In making a "drill here decision" some form of risk analysis is always performed. However, how many times do we produce a structure map with a corresponding spatially varying error estimation? Decisions are made at all steps in the chain of seismic data acquisition, processing, imaging, and interpretation that will affect the end result. Each step and decision have inherent errors. We could capture some of these decisions and errors if we wanted to, or if our funding source required us to, in order to provide error bars on our interpretive results. Take prestack migration as an example. Often an interpreter will be provided with full, near, mid, and far offset prestack migrated volumes. Rarely will structure maps on the same horizon from these four volumes be identical. The differences could be used to produce error maps. How often are such maps made? In many of our processing procedures error volumes are generated to guide workflows. Are these error volumes provided to the interpreter so that they can be used to risk a prospect? The goal of this workshop is to discuss both what is currently done, but more importantly, what could we do better to identify uncertainties in our final interpretive products.