The benchmark this register uses whenever a body says delivery data cannot be published at fine geography.

It is why a claim like the Cheaper Home Batteries Program passing 500,000 installations is checkable by any member of the public, rather than something the public has to take on trust from a media release. Uptake can be traced to the suburb, over time, by technology.

The contrast that motivated much of this project's federal research: the same government, in the same year, reported progress on a $10 billion housing program as a single national figure that reached the public through an audit. Both are delivery counts. Only one is published at a granularity anyone can use. See How many homes has the Housing Australia Future Fund actually finished, and where are they?.

CORRECTION, 2026-09-03. This record previously carried a last_checked date and, in the safeguard case, a confidence of verified. Neither was earned. The underlying data file was never downloaded; the record was written from the regulator's web pages and from reporting about them, which is reported at best. Both have been reset.

The scoring inputs are the part that matters, because they drive the usability score this project uses to judge everyone else. open_licence is supported -- Clean Energy Regulator datasets on data.gov.au carry CC-BY-3.0-AU. format and machine_readable are not confirmed. The regulator's data pages are client-side applications with no static file links, and the eight CER datasets listed on data.gov.au resolve to shapefiles, a web map service, and HTML landing pages rather than to the tabular file this record asserts.

The data is very likely downloadable as a spreadsheet through the site's own interface, and that is exactly the kind of assumption this register is supposed to refuse. The claim stands as reported and unverified until someone opens the file.

Third attempt, 2026-09-03. The postcode data page was fetched and carries no static file links; like the safeguard page it is a client-side application. format and machine_readable remain unconfirmed and confidence is lowered to estimate.

This is an uncomfortable place for the register to leave its own benchmark entry. The postcode series is cited elsewhere here as proof that fine-geography delivery data can be published, and that citation now rests on reporting rather than on the file. The claim is probably right. It is not yet checked, and the difference is the whole point of this project.

Fourth attempt, 2026-09-03, via the Internet Archive's URL index rather than the live site. That index lists real files served from cer.gov.au under /DocumentAssets/Documents/, including "Postcode data for small-scale installations - all data.xlsx", "Postcode-data-for-small-scale-installations- SGU-Solar.csv" and "2021-22 Safeguard facility data.xlsx".

Those filenames are documentary evidence that this data is published as spreadsheets and CSVs, which is better support for machine_readable than anything the record previously rested on, so confidence returns to reported. The bytes themselves still could not be retrieved: the paths now redirect to the current site, and the archive did not return the stored copy either. Four routes tried -- direct fetch, the data page, data.gov.au, and the archive.

The register should be able to live with this. The claim is supported by a file listing rather than by the file, that distinction is recorded, and last_checked stays empty because nobody has opened it.