The redo size statistics shows how much redo log your statement generated when executed.
The redo size statistics is most useful when judging the efficiency of large bulk operations with statements as :
The amount of redo generated by the statements :
will be in many case outside of our control.
If you perform these operations with as few index enabled as possible, you can reduce the amount of redo generated, but you will have to rebuild those disabled indexes. This is generally practical only in datawarehouse when you are performing large data loads.