To improve problem-solving performance, individuals can rely on social learning. This approach is constrained by an individual's social network, which influences the efficiency of the problem-solving process. To date, research disagrees on what kind of network structure is preferable, providing support for efficient network structures, as well as for inefficient networks. However, studies implicitly assume that solvers always imitate superior solutions, an assumption that lacks empirical grounding. We propose a simple derivation of an existing simulation framework by incorporating a known cognitive bias (‘IKEA effect’), whereby individuals are assumed to prioritize individual information. This effect allows inefficiencies to be embodied at the individual microlevel, reducing the need for inefficiencies at the structural macrolevel. Simulation results explain discrepancies in previous results, illustrating how more realistic microlevel assumptions substantially impact macrolevel outcomes.