Methodology

How it works

ABN Signal runs on a weekly comparison model. A current snapshot of publicly available ABN records sits alongside a history of prior snapshots. Fresh data arrives each week, and any differences between the two snapshots get recorded.

The sections below explain how the process works and what it can surface.

Where the data comes from

All data originates from publicly available ABN records published by the Australian Business Register (see data sources). Fields include ABN, entity name, entity type, ABN status, GST status, and related identifiers.

ABN Signal operates independently, reading and comparing public record snapshots as an independent data-intelligence product.

How often data refreshes

Data refreshes happen weekly. The current snapshot is compared against the one from the previous week.

What does "detected" mean on this site? It means the change showed up in the most recent weekly comparison. The underlying event at the source register may have happened any day during the preceding week.

Across the site, you will see language like "weekly-detected" and "seen in the latest weekly update." ABN Signal uses precise terminology that reflects the weekly refresh cadence.

Change detection

Each week, four steps run:

1. ABN Signal loads the latest data snapshot.
2. Every record is compared field by field against the previous week.
3. Where a monitored field has changed, the difference is logged.
4. Logged changes are stored with the detection date.

For example: if an ABN record's GST status moves from Active to Cancelled between two snapshots, that difference is captured and timestamped. See coverage for which fields are monitored.

The same weekly refresh also detects entirely new ABN records entering the dataset. These appear on the new registrations page, separate from field-level changes on existing records.

The process is systematic — it captures every field-level difference between two points in time and records exactly what changed and when.

Interpretation

Raw field data alone can be hard to read. Each change comes with a plain-language explanation next to the raw numbers and status codes.

When a record shows a status change, the explanation separates three layers:

• Direct facts: what the fields show (e.g., "ABN status is Cancelled from 1 Jan 2026")
• Interpretation: what the data may suggest (e.g., "This may indicate the entity is no longer actively registered")
• Boundaries: what the data does not prove (e.g., "This does not confirm whether the business has ceased all operations")

The aim is to be clear about what the data shows without overstating what it proves.