How reliable does a delivery drone have to be?

The advent of autonomous drones has initiated serious discussions about safety, accuracy, and reliability. Teams around the world are working on systems that enable unmanned flying delivery systems that can deliver the morning newspaper, your morning coffee, or packages from your favorite retailer directly to your current location. If you were assigned to one of these teams as the reliability engineer and asked to make sure it was reliable enough, what would you do? There are no standards, no similar products, nor history. How would you establish a reliability goal? In this paper, we will explore this hypothetical case and step though the logic that would allow you to draft a meaningful reliability goal. We will explore the combination of return on investment, repair and replacement costs, and the impact of failures. Making assumptions that can be measured and refined allow us to draft a framework to create a viable set of targets for guiding system development, checking product verification, and achieving market acceptance.