1 00:00:00,740 --> 00:00:07,519 Warning: This video clip contains material intended for an audience of 18 years and older! 2 00:00:07,519 --> 00:00:12,480 Hi! Now that I have your attention, let's talk about problem demarcation. 3 00:00:12,480 --> 00:00:17,449 In this short video clip I will explain to you how you can demarcate your client's problem 4 00:00:17,449 --> 00:00:22,779 in such a way that the subsequent analysis of the system will deliver the most actionable 5 00:00:22,779 --> 00:00:24,560 information to your client. 6 00:00:24,560 --> 00:00:28,759 It will be a rather abstract presentation, focusing on methods. 7 00:00:28,759 --> 00:00:33,070 If you are looking for a concrete example, then you should watch the second video clip 8 00:00:33,070 --> 00:00:35,110 on this same topic. 9 00:00:35,110 --> 00:00:39,720 Let's first motivate why problem demarcation is important. 10 00:00:39,720 --> 00:00:44,650 First and foremost, it is in your client's interest that you establish what is the problem 11 00:00:44,650 --> 00:00:47,370 that is most relevant to analyze. 12 00:00:47,370 --> 00:00:52,620 Secondly, it helps you to be efficient: a good problem demarcation makes that you will 13 00:00:52,620 --> 00:00:58,760 look only into issues that matter, and that you do your analysis in adequate detail. 14 00:00:58,760 --> 00:01:02,600 Thirdly, it helps you to be accountable for your findings. 15 00:01:02,600 --> 00:01:07,280 By demarcating the problem you make clear to your client what will be the scope of your 16 00:01:07,280 --> 00:01:13,020 analysis; in other words, the things that you will consider, and also the things that 17 00:01:13,020 --> 00:01:16,499 you might have considered, but decide to ignore. 18 00:01:16,499 --> 00:01:22,219 This allows both you and your client to reflect on how this will eventually limit the conclusions 19 00:01:22,219 --> 00:01:27,159 that you can draw, and the recommendations that you can make. 20 00:01:27,159 --> 00:01:30,549 So how do you do problem demarcation? 21 00:01:30,549 --> 00:01:36,390 In principle, you go through five steps: You begin by choosing one issue that appears to 22 00:01:36,390 --> 00:01:39,759 be of immediate interest to your client. 23 00:01:39,759 --> 00:01:43,159 Something that your client wants to change or considers doing. 24 00:01:43,159 --> 00:01:46,920 From this starting point you perform a means-ends analysis. 25 00:01:46,920 --> 00:01:50,200 This will give you an overview of related issues.. 26 00:01:50,200 --> 00:01:55,079 You then develop alternative problem statements, and for each problem you make an objective 27 00:01:55,079 --> 00:01:58,729 tree and outline the associated system boundary. 28 00:01:58,729 --> 00:02:03,280 Comparing these alternatives will allow you to decide what problem should become the focus 29 00:02:03,280 --> 00:02:04,479 of your analysis. 30 00:02:04,479 --> 00:02:10,880 To give just one example: if your client is the Port of Rotterdam, several issues might 31 00:02:10,880 --> 00:02:17,880 surface during your first talks: The port might, for example, want to deepen the harbor, 32 00:02:17,980 --> 00:02:23,530 increase the capacity of its container stacks, or reduce its CO2-emissions. 33 00:02:23,530 --> 00:02:28,000 Each of these issues would be a good starting point for your means-ends-analysis. 34 00:02:28,000 --> 00:02:32,440 By constructing a means-ends diagram, you obtain an overview of issues that you could 35 00:02:32,440 --> 00:02:34,220 focus on. 36 00:02:34,220 --> 00:02:40,590 The building block for this diagram is a means-ends box: a rectangle that you label with a single 37 00:02:40,590 --> 00:02:46,760 verb phrase that expresses something that your client wants to achieve (an end). 38 00:02:46,760 --> 00:02:51,250 You use a verb phrase because a means-ends box also represents something your client 39 00:02:51,250 --> 00:02:54,480 can do (a means). 40 00:02:54,480 --> 00:02:58,870 Starting from your first box, you develop your means-ends diagram by repeatedly asking 41 00:02:58,870 --> 00:03:00,870 two questions. 42 00:03:00,870 --> 00:03:07,170 You begin by asking "why?" to discover higher level ends that explain why your client wants 43 00:03:07,170 --> 00:03:09,680 to reduce CO2-emissions. 