Over the holidays, I visited my sister and brother-in-law. While we were chatting, he mentioned he wanted to refurbish some audio equipment and was trying to decide between fixing the electrical components himself and hiring an expert to do it for him. He said he was concerned with the potential cost and quality of the repairs, as well as how long it would take to complete the repairs, but he was uncertain about the range of values these three factors could take under the two alternatives. I must have been smiling from ear to ear. Without even realizing it, he’d described a simple, yet interesting multi-objective decision under uncertainty, and I wanted to show him how I could help him with his decision situation.
I pulled out a sheet of paper and quickly drew a rough influence diagram to represent his decision, the three uncertainties, and the resulting outcome “node” of his confidence or pride in the refurbished system. Then I asked him about the range of costs the components could take and the range of time and quality the repairs could take. After drafting probability distributions for the three factors, we chatted about his value functions and weights for the three factors and assigned scores and probabilities to the two alternatives across all three factors. I transferred all of this information, which only took a few minutes to collect, into a spreadsheet and quickly churned out preliminary certainty equivalents for the two alternatives. After performing a quick sensitivity analysis on some of the model’s inputs, including the weights and multi-attribute risk tolerance, we determined we had robust certainty equivalents for the alternatives.
My brother-in-law was pleased with the whole exercise. It helped him frame and understand the decision situation. By breaking the problem down into objectives, uncertainties, scores, and values, he could clearly see how changes in those factors would affect which alternative was preferred. Best of all, in this case, the analysis was quick and free…unless you count the cost of the sandwiches we ate while we worked on the problem. I enjoyed conducting this decision analysis with my brother-in-law and was very pleased that he thought it had been a valuable use of his lunchtime.
Do you have a similar story? Have you ever helped a friend or family member with an interesting decision problem? Are you working on an interesting problem at work? When you get a chance, please take a few minutes to write about and share your experience. You can think of it as your chance to tell the next “You might be a decision analyst if…” story.