Conventional analyses used to interpret filtration data assume a constant internal liquid flow rate (q(L)) and a negligible solid velocity (the constant-q(L) approximation). To interpret such data, the upper bounds of errors to estimate alpha(av) and k(av) were analytically derived by the constant-q(L) approximation. The solution of the filtration model that incorporates a nonzero solid velocity was then analytically derived. When filtering a highly compactible filter cake, the cake was first compacted toward the filter medium that forms a skin layer and, in doing so, rapidly reaches the steady-state distribution predicted by the constant-q(L) approximation. Except at the first stage of the filtration, the approach of Tiller et al. (1999) is valid for interpreting filtration data of a highly compactible filter cake. Furthermore, the constant-q(L) approximation provides the upper limit of errors in terms of estimating cake characteristics regardless of the solid velocity effect.