Due to physical orientations and birefringence effects, practical quantum information protocols utilizing optical polarization need to handle misalignment between preparation and measurement reference frames. For any such capable system, an important question is how many resources -- e.g., measured single photons -- are needed to reliably achieve alignment precision sufficient for the desired quantum protocol. Here we study the performance of a polarization-frame alignment scheme used in prior laboratory and field quantum key distribution (QKD) experiments by performing Monte Carlo numerical simulations. The scheme utilizes, to the extent possible, the same single-photon-level signals and measurements as for the QKD protocol being supported. Even with detector noise and imperfect sources, our analysis shows that only a small fraction of resources from the overall signal -- a few hundred photon detections, in total -- are required for good performance, restoring the state to better than 99% of its original quality.