The steady state diffusion cell test is often used to determine the chloride ion diffusion coefficient in cementitious materials. It involves the measurement of the flux of chloride ions through a specimen under near steady state conditions. It has been noted that such a test may also provide data which characterises the chloride binding capacity of the specimen. In this work a numerical model of chloride diffusion subject to the effects of chloride binding is used to assess the effect of deviations from the steady state on the data obtained from a diffusion cell test. It is noted that there will be a tendency to underestimate the diffusion coefficient, although good practice should limit this error. The predicted error in the chloride binding isotherm is smaller than that in the diffusion coefficient. Furthermore, the influence of errors in the effective porosity on model predictions is limited as the resulting effect on the values of the calculated parameters describing chloride diffusion and binding counteract each other.