Research summary: In this study, we revisit the relationship between corporate social performance (CSP) and corporate financial performance (CFP) by conducting a replication of Waddock and Graves (1997). Using 1990 KLD ratings as the CSP measure, the original study reports a positive bidirectional relationship between CSP and CFP. However, our replication analyses with a larger sample over a longer time period indicate that the findings of the original study may not be generalizable to different samples. We argue that our replication casts doubt on the original study and can serve as a starting point to reconsider the CSP‐CFP relationship. Based on the findings of our replication, we discuss the differences between the replication results and the original findings, and then suggest several approaches to revise and extend the original study. Managerial summary: Advocates of corporate social performance (CSP) have long argued that “doing good leads to doing well.” However, the evidence to support this argument is not strongly convincing, and managers hence doubt whether better CSP leads to improved corporate financial performance (CFP). In this article, we directly examine the relationship between CSP and CFP. Our article reports that CSP may not have a positive influence on CFP. Instead, our article shows the complexity of the relationship between CSP and CFP. Therefore, we cannot simply argue that doing good will necessarily lead to doing well. Copyright © 2016 John Wiley & Sons, Ltd.