Most estimates of the free rider fraction are based on ex-post surveys of program participants. Program participants who indicate that they would have made the supported changes without the program are labeled as ＇＇free riders.＇＇ This paper provides an estimate of the free rider fraction based on consumer actions. A set of energy-use clusters acts as a base against which the likely behavior of consumers in the absence of an efficiency program can be assessed. The clusters can also be used target spending at customers who are least likely to invest in efficiency on their own. The Actions-Based estimate does not suffer from the biases implicit in the standard ex-post survey estimates.