This PhD project investigates the pricing of full-service maintenance contracts, which cover all preventive and corrective maintenance activities over a predefined contract horizon for a fixed, upfront agreed price. Such contracts play a central role in industrial service operations, yet their pricing remains complex due to uncertainty in future maintenance demand and heterogeneity in customer preferences.
Existing pricing approaches typically follow either a cost-plus strategy, based on predictions of future maintenance costs, or a value-based strategy, reflecting customers’ willingness to pay. While both approaches are well established, they do not explicitly quantify the causal effect of pricing decisions on customer acceptance.
The core objective of this PhD project is to develop and apply causal machine learning methods to estimate the probability that a customer accepts a given price offer. Causal machine learning integrates ideas from econometrics and machine learning to estimate treatment effects while controlling for confounding factors. In contrast to purely predictive models, these methods allow researchers and practitioners to reason about counterfactual outcomes, such as how acceptance probabilities would change under alternative pricing strategies, making them particularly suitable for pricing and operational decision-making problems.
Empirically, the project will use a rich dataset consisting of historical price quotations combined with detailed information on customers, machines, contract characteristics, and other relevant contextual features. Based on these predictors, the PhD candidate will develop causal models to estimate acceptance probabilities for alternative price offers and to generate insights into the drivers of customer decision-making in service contracts.
The project is situated in the field of Operations Management, with strong links to service operations, pricing, and data-driven decision support. The PhD candidate will gain advanced methodological training in causal inference and machine learning, while working on a problem with clear managerial relevance.
