Prior research has indicated that frequent feedback could reduce residential electricity consumption by 10% to 15%. However, because feedback was primarily given in written form, this procedure might not be practical. The present study evaluated a potentially more practical feedback procedure during peak-use periods with high electricity consuming households. The study was conducted during the winter in an upper-middle class neighborhood of almost identical, all-electric townhouses (N = 71) that averaged about 170 KWH per day per household for a monthly bill of over $200. Twelve households received daily written feedback. Sixteen households (self-monitoring) were taught to read their outdoor electricity meter and to record KWH used every day. A comparison group was composed of 14 households that had volunteered to participate and 29 others that had only given permission to have their meters read. During a 1-month period that the procedures were in effect, the feedback group reduced consumption by 13% and the self-monitoring group by about 7%. These reductions, relative to the comparison group, were maintained during an early spring 1-month follow-up period and, to a lesser extent, during a 6-week warm spring period. Self-monitoring participants were highly reliable and persistent meter readers. Reductions in electricity use were reported by households to be largely attributable to lowering of the heat thermostat, and large monetary and KWH savings were found. Techniques to make self-monitoring cost-effective important components of the self-monitoring procedure, methods to apply self-monitoring more broadly, and plans to combine behavioral procedures with physical technology are discussed.