Why do companies post ghost jobs? The stated reasons, the accidental ones, and what the data shows

The honest answer to why companies post ghost jobs is that most of them do not decide to. A minority post placeholders on purpose. The majority produce ghost-like postings as a side effect of how requisitions, applicant tracking systems and agencies work. The data can tell the two apart, because deliberate and accidental ghosts leave different traces, and they cluster on different boards.

Why do companies post ghost jobs?

Companies post ghost jobs to build a candidate pipeline before a role is approved, to appear to be growing to customers and investors, to hold a headcount budget line that would otherwise be cut, to satisfy a policy that every hire be publicly advertised when an internal candidate is already chosen, and, for agencies, to attract resumes they can sell into other searches.

Those are the motives employers admit to in surveys, and each has a business logic behind it. Pipeline building is cheapest when a posting is live all the time. Headcount that is visibly “in market” is harder for a finance team to claw back. A public posting for a pre-decided hire is often required by internal policy or, for some employers, by immigration rules that oblige them to test the labour market. None of these is malicious, and all of them produce a listing with nobody behind it.

The trace they leave is a thin, long-lived posting. A placeholder does not get a real description, does not get a salary band, and does not get closed, because closing it would mean someone owned it.

How rare is that trace? Only 1.3% of scored active postings have a description under 400 characters, and only 1.4% have been open 45 days or more. The deliberate placeholder exists, but it is a small minority of what applicants experience as ghost jobs. The bulk of the frustration comes from the mechanisms in the next section, none of which anyone chose.

Why do ghost jobs exist even when nobody intends them?

Ghost jobs exist without intent because the systems that publish postings are better at creating them than removing them. Four mechanisms produce most of the accidental ghosts JobsPipe sees, and each shows up in the data as a specific signal.

Evergreen requisitions

High-volume employers keep a permanent posting open for roles they hire continuously: nurses, drivers, warehouse associates, customer support, sales development. The posting is real in the sense that people are hired against it, and ghost-like in the sense that the specific listing you applied to may correspond to no current vacancy. Evergreen reqs are what the mass poster signal catches: 19.2% of scored active postings belong to a company with 500 or more open roles at once.

Applicant tracking system hygiene

An ATS closes a posting when someone closes it. When the hire is made through a referral, the req is quietly cancelled, or the recruiter leaves, the listing stays up. Some systems re-date a posting on every edit, which is why 5.8% of scored postings carry a posted date that drifts 30 days or more from when they were actually sighted. The posting says it is new. The sighting log says otherwise.

Agency reposts

Staffing agencies repost the same search repeatedly, sometimes on behalf of several clients, sometimes to refresh its position in the results. The repost signal, three or more sightings with a gap of ten days or more, is on 7.3% of scored postings overall and concentrates heavily on a few boards, as the next section shows.

Pipeline building that outlives the pipeline

A recruiter opens a posting to gauge the market for a role that is being scoped. The scoping ends, the role changes shape, and the original posting is still there. 28.6% of active postings across the ten boards were first sighted more than 30 days ago, and only 1.4% of scored postings carry the 45-day signal, but almost every one of those started as something legitimate.

Which boards carry the most ghost-like postings?

The share of scored active postings with a ghost score of 25 or higher, by board. A score of 25 requires at least one strong signal (age, reposts or mass posting) plus a supporting one, so this is the “deserves a second look” line, not a ghost count.

The ordering is the story. The boards at the top are applicant tracking systems used by large employers and a startup board where founders leave roles open indefinitely. Both are evergreen territory: 94.9% of scored Workday postings belong to a company with 500 or more open roles, and 80.1% of scored Y Combinator postings have been open 45 days or more. The boards at the bottom either expire free listings on a timer or serve small employers with a handful of roles each. The general-purpose boards sit in between: LinkedIn at 21.2% and Indeed at 11.7%. The full table with age and salary columns is on the ghost jobs page.

What a reposted job looks like in the data

A repost is invisible on the board, because the board shows you one posting with one date. It is obvious in a sighting log. JobsPipe sees each board on a schedule and appends a sighting every time a posting is present, so a listing that was live, disappeared for two weeks and came back reads as a cluster of sightings, a gap, and another cluster. The repost signal fires at three or more sightings with a maximum gap of ten days or more.

Overall, 8.0% of active postings have been reposted by that definition. On the boards at the top of the chart it is closer to every other posting. That is not fraud. It is the normal rhythm of a large employer refreshing hundreds of requisitions, and it is exactly why a repost on its own is a yellow flag rather than a red one.

For a developer the practical move is not to reject reposted jobs but to date them correctly. Use discovered_at, the first time JobsPipe saw the posting, rather than date_posted, which is whatever the board currently claims, when you decide how fresh a listing is. Then let max_ghost_score handle the stacked cases. The ghost jobs definition post has the full field list and a request to copy.