Unwind the clock? Temporal distance and user interactions on a digital platform
利用StackOverflow数据，研究发现时区差异导致的时间距离显著减少跨区域用户互动，且对冷门知识社区影响更大，平台可通过调整内容呈现顺序来缓解这一障碍。
Abstract Research Summary Digital platforms provide arenas for global knowledge diffusion, but their underlying architecture often relies on synchronous exchange, inadvertently siloing users into distinct “time pockets” based on time zones. Using proprietary data from StackOverflow, we implement a regression discontinuity design to causally estimate that a 1‐h increase (decrease) in the temporal distance between two regions leads to a 13.6% decrease (9.5% increase) in views and a 20.9% decrease (21.0% increase) in votes between those regions. These temporal frictions disproportionately penalize interactions in niche knowledge communities (e.g., sudo, slack‐api) over those in popular ones (e.g., Python, JavaScript). Importantly, we show that a platform can mitigate temporal barriers by shuffling content representation. This paper contributes to our understanding of platform strategy, temporal distance, and global knowledge diffusion. Managerial Summary While digital platforms deliver global connectivity, time zone differences create invisible barriers that stifle user interactions and knowledge exchange. Our research shows that temporal distance between regions on StackOverflow, a global knowledge community for computer programming, significantly reduces cross‐region views and votes. These temporal silos disproportionately harm niche communities, where valuable content is less likely to be shared among users who are not online simultaneously. Popular communities, however, remain largely unaffected. For platform designers, relying solely on chronological feeds inadvertently fragments their global user base. To unlock cross‐border exchange and support specialized communities, managers could adjust algorithms to “shuffle” content representation based on dynamic triggers rather than just the posting time. Doing so bridges temporal gaps and connects out‐of‐sync users.