Currently, segmentation ignores None/NA values but it would be good for that to be included. Including None/NA as valid segments could be useful for workflows where missing values have special significance and need a separate step with inherited validation rules.