GPT-Driven Value-Based Pricing Research: Comparing Synthetic Data versus Human Survey Results
(A2024-119685)

AUTHORS

Anett Erdmann, ESIC University, ESIC Business & Marketing School; José Ramos-Henriquez, Universidad de la Laguna, Instituto Universitario de la Empresa

ABSTRACT

Effectively managing prices relies on understanding consumer perspectives, including their psychological price thresholds and the necessary price stimuli to trigger demand responses. Traditional methods like surveys and experiments are costly and time-consuming. This study explores using ChatGPT (Chat Generative Pre-trained Transformer) for effective value-based price research, leveraging the power of synthetic data. We analyze acceptable price ranges (Van Westendorp price Sensitivity) and necessary price stimuli (Weber-Fechner law) for four consumer technology products, differentiating by expenditure categories, product lifecycles, and consumer segments. Our GPT-based experiment, coded in R using the OpenAI API and samples of 1000 cases is statistically compared with conducted human surveys. GPT-driven value-based pricing research provides realistic results, allowing for price segmentation and price adjustments, optimizing resources and potentially challenge current practice.