The present paper addresses the problem of decomposing the risk of a multi-factor credit portfolio into marginal contributions through a fast analytical approach. It is based on Taylor polynomial expansion of the overall risk and on the subsequent partial derivatives with respect to the single exposures, exploiting the Euler principle. The proposed approximation, which also accommodates an efficient treatment of obligors with similar risk profile, is suitable for large and complex bank portfolios. Furthermore, it performs quite well if tested against numerical techniques, among which the authors chose the Harrell-Davis estimator. The latter, aside from representing a benchmark measure, should however be applied only in the case of very small and concentrated portfolios. In addition, a comparison with the most usual variance-covariance approach is drawn, emphasizing its drawbacks in the correct representation of risk allocation.