A system and method for efficiently executing a plurality of live parallel Internet content usage experiments using a large pool of randomly selected participating users (separated into control and test groups) while minimally disturbing offered content is provided. In an illustrative implementation, a content provider communicates the parameters for a content usage experiment to a content server. A complimentary test subject verification script operates on the content server that determines which client computing devices are to be included in the experiment group. In operation, a participating client computing device requests content from the content provider's server. The content provider's server executes the test subject verification script to determine if the client computing device is to be considered as a test subject. The content usage of the experiment content and non-experiment content is tracked and compared to each other to determine if the experiment content impacted content usage.