Early Signals of Algorithmic Content Spread on TikTok: Predicting Reach Through View Trajectories

Analyzed high-frequency panel data collected at 10-minute intervals from 950 TikTok videos published by 86 fashion and apparel brands to investigate the relationship between algorithmic content distribution and post performance. Applied data analysis and clustering techniques to identify three distinct reach archetypes and evaluated how early view trajectories could predict long-term content success. Developed predictive models demonstrating that eventual high-reach posts can be identified within the first two hours after publication, significantly outperforming follower count as a performance indicator. The findings provide marketers with practical tools for optimizing content promotion, reallocating advertising resources, and improving decision-making in algorithmically curated social media environments.