Demand estimation and algorithmic price recommendations
The paper “Demand Estimation Using Managerial Responses to Automated Price Recommendations” by Alexander K. Wagner (PLUS) and colleagues Daniel Garcia (U Vienna) and Juha Tolvanen (U Tor Vergata Rome) has been published in the November 2022 issue of Management Science, one of the leading journals in Management.
The paper addresses the fundamental question of how to learn about the impact of a firm’s prices on its own sales in a setting where prices were influenced by algorithmic recommendations. The authors come up with a novel way of estimating price elasticities, that is, by how much hotel customers change their purchases in response to a price change. The machine-learning based method for causal inference uses price recommendations and the time it takes for hotel managers to implement recommendations. Although the paper focusses on dynamic pricing of hotel rooms in a revenue management context, the approach can be applied in other dynamic decision-making problems where a human decision maker has access to algorithmic recommendations but responds to recommendations with delay, e.g. in staffing or re-stocking problems.
The full article can be read here: https://doi.org/10.1287/mnsc.2021.4261