A true-up is not a failure in itself — some variance is inevitable. The question is magnitude. A programme accruing from transaction detail settles within a couple of percent; a programme accruing from an estimated rate routinely settles 10–30% away, which means the number the business planned against was materially wrong for several quarters.
Large true-ups also concentrate: they land in one period, usually at year end, distorting a period that had nothing to do with the activity that caused them.