Data from 929 product tests inConsumer Reportsbetween January 1975 and November 1984 were used to estimate the potential payoffs to consumer search. A “best choice’ and a “worst choice’ were identified in each product test using three different selection methods. In methods 1 and 2, the worst choice was qualitatively inferior to the best choice. However, the method 1 worst choice was the most expensive lower‐quality option, while in method 2 it was the least expensive alternative. In method 3, both the best and worst choices were selected randomly without regard to quality rankings or price. When the worst choice in a product test was more costly than the corresponding best choice, the price difference (in percent) was computed to estimate the returns to search in 12 product classes. Implications for consumers and consumer educators are also i