Construct design and sample formulation are critical in structural biology projects. Large numbers of sample variants are often produced and analyzed for a single target and significant effort is dedicated to sample characterization in order to identify at an early stage the most promising samples to help save manpower and time. Here, we present a method based on a thermal stability assay that can help estimate the likelihood of biological samples to produce crystals. This assay is rapid, inexpensive and consumes very small amounts of sample. The results can be used to prioritize certain constructs at an early stage or as an objective test to help decide when to undertake other type of approaches addressed at improving sample properties.