Yiangos Papanastasiou

 

Faculty Speaker Yiangos Papanastasiou [London Business School]
Title Crowdsourcing Exploration
Date & Time Friday, February 27, 2015 at 10:30am
Place UCLA Anderson School of Management 
Room D-310 

 

Abstract
In an online review platform, consumers observe information regarding the past experiences of their peers, before choosing between alternative products or services whose quality is uncertain. In turn, their choices result in the generation of new experiences which they then report back to the platform. We investigate the problem of designing optimal information-provision policies, when the platform's goal is to maximize aggregate consumer surplus. We develop a decentralized multi-armed bandit framework, where a forward- looking principal (the platform designer) commits upfront to a policy which dynamically discloses information regarding the history of outcomes to a series of short-lived rational agents (the consumers). We demonstrate that the platform's performance is non-monotone in the accuracy of its information-provision policy. Because consumers are constantly in "exploitation" mode, policies that disclose full and accurate information on past outcomes suffer from inadequate "exploration." We find that the designer can (partially) alleviate this inefficiency by employing a policy that strategically obfuscates the information at the platform's possession; interestingly, such policies are beneficial despite the fact that consumers are aware of both the platform's objective and the precise way by which information is being disclosed to them. The problem of explicitly designing optimal information-provision policies is found to be generally intractable; nevertheless, we establish the main structural properties of optimal policies and use them to propose an effective and computationally-tractable heuristic approach.

Yiangos Papanastasiou