Managing Risk Through Simulation

Uncertainty and risk is ever-present in our everyday business transactions. Whether we're estimating productivity, forecasting product demand, calculating utilization of facilities, or any other business calculation, the use of estimates and uncertainty of critical variables can produce a wide range of outcomes. Which one we pick often relies on our ability to intuit the most likely outcome. As most managers will admit, even the best-designed forecasts are rarely realized exactly.

So what is a savvy 21st Century manager to do when confronted with uncertainty in forecasting the results of business processes? If you're like most managers, you'll go through a series of scenarios to project what you expect will happen. You may also project "best" and "worst" case scenarios. This is typically how we predict uncertain events. But these three forecasts (expected outcome, best outcome, and worst outcome) lack the information necessary to make good business decisions.

To make more informed decisions, managers need to know how likely it is for each of these scenarios to occur. For example, a worst-case scenario that involves massive revenue losses may derail a business opportunity. However, had the manager known that the chance of the worst-case scenario's occurring is one in a million, the potential for revenue losses would be put in a better perspective.

To gain this additional perspective on probabilities of forecasting outcomes, managers can simulate their business decisions based on the variables involved in each business decision. A forecast simulation is a technique that runs through a large number of possible outcomes for a single decision. To perform an effective business simulation, managers should go through four steps: identify the variables affecting the outcome, collect data for each variable, develop and run a simulation, and analyze the simulation for decision-making information.

1. Identify the variables in the forecast. What makes forecasting uncertain? If you are forecasting sales, is your market share uncertain? Is your production volume uncertain? Is your distribution network uncertain? Each of these variables will have an effect on how much your company can sell. Identifying each variable that will impact your bottom line is critical to making an informed forecast.

2. Collect data for each variable. This can be difficult and time consuming, but this step will determine the range of outcomes in your forecast. When collecting data, pay special attention to the range of possible values that each variable can take. For example, if production volume varies between 1,800 and 2,200 tons monthly, use this range in your simulation rather than a simple average or "best guess". Incorporating the entire range will make your forecast much more informative than a single data point.

3. Set up and run a simulation. Once you've identified the variables for your forecast and established the range of values for each variable, you are ready to formulate your simulation. Most spreadsheets can accommodate simulations, and there are also several off-the-shelf packages to assist in simulation and forecasting. It would be a good idea to run at least 500 iterations of the forecast to generate a distribution of outcomes.

4. Analyze the simulation results. Use the results of the simulation to inform the forecast decision. Is the average of your simulation a reasonable target for "most likely" outcome? How likely is your worst-case scenario to happen? How likely is the best case? With a complete simulation, you can also create "what-if" scenarios. By changing the range of values that a variable can take, a manager can simulate a business scenario. What if the northeast plant closes? What if our market share increases by 15%? Finally, study the simulated outcome to determine which variables contribute the largest amout of uncertainty to the simulation. If one or two variables contribute the majority of uncertainty to your simulation, focus on quality controls around these variables to mitigate your risk in forecasting.

Simulating forecast decisions can be difficult and time consuming. But the resulting wealth of management information will make any business decision more informed and easier to make.

Matt Stanczak specializes in business modeling and operations research at THE HILL GROUP. If you would like further information about business modeling, contact Matt at 412-343-9393 or mstanczak@hillgroupinc.com.  

Insight and Innovation Home

These materials have been prepared for educational and information purposes only. They are not consulting advice or opinions on any specific matters. Transmission of the information is not intended to create, and receipt does not constitute, a consultant-client relationship between The Hill Group, Inc. and any recipient of this material. Readers should not act upon this information without seeking professional advice.