diff --git a/doc/jsk_pcl_ros/nodes/organized_statistical_outlier_removal.md b/doc/jsk_pcl_ros/nodes/organized_statistical_outlier_removal.md
index b6b78087d7..5e1ee4b7d2 100644
--- a/doc/jsk_pcl_ros/nodes/organized_statistical_outlier_removal.md
+++ b/doc/jsk_pcl_ros/nodes/organized_statistical_outlier_removal.md
@@ -28,7 +28,10 @@ Organized version of statistical outlier removal.
 ## Parameter
 * `~use_cluster_point_indices` (Boolean, default: `False`)
 
-  use cluster point indices or not
+  use cluster point indices or not.
+  this option is designed to use after clustering (i.e. euclidean clustering),
+  because we can consider the distribution of each cluster with cluster point indices
+  information.
 
 * `~keep_organized` (Boolean, default: `True`)