There is increasing interest in approaches that leverage “real world evidence” in clinical studies, especially for regulatory decision making. FDA Commissioner Rob Califf has written on the topic multiple times in the FDAVoice blog, and presented the topic at external meetings. Some view “real world evidence” as reflecting data that has already been collected, but the term can also apply to data that is generated during the course of usual health care. In this presentation I will discuss three pertinent aspects of real world data: generation, analysis, and interpretation. Generation of such data can be unsupervised (in which case access and harvesting are important issues) or directed (so that data is generated more from the perspective of designed experiments). For analysis, interesting issues concern re-use of data and linking data from different sources. Interpretation of real world data involves the nature of conclusions that can be drawn from non-research grade information, and accommodating changing transparency requirements to facilitate validation of conclusions. All three aspects of generation, analysis, and interpretation are of interest to statisticians as we grapple with the challenges and complexities of working in this fast evolving space.