This paper sets out a method to extract Maxwell model parameters from experimental compression-relaxation tests and investigates common experimental sources of bias when dealing with viscoelastic materials. Particular attention was given to viscoelastic materials that relax stress quickly. The proposed method differs from the methods usually used in that it takes into account the stress that can relax when a material is submitted to compression before proper relaxation. Among the experimental biases that can affect the tests, this study investigated the impact of the geometry defects of the samples, of the sensitivity of the rheometer used and of the compression speed on the characterization of the material. The uncertainties caused by these biases were then propagated in the proposed method. The proposed method was used to study the evolution of the viscoelastic properties of semihard cheese during ripening. Variability between cheeses proved to be greater than the uncertainty of the proposed method, and no tendency could be established, meaning that the viscoelastic parameters were considered constant during ripening. (C) 2015 The Society of Rheology.