Atom probe tomography is often introduced as providing atomic-scale mapping of the composition of materials and as such is often exploited to analyse atomic neighbourhoods within a material. Yet quantifying the actual spatial performance of the technique in a general case remains challenging, as they depend on the material system being investigated as well as on the specimens geometry. Here, by using comparisons with field-ion microscopy experiments and field-ion imaging and field evaporation simulations, we provide the basis for a critical reflection on the spatial performance of atom probe tomography in the analysis of pure metals, low alloyed systems and concentrated solid solutions (i.e. akin to high-entropy alloys). The spatial resolution imposes strong limitations on the possible interpretation of measured atomic neighbourhoods, and directional neighbourhood analyses restricted to the depth are expected to be more robust. We hope this work gets the community to reflect on its practices, in the same way, it got us to reflect on our work.