Demand modeling for mobile app stores

Abstract

Smartphones have reached a relatively high market share of the mobile market, creating new market opportunities. As a result, different stakeholders are investing in the mobile industry attempting to generate a higher revenue share. Hence, competition between various mobile device manufacturers has increased, as they compete for customers. These device manufacturers have created their own ecosystems, trying to lock-in their customers. These ecosystems include the application (app) stores providing services for mobile users. Currently, the two leading app stores are the Apple App Store and Google Play. Similarly, the competition exists among app developers of both stores. Therefore, it is vital to understand the user demands to design a successful app is popular in these stores. This thesis identifies successful app categories for both app stores from the perspective of an app developer. It adopts basic descriptive analysis for the dataset provided during September and October 2013 regarding the US and Finnish markets. Furthermore, it introduces a probabilistic graphical model based on Bayesian Network, aiming to understand the dynamics of mobile app stores. The thesis defines the success indicator for each category of apps, and then compares the results of both app stores. The top successful app categories in the US market include Social Networking, Productivity, Music, Finance, Education, Sports, Entertainment, and Travel. The corresponding app categories in Finland include Social Networking, Finance, Education, Music, Productivity, Entertainment, Photos and Video, Lifestyle, Games, and News. The thesis concludes that Google Play has higher success indicators than Apple App Store both in US and Finnish markets. Additionally, the success indicator is higher for free apps compared to paid apps. The results of this research contribute to recommendations for developers, during the development and publishing stages of an app, as well as building marketing strategies for mobile apps. Furthermore, it suggests a framework to identify successful apps in mobile app stores.