Despite the clear theoretical link between promotions and job performance, the few studies that have tested this relationship have instead found that the role of job performance level in determining promotions is much less than might be expected. In 4 studies, we propose and test a different way of thinking about the performance-promotion relationship. Prospect theory, spiraling theory, and sponsored and contest mobility were used to support the notion that change in performance is at least as important to the prediction of promotion decisions as is absolute level of performance. In Study 1, performance and promotion data were collected for 563 white-collar employees at each of 4 time points spread over 6 years. As hypothesized, change in job performance significantly predicted change in workers' hierarchical level (i.e., promotion) beyond previous performance level. In Study 2, we found that upward trends are associated with ratings of future performance expectation and promotability through their effects on attributions of conscientiousness, proactive personality, and job dedication. In Study 3, we replicated the findings of Study 2 and found no evidence of a performance trend by performance level interaction. In Study 4, we showed that those with upward trends are preferred even to those with performance that is consistent and strong. Taken together, our results suggest that the reason for the modest performance-promotion relationships found in previous research may be that performance trends are seen by decision makers as containing at least as much promotion-relevant information as do performance averages or recent performance levels. (PsycInfo Database Record (c) 2021 APA, all rights reserved).