Section 5: Scenario analysis

Section 5: Scenario analysis

“Introduction … Scenarios … Estimates … Score cards and SMART”
(Source URL)

Summaries

  • Step 5: Scenario analysis > Scenarios > Video uncertainty and scenarios
  • Step 5: Scenario analysis > Estimates > Video estimates
  • Step 5: Scenario analysis > Score cards and SMART > Video score cards

Step 5: Scenario analysis > Scenarios > Video uncertainty and scenarios

  • Here is how a scenario analysis can help in dealing with these external factors.
  • Imagine you are a car dealer and you want to build and open a new shop for your company somewhere in a densely populated and industrial area.
  • In case of the car dealer, important external factors are for example the oil price and the land prices of empty building plots.
  • The land price is a factor with quite a high level of uncertainty and also a high impact, because the price the car dealer pays for the building ground covers a large part for the total investment costs for a new shop.
  • Again, the impact is also high, because the oil price influences the attractiveness of car driving and the demand for new cars As both these factors are uncertain and have a high impact, they provide very interesting material for a scenario analysis.
  • Based on these two factors, for example, I will consider four scenarios: Scenario 1 has a high oil price and the prices for land are also high as well.
  • In Scenario 2 the oil prices are high as well, but buying land to build on is more profitable than in Scenario 1.
  • A final scenario that I will use to test the alternatives is a scenario in which both oil price as the land prices are low.
  • Thinking about these scenarios provides me with more insight about how the future could possibly develop and helps to draw some conclusions.
  • What I will do for the car dealer is to create a score card for each scenario.
  • Implementing an alternative in one scenario will have a different effect on the criteria than implementing that same alternative in another scenario.
  • In the eyes of the car dealer, scenario 4 is a future he would like to see happening.
  • In this scenario, car driving is highly attractive and constructing a new shop is relatively cheap.
  • On the other hand, he hopes that scenario 1 will not become real.
  • What he is looking for is an alternative that is not just performing well in one of the scenarios, but is able to deal with as many different scenarios as possible.
  • I am Bert Enserink and I am the director of studies of the Engineering and Policy Analysis master program and I am a researcher here at Delft University of Technology and I am a specialist in Policy Analysis and dealing with uncertainty.
  • One of the methods we use a lot is scenario analysis.
  • What will happen with the EU, what will happen with the Euro? What exchange rates? What is the political future in the Eastern part of Europe? A lot of uncertainty and what we did was make a couple of scenarios, combine different trends, and see how this company would have to behave in these different futures.
  • Scenario studies can be very helpful in these kind of situations where there are a lot of uncertainties, external events happening that people don’t have any influence on and where you have to decide on strategic important issues for your company, for your government, for anything for policy making.
  • In this course, we learn the basic method for scenario analysis.
  • These are typical models that are used in a scenario environment with different futures, with extensive modelling.
  • This video has introduced you to the basics of scenario analysis.
  • Chapter 4 in the book describes the essence of scenario analysis.
  • Please, before you start making scenarios for your own case, study this chapter.
  • How do you categorize them using the axes certain-uncertain and high impact-low impact? You categorization determines which of these factors are a candidate for scenario analysis.
  • If they are very certain, there is no need to make scenarios as you already know what this factor will look like in the future.
  • Now you have your scenario factors, what useful scenarios can you create for your analysis? In this video we showed a quantitative description of such a scenario, but we also tried to make a little story of it.
  • In the end, the numbers don’t say all that much, what is much more important is what the scenario looks like.
  • Scenarios are possible future states of the situation you are analyzing.
  • When you have your scenarios, you link them to your analysis by making a score card for each of them.
  • We have seen many mistakes in scenario analysis over the years.
  • Here are the two most common mistakes when making a scenario analysis.
  • People around you need an argument why that factor needs scenario analysis.
  • Making more scenarios is more work and is more expensive.
  • Clients therefore tend to force you to reduce the number of scenarios.
  • They will challenge the explanations of your external factors, so you’d better have good ones! Another common mistake is that some consultants or managers make a prediction of the future and then use that prediction as a scenario.
  • Apply the scenario technique and see what it brings you! “.

