The spatial distribution of Milky Way (MW) subhaloes provides an important set of observables for testing cosmological models. These include the radial distribution of luminous satellites, planar configurations, and the abundance of dark subhaloes whose existence or absence is key to distinguishing amongst dark matter models. We use the COCO $N$-body simulations of cold dark matter (CDM) and 3.3keV thermal relic warm dark matter (WDM) to predict the satellite spatial distribution. We demonstrate that the radial distributions of CDM and 3.3keV-WDM luminous satellites are identical if the minimum pre-infall halo mass to form a galaxy is $>10^{8.5}$$mathrm{M}_{odot}$ The distribution of dark subhaloes is significantly more concentrated in WDM due to the absence of low mass, recently accreted substructures that typically inhabit the outer parts of a MW halo in CDM. We show that subhaloes of mass $[10^{7},10^{8}]$$mathrm{M}_{odot}$ and within 30kpc of the centre are the stripped remnants of larger haloes in both models. Therefore their abundance in WDM is $3times$ higher than one would anticipate from the overall WDM subhalo population. We estimate that differences between CDM and WDM concentration--mass relations can be probed for subhalo--stream impact parameters $<2$kpc. Finally, we find that the impact of WDM on planes of satellites is likely negligible. Precise predictions will require further work with high resolution, self-consistent hydrodynamical simulations.