Neutron imaging had a breakthrough when electronic camera detectors became available 30 years ago, which allowed for digital image processing, but also for sacrificing intensity for higher resolution and sharper images. Detectors have become so sensitive that neutron computed tomography can be performed at 500 W (!) reactor power, and installations are now possible at low-power research reactors that were previously regarded as providing insufficient neutron flux, but such requires an optimal design of the neutron beam for imaging. This article provides an overview about lessons learned from experiences designing many neutron beam lines and facilities for imaging. It will not save a new user from looking up many references, but it highlights hidden but crucial details that are essential for a well-performing neutron imaging facility.