Step 5: Scenario analysis > Estimates > Video estimates

  • In a situation such as this you, as an analyst or consultant, can quickly check people’s arguments with so called estimates, which are in essence quick and dirty calculations.
  • If we want to make an estimate, we start with questions about the order of magnitude of a certain unknown factor.
  • How much electricity is used in a small city of around two hundred thousand inhabitants by all devices that are in stand-by modus? This factor cannot be estimated immediately.
  • If you want to estimate this number order of magnitude, you decouple this factor in sub-factors, until you can make an easy estimate of such a sub-factor that you can know.
  • Also that factor cannot be estimated, i.e. determined within a range of ten times smaller, ten times larger.
  • What you can do is estimate the average house hold size, say two.
  • Hello, I am Bert van Wee, professor of transport policy at Delft University of Technology in the transport and logistics group.
  • My main research interests are in long term policy relevant societal changes in the areas of accessibility, the environment, and safety.
  • Suppose you want to estimate the impact of policies to stimulate people purchasing and using electric vehicles on the concentrations of pollutants and next the health effects.
  • So what you do is you make estimates for all those pieces.
  • What you for example do is you look at how many people do you think will buy an electric car and use it because of the policies.
  • Combining that estimate with estimates for emission per kilometer for conventional cars as well as electric cars, you can estimate total car emissions.
  • If you have estimates for other emissions like industries or power plants, you know total emissions.
  • Then you can look at, what is the distance between emissions and where people stay? You look at the distance between cars and where people stay and power plants that also emit.
  • More policy decisions are based upon these kind of estimates than on the outcomes of one model, because there is not one model predicting in one go what policy makers want to know.
  • How many trains arrive and depart then? Say, twelve per hour, six in one direction , six in the other direction.
  • How many people can travel in one train? Say two-hundred-fifty.
  • As analyst you quickly have to make an estimate on the back of an envelope So, how many carriages does an average train have and how many seats are in such a carriage? eight and forty respectively? Fine, then we are done.
  • My city has about one-hundred thousand inhabitants and, as the trains also contains people from other cities and villages, this number could be ok, when we look at the order of magnitude.
  • Formally you should even write down eight times ten to the fifth power seats per day, as your estimate is for sure not more accurate than one significant digit.
  • Dear analysts, managers and consultants, welcome back to the part in which we work on our own case! In chapter 5 in the book there is a section on estimates.
  • What you also can do, is make estimates, just like you have seen in this video.
  • For your score cards, have a look at what cells cannot be filled these you can estimate.
  • Try it out in multiple ways, so you can compare different estimates see if you, order of magnitude, end up roughly with the same number.
  • What do you say to people that claim estimates are useless because the number you come up with is not correct anyhow?’ There are two very common mistakes many people make when estimating.
  • People might disagree and come up with another number.
  • This needs to be further decomposed into subfactors like the number of carriages in a train and the number of seats per carriage.
  • When making estimates you are not very accurate, as you are only accurate in the order of magnitude.
  • Final estimates often look far more accurate than they actually are.
  • Stating the estimate is 6.34532 on average, gives the idea that you know this number very precise.
  • Enjoy filling your score cards, with scores from many different sources, also from the estimates as you have seen in this video.

