Abstract: An invention is disclosed for scalable billing usage data storage. In an embodiment, users are billed on a monthly cycle. The data storage is divided into two partitions. First, the data is divided into an “L1” partition for each day of the billing cycle. Within each L1 partition, data is further divided into “L2” partitions based on a customer's subscription ID. Through such partitioning, each customer's billing usage data for a billing cycle is entirely contained within a single L2 partition, and that data may be retrieved without spanning partitions.