How these figures are made

The figures come from the Internet Vacancy Index, published monthly by Jobs and Skills Australia. It counts job advertisements appearing on a defined set of online job boards, by occupation and by region.

An advertisement is the unit. Not a vacancy, not a hire, not a person: an advertisement, appearing on one of the boards the index covers, during the reference month. That distinction runs through everything on this site and is the single most important thing on this page.

The current release covers 50 regions and 56 occupation groups, for July 2026. Australia is divided into the eight capital cities and the regions outside them, which together cover the country exactly once.

Nothing on this site is a count of vacancies in Australia. In particular:

Boundaries are the Australian Statistical Geography Standard, Edition 4, published by the Australian Bureau of Statistics. The index reports the eight capital cities as Greater Capital City Statistical Areas and the rest of the country as Statistical Areas Level 4.

Those two levels sit side by side rather than one inside the other, and their figures must never be added together. A capital city and the SA4s that make it up describe the same ground twice.

A state total on this site is the sum of the regions the publisher reports on within that state. The publisher releases this index by region, not by state, so the state figure is arithmetic done here. Every page showing one says so. Where a region carried no figure for the month, the state total is the sum of those that did, and the page says how many reported.

Boundaries are simplified for drawing. That is a display decision and it affects the shapes on the map, never the figures beside them.

Occupation groups are reported at the level Jobs and Skills Australia publishes them, with the codes and names the publisher uses. Where the publisher names a code inconsistently across regions, no name is shown rather than one of them being chosen.

Occupation groups must not be added together. They nest and overlap, so summing them produces a total nobody published. The all-occupations figure comes from the publisher’s own all-occupations row, never from adding the groups up.

Job advertisements are a separate matter. A listing is mapped to an occupation classification only where the mapping is established; an unmapped listing stays unmapped rather than being guessed into a plausible code.

Every figure belongs to a reference month, and the month is stated wherever the figure appears. The current release is July 2026.

2 reference periods are held: the current month and the one before it, currently June 2026. A change shown on this site is that comparison and nothing more.

Two points are a comparison, not a trend. Nothing here is described as a trend, because a direction cannot be established from two observations. The publisher’s own workbook holds the full history, and trends will appear on this site when enough periods have accumulated to support them honestly.

Where no prior period is held, a page reports that there is no comparison. It does not report a change of zero, which is a different statement.

A figure that does not exist is never shown as zero. Absence has kinds, and they are distinguished:

Alongside the statistics, this site republishes individual job advertisements from the sources it is licensed to carry. These are a different kind of thing from the index figures and are never mixed with them.

They are shown and never counted. The listing corpus mixes sources, and one of them reserves aggregate figures for a written licence, so there is no count of advertisements anywhere on this site. The figures on the map and the location pages come from the Internet Vacancy Index; the listings do not add to them and are not a sample of them.

Three dates travel with a listing and they mean different things:

The original advertisement is authoritative. Every listing links to it, and where this site and the source disagree, the source is right. What is and is not reproduced from an advertisement is recorded field by field.

This site is a search for work in regional Australia, so every listing is placed against the published definition of a designated regional area: the Migration (Designated regional areas for certain skilled and temporary graduate visas) Instrument (LIN 22/022) 2022, in force since 5 March 2022. The instrument is three pages long and is written entirely in postcodes. Two tables list the postcodes that are inside the definition; a postcode in neither is outside it, and in practice that means Sydney, Melbourne and Brisbane. The instrument never names those three.

Nothing here interprets it. Every answer is a lookup against a table a reader can open, and each listing states the postcode and the rule that placed it so the lookup can be repeated. These labels describe where a job is. They are not a statement about any person’s visa position, and nothing on this site decides whether anyone may apply for or hold any visa.

Sources do not all publish a postcode, so three rules are used and the one that applied is shown on the listing. They are not equally strong, which is why the difference is published rather than smoothed away.

Where a source publishes map coordinates rather than a postcode, the postcode is found by locating those coordinates inside an Australian Bureau of Statistics postal area, and the listing says so. Postal areas approximate Australia Post postcodes rather than reproducing them, which is the Bureau’s own caveat and belongs beside any label resting on one. No postcode is ever invented: a listing whose coordinates fall in no postal area falls through to a weaker rule, or goes unplaced.

Each advertisement carries a label describing what it says about visa sponsorship, together with the words that produced it. The label reports wording. It is not advice, and it says nothing about any reader’s visa position or eligibility.

The three affirmative labels record how firmly an advertisement put it, because an employer stating sponsorship is available and one who might consider it for the right person are making different statements. Where wording could be read at more than one strength, the weaker reading is published. That is deliberate: telling a reader an employer offers sponsorship when the advertisement only raised the possibility may send them to relocate on a false basis, whereas the reverse understates an employer whose own sentence is printed directly beneath the label. A statement that sponsorship is not available outranks everything else in the same advertisement.

Labels are matched on phrases rather than keywords, because “visa” alone appears in job titles and in descriptions of who may apply, and “sponsorship” alone appears in marketing and events roles that have nothing to do with visas. Neither is an offer to sponsor anyone. For anything about a visa, see the Department of Home Affairs.

Some advertisements carry a line naming things their own text mentions: a credential, a licence, a named piece of software. The line says the advertisement mentions the thing. It does not say the employer treats it as mandatory, and it is never an assessment of a reader.

Nothing is inferred. A skill is attached only where the advertisement writes it, never from a job title and never because roles of that kind usually want it. That rule is stricter than the one the sponsorship labels follow, and deliberately: a wrong sponsorship label misstates an employer in public, underneath the employer’s own quoted sentence, where a reader can see it. A wrong skill quietly removes a real job from a real person’s search and nobody ever sees it happen.

Four things are deliberately not read. General working qualities such as leadership and communication, because they appear in more than a quarter of advertisements as description of the role rather than as anything asked of an applicant, and a filter matching a quarter of an index separates nothing from nothing. Occupations, because an occupation is the job rather than a skill and classifying one is a separate problem with an unresolved licence question behind it. Qualification levels, which say how much study a role wants rather than what the holder can do. And vaccination status, which is the most common requirement in this index and is health information about a person rather than a skill.

No count of skills is published anywhere on this site. The licence covering part of this index reserves aggregate figures, and a count drawn from the rest would describe one state’s public service rather than a labour market. Showing what one advertisement says, and selecting advertisements by it, is not an aggregate.

The things most likely to mislead a reader who skips the rest of this page:

If a figure here looks wrong, an advertisement is out of date, or a listing carries information it should not, it should be corrected. Every figure is traceable to the release it came from, so a correction can be checked rather than argued about.

A reporting route is not yet published, and this page will name it when it is. In the meantime the original publisher is authoritative for anything they released, and every listing and figure on this site links back to its source.

Regional Sponsor is not affiliated with, endorsed by or sponsored by any organisation whose advertisements or data it publishes. Each source is named beside the material it supplied.