This document discusses how Integral Ad Science helps brands ensure their marketing messages are effectively delivered to humans through digital advertising. It covers how the company detects and prevents non-human (bot) ad fraud using a combination of offline and streaming data processing. It also examines how to determine which human interaction metrics are meaningful, such as viewability versus length of ad exposure, through randomized testing and causal inference methods. The goal is to identify environments and metrics that maximize a brand's message being heard by humans.