How to spot a ghost job before you send the outbound message

A live job posting is not the same thing as active hiring intent. Postings sit unfilled for months, reappear on a repost cycle, and go up during hiring freezes with no real requisition behind them. On Ashby's own applicant-tracking-system data, 18% of postings analyzed in 2024 were never filled: paused, offer-made-then-withdrawn, or closed with no reason given (Ashby, 2025-04-08). Before you spend an outbound message on a "they're hiring" signal, five corroborated signals separate the postings that reflect real intent from the ones that are noise, reposts, or leftovers from a freeze: how long it's been live, how specific the description is, whether it exists anywhere besides third-party boards, whether it's on a repost cycle, and whether direct outreach gets any response at all.

Why "it's live" isn't "they're hiring"

Four datasets get cited most often for this question, and they don't carry equal weight.

The strongest evidence comes from Revelio Labs, which matched actual postings to actual subsequent hire records rather than asking anyone to self-report. Revelio found the hires-to-postings ratio fell from 0.75 in 2018 to below 0.5 by 2023 (Revelio Labs, 2023-10-31), in the firm's own phrasing, only every other posting results in an actual hire. The same dataset tracked recruiter-ghosting incidents, pulled from candidate interview-review text, up 120% since 2018.

One tier down sit two platform-data studies, drawn from applicant-tracking-system records instead of matched hire outcomes. Ashby's dataset spans January 2021 through December 2024 across more than 22,000 postings on its own ATS, skewed toward tech and startup employers. The 18% never-filled figure is specific to 2024 (Ashby, 2025-04-08) and breaks into 5.5% paused, 1% offer-made-but-no-hire, and 3.5% closed with no reason given. Those three categories sum to 10%, not 18%, and Ashby's report doesn't name what makes up the remaining eight points. Worth flagging before you treat the number as fully accounted for. The same report also found that 97.5% of jobs at companies with 50 or more employees move at least one candidate to an interview, which argues against a picture of postings as mostly theater. Greenhouse's platform data, separately, puts ghost postings at 18-22% per quarter (Greenhouse, 2024-12-10). Greenhouse's report also carries a survey figure, 60% of full-time US workers said they'd encountered a suspected ghost job, versus 56% in the UK and 59% in Germany, but that's a perception survey, not the platform number, and the 60% figure is a US-only statistic rather than a global one.

The weakest of the four is a survey run by Pollfish for ResumeBuilder, fielded starting 2024-05-22: 1,641 hiring managers were screened, 649 completed it, and 40% admitted their own company had posted a fake listing in the past year, with 30% saying one was currently active. That's a self-report on a reputationally damaging admission, filtered through a large screening funnel. It's a real data point, but it doesn't belong in the same sentence as Revelio's matched-outcome data or Ashby's and Greenhouse's platform records as if all four carried equal certainty.

The five corroborated signals

These five show up across independent sources, including Forbes, Fast Company, and Fortune alongside vendor blogs, often enough to treat as a working checklist rather than a hunch.

• The posting has been live for a long time, or its timestamp looks stale.
• The description is vague or generic, with no real day-to-day detail.
• The role only appears on third-party boards, never on the company's own careers page.
• The same listing reappears on a cycle, reposted every few weeks or months.
• Direct outreach about the role gets no response at all, from anyone.

Any one of these on its own is thin. Two or three together, on the same posting, are worth pausing over before you write an outbound message that references it.

Two signals worth less certainty

A smaller set of signals shows up in only one or two sources, which makes them worth noting but not worth treating as settled. The first is timing: a posting that went up during or shortly after a company disclosed layoffs or a hiring freeze deserves a second look, though a freeze in one division doesn't rule out real hiring in another. The second is visibility: no hiring manager or company account promoting the role anywhere, on any channel. Related but thinner still are a role whose seniority or scope doesn't match the company's size or stage, and a total absence of any named contact in or around the listing. None of these four is confirmed the way the five above are. Use them to lower your confidence in a posting, not to rule one out on their own.

What this means for using hiring signals at all

None of this means job postings are a bad signal to build outreach around. It means a single live posting, on its own, tells you less than it looks like it tells you. The studies above disagree on how big the ghost-job problem is: Ashby's own numbers argue both that 18% of 2024 postings went nowhere and that 97.5% of postings at larger companies produced at least one interview. Both can be true at once, because the question worth asking isn't how common ghost postings are in general. It's whether this specific posting, on this specific company, corroborates against the other signals available: growth activity, funding, leadership changes, hiring across multiple roles at once. A posting that clears two or three of the five signals above, and lines up with other independent evidence that the company is actually growing, is worth a message. One that clears none of them and stands alone is a coin flip you're spending your limited outbound time on. There is a second failure mode worth knowing alongside this one: a posting can be entirely real and still be closed to an outside agency, depending on who controls the requisition.