# The Live Posting Standard, and the Ghosts That Fail It

*You can take any single posting and return a defensible verdict of live, stale, or ghost from a fixed checklist in minutes.*

- Canonical URL: https://www.refolk.ai/candidates/guides/live-posting-standard-ghost-jobs
- Pillar: Applying at volume
- Format: Standard
- Published: 2026-08-05
- Last reviewed: 2026-08-05
- Reading time: 17 min

You have a list of postings and finite hours. This standard tells you which of those postings are live openings a company is actually hiring for, which are stale, and which are ghosts, using criteria weighted so that two people grading the same posting reach the same verdict. It is written for someone applying at volume who needs to spend application effort only where it can convert.

Most advice on this reads as an unranked pile of "red flags" left for you to weigh by feel. This is a pass/fail standard instead. It keys on one anchor a company cannot fake - the date a posting first appeared in the employer's applicant tracking system - and one cross-check that has become the only reliable liveness test - the employer's own careers page. Run it in minutes, attach the evidence, and defend the call.

## What "live," "stale," and "ghost" mean, precisely

A live posting is one where the employer is actively hiring for this specific seat right now; a stale posting is a real role that has aged past the point where activity is likely; a ghost posting is a listing that does not correspond to an open seat the company intends to fill. The standard outputs one of these three, never a yes/no.

The three-way split matters because "real company," "open role," and "actively interviewing" routinely diverge. As one source in the record puts it, real, open, and actively interviewing are three different conclusions. A posting can belong to a genuine company, describe a genuine role, and still not be a seat anyone is filling this quarter. A binary "legit or scam" test collapses those cases and sends you applying into a pipeline that goes nowhere.

> Real company and actively hiring for this seat are different claims, and the standard grades the second one.

Ghost postings are not rare enough to ignore. In Clarify Capital's February 2026 scrape of 176,268 Indeed listings across 49 industries and every U.S. state, 1 in 7 active posts were ghost jobs, and 4% had been active for four months or longer. That 4% is the dead tier: postings that outlived their intent by a full season and are still collecting resumes.

**1 in 7 - Active Indeed posts flagged as ghost jobs**

From Clarify Capital's February 2026 scrape of 176,268 listings across 49 industries and every U.S. state.

## How common are ghost jobs, and why the numbers disagree

Prevalence estimates cluster between 14% and 27% for scrape-based studies and jump to 39 to 81 percent for employer surveys, because the two methods measure different things. Scrapes count the share of live listings that look like ghosts. Surveys count the share of companies that ever posted one. Never blend them: 40% of companies is not 40% of postings.

| Study | Figure | Method |
|---|---|---|
| Clarify Capital 2026 | 1 in 7 (~14%) | Indeed scrape, 30-day flag |
| Greenhouse | 18-22% | ATS/platform data |
| ResumeUp.AI | 27.4% | LinkedIn listing analysis |
| ResumeBuilder 2024 | 40% posted one | Employer survey (n=649) |
| MyPerfectResume | 81% of recruiters | Recruiter survey |

Read the method column before the figure. Clarify's 14% and ResumeUp.AI's 27.4% describe the pool of postings you are grading, which is the number relevant to your list. ResumeBuilder's 40% and MyPerfectResume's 81% describe employer behavior across a whole company's history. Both are true; they answer different questions. When you cite a number to justify skipping a posting, cite a scrape figure, because that is the one that estimates what fraction of your list is at risk.

One structural reason stale posts persist: nobody owns closing them. In Refolk's index of professional profiles, the U.S. recruiter and talent-acquisition pool numbers 131,261, roughly 14.3 times the United Kingdom's 9,184 for the same title set. Yet senior TA leaders in the U.S. with Manager, Director, or VP titles number just 754, about 0.57% of the U.S. pool. Few people own the hygiene of closing dead requisitions, so postings outlive their intent as a matter of course, not conspiracy.

**754 - U.S. senior TA leaders in Refolk's index**

About 0.57% of the 131,261-person U.S. recruiter pool, so few people own closing dead requisitions.

## The clock is the exploit: recovering the true first-appearance date

The single most misleading field on any posting is its age, because reposting resets the counter. The date that actually grades a posting is the one an applicant tracking system, or ATS, records the first time the role is published - the one timestamp a repost cannot reset.

Here is the mechanism. When a company reposts, the original listing is removed and replaced with a new one, paid promotions may end, the position appears as new to job seekers, the reposted job keeps its internal candidates and stages, and the repost date resets the counter. On aggregators the distortion compounds: LinkedIn's "posted X ago" is often a refreshed or repost date, not the original, and the same role can be relisted repeatedly, so the shown date can be months newer than when the job was opened. A company running a permanent pipeline can keep a role reading "just posted" indefinitely.

> **Rule:** Grade on the ATS date, never the displayed date
>
> The only defensible age is the ATS first-publish timestamp. The "posted X ago" stamp resets with every repost, so a 3-day-old stamp on a role that first appeared in February is a fundamentally different object than one genuinely new.

The recovery procedure is mechanical. Read the company name from the posting, look up that company's ATS board (Greenhouse, Lever, Ashby, or Workday), match the role by title, and read the date the ATS recorded the first publish. If you cannot find the role on any ATS board, that is itself a finding: document it as "not found on ATS" and lean harder on the careers-page check. Reposting caps make this legible - a job can only be reposted after 30 or more days, and each repost resets that 30-day counter, so a role showing three or more unchanged cycles over several months is the ghost pattern.

Once you have a real date, the brackets do the grading.

| Bracket | Days since true first-appearance | Verdict tier |
|---|---|---|
| Fresh | 0-7 | Live candidate |
| Aging | 8-29 | Verify before applying |
| Stale | 30+ | Ghost flag (Clarify threshold) |
| Dead | 120+ | Ghost pattern (4% of posts) |

The 0-7 day rule comes from a career coach's long-standing practice of applying only to roles posted within the past couple of days, or a week at most. The 30+ flag is Clarify Capital's ghost-job threshold. Clarify's project manager frames age as the primary tell directly: a job posted 48 hours ago is more likely to be actively hiring than a job posted three months ago.

#### Recovering the true first-appearance date

1. **Read displayed age** - Note the "posted X ago" stamp and mark it unverified
2. **Find the ATS board** - Look up the employer's Greenhouse, Lever, