In this note we analyze the sustainability of collusion in a game of repeated interaction where firms can price discriminate among consumers based on two types of customer data. This work is related to Liu and Serfes (2007) and Sapi and Suleymanova (2013). Following Sapi and Suleymanova we assume that consumers are differentiated both with respect to their addresses and transportation cost parameters (flexibility). While firms have perfect data on consumer addresses, data on their flexibility is imperfect. We use three collusive schemes to analyze the impact of the improvement in the quality of customer flexibility data on the incentives to collude. In contrast to Liu and Serfes in our model it is the customer flexibility data which is imperfect and not the data on consumer addresses. However, our results support their findings that with the improvement in data quality it is more difficult to sustain collusion.