Studies both of entire galaxies and of local Galactic star formation indicate a dependency of a molecular cloud's star formation rate (SFR) on its dense gas mass. In external galaxies, such measurements are derived from HCN(1-0) observations, usually encompassing many Giant Molecular Clouds (GMCs) at once. The Andromeda galaxy (M31) is a unique laboratory to study the relation of the SFR and HCN emission down to GMC scales at solar-like metallicities. In this work, we correlate our composite SFR determinations with archival HCN, HCO+, and CO observations, resulting in a sample of nine reasonably representative GMCs. We find that, at the scale of individual clouds, it is important to take into account both obscured and unobscured star formation to determine the SFR. When correlated against the dense-gas mass from HCN, we find that the SFR is low, in spite of these refinements. We nevertheless retrieve an SFR - dense-gas mass correlation, confirming that these SFR tracers are still meaningful on GMC scales. The correlation improves markedly when we consider the HCN/CO ratio instead of HCN by itself. This nominally indicates a dependency of the SFR on the dense-gas fraction, in contradiction to local studies. However, we hypothesize that this partly reflects the limited dynamic range in dense-gas mass, and partly that the ratio of single-pointing HCN and CO measurements may be less prone to systematics like sidelobes. In this case, the HCN/CO ratio would importantly be a better empirical measure of the dense-gas content itself.