Household-specific feedback on the microbiological safety of drinking water may result in changes to water management practices that reduce exposure risks. We conducted a randomized, controlled trial in India to determine if information on household drinking water quality could change behavior and improve microbiological quality as indicated by Escherichia coli counts. We randomly assigned 589 participating households to one of three arms: (1) a messaging-only arm receiving messaging on safe water management ( n = 237); (2) a standard testing arm receiving the same messaging plus laboratory E. coli testing results specific to that household's drinking water ( n = 173); and (3) a test kit arm receiving messaging plus low-cost E. coli tests that could be used at the household's discretion ( n = 179). Self-reported water treatment increased significantly in both the standard testing arm and the test kit arm between baseline and follow-up one month later. Mean log10 E. coli counts per 100 mL in household stored drinking water increased in the messaging-only arm from 1.42 to 1.87, while decreasing in the standard testing arm (1.38 to 0.89, 65% relative reduction) and the test kit arm (1.08 to 0.65, 76% relative reduction). Findings indicate that household-specific water quality information can improve both behaviors and drinking water quality.