The turbidity assay is commonly exploited to study protein liquid-to-liquid phase separation (LLPS) or liquid-to-solid phase separation (LSPS) processes in biochemical analyses. Herein, we present common pitfalls of this assay caused by exceeding the detection linear range. We showed that aggregated proteins of high concentration and large particle size can lead to inaccurate quantification in multiple applications, including the optical density measurement, the thermal shift assay, and the dynamic light scattering experiment. Finally, we demonstrated that a simple sample dilution of insoluble aggregated protein (LSPS) samples or direct imaging of liquid droplets (LLPS) can address these issues and improve the accuracy of the turbidity assay.