Matching Mobile Applications for Cross-Promotion

Abstract

As the mobile app market grows rapidly, with millions of apps and billions of users, search costs are increasing tremendously. Similar to the case of recommender systems, the challenge is how apps can be recommended to the right users and how consumers can find the right apps. This paper studies a new mobile app ad framework, cross-promotion (CP), which is to promote new “target” apps within other “source” apps. With unique random matching experiment data, we empirically test the important determinants of ad effectiveness. We then propose a machine-learning-based framework to optimally match source apps to target apps to improve ad effectiveness in terms of app downloads and postdownload usages. The simulation results show that app analytics capability is essential in building accurate prediction models and in increasing ad effectiveness of CP campaigns and that, at the expense of privacy, individual user data can further improve the matching performance. The paper has important managerial implications because it provides direct guidance to better utilize CP for app developers and to leverage data analytics and machine-learning models for platform managers. It also provides policy implications on the trade-off between utility and privacy in the growing data economy.