Product Development and Platform Fees Design

Abstract

Digital platforms often adopt a 30% commission rate on all transactions as the rule-of-thumb rate. Such commission scheme has raised public and antitrust concerns in recent years, fearing that a 30% rate hurts innovation incentives of software developers. Nevertheless, empirical evidence of on this topic remains limited. Using a unique dataset of product development in a public beta testing program at a leading video game platform, we study how commission schemes would change the product update decisions and the full-release decisions of developers. We use pre-trained language models and supervised machine learning algorithms to classify updates into Feature, Regular, and Bug fix updates, and document positive correlations between past product updates and consumer demand and between past demand and updates and current release decisions. Motivated by these facts, we construct and estimate a structural model of developer and consumer behaviors on platforms and find substantial costs for product development. The counterfactual analysis shows that lowering commissions can lead to more updates, fewer games will be released, leading to an over-investment in beta testing. The platform revenue will decrease as the increase in future consumer demand driven by innovation cannot compensate for loss in commissions.