A behavioral study of capacity allocation in revenue management

We investigate how human decision-makers handle the two-class capacity allocation problem—a foundational revenue management problem and a cornerstone for more sophisticated scenarios. We consider the problem with ordered and unordered arrivals, and then introduce a simplified version with single up-front decision. We explore the decisions across two price ratios, symbolizing heterogeneous vs. more homogeneous customer bases. We find that decision-makers ineffectively allocate capacity, resulting in suboptimal revenues and underutilization of RM potential. Theoretically, the decision-making format for the capacity does not matter, whereas we find that decision-makers set higher protection levels when making an up-front decision rather than sequential decisions and earn higher revenues for heterogeneous customers. Decision-makers perform poorly in sequential decision-making, and heuristics can be useful in improving performance. Moreover, we observe several regularities in decision-making. With higher customer heterogeneity, decision-makers tend to accept customers who would have been declined by the optimal policy, indicating they are less demanding. Conversely, with more homogeneous customers, they tend to turn away customers who would have been accepted by the optimal policy, implying they are more demanding. Lastly, we investigate whether the decision-making is affected by feelings of regret and find strong effects for both winner’s and loser’s experienced regret.