Last updated: 2026-08-30
Every figure Qarera publishes comes from one live index of job postings. This page explains where the postings come from, how each published metric is defined, and — the part that matters most if you are citing us — what the data cannot tell you.
Postings are collected directly from employers rather than bought from a feed reseller. Two kinds of source feed the index: applicant tracking systems — the hiring software a company publishes its own careers page with — and public job feeds operated by employment agencies and job platforms. Collecting at the employer end is what lets us read the original description text, which is where the language, sponsorship and skill signals come from.
New postings are collected daily. A separate sweep re-checks postings already in the index and retires the ones that have disappeared from the employer's board, so counts reflect what is currently advertised rather than everything ever seen. Published pages carry the date of the data they show; the figure and its date always move together.
Every source of job-posting data has structural blind spots. Ours are these. If you are citing a figure, these are the caveats worth carrying into your sentence.
The index counts advertised openings that we can see. It is not a vacancy count for a national economy, and it should not be read as one. National statistical offices publish that figure; ours is a measure of what employers are advertising, publicly, in the places we collect from.
Because we collect from applicant tracking systems and structured career pages, the index skews toward companies large or modern enough to run one. Small businesses, family firms, agency-mediated hiring, and roles filled by word of mouth are under-represented. The skew is strongest in markets where SMEs dominate hiring.
We mark a posting live until it disappears from the employer’s own board. Employers routinely leave filled roles up, and some repost the same role repeatedly. Counts should be read as advertised openings, not as unfilled positions.
Employers do not tag the language of a posting, so we classify it from the text. Classification is reliable on long descriptions and least reliable on short ones, on bilingual postings, and on listings that mix an English job title with a local-language body. Treat the English share as a close estimate, not an exact count.
This metric counts postings whose text refers to sponsorship, relocation support, or work-authorisation assistance. It captures employers who say something about it. It does not verify that sponsorship is genuinely available, and it misses employers who sponsor without mentioning it.
City is resolved from free-text location fields, which are inconsistent across employers; remote and multi-city postings are hard to place. We remove duplicates where the same role appears through more than one source, but some survive, and some distinct roles at the same employer look like duplicates.
If a figure looks wrong, tell us and we will check it: hello@qarera.com. Where we find an error, we correct the number and note the change here.
Covering 8 countries and 127,032 live postings as of 2026-08-30. Current figures · How to cite this data