44 00:03:09,680 --> 00:03:15,230 Then you start asking "how?" to identify for each end the means that could help your client 45 00:03:15,230 --> 00:03:17,420 achieve this end. 46 00:03:17,420 --> 00:03:23,520 Let me quickly show you how these why and how questions lead to a means-ends diagram. 47 00:03:23,520 --> 00:03:30,500 By answering your "why" question for your starting point S, you get a higher level "means-ends box". 48 00:03:30,680 --> 00:03:37,680 You draw an arrow to indicate that using means S will help achieve the higher level end H. 49 00:03:38,160 --> 00:03:45,160 You then ask "why does my client want to achieve H?", which will produce again a higher level 50 00:03:45,780 --> 00:03:47,280 means-ends. 51 00:03:47,280 --> 00:03:54,300 You repeat this until you reach your clients' fundamental objective -- let me label that with F. 52 00:03:54,590 --> 00:04:01,250 You then start asking "How can my client achieve this?" for each means-ends you have discovered. 53 00:04:01,250 --> 00:04:06,270 This forces you to consider alternative ways in which your client may achieve its ends. 54 00:04:06,270 --> 00:04:10,790 There may be other means, in addition to H, by which your client can achieve F. 55 00:04:10,790 --> 00:04:17,790 Likewise, there may be more ways, in addition to S, to achieve H, and possibly your client 56 00:04:18,019 --> 00:04:22,370 should focus on these. 57 00:04:22,370 --> 00:04:27,499 By continuing to ask "how?", you may discover that your client has a wide range of means 58 00:04:27,499 --> 00:04:29,759 to achieve its ends. 59 00:04:29,759 --> 00:04:35,889 At some point, you should stop, typically when the means become very detailed and operational. 60 00:04:35,889 --> 00:04:41,629 By that time, the boxes and arrows that you have drawn form a means-ends diagram. 61 00:04:41,629 --> 00:04:47,310 The next step is to use this diagram to systematically develop alternative problem statements. 62 00:04:47,310 --> 00:04:52,400 In principle, you should consider each box in the diagram that has boxes connected to 63 00:04:52,400 --> 00:04:58,600 its bottom, because that indicates that your client has the means to achieve the end in this box. 64 00:04:58,979 --> 00:05:02,900 In my example, there would be six. 65 00:05:02,900 --> 00:05:10,639 These six then give you six alternative problems. 66 00:05:10,639 --> 00:05:15,060 A good problem formulation expresses a dilemma or trade-off. 67 00:05:15,060 --> 00:05:19,960 In a means-ends diagram, however, every box is an end, that is, something desirable for 68 00:05:19,960 --> 00:05:21,110 your client. 69 00:05:21,110 --> 00:05:24,050 So we must now also look for bad things. 70 00:05:24,050 --> 00:05:29,889 The way to do this is to ask yourself "what are the undesirable side-effects of my client's 71 00:05:29,889 --> 00:05:36,340 means?" Let me show you how this works: First, you pick a box in your means-ends diagram 72 00:05:36,340 --> 00:05:41,289 and you take this (for now) to be your client's focal objective. 73 00:05:41,289 --> 00:05:47,289 The diagram will then give you the means that your client may use to reach this objective. 74 00:05:47,289 --> 00:05:53,500 For each means, you then consider what side effects may occur when your client would use it. 75 00:05:53,900 --> 00:05:59,529 Means usually cost money, but you will often find other side effects as well: there may 76 00:05:59,529 --> 00:06:03,719 be safety issues, environmental issues, and so on. 77 00:06:03,719 --> 00:06:08,569 The focal objective together with the undesirable side effects of the means to achieve this 78 00:06:08,569 --> 00:06:15,569 objective will produce a good problem formulation: "How can your client achieve one without causing 79 00:06:15,770 --> 00:06:21,629 too much of the other?" You can do this for each box in your diagram, provided that it 80 00:06:21,629 --> 00:06:24,169 has boxes connected to its bottom. 81 00:06:24,169 --> 00:06:31,169 So if, instead of A, we focus on B, we see that our client has two means to achieve B. 82 00:06:31,629 --> 00:06:38,400 Both again have side effect X, and one also has side effect Z. 83 00:06:38,400 --> 00:06:44,599 This would then give as problem statement "How can our client achieve B without too 84 00:06:44,599 --> 00:06:51,599 much X or Z?" We can do exactly the same for the other means to achieve A. 85 00:06:51,809 --> 00:06:58,300 Each time we look for undesirable side effects to identify the tradeoffs involved. 