1
Introduction

Welcome to Smart Decisions where you will learn about many ways to model and solve complex decision problems using Multi-Criteria Decision Analysis (MCDA). Each chapter will take you through a simple and common decision problem that most people face in their lifetimes, showing you how to solve it with a structured decision-making methodology in both Microsoft Excel and R. Microsoft Excel is the primary tool used in business and education for making spreadsheets and interactive formulaic templates while R is a popular tool used by specialists in analytics, data science and statistics. If you are not familiar with either Excel or R, don't worry! Chapter 2 is dedicated to telling you all about these software packages and how to use them. Trust me you will be an Excel and R whizz-kid in no time. That being said, this is not the primary aim of this book. The core of this book is to teach you about all of the wonderful and powerful decision-making methodologies out there which will help you justify and make better decisions at work and at home.

Let's start by discussing a hypothetical decision problem. Imagine you are a student who has just graduated and after years of partying and hard work you now have to decide what to do with your life. You are considering three distinct options; (1) find a job, (2) do further studies, or (3) start your own business. Your parents tell you to find a job related to your degree but your university professor suggests that you should do further studies before starting a professional career. To make matters more difficult, your friend, a recently established entrepreneur, has a great business idea and suggests you become a partner in their business. How do you decide what to do?

There are lots of interesting behavioural aspects to consider here, such as do you unconsciously favour the first option you discussed with your friend (this is referred to as anchoring) or do you prefer to stay in education as you haven't worked or started a business before (referred to as familiarity bias)? Although it is very important to acknowledge and understand the behavioural aspects of human decision-making, this will be rarely discussed in this book as the focus here is on the process of decision-making and not on the psychology of decision-making. That being said, it is important to understand both areas of decision-making and therefore if you want to read more about behavioural decision-making we recommend reading Thinking Fast and Slow by Daniel Kahneman, Psychology of Judgment and Decision Making by Scott Plous and Preference, Belief, and Similarity by Amos Tversky.

Most people who study decision-making now agree that there are two systems of decision-making, system 1 which is based on intuition or gut, and system 2 which is controlled, conscious and requires considerable effort (and time). We will be focusing on the methods and techniques that can be used for important and complex decisions that fit within system 2, or as we like to call it methods for structured decision-making.

Going back to the example we discussed about the recent graduate choosing what to do with their life. This is a particular problem where alternative options are already known to the decision-maker and the decision problem is to evaluate each of these options in order to select one of these alternatives based on several different criteria. This is referred to as an evaluation problem where the ultimate goal is to obtain a ranking or rating for each alternative. Other examples of such problems would be a regulatory authority wishing to publish the ranking of all universities in a country or a food agency seeking to rate all food shops from 1 (inadequate) to 5 (outstanding).

Another category of decision problems would be those where we search for a feasible solution which is not explicitly known to us. This category of problems can be labelled as design problems, for example, product pricing or time-tabling problems. A product can be priced with any value on a continuum but the decision-maker is interested in finding the most feasible price on this continuum. Similarly, one may find it difficult to manually schedule a timetable with no conflicts whatsoever, and therefore seek to find a feasible solution. In these two examples, the former one involves selection of a single value while the latter can be visualised as the selection of multiple values. The problems involving multiple values (sometimes called multivariate) can also be referred to as allocation problems. Table 1.1 summarises these categories of decision-making problems along with an example for each possible type of problems.

Table 1.1 Examples of evaluation and design decision problems.

Another important categorisation is to consider the number of decision-makers. In the case of two or more decision-makers, conflicts may occur, usually due to different preferences, interests, or each person's level of knowledge. By contrast, those decisions involving a single decision-maker may not face these kinds of conflicts, although conflicts among different objectives/criteria may still occur, for example, searching to buy the cheapest house but wanting to live in a nice area.

1.1 How to Structure Your Decisions?

Whilst there can be many ways to structure a decision problem, Hammond J.S., Keeney R.L., and Raiffa H. (2002) proposed a very nice and simple framework known as PrOACT, which is an acronym for Problem, Objectives, Alternatives, Consequences, and Trade-offs.

Problem: The first element in this structure is the "Problem" itself. A good decision maker always questions the definition of the problem before identifying the ways to solve it. It is useless to find or suggest a wonderful solution to the wrong problem; therefore, it is important to identify the real problem before any further action.

Objective: After clearly identifying the real underlying problem, it is important to set objectives that can make us evaluate all the possible solutions. Some people may refer to these objectives as criteria. Linguistically, these two terms have slight difference, for example, a criterion can be "price", but the objective can be "finding cheapest price" or, if you are a rich person who wants to show off, the objective can be "finding the most expensive". In this example, both objectives are about the same price criterion, but one is trying to minimise while the other is trying to maximise. In most real-life problems, there are multiple criteria/objectives, and often these objectives are in conflict with each other. For example, one may dream of getting the cheapest item but with the highest quality.

Alternative: Consider a situation where there is only one possible course of action or choice. Will it be a decision problem? Not really, because the decision problem becomes a problem only when you have multiple alternative options, and you have to pick only one of these choices. Therefore, it is important to realise that identifying multiple alternative options is a very important step. One may miss a great (or best) solution without realising that it exists.

Consequences: After generating a good set of alternative options, it is important to evaluate the consequences of each of these alternatives. Usually this is done by creating a table where alternatives are placed in rows, objectives are places in columns, and then the consequences are penned down in each cell of this table. When describing the consequences in these cells, it is important to make sure they are as accurate as possible. Bad assessments will generate bad results, as they say "garbage in, garbage out".

Trade-offs: As we mentioned earlier, multiple objectives are often in conflict with each other, and therefore, you might end up with a table of consequences where there is no single clear winner. In this case, a trade-off analysis will be required where we need to sacrifice (or trade) something from one objective if we want to gain at least a little on another one.

In ProACT, there are three more elements that decision makers should consider in their decision-making exercise. These are Uncertainties, Risks, and Linked Decisions; but we will not discuss them here in detail. If you are interested in reading more about ProACT, we recommend reading the original book from Hammond J.S., Keeney R.L., and Raiffa H. (2002).

1.2 Different Stages in Decision-making

It is fair to say that every decision process starts with a need or desire to decide. The decision-maker realises this need when multiple alternatives appear viable to them whether known or unknown to them at the time of realisation. This realisation leads them to the two often-intertwined phases of structuring and exploration. For example, you have just moved to a new house and you realise you really need an air-conditioner. While exploring the various air-conditioning products on a popular shopping website, you make yourself aware of the criteria that must be taken into consideration when purchasing an air-conditioner. This is mostly true when the...