Not every supersession costs the same

When a training product is superseded, the register records one more thing about the replacement: whether it is equivalent or not equivalent. That single flag is the difference between a paperwork exercise and a teaching problem.

Across the whole national register — every superseded qualification, in every training package, current or retired — 40% carry the not-equivalent flag: 1,857 of 4,619. Restricted to the packages measured in the table below — current packages with at least 30 superseded qualifications — it is 38% (1,337 of 3,492). But either average describes almost nobody. Depending on which training package you deliver, the rate runs from 8% to 70%.

The rate, by training package

Qualifications are flagged more often than units

The same measurement, split by what kind of product was superseded rather than by which package it sits in. A qualification is flagged not equivalent at roughly one and a half times the rate of a unit of competency.

This is worth knowing when you read your own scope. A list of forty superseded units and three superseded qualifications is not forty-three equal problems: the three qualifications are both more likely to be flagged not equivalent and the only ones that carry enrolments of their own.

How this was measured

Counted from the register's own catalogue of 125,996 training products, read on 13 August 2026. For every superseded qualification, the register publishes its status as either supersededEquivalent or supersededNonEquivalent; the rate is simply the second as a share of both. Products are grouped by the first letters of their code, which is how the register itself groups them.

Three denominators appear across our measurements, so here is the bridge. The register holds 4,619 superseded qualifications in total (the figure at the top of this page); 4,590 of them declare a replacement, and that set is the base of our supersession-chain study; and 3,810 sit in currently-registered training packages, which is the base of the by-type table above. Different questions, different bases — each named where it is used.

Two exclusions, both deliberate. First, only training packages whose own register entry is current. Without that filter the two ends of the table were artefacts — THH (hospitality, retired) showed 84% and ICA (information technology, retired) showed 2%, because a retired prefix stops receiving new supersessions and its rate freezes. Second, a minimum of 30 superseded qualifications per package: below that a handful of cases swings the rate by twenty points. That threshold excludes 20 current packages from the table. Both filters are stated so the count can be reproduced or disputed.

What this does not say

• It does not say how many RTOs are affected. This measures the catalogue, not anyone's scope of registration. How many providers actually have one of these on their live scope is a different measurement — that one is in our study of RTO scopes.
• It does not explain why any package has the rate it has. That would need the decisions behind each training package review, which we have not read. This is what the register records, not why.
• It does not say a learner has to repeat anything. The flag records that outcomes do not carry across one to one. What that means for a specific enrolment sits with the provider and the regulator.