Online consumer ratings provide important feedback for businesses and yield essential purchase information for consumers. Extant literature has recognized the importance of sequential and temporal dynamics of consumer ratings, but has shed light upon short-term dynamics (e.g., an initial decreasing rating trend) and lacks analyses of long-term dynamics. Existing findings thus cannot explain these long-term dynamics, which are particularly important as many items receive ratings over the long term. In this paper, we therefore examine long-term sequential and temporal dynamics in consumer ratings and in particular whether initial rating dynamics influence average ratings in the long-term. To do so, we apply regression models to an extensive long-term review dataset. First, we find and explain a new long-term sequentially increasing rating trend which leads to a U-shaped relationship between ratings and their order. Second, we reveal that strong initial rating dynamics have significant negative impact on long-term average ratings.