Summary

  • ACMA recorded two distinct failures associated with Optus's 8 November 2023 outage: emergency-call access denied to 2,145 people and 369 welfare checks not conducted. Those measures must not be added together. ACMA's November 2024 findings.
  • A penalty payment and announced safeguards do not, by themselves, demonstrate reliable emergency-call fallback or completed caller follow-up. Equally, the absence of subsequent test results from these announcements does not prove that repairs were not made.

An emergency-call failure presents two different recovery tasks. One is to make the service available. The other is to establish what happened to people who tried to reach help while it was unavailable. Restoring connectivity cannot, on its own, close the second task.

The historical enforcement case involving Singtel Optus Pty Limited makes that distinction concrete. On 8 November 2024, the Australian Communications and Media Authority said Optus subsidiaries had paid penalties totalling more than A$12 million for emergency-call rule breaches arising from the nationwide outage of 8 November 2023. That was a reported enforcement outcome, not a certificate that every operational weakness had been corrected. ACMA's announcement.

Two measures, not one failure rate

ACMA found that Optus failed to provide access to the emergency call service for 2,145 people during the outage. The unit matters: this is a count of people in the regulator's finding, not a count of calls or repeated attempts. ACMA separately found that Optus failed to conduct 369 welfare checks on people who had tried to make an emergency call. The regulator's access and follow-up findings.

The cited account does not establish the overlap between those measures or a common denominator. Adding them would not produce a defensible total of affected people. Dividing one by the other would not produce a defensible welfare-check failure rate. Nor do the figures establish injuries, deaths or the circumstances of any particular caller.

They do establish why the repair question needs more than a network-availability measure. A technical assessment asks whether emergency calls can connect under the relevant conditions. A follow-up assessment asks whether the affected population has been identified and the necessary actions accounted for. Someone who could not reach help should not disappear from an organisation's response merely because its network monitoring has returned to normal.

What the 2024 response established

On 30 April 2024, the Telecommunications Industry Ombudsman welcomed the Australian Government's response to the review into the Optus outage. That was an institutional response to the proposed way forward; it was not a technical finding that Optus's repairs had passed testing. The Ombudsman's statement.

In its November 2024 announcement, ACMA said Optus had failed in network management in several areas and that the outage should have been preventable. The preventability assessment belongs to the regulator. Its announcement does not provide the detailed technical sequence needed to assign a particular fault to a router, software change or supplier. ACMA's network-management assessment.

ACMA also described a broader response: better communications during outages, greater oversight of the Triple Zero system, regular systems testing and arrangements to ensure emergency calls could be carried by other carriers. It said it had updated the Telecommunications (Emergency Call Service) Determination 2019 and was developing a new industry standard for outage communications. Those descriptions are anchored to November 2024; they should not be read as a statement that the standard remained under development in September 2026. The safeguards described at the time.

The distinction is between a finding, a regulatory response and an operational result. Each can be important without proving the next. A welcome for reform does not verify its implementation. A requirement to test does not disclose the test conditions or results. A payment does not show whether an emergency call can complete during a future disruption.

What would demonstrate repair?

For a claim about technical correction, useful evidence would identify the failure scenario, the date and scope of testing, the results and any exceptions. It would also show whether material network changes require a new test. A successful test answers the question it was designed to test; it should not silently become assurance for conditions outside its scope.

For a claim about cross-carrier fallback, the question is more specific than whether another network exists. What end-to-end emergency-call path was tested? Under which operating conditions? Which dependencies and organisational hand-offs were included? The assurance should describe the tested route and its limitations rather than rely on a general statement about redundancy.

For caller follow-up, the evidence has a different shape. A reconciled account should distinguish the population requiring attention, completed actions and unresolved cases. It should permit scrutiny of the process without exposing personal information. Network test results cannot substitute for that account, just as a completed follow-up process cannot demonstrate that a technical defect has been removed.

Outage communications require another distinction: a written procedure is not the same as evidence that usable information reached the intended audience during a disruption. An assessment should ask how the communication channel was expected to work when ordinary services were impaired and what evidence shows that it did.

These are proposed tests for evaluating assurance claims, not additional legal duties attributed to Optus. They follow the practical difference between preventing a failure, detecting its consequences and accounting for the response.

A bounded conclusion, not a present-day compliance finding

This briefing examines the dated enforcement and institutional announcements. They do not supply a subsequent independent report establishing the completion or failure of every safeguard. That limit cuts both ways: it prevents a declaration of comprehensive repair, but it also prevents a claim that repair did not occur.

A later assessment would need to examine implementation evidence, including the scope and conclusions of any independent review and the operator's account of what changed. Neither an endorsement nor a broad assurance statement should be made to answer questions it did not investigate.

The historical record establishes serious regulatory findings and a described reform response. The defensible next question is not whether one announcement sounds conclusive. It is whether the evidence matches the particular outcome being claimed: payment, technical correction, usable fallback, effective communications or completed caller follow-up.