Segregation in industrial settings is responsible for a significant amount of lost product due to poor quality issues. In the pharmaceutical industry, segregation of the active ingredient is a critical issue that can lead to loss of life or physical harm if not closely monitored and controlled. Therefore, finding a way to control or predict segregation is critical to optimizing product design or to mitigate quality issues with bulk powders and granules. Obviously, the best way to handle segregation is to create a product consisting of a mixture of key ingredients that does not tend to separate when subjected to typical stimulus in handling processes and distribution networks. While this is the best alternative, it is often difficult to fully achieve in practice. One of the needs to accomplish this goal is to find a method of easily characterizing a mixture to measure segregation potential. This paper addresses that need. It describes an automated methodology used to measure segregation and evaluates that method for consistency, repeatability, and correlation to previous methods. The method first forms a pile of material in a controlled manner and then uses reflectance spectrum to differentiate between components in a mixture along the pile. The method of computing the concentrations and segregation intensities from reflectance measurements is presented. Repeat experiments are done to determine the expected error of the method. This error is found to be within 7% from test to test for a badly segregating material and within 0.5% for a moderately segregating material. The method also uses a complex data acquisition scheme and numerical analysis of large amounts of data. We measured the error of the data collection and subsequent numerical analysis and found the error for computation to be within 0.3%. We also compared this to other manual methods and found good correlation to these methods of segregation measurement generating data within 7.8% of other methods. (C) 2014 Elsevier B.V. All rights reserved.