On determining the cluster abundance normalization