An examination is conducted on a commonly used procedure for extracting (un)polarized γn → π − p and γn → π 0 n observables from d(γ, π − )pp and d(γ, π 0 )pn data, using a model that consists of the impulse term and the final-state interaction (FSI) terms due to nucleon-and pion-exchange. Recent experimental and theoretical analyses used an extraction method that does not impose a cut on the final πN invariant mass W . I demonstrate that the use of this method can result in the γn → πN observables that are seriously distorted by the nucleon Fermi motion, and that one can efficiently avoid this problem by imposing a cut on W . It is also shown that the use of kinematical cuts of recent experimental analyses can still leave in the selected samples substantial FSI effects that must be corrected in extracting the γn → πN cross sections. In terms of the nucleon-and pion-exchange mechanisms, I give the first qualitative explanation of the FSI corrections, obtained in a recent MAMI experiment, for extracting γn → π 0 n cross sections.