Abstract: This Essay proposes a new method of monitoring regulatory compliance by a firm that operates multiple sources of risk, such as air polluting smokestacks. The expense of individually monitoring such sources may consume a large share of the agency's enforcement budget, undermining deterrence objectives. Under our approach, regulators would instead randomly select one of the firm's sources of risk, determine the firm's liability at that source, and apply that outcome perforce as determinative of liability at all of the sources. This method, which we call single-outcome sampling (“SOS”), replicates or improves deterrence generated by the current source-by-source enforcement model, but at a fraction of the cost. To demonstrate these benefits, we apply SOS to the EPA's monitoring of compliance with Clean Air Act regulations. We also address potential risk-bearing and judgment-proof costs associated with our proposal and explain how both problems can be solved.