Measuring and Maximising Clarity
25 Apr 2025
12:00 -13:30
Times are shown in local time.
Open to: All
Room W2.02 (Cambridge Judge Business School)
Trumpington St
Cambridge
CB2 1AG
United Kingdom
Join our Marketing seminar
Speaker: Dr Stefan Bucher, Faculty of Economics, University of Cambridge
About the seminar topic
Information design powered by machine learning bears the potential of improving human decisions by making choosers’ cognitive constraints less binding. A wide range of important applications, from consumer search to mandated disclosures, call for the maximisation of objectives that are not directly observable in behaviour, such as the comprehensibility of textual information. Bringing machine learning to bear on such objectives thus requires innovative behavioural measurements that can reveal them, promising better-informed decisions as a result.
In this talk, I will develop such methods for measuring and maximising clarity. I will first introduce an algorithmic approach to choice architecture that combines reinforcement learning with a cognitive model of choice behaviour. Specifically, I develop a rational-inattention model of multi-attribute choice to describe the behaviour of a consumer facing information costs. I then use reinforcement learning to solve the problem of a sender nudging or persuading the consumer by tailoring information to their revealed preferences. I experimentally show that algorithmic information filtering significantly enhances subjects’ choice accuracy.
I will conclude by illustrating how this framework can be extended to less structured data, discussing an experiment designed to measure and improve the comprehensibility of privacy disclosures. This underscores the broad relevance of the approach in designing information that is digestible and effective in inducing desirable behavioural outcomes.
Speaker bio
Stefan Bucher is an Assistant Professor at the University of Cambridge’s Faculty of Economics. His research examines cognitive frictions in economic choice behaviour and seeks to reduce information overload by developing algorithmic tools for choice architecture. His work integrates methods from behavioural and information economics, cognitive science and machine learning.
Following his PhD at New York University, Stefan was a postdoctoral scholar at the Max Planck Institute for Biological Cybernetics and the Tübingen AI Center, as well as the MIT Sloan marketing group. He holds a BSc and MSc in Computational Science and Engineering from ETH Zürich and a MSc in Economics from the London School of Economics and Political Science.
Register
No registration required. If you have any questions about this seminar, please email Luke Slater.