How Netflix’s Recommendation Algorithms Function in Small Markets – The Case of Serbia

This study examines the functionality of Netflix's recommendation algorithms in smaller markets, focusing on Serbia. Through a reverse-engineering experiment involving user profiles with diverse viewing habits, the research highlights the mechanisms of algorithmic personalization and its limitations. The findings reveal that while Netflix algorithms effectively adapt to individual preferences, they rely heavily on global trends and widely consumed content, often struggling with niche or regional preferences. The study further explores how algorithms manipulate user behavior by promoting certain content through tailored visuals and cognitive strategies, and it discusses the challenges of personalization in markets with limited local content.