This study explores the evolution of guest experiences on Airbnb by analyzing repeated reviews using sentiment analysis, structural topic modeling, and linguistic characteristics examination. Drawing from over 57,000 repeated reviews across eight U.S. cities, the research reveals that guests’ reviews become more concise and emotionally moderate over time. Rather than signaling dissatisfaction, these changes suggest stabilized satisfaction and growing loyalty. To systematically analyze these changes, this research employs Systemic Functional Linguistics (SFL) as a theoretical framework, conceptualizing language as a means of communication through text. Specifically, the study focuses on three linguistic dimensions—ideational, interpersonal, and textual—to investigate how review content evolves. These findings align with Expectation–Confirmation Theory and Psychological Contract Theory, emphasizing the importance of first impressions and consistent service delivery, while offering insights for hosts and platforms aiming to strengthen long-term guest relationships. The study contributes to understanding guest behavior dynamics in peer-to-peer accommodation platforms.