Methodology

What these figures are

They are counts of job listings we observed on a specific job board, on a specific date. They are not a census of the Gulf labour market, not an official statistic, and not a measure of how many jobs exist in the region. Any role never posted to the board we monitor is absent from these numbers by construction.

Where the data comes from

What is counted, and what is excluded

• We count listings marked active in our index at the snapshot date — one row per listing as published, not per employer and not per distinct real vacancy.
• Figures cover the six GCC states only: the United Arab Emirates, Saudi Arabia, Qatar, Kuwait, Oman and Bahrain. Our index holds a small number of listings for markets outside the GCC; those are excluded from these counts entirely rather than reported under a Gulf heading.
• If the same role is advertised twice, it is counted twice. We do not de-duplicate across sources, so figures for one source are never additive with another.
• Figures from sources we have not approved for publication are excluded entirely, including from any total. A published total covers only the named source.
• Listings the board itself has removed are excluded where we can confirm removal by re-checking the board directly. Where no such confirmation is available for a source, we do not publish freshness or “live now” claims for it at all.

Limitations we know about

These are the reasons we publish less than we could:

• Listing age is not verifiable. The posting date we hold is whatever the source reported the last time we read the listing, and it is overwritten on every re-read. We cannot tell a genuinely new posting from a refreshed old one, so we publish no age or “stale listing” statistics.
• Job titles are not a taxonomy. Matching titles by text is wildly imprecise: on our own index, searching for one common engineering term returns roughly 900× more listings than an exact title match. We therefore publish no role-level or job-title counts.
• We have no salary data. Structured pay fields are present on a negligible fraction of listings, and prose mentions skew toward lower-paid roles. We publish no salary figures, averages or comparisons — and our product does not either.
• We cannot see hiring outcomes. We do not know who was interviewed, hired, or ignored. We publish no response rates, hiring rates, or success statistics, and we never promise a job or an interview.
• Job-description coverage is partial. Where we report it, it describes our own dataset — the share of listings for which we hold substantive description text — not a property of the market.

How the figures are produced

Figures are not computed on page load. An internal audit runs against our database, classifies every candidate figure by how defensible it is, and records the result as a dated, immutable snapshot. Only a snapshot that has been reviewed and deliberately published becomes visible here. Every public surface — this page, and any machine-readable interface — reads that same snapshot, so they cannot disagree with each other. If nothing is published, nothing is shown; we do not fall back to an unreviewed live number.

Corrections

If a figure here is wrong, we want to know and we will correct it. Email unfairjobshq@gmail.com with the page and the figure. Corrections are made by publishing a new snapshot; the snapshot date and audit-run identifier shown on each page make it possible to tell which version of the figures you are looking at.