Latticing and Device-Histories: Dynamic Customer Profiling for Mobile Advertising Campaigns

Customer targeting for mobile advertisements is a high end exercise in big data. The universe of users is on the order of 300M while active advertising campaigns are typically several hundred in number. Matching users with the campaigns they are most likely to engage with in extreme real-time environments requires adaptive model management, advanced parallel processing hardware and software, and the integration of multiple very large databases. A key component in this overall process is the generation of a mobile device-driven customer targeting profile. We present a dynamic customer profiling technique using latticing and device-histories which maps mobile devices to specific lattices (geographic locations) and tracks user behavior via device-histories. This targeting methodology provides a distinct advantage over existing approaches by profiling on an individual, rather than a customer segment, basis which enables a much finer spatial and temporal parsing of the user marketplace.