While counting cars in parking lots is not a new concept, the extent of such an operation for more than a single retail location was beyond the scope of feasibility just a few years ago. Information that once required thousands of labour hours to acquire can now be derived from a single dataset of satellite imagery. With advances in machine learning technology, and the availability of 30 cm resolution imagery, cars can now be detected by satellites 400 km above them. With this identification comes the ability to track and model the parking lot traffic of massive retailers. Fluctuations in this traffic appear to correlate accurately with store revenue. Access to this kind of information can help you make more guided decisions in your strategy for the upcoming quarter and beyond.