The purchase of the right vehicles in the right quantity and the matching of potential customers with the most convenient available vehicles. As for every data intensive project, using state of the art machine learning techniques is very tempting. Unfortunately, in theory this is often easier said than done. Even though we have the necessary data infrastructure and machine learning expertise, the amount of data collected over the two years of Cluno existence and its quality prevents us from leveraging the fanciest techniques. Under these conditions, our Data Science team had to get extra creative in order to distill as much intelligence as possible from the existing data and to support our different departments enabling Cluno to be one of the car subscription leaders it is today – without losing ourselves in the depths of the most complex machine learning algorithms.