We have been exploring an O2O (online-to-offline) service that leverages an electronic coupon on a smartphone and digital signage on a vehicle as a form of push media. In the O2O service, the digital signage on a vehicle, such as bus or train, stimulates users to obtain the coupon by advertising it in the vehicle, and drives conversion in the store, i.e., encourages users to visit the store and purchase the goods the coupon is promoting. In this service scenario, it is necessary to estimate the purchase conversion in order to quantitatively evaluate how effective the service is. Thus, in this paper, we propose a purchase conversion estimation method using cellular radio characteristics. The challenge is to distinguish correct coupon usage in the store from incorrect usage in other places such as the home. The method assumes that the percentage of correct coupon usages in the store is much higher than that in other places. Base on this assumption, the proposed method, which relies on a server, gathers the cellular radio characteristics such as base station ID and signal strength at the time of coupon usage, clusters the radio characteristics, and finally estimates the coupon usages belonging to a cluster whose size is large as the purchase conversion. The conversion estimation method is independent of the phone settings such as switch-off in the GPS module. In addition, the proposed method does not require both the introduction of dedicated infrastructure or prior surveys of the purchase site. Based on the preliminary evaluation, we found that the method works well if the location of incorrect usage of the coupon is more than 500 m away from the store.
展开▼
