This study extends earlier work by [Kahneman, D., and Snell, J. (1992). Predicting a changing taste: Do people know what they will like? Journal of Behavioral Decision Making, 5, 187–200.]. suggesting that people are poor at predicting changes in liking. This is an important issue because an absence of this ability would make it difficult for people to optimize their own choices. Twenty undergraduates and 20 of their parents sampled four relatively unfamiliar consumer products, two foods and two body products, for 8 days. On Day 1, participants rated their initial liking and predicted their liking after seven daily uses of the products. Predictions were compared to actual liking on Day 8. Consistent with prior work, participants were poor at predicting their actual hedonic trajectories because they underestimated the degree to which their preferences would change. Contrary to predictions, parents were no better than students at this task, even though they had some 20–39 years more experience in observing their own hedonic trajectories. There is no evidence for any parent–child resemblance in either liking for the products or ability to accurately predict hedonic trajectory, and no evidence for consistency in ability to predict trajectories across the four different products. In general, participants underestimate the degree to which their preferences will change.