Guest Post: Connecting Decisions to Data: A Case Study, Part 6: Maximizing Opportunity while Limiting Risk

In my prior post, Simulating Actions to Outcomes: The Opportunity Envelope, I showed you a graph of the simulation result for the First Manufacturing Run case study. In this post I’ll show you how to take the next step beyond basic simulation: analyzing opportunity vs. risk. 

You may be surprised to learn that, simply, aiming for the highest attainable profit (or, more generally, making a decision based solely on maximizing the potential outcomes) may not always be the beset choice. This is because the most profitable outcome may be very “close” to outcomes that result in very large losses. “Close” means that, if the values of the model elements vary slightly (equivalently, if the values used in the model, or the relationships they participate in, are slightly inaccurate), a large profit can quickly turn into a large loss.

Decision Intelligence software can measure the risk associated with a decision by exploring the behavior of the Outcome as the decision Levers, Externals, and model Dependencies are varied slightly around the chosen decision Lever values. So, rather than simply providing you with evidence for which decision leads to the optimal outcome, the simulation graph tells a much more in-depth story. 

Specifically, if the Levers, Externals, and Dependencies are varied slightly, say within an interval of ±5% (call this the Risk Interval) of the chosen decision value, then the data generated by the simulation will show how the Outcomes also vary. This can be used to gain a deeper understanding of the risk associated with the chosen decision, or search for decisions that not only maximize the Outcomes but do so while limiting the risk of loss (or other negative outcomes).

In the product launch scenario, although we are trying to maximize profit, we also want to avoid any decisions where making a large profit comes at the risk of incurring a large loss if circumstances change only slightly. Examining the risk intervals that surround various decisions for highly profitable results provideslets you make a decision that not only leads to a favorable outcome, but balances this against whatever is determined to be an acceptable level of risk. It may be the case that the optimal risk-adjusted outcome is not the one that yields the highest possible profit, but the highest expected profit that takes all of the possible outcomes in each Risk Interval into account.

Analyzing decision outcomes using Risk Intervals gives: 

• The expected value of the profit, or loss if this is negative (this is the average of all outcomes in the Risk Interval). 
• The percentile of outcomes that show a profit above or below a certain value. These percentiles can be interpreted as a probability. For example, if 20% of all results in the interval around a decision point show a loss, it is reasonable to assume that there is a 20% chance that taking the actions associated with that decision will result in a loss. These characteristics and the way they are obtained from the Risk Interval are shown in the graph at the top of this post. 
• The difference between the maximum outcome and the expected outcome as a measure of the volatility of the decision, that is, how sensitive the outcomes are to small changes in the dependency chains that lead to them. 

Combining the opportunity and risk information and running it through Decision Intelligence software, we can create a dashboard to show the information for supporting our decision. Unlike traditional DI dashboards, this decision-centric dashboard that looks different from the data-centric one produced by our fictitious analytics team. I’ll show you the Decision-centric dashboard in my final post of this series. 

Mark Zangari

Co-Founder and CEO at Quantellia, LLC | Website

In addition to his duties as CEO, Zangari leads Quantellia LLC's Scalable Solutions division, where he is responsible for financial technology, telecom, and Covid-19 solutions. Zangari is also the architect of the company's World Modeler solution suite.  Before joining Quantellia, Zangari spent 15 years as CTO for a spatial GIS company, where he specialized in providing solutions to utilities and telecoms.

A physicist by training, Zangari's papers on cosmology are still referenced today.