Current and future wide field surveys such as the Dark Energy Survey, eROSITA and Euclid are predicted to detect hundreds of thousands of galaxy clusters. With these numbers, we have the potential to dramatically increase cosmological constraints from galaxy cluster surveys. Though a variety of techniques exist to detect clusters, their masses cannot be directly measured, but only indirectly inferred from observed properties that correlate with mass. While cluster galaxies contribute only ~1-3% to the total mass budget of a cluster, they trace the gravitational potential well, providing an inexpensive mass proxy. To maximise the constraining power of clusters for upcoming wide-field surveys, it is essential to characterise the level of scatter and systematic bias associated with these mass estimation techniques. In this talk I will present new results on the outcome of an extensive blind study, The Galaxy Cluster Mass Reconstruction Project, which was created in order to ascertain how accurately we can measure cluster masses using techniques that rely upon the positions, velocities, colours and magnitudes of galaxies. I will describe the impact of various member galaxy selection procedures on the final mass estimate, focusing on the friends-of-friends, phase space and red sequence -based approaches. I will then explore the accuracy of a diverse set of galaxy-based mass estimation techniques: velocity dispersion, caustic, richness, abundance matching and radial -based methods. Finally, I will address the implications of the magnitude of scatter in the recovered masses for future cosmological surveys relying on cluster masses.