Summary

  • IRINN’s APNIC 62 presentation says thirteen IPv6 webinars held from January 2024 through February 2025 reached more than 3,560 attendees. The thirteen displayed values sum exactly to 3,567.
  • The same slide closes with “5000 plus” attendees. The gap is at least 1,433, but the deck does not define whether it comes from registrations, repeat attendance, recordings, other sessions or a different period.
  • A training-reach receipt should publish the unit, event roster, time window, deduplication rule, channel coverage and correction version before attendance is used as evidence of programme reach—still less of learning or IPv6 deployment.

Thirteen bars make a ledger. IRINN’s September 2026 update to the APNIC National Internet Registry Special Interest Group gives each webinar a topic and a number. “IPv6 Security by APNIC/IRINN” has 230. “DNS Security, Threat & Mitigation” has 657. The other eleven values are 240, 482, 352, 264, 176, 368, 320, 144, 151, 95 and 88.

Add them and the answer is 3,567. That is satisfyingly close to the sentence above the chart: thirteen webinars between January 2024 and February 2025 “reached over 3,560 attendees”. The rounded claim and the displayed series fit.

Then the last line says: “Total attendees across all 13 sessions: 5000 plus..” At its minimum, that is 1,433 above the bar sum, a difference equal to 40.2% of 3,567. The problem is not that a round number exists. It is that the slide offers no field telling a reader what changed between the chart and the footer.

Give the larger number its strongest defence

There are ordinary ways for both totals to be true. One could count live attendance and the other registrations. One could include replay views, partner platforms or late reporting. The chart might stop in February 2025 while the footer covers a longer window. A participant attending three sessions might appear three times in a session total but once in a distinct-person count—or the reverse, if one figure counts enrolments across a larger programme. The deck may also have been assembled from two reporting extracts whose labels did not survive consolidation.

None of those explanations is established by the published file. They are reasons to ask for a reconciliation, not reasons to choose a preferred number. The public NIXI Academy page distinguishes online courses, training videos and curated video material, which makes channel boundaries more important, not less. The preceding APNIC 61 NIXI update described regular webinars and online training but did not publish this thirteen-session table or either total.

The most defensible public fact is therefore narrow. One APNIC 62 slide contains a coherent 3,567 series and a separate 5,000-plus statement. It does not show whether the second statement has the same unit, scope or extraction date.

Attendance is not a population—and not an outcome

The word “attendees” can hide several denominators. Event systems commonly retain registrations, check-ins, peak concurrent viewers, connections, completed sessions and recording views. A person reconnecting can create more than one technical session. One person can attend several webinars. An institutional group can watch one connection. Privacy-preserving reports may intentionally avoid joining identities across events.

Those choices are not inherently defective. Each answers a different question. Session attendance measures activity delivered. Unique people measure reach. Registrations measure demand before no-shows. Completion can indicate sustained participation. A survey or assessment may provide evidence of learning. A subsequent network measurement may say something about deployment. Moving from one noun to another without a join is where a reporting number acquires authority it did not earn.

That boundary matters in an IPv6 programme. The APNIC 62 deck describes topics spanning fundamentals, subnetting, routing protocols, transition mechanisms, deployment case studies and DNS security. The breadth is visible. The chart does not establish that 3,567 different people attended, that 5,000 people were trained, or that any participant later enabled IPv6. It should not be asked to do so.

The same restraint protects the programme. An unresolved footer should not erase thirteen granular values. Nor should a large attendance claim be forced to carry learning, certification or deployment outcomes that require separate instruments.

Publish a reach receipt, not a list of learners

A useful correction would be small. Name the programme and reporting window. List the thirteen event identifiers and dates. Define the unit for every value: registrations, live attendance instances, distinct accounts, completed sessions or replay views. State whether repeat attendees are deduplicated within an event, across the series or not at all. Name included delivery channels and the extraction date.

Then reconcile the totals. If 3,567 covers live session entries and 5,000-plus includes recordings, say so and publish both subtotals. If one is an older number, identify the version it replaced. If the footer is a broader programme total, separate it from the thirteen-event series. Corrections should retain the previous slide and attach a dated note rather than quietly exchanging one figure for another.

No learner roster is required. A privacy-safe receipt can use aggregate counts, suppressed small cells and a declared deduplication method. The point is provenance, not surveillance.

The APNIC 62 report correctly presents the file as an IRINN update within the NIR SIG. That makes the deck a useful public record of what a national registry chose to report, not proof that APNIC independently audited each figure. Attribution belongs in the receipt too: reporter, publisher, measurement owner and reviewer are different roles.

Capacity building is difficult to measure well. That is precisely why the visible arithmetic deserves preservation. The thirteen bars already supply a reproducible subtotal. One line explaining the extra 1,433 or more would turn a puzzling slide into evidence that operators, funders and participants can compare without mistaking activity for impact.

Sources