Diminishing Returns (Concept)

Diminishing returns refers to the case where the introduction of an additional unit of resource results in a smaller increase in output. For example, a single nurse is able to screen 20 patients per hour. However, two nurses will not be able to screen 40 patients each hour (e.g., waiting for the other nurse to finish using the blood pressure monitor).

This is also observed when constructing prediction models (e.g., doubling the amount of data does not double the model accuracy). In practice, the desired prediction model should take into account the cost of acquiring data (e.g., automated or manual) and the decisions to be made (e.g., no point having a highly accurate prediction when orders are placed in lot sizes of 100). This is discussed in IE3120 Manufacturing Logistics, IE4213 Learning from Data and IE5107 Material Flow Systems.

Diminishing returns is also discussed in IE4243 Decision Modeling & Risk Analysis where a $1,000 reward is likely to result in greater joy to a poor individual rather than a wealthy individual. This is commonly referred to as diminishing marginal utility and is consistent with risk-averse decision making behavior.