In various embodiments, strategies, tools and techniques are provided for processing and analyzing data in an advertising measurement system accessible through a software-as-a-service model or a client-downloaded computer program. The system may be configured to receive and process household media exposure data, product purchase data, advertising data, program data, and demographic data, wherein the data may be obtained from various sources, including from a program delivery source (e.g., a television set-top box) located in a household of a consumer. The data may be matched and cross-correlated for calculating return on advertising investment measurements and other metrics. The data may be used to optimize media placement generally or to address advertising content to specific households. Methods for storing and accessing data in the advertising measurement system by distributing the data across multiple shards are also provided. In addition, methods for protecting privacy of communicated data are disclosed.