we re-opened 1,165 job listings. one in four was already gone.
every listing in pirch's public pool was confirmed live — a human-grade check, not a scrape — within the past month. this week we went back and re-opened all 1,165 of them. 27% had already vanished. here's the decay curve, and why “posted 3 days ago” won't save you.
you've lived this: you find the job, you spend an evening on the application, you click through to submit — and the posting is gone. not “position filled.” gone gone. 404.
we wanted to know how often that actually happens, so we measured it on our own inventory. pirch's verified-jobs pool only contains listings that were opened and confirmed live and accepting applicants — every one within the last month, most within the last two weeks. this week we went back and re-opened all 1,165 of them.
method, in one breath: we fetched every listing politely (with retries), and only called a job dead when the page itself was deleted — an http 404 or 410, the server saying “this posting no longer exists.” 432 listings sat behind bot walls that wouldn't let an automated check through, so we excluded them entirely rather than guess. that left 733 listings with a definitive answer. the raw results are here — every verdict, every timestamp.
the decay curve
the longer it had been since we verified a listing, the likelier it was to be dead — no surprise. what surprised us was how early the dying starts. these are real jobs, remember: each one had a live page, a real employer, an application form that worked. this isn't a ghost-job story about postings that were never real. it's the other problem — real listings just don't live very long.
share of listings already deleted, by time since we verified them live
where the bodies aren't
one more honest wrinkle: the boards that blocked our checker (403s, rate limits) might die faster or slower than the ones that didn't — we can't know, so they're out of the numbers. and sources differ: of the listings we could check, one major aggregator's jobs were dead 30% of the time while one smaller remote board's were dead 15% of the time. where a listing lives changes how long it lives.
if you want the companion problem — postings that were never really open in the first place — that's a different pathology with its own field guide: how to spot ghost jobs. today's data is about honest jobs dying fast. the research on never-real jobs (1 in 7, per the study Forbes covered in april; up to 21% in the academic estimate) stacks ON TOP of this decay. the pool you're fishing in is smaller than it looks, twice over.
what a listing's age means for your odds
| time since last confirmed live | still reachable | already deleted |
|---|---|---|
| under 3 days | 88% | 12% |
| 3–7 days | 79% | 21% |
| 1–2 weeks | 69% | 31% |
| 2–3 weeks | 67% | 33% |
| 3+ weeks | 66% | 34% |
a job listing doesn't age like wine. it ages like milk — and the sell-by date isn't printed on the carton.
what to do with this
the practical reading isn't “apply faster, panic more.” it's that freshness is a real variable and you mostly can't see it. a listing's posted date is public; whether anyone re-confirmed it's alive is not. so put your effort where decay can't eat it: apply the same day you find something good, front-load the roles that genuinely fit instead of queueing twenty maybes for the weekend, and treat any application older than a week as spent — follow up once, then let it go.
and when a posting 404s mid-application: it was never about you. the door didn't slam. it was already gone — you just couldn't see it from the hallway.
worth your evening
- a listing verified live today (not “posted today” — checked today)
- roles you'd genuinely take, applied to the day you find them
- one follow-up on week-old applications, then release
quietly spent
- the “save for the weekend” queue — a third dies before saturday
- sorting by posted date and calling it freshness
- re-reading a two-week-old application for signs of life

the pool is smaller than it looks, and shrinking daily.
of 733 job listings we could definitively re-check — every one verified live within the past month — 27% were already deleted. one in eight died within three days. and the posting date barely predicted any of it.
disclosure, because it's the honest thing: pirch sells the antidote to exactly this. every job a hunt returns was opened and confirmed live within the last 72 hours, and our re-verify pass keeps re-opening listings for as long as they're in the pool — this study is literally our maintenance process pointed at itself. distrust this paragraph accordingly; the data above is downloadable and the method is stated. milk goes off. check the carton.
want the 12%-dead shelf, not the 34% one?
every job in a pirch hunt was opened and confirmed live and accepting applicants within the last 72 hours — the decayed ones get thrown out before you ever see them. first hunt's free.
start your free huntsources
- pirch — raw study data: 1,165 listings re-opened july 28, 2026 (verdict + http status + age per listing, JSON)
- Forbes — “1 in 7 job postings are ghost jobs, new study reveals” (April 2026)
- Hunter Ng — “Why is it so hard to find a job now? Enter Ghost Jobs” (arXiv; Glassdoor data; up to 21%)
- MyPerfectResume / PRWeb — the “ghost job economy”: JOLTS openings-vs-hires analysis
- CNBC — ghost job postings add uncertainty to the stalling jobs picture (Nov 2025)