Blog · Published August 1, 2026
How many job listings are already dead when you click apply?
Most writing about stale job listings measures how long a posting has been up. That is a proxy. The direct question is simpler: if you click apply right now, is there anything on the other end? Almost nobody measures that, because it means actually following the link rather than reading the posting.
So we follow it. Every listing gets its apply URL fetched, and the destination is recorded: a normal page, a 404 or 410, or a page whose own text says applications are closed. A request that fails for our own reasons, a timeout or a block, is recorded as unknown rather than dead. We would rather understate the problem than hide a real job.
The important part of the method is which listings get counted. Checking runs oldest-first, and old listings are dead listings, so any headline drawn from everything checked would say more about our processing order than about the market. The published figure is therefore restricted to listings a source was still advertising within the last seven days. These are not archives anyone forgot about. They are current inventory.
The rate climbs sharply with age even inside that window, which is the practical lesson: a listing that has been sitting in a feed for a fortnight is a materially worse bet than one that appeared this week, and the gap is larger than most people assume.
None of this means a working apply link is a good sign. It only rules out the most basic failure. A role can accept applications for months while the team has already hired internally, and no amount of link-checking will reveal that. What link-checking does is remove the listings that could never have worked, which turns out to be a meaningful share of what you are being shown.
Cite this
JobiQueue. "How many job listings are already dead when you click apply?." August 1, 2026. https://jobiqueue.com/blog/how-many-job-listings-are-already-dead
What this does not measure
This counts listings JobiQueue ingests, not the whole job market, and only those whose apply link we could reach. A blocked or timed-out request is recorded as unknown, never as dead, so the figure is a floor rather than a ceiling. It also cannot see roles that are technically live but already filled internally: a working apply link is not evidence anyone is reading applications. Per-source rates are deliberately not published, because links are checked oldest-first and any board-by-board comparison would mostly reflect which listings happened to be checked rather than how that board behaves.
Figures are free to reuse with attribution (CC BY 4.0) and are current as of August 10, 2026. The corpus is re-measured continuously, so always carry the measurement date with the number. Ghost Job Index · Methodology
Check a listing before you apply
Paste any job URL into the free trust scan for scam risk, ghost signals, and apply-link checks. No account needed.
Run a free trust scan