Harvard Business Review LogoMay 4, 2026Gandee Vasan/Getty ImagesResearch into organizational learning suggests companies often squander the benefits of diverse talent not because of who they hire, but because of how employees learn from one another. Drawing on computational modelingIn 2012, journalist Kurt Eichenwald published a devastating exposé in Vanity Fair. Microsoft, he reported, had become a place where talented people fought each other instead of the competition. The culprit was stack ranking. Every review cycle, managers sorted employees into a fixed curve of winners and losers. Rankings were updated once a year. Between reviews, everyone knew who the previous cycle’s winners were and copied them.
Stop Trying to Replicate a Single Star Performer
Research into organizational learning suggests companies often squander the benefits of diverse talent not because of who they hire, but because of how employees learn from one another. Drawing on computational modeling inspired by James March’s seminal work, the authors show that traditional performance systems—such as Microsoft’s former stack ranking—create two hidden problems: employees over-copy established stars (“overshooting”) and rely on outdated signals of who is performing best (“frozen targets”). These dynamics prevent the recombination of distinct knowledge into better solutions. The fix is to replace fixed rankings with continuously updated “live targets,” allowing new high performers—often underdogs who reach the frontier via different paths—to become visible and influential immediately. Combined with a “mind-the-gap” rule, where those far behind learn quickly and those near the frontier learn selectively, this approach accelerates innovation on complex problems. The implication: competitive advantage depends less on assembling diverse talent than on designing systems that continuously surface and recombine it.
105 words~1 min read
