The impact of personnel turnover on an organization's ability to learn, and hence on its ultimate performance, is explored for organizations with different structures and different tasks. A model of organizational decision making is presented where: (1) the organization is faced with a continuous sequence of similar but not identical problems; (2) each problem is so complex that no one person has access to all of the information nor the skill to comprehend all of the information necessary to make the decision; (3) individual decision makers base their decisions on their own previous experience; and (4) there is personnel turnover. Using simulation the impact of turnover on the rate and level of learning for hierarchies and teams is examined. This research suggests that while teams in general learn faster and better than hierarchies, hierarchies are less affected by high turnover rates particularly when the task is nondecomposable. Institutionalized memory, as embodied in the memories of distributed individuals and in the advisory relationships between individuals, determines the consequences of personnel turnover.