The question of why individuals choose to explore or exploit as their learning accumulates remains largely unexplored in organizational literature in a strictly causal sense. To bridge this gap, we conducted an experimental laboratory study of individual decision-making sequences using a real-effort task that involved a training phase and an active phase. In the training phase, the participants used their skills to solve the same task in eight rounds to simulate the development of individual-level learning. In the active phase, we observed sequential choices over exploring or exploiting. The participants were financially incentivized to abandon a familiar task (that they learned in the training phase) by providing higher performance-related payoffs for exploring novel task environments. Interestingly, we not only found that different kinds of performance feedback affected the exploration-exploitation choice, but that the feedback-choice linkage is contingent upon the initial conditions of the task environment in terms of its simplicity or complexity. We found that when individuals are initially exposed to simpler tasks, they are more likely to continue exploiting a familiar task; and when they are initially exposed to more complex tasks, they are more likely to explore new and more profitable tasks and then continue exploiting the new tasks they learned. These findings contribute to the literature on individual search by demonstrating the important role of initial conditions and path dependence in exploration and exploitation behavior.