Frequent Flier Programs (FFPs) are said to impact airline consumer behaviour such that revenue of sponsoring airlines increases. Prior research relies on aggregate industry data to study FFPs. We examine the impact of FFPs on individual consumer behaviour in a quasi-natural experimental set-up using a combined discrete choice and count data model. We exploit an unanticipated change in the FFP to avoid self-selection bias. We derive the causal effect of redesigning a frequency reward program into a customer tier program on average transaction size, purchase frequency, revenues of the sponsoring airline, and compensating variation. We find that, on average, revenues increased by 8$ per member over a 16 month period. The welfare impact is small but positive. We find that, on average, consumer surplus increased by 5$ per member over a 16 month period. The results vary su bstantially across individuals. In line with previous studies, our results suggest that moderate buyers increase their average transaction size and purchase frequency most due to the introduction of the customer tier program.