What People Focus on When Reviewing Your App: An Analysis Across App Categories

Abstract

User-generated app store reviews of mobile apps provide valuable information for app developers. However, making sense of this large pool of data across many apps in a structured way remains a challenge today due to the large data volumes and because app stores lack systematic, built-in analysis tools for user reviews. We obtained and analyzed a large-scale dataset of more than 25 million sentences from user-generated reviews belonging to almost 600'000 apps across 34 categories. We provide detailed insights into users' expectations and how different aspects of mobile apps (e.g. their user interface, privacy aspects) impact star ratings. We also discuss how this differs across app categories and between free vs. paid apps. This analysis goes beyond prior work in breadth and detail: It allows developers to gain highly detailed, data-driven insights to design and improve on their apps.