Structured, uncertainty-driven exploration in real-world consumer choice.

Area of Science:

• Cognitive Science
• Behavioral Economics
• Computational Modeling

Background:

• Effective decision-making relies on exploration and generalization.
• Existing computational models of exploration are often limited to constrained laboratory settings.
• Generalizability of these models to complex, real-world scenarios remains unclear.

Purpose of the Study:

• To investigate the factors guiding exploratory behavior in a large-scale real-world dataset.
• To analyze customer exploration and generalization strategies using computational models.
• To determine if adaptive exploration and generalization principles apply to complex choice environments.

Main Methods:

• Analysis of a dataset comprising 195,333 customers and 1,613,967 orders from an online food delivery service.
• Application of computational models to analyze observed customer behavior.
• Examination of exploration patterns and generalization strategies within the dataset.

Main Results:

• Customers exhibit hallmarks of adaptive exploration and generalization.
• Evidence suggests customers engage in uncertainty-directed exploration.
• Feature-based generalization strategies appear to guide customer exploration.

Conclusions:

• People employ sophisticated strategies for exploration in complex, real-world environments.
• Findings support the applicability of computational models of exploration to practical choice problems.
• Customer behavior in online food delivery reflects adaptive learning and decision-making principles.