1. What's New in Forward-Looking Models


This lesson explains the characteristics and challenges of building forward looking models, and introduces the SupraChem case.

To explore more Kubicle data literacy subjects, please refer to our full library.


  1. Understanding Forward-Looking Models (00:04)

    Forward-looking models aim to predict future outcomes based on a set of assumptions. They aim to improve decision-making where the outcome of a present decision is based on future events. For example, many investment decisions require the use of forward-looking models to understand the size and timing of the investment returns.

    The main benefit of forward-looking models is that they can deliver better insights than models based on historic data. The main disadvantage is that model predictions can differ from reality, especially for predictions that are further in the future. As a result, you need to qualify the recommendations you make from a forward-looking model.

  2. Course Case Study (01:47)

    In this course, we consider a large chemical manufacturer called SupraChem. They have developed a renewable fuel product and built a small test facility to manufacture it. They must now decide if they should build a larger reference plant to manufacture the fuel at a larger scale.

    The outcome of the investment depends on the production cost for the fuel achieved by the reference plant. There is a high probability that the target price will not be achieved. However, if the target is achieved, the potential revenues are enormous. This is a high risk investment, because it cannot be undone once SupraChem commits to it. Therefore, high quality analysis is essential to decide if the company should build the plant.

    For more details about the company and the investment, you can download the lesson file for this lesson.


The reason we build models is to help us make better decisions. Sometimes we do this by analyzing historical data for insights and other times we try to predict the future with forward-looking models. The first modeling in case of Zippy Airways examined historical flight information to see if an overbooking policy could deliver additional profitability. In this course, our case study SupraChem will need a forward-looking model to help the company make an investment decision on a new technology. So why build forward-looking models? Well unsurprisingly, forward-looking models can provide better insights on the future than models based on historic data. Indeed forward-looking models typically utilize historical data when developing their projection assumptions. Certain types of decisions simply require forward-looking models. In business, many of the decisions management make are investment decisions. When you make an investment, forward-looking models are essential because you'll want to know how the money will be spent, what returns you will get on your investment and when you will realize these returns. Forward-looking models do come with a big health warning, however. Firstly, they require a lot more assumptions often many years into the future. As a result, forward-looking models have the potential to deviate hugely from reality. When presenting to a client or to superiors, you should always be acutely aware of the accuracy of your assumptions and qualify your recommendations accordingly. In this course, we'll introduce a number of new concepts and tools for forward-looking models. This includes decision trees, tornado charts, scenario analysis and much more. With that introduction to forward-looking models out of the way, let's introduce the case. Our client SupraChem is a large chemical manufacturer in the US. For the past five years, a team of engineers have been working on a new renewable fuel product. This is a new market for SupraChem outside their core competency and the executive team are anxious and excited by the opportunity. The engineering team have built a test plant that currently produces the renewable fuel at $7 dollars per gallon. The market price however is currently $2.80 per gallon. The next phase of the project, a larger reference plant will incorporate new technology and confirm the exact production cost that Supra could hope to achieve at commercial scale. This is a high-risk investment. The reference plant has only a 15 percent chance of succeeding and hitting the target price of $1.80 per gallon. There is an 85 percent chance that the reference plant will be written off as a bad investment. However, if the reference plant is a success, the subsequent commercial plant will be able to generate 200 million gallons per year and the potential revenues are huge. So the question is, should SupraChem proceed with the reference plant? To help us get started, I've included some additional numbers with regard to the project. I suggest downloading the file beneath this video and keeping this slide at hand as we walk through the model. In contrast to our previous case Zippy Airways, there's much more at stake in this decision. The investment numbers are larger, the risks are larger, particularly because this is a go, no-go decision. While Zippy Airways could tweak their overbooking policy over time and adjust the model with new data, this investment once committed cannot be undone. In companies, investments such as these can make or break careers, so it's critical that excellent analysis is conducted before making a recommendation. Let's get started in the next lesson with diagraming the problem.