It All Starts with “What If…?”

How many times have you uttered or typed those two little words as a supply chain practitioner? Hundreds or even thousands of times? Asking “what if…?” is the genesis of any strategic supply chain decision-making process: What if we introduce the new product line next year? What if this facility can only run at 60% capacity for the next few weeks? What if I add a third shift to my operation?

While the practice of asking questions to explore potential choices and outcomes is as old as humanity, technology continues to allow more precise, insightful analysis into how intentional and unintentional actions affect the behavior of complex systems.

You understand the importance of looking at a myriad of scenarios as part of the supply chain decision-making process, but some of the terms technology vendors use nowadays can be confusing. Good examples are scenario modeling, what-if analysis, and simulation. What’s the difference between each approach?

The Difference Between Scenario Modeling and Simulation Modeling

The terms “scenario modeling” and “simulation” are often used interchangeably yet they are not the same thing. As a simulationist who is passionate about this modeling approach and its benefits, I feel compelled to ensure that the true value of supply chain simulation is not overlooked by someone who believes performing what-if analysis with any type of model is the same as simulation modeling.

Most of the confusion comes from the fact that people do not understand what simulation is, what type of business decisions can be made with this modeling approach and the value it can bring that other models cannot. Let’s dive in and break down each component.

What Is Scenario Modeling?

Merriam-Webster defines a scenario as “an account or synopsis of a possible course of action or events”. You can think of scenario modeling as using a model to examine potential futures and predict the various results and potential outcomes. It can be used interchangeably with the term “what-if analysis”.

The terms ‘Scenario modeling’ and ‘what if analysis’ can be used interchangeably.

Determining what scenarios to run is only one step in the process. Defining what outputs, KPIs and metrics to measure is often a more difficult step, along with deciding what type of model will provide the information needed to answer the business questions.

Defining Different Types of Supply Chain Models

Consider a model as a representation and abstraction of a real system or a proposed system. There are many classifications of models but for this discussion, we’ll touch on the difference between static and dynamic models, along with deterministic and stochastic models:

• Static models describe a system and relationships that do not change with respect to time.

• Dynamic models represent time varying relationships and interactions.

• Deterministic models are the most common type of model. There is no randomness in deterministic models and therefore a future event can be calculated exactly.

• Stochastic models—most simulation models are stochastic, which means they have one or more random variables as inputs, hence leading to random outputs. This often more closely represents a real system since variability is present in every system.

How to Decide Which Model Type to Use

One of the most difficult but most important decisions a modeler will make is what type of model to use for the business decisions and analysis needed.

Consider you’d like to analyze the scenario, “What if demand grows by 25% next year?” If the output you are analyzing is whether your current supply chain network can handle this increased volume or if you’d need to expand your network (and how), you would build a certain type of model, with a certain level of detail, with outputs focused on answering those questions.

However, you might want to analyze that same scenario to answer additional questions like:

“What should the inventory stocking and replenishment policies be at each location, given the increased demand?”

“Will our customer service levels remain the same?”

The model, the detail and the outputs needed for this analysis would be different. And therefore, it’s critical to choose the correct model and modeling approach to provide the correct information for those outputs.

For example, you might consider performing scenario analysis with a simulation model to address the second two questions whereas a robust optimization model might provide the best results needed to address the first question. Scenario modeling would be done on both models, but it is only accurate to consider one of the two approaches to simulation modeling.

From M&A analysis to inventory strategy, supply chain design can answer myriad supply chain what-if questions. → See use cases for supply chain design

What Is Simulation?

What is simulation modeling and why is it different from many other mathematical modeling approaches? A simulation model is used to understand the behavior of the system or to evaluate various strategies for the operating of the system. While there are some static simulation models, such Monte Carlo models, we will focus on dynamic, discrete event simulation models, which represent systems as they change over time.

A simulation model is used to understand the behavior of the system or to evaluate various strategies for the operating of the system.

Typically, there are many complex interactions and implications of decisions that can only be accurately modeled with the element of time. While a simulation model might be run deterministically (without variability) to model the baseline (or current system), most simulation models are stochastic so that the randomness and variability present in the real system can be incorporated into the analysis. What real system does not have variability?! Think about travel times, order quantities, or processing times, to name just a few.

Scenario Modeling with Simulation

It is rare that a simulation model is run without running multiple scenarios, so scenario modeling is almost always part of the simulation modeling process. Therefore, it is appropriate to use the terms together in this case, which helps clarify that they are indeed two different concepts. You’ll often hear simulation modelers talk about creating experiments instead of using the term scenario modeling – an experiment is running several different scenarios.

Hopefully this has clarified that the term scenario modeling is not the same as simulation modeling. If the mathematical model that you are running scenarios with is not simulating the behavior of a system but instead just solving an equation or objective function, it should not be considered a simulation. You might be missing out on the true value that simulation can bring to your decision-making.

How to Use Supply Chain Simulation

I encourage you to learn more about simulation and its benefits and applications and incorporate this technology into your modeling toolbelt. In most cases, both optimization and simulation should be used for analysis, as they both have their strengths and complement each other well.

Look for a modeling solution like Cosmic Frog that gives users access to both technologies and creates an intuitive workflow for the user. And then you might find yourself using scenario modeling in both optimization and simulation.

Did you know? Cosmic Frog is a 3-in-1 supply chain design solution that incorporates optimization, simulation, and risk. Take it for a test drive → Create your free account

If you’d like to learn more about discrete event simulation, how it can improve your decision making and how to begin working with it, you’re speaking our language! Contact us.

About the Author

Renee Thiesing

Renee has over 20 years of experience working in the software industry, most of which has included a focus on simulation and applying this technology towards process improvement and enablement of digital transformations. With a focus on manufacturing and supply chain, she has worked with customers across verticals, such as CPG, Pharma, Aerospace & Defense, and Food & Beverage. Renee has held a variety of roles throughout her career in software and is currently the VP Strategy and Risk Solutions at Optilogic.