86 00:06:58,300 --> 00:07:04,759 So we imagine what would happen if our client decides to use means G and H to achieve C. 87 00:07:04,759 --> 00:07:07,900 We do the same to find the tradeoffs involved to achieve D. 88 00:07:07,900 --> 00:07:14,900 And let me give just one more example: when we consider G as focal objective, we imagine 89 00:07:17,069 --> 00:07:24,069 what would happen if our client would use means J, K or L, and if that leads us to identify 90 00:07:24,069 --> 00:07:31,069 also U and V as undesirable side effects, a fifth problem statement would be: "How can 91 00:07:31,930 --> 00:07:38,930 our client achieve G without too much U, X or V?" What you do then is consider each of 92 00:07:40,599 --> 00:07:43,870 the problem statements you have formulated. 93 00:07:43,870 --> 00:07:48,430 Some may be more pertinent than others, and possibly you can already in this stage put 94 00:07:48,430 --> 00:07:54,849 some aside for being not (or not yet) the most relevant issue for you to analyze further. 95 00:07:54,849 --> 00:07:59,729 For each of the remaining problem statements you then develop an objectives tree. 96 00:07:59,729 --> 00:08:00,990 This is fairly easy. 97 00:08:00,990 --> 00:08:06,719 Remember that the sole purpose of an objectives tree is to define crisp, measurable decision 98 00:08:06,719 --> 00:08:08,490 criteria for your client. 99 00:08:08,490 --> 00:08:14,150 A good objectives tree thus operationalizes the desirable and undesirable factors in the 100 00:08:14,150 --> 00:08:15,349 problem statement. 101 00:08:15,349 --> 00:08:19,400 In this example: A, X and Y. 102 00:08:19,400 --> 00:08:24,330 Since an objectives tree needs a root, we add it: since the overall objective is to 103 00:08:24,330 --> 00:08:31,289 achieve A without X or Y, we can express this as "good A" -- good, because it is A without 104 00:08:31,289 --> 00:08:33,630 any negative side effects. 105 00:08:33,630 --> 00:08:37,380 We must then operationalize A, X and Y. 106 00:08:37,380 --> 00:08:42,000 If A cannot be measured directly, we must find proxies that can. 107 00:08:42,000 --> 00:08:47,730 Let's assume that these are a1 and a2, and that for A to be good, a1 should be small, 108 00:08:47,730 --> 00:08:50,360 while a2 should be large. 109 00:08:50,360 --> 00:08:54,790 If X is monetary cost, then it can me measured immediately. 110 00:08:54,790 --> 00:09:00,360 And if Y needs further operationalization, this will -- like A -- involve finding good 111 00:09:00,360 --> 00:09:02,170 proxies. 112 00:09:02,170 --> 00:09:06,800 By defining the criteria, we actually define part of the boundary of the system that we 113 00:09:06,800 --> 00:09:12,509 should consider when analyzing this problem: In our analysis, we should establish how strongly 114 00:09:12,509 --> 00:09:16,029 the means of our client will affect the criteria. 115 00:09:16,029 --> 00:09:21,720 So our objectives tree provides the right-hand side of a system diagram. 116 00:09:21,720 --> 00:09:28,720 Likewise, the means-ends diagram provides the left-hand side. 117 00:09:29,940 --> 00:09:34,670 If we do this for each of the alternative problem statements, we get an overview of 118 00:09:34,670 --> 00:09:38,600 alternative ways of demarcating our client's problem. 119 00:09:38,600 --> 00:09:44,690 In this case three different problems with three different demarcations. 120 00:09:44,690 --> 00:09:49,680 This illustrates that, depending on how we demarcate the problem, we will focus our further 121 00:09:49,680 --> 00:09:56,680 analysis on different measures that our client may take, and on different outcomes of interest. 122 00:09:57,480 --> 00:09:59,089 Okay. 123 00:09:59,089 --> 00:10:05,700 The purpose of this video clip was to explain how you can do problem demarcation in a systematic way. 124 00:10:05,730 --> 00:10:10,750 Starting from one salient issue, means-ends analysis reveals a broad range of problems 125 00:10:10,750 --> 00:10:12,889 that you could focus on. 126 00:10:12,889 --> 00:10:18,079 By doing a quick scan of each problem, you can then make an informed decision about what 127 00:10:18,079 --> 00:10:20,970 is most relevant for your client. 128 00:10:20,970 --> 00:10:24,720 Remember: you do your analysis for your client. 129 00:10:24,720 --> 00:10:27,560 After all, it is all about your client's problem!