Step 5: Scenario analysis > Score cards and SMART > Video score cards

  • We started by listing the actors and making their problem explicit and later came up with many alternatives.
  • The next step is to help yourself or your client to make a choice between all these different alternatives.
  • The most common form is a table with all alternatives in the first row and all criteria in the first column.
  • At that time, the government presented three different alternatives for this rail line: The first, a completely new line.
  • Even though the second alternative was declared the cheapest and easiest, the track was considered too slow to be competitive with cars and planes.
  • The parliament preferred alternative 1, an entirely new line as it would meet environmental and transportation aspects better than the other alternatives.
  • Many actors had different alternatives as their preferred solution, for many different reasons.
  • So how can we, as analysts or consultants, help these people to make a clear choice among the different alternatives? What you want to do is to make a clear overview of the analysis so far that enables to compare the alternatives and see how they score on the different criteria.
  • It does not tell us whether an alternative is good or bad. However, is does tell something about how an alternative scores on the different criteria.
  • So it makes a comparison of the different alternatives possible.
  • In this course we provide two techniques to perform a multi-criteria-analysis: The Score Card technique and the Simple Multi Attribute Rating Technique, SMART in short.
  • The first thing to do is to choose for what context scenario you will make a Score Card Alternatives will score differently in different potential futures.
  • As consultant or manager, you do not only want to compare the alternatives on criteria, but you also want to know what alternatives are robust, i.e., what alternatives perform relatively well in many different potential futures.
  • The criteria you put in this score card should be the same criteria you used in your causal diagram and in your problem diagram.
  • Put all the alternatives that you want to compare in the first row.
  • In principle they are the same alternatives you have identified and used in your causal diagram and problem diagram.
  • There, for practical reasons, you might have made summaries and listed not all, but a set of alternatives.
  • Here you list all alternatives, whether there are 4 or 400.
  • Now, by taking values from the literature, by doing experiments, consulting experts or making estimates, determine the scores and fill in these scores in the table.
  • The SMART builds upon the score card and is a more extensive technique for comparing alternatives.
  • We present here normalisation of scores into values between 0 and 1, 0 being the worst and 1 being the best.
  • First is to determine for every criterion what the highest and lowest value is for all alternatives.
  • Add another row to the table for the weighted sum of the normalised values for each alternative.
  • You get these by multiplying each value with the weight and add the scores.
  • When we want to do that, we need a couple of alternatives to make sure that we can store and use the energy at a required time.
  • In order to study these alternatives within my research team, we use score cards.
  • The score cards also help us to understand exactly what criteria are important to study and understand the viability of these alternatives.
  • That we start to think about how we can measure these criteria and what kind of modelling and analysis we still need to do to fill out these score cards.
  • So this combination of ease of communication, helping us or maybe even forcing us to understand what kind of criteria are there and how we can measure and model them helps us a lot in getting a fast and good grip on the entire quite complex problem.
  • The use of problem diagrams in combination with score cards helped us to evaluate the alternatives and to compare them and it forced us to make the criteria and their measurement and modelling very explicit.
  • After doing all analyses of the problem situation and brainstorming for different alternatives, it was time to consider the impact of the possible alternatives both on her study results as on her private life.
  • By looking at the consequences of several alternatives like dropping one of her hobbies or sports, skipping some of the courses, taking a less active role in her committee, sleeping less and many others, the student was able to come to a conclusion about a set of actions to improve the study results, without harming her private life too much.
  • Let me help you by applying Score Cards and SMARTs to your own case.
  • Score the different alternatives you identified on your chosen well-balanced set of criteria from your goal trees.
  • When a certain alternative has higher costs, this is obviously a bad score for the one who has to pay, but not necessarily for the one who sells.
  • In SMARTs you do not see the raw scores anymore, only a score between 0 and 1, thus it is very important that you do the normalisation well.
  • The second is about weight factors, remember that the weight factors have a huge impact on the rank order of alternatives.
  • The analytic ranking order of alternative can suddenly be completely different after a very small change in weight factors.
  • This information can be even more interesting than the specific scores of the alternatives, as it says something about the usefulness of the method! Finally, don’t forget that the highest ranked alternative of the SMART table is not necessarily the best solution for the problem; SMART was just a tool to make the views of actors on what is important explicit.
  • Use Score Cards and SMART’s as a group tool, discuss and come to a decision on what to do together.

Return to Summaries

(image source)

 

Leave a Reply

Your email address will not be published. Required fields are marked *