/or management.  Like the ‘lines of defence’ described in [PRM]Chapter 8, the approvers would act as yet another check on the decision made and the quality of the process used to make that decision.  But by considering carefully the invisible pressures at each stage of the decision process and by documenting them, ie making the decision visible, approvers can be assured that a sound decision process had been followed.

It is highly probable that when significant decisions are made, that some information will be unknown, ie there will be some degree of uncertainty, or risk.  Provided that the unknowns are documented, management can put in place the monitoring and information-gathering processes to make the unknown better understood, ie to reduce risk.

To some, following a formal decision-making process, such as that described here, will seem overkill, and to others it will appear to be merely a ‘tick in the box’ exercise.  In some situations this may indeed be the case, but a formal process with challenges, pauses, reflection and cross-checks is more likely to identify concerns that may lead to serious problems than an over-confident decision-maker relying on their instinct. Just Do It is an appealing advertising slogan but the athletes that front such advertisements have not arrived at the pinnacle of their chosen fields by ‘just doing it’ but have put in years of disciplined hard work beforehand.  The athletes only make it look easy because of the effort and thought that they have already put in.  Decision-making too needs hard work to excel in – ‘just doing it’ is a recipe for disaster.

[1] See Kahneman (2011)

[2] This does not diminish the day-today decisions made by ordinary people, such as when to plant a particular crop, but these were often pre-programmed using tools such as a ‘farmer’s almanac’.

[3] This is also known as ‘Monday Morning Quarter Backing’

[4] See Mintzberg (1978)

[5] A discussion of such heresy is beyond this book other than to note that as holders of advanced Business Administration degrees and significant managerial experience, the authors have a lot of sympathy for Professor Mintzberg’s point of view.

[6] See Weharne et al (2103)

[7] See Ariely 2013) and Heffernan (2011) for examples of self-deception and ‘wilful blindness’

[8] See Janis (1971)

[9] See Heffernan (2011)

[10] See Bohns, V, Roghanizad, M and Xu, A (2013) Underestimating Our Influence Over Others’ Unethical Behavior and Decisions, Personality and Social Psychology Bulletin XX(X) 1–15 published online

[11] Care would have to be taken that this process does not result in ‘paralysis by analysis’.

[12] See for example Flight Fitness (undated) The “I’m Safe” Checklist http://www.leftseat.com/imsafe.htm which provides references to the FAA (Federal Aviation Authority) regulations in this area

[13] See Martin (2013) and Fraser (2014) for a description of the chaotic decisions made by the RBS Board and management during the ABN-AMRO takeover.

[14] See McConnell, P (2014) Dissecting the JPMorgan whale: a post-mortem Journal of Operational Risk 9(2), 1–42

[15] See Janis (1971)

[16] In Economics, the more acceptable term, Contrarian Thinker, is often used for the same role.

[17] See Mintzberg (1978)

[18] See ISO 31010 Risk management — Risk assessment techniques is a standard developed ISO as a part of the ISO 31000 set of standards.  In mid-2014 it is currently under review.

[19] See Kahneman (2011)

[20] For a description of the Planning Fallacy, see Kahneman (2011)

[21] See Flyvbjerg et al (2003)

[22] See Slywotsky (2007)

[23] See Vose (2008)

 

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s