Purpose – We present a simulation study using a model of a new optically pumped magnetometer sensor for the application in the field of magnetoencephalography. The effects of sensor distance and orientation on the measurement information and the sensitivity to neuronal sources are investigated. Further, we use a combinatorial optimization approach for sensor placement to measure spontaneous activity in the region of the occipital cortex. Design/methodology/approach – We study the effects of sensor distance and orientation on the sensitivity to cortical sources and the measurement information. A three-compartment model of the head using the boundary element method is applied. For sensor setup optimization, a combinatorial optimization scheme is developed. Findings – The sensor distance to sources considerably affects the sensitivity and the retrieved information. A specific arrangement of four sensors for measuring spontaneous activity over the occipital part of the head is optimized by effectively avoiding position conflicts. Research limitations/implications – Individual head models as well as more detailed noise and signal models will increase the significance for specific use cases in future studies. Originality/value – Effects of sensor distance and orientation are specifically evaluated for a new optically pumped magnetometer. A discrete optimization scheme for sensor optimization is introduced. The presented methodology is applicable for other sensor characterisation and optimization problems. Our findings are helpful for the development of new sensors.