Data classification

You cannot protect everything equally, so you have to know which data actually matters.

The failure mode is schemes with too many levels and no practical difference between them. Four tiers, elaborate definitions, and everything ends up labelled Internal because nobody can tell the categories apart under time pressure. A classification is only worth having if the label changes what happens: who can reach it, where it may go, how long it is kept. Otherwise it is an exercise that produced a document.