Under complicated billing structures, the price to which consumers respond remains inconclusive. In this paper, I exploit a quasi-experiment to estimate a causal effect of price for residential water customers during the introduction of increasing block rates for a North Carolina utility. Perceived price is identified through a billing anomaly in which changes in marginal and average prices move in opposite directions. Empirical results contribute evidence that residential water customers respond to average price. Average price elasticity estimates vary from −0.43 to −1.14 across the distribution of consumption in triple-difference models, with an estimate of −0.31 in the tightest bandwidth of regression discontinuity specifications.