Understanding the impact of video quality on user engagement

摘要: As the distribution of video over Internet becomes main- stream and its consumption moves from computer to TV screen, user expectation for high quality is constantly increasing. In this context, it crucial content providers understand if how affects engagement best invest their resources optimize quality. This paper a first step towards addressing these questions. We use unique dataset that spans different types, including short on demand (VoD), long VoD, live popular con- tent providers. Using client-side instrumentation, we measure metrics such as join time, buffering ratio, average bitrate, rendering quality, rate events.We quantify both at per-video (or view) level per-user viewer) level. particular, find percentage time spent in (buffering ratio) has largest impact across all types content. However, magnitude depends type, with being most impacted. For example, 1% increase ratio can reduce by more than three minutes 90-minute event. also see bitrate plays significantly important role case VoD