We rely heavily on factorisation formulae to make useful predictions at colliders such as the LHC. But when do such formulae work, and why? When do they not work? I will discuss this issue in the context of the production of a colour singlet V at hadronic colliders. Observables discussed include the total cross section for the production of V, transverse momentum of V (including spin-dependent observables), and hadronic transverse energy E_T accompanying V. Based on arXiv:1405.2080 and arXiv:1709.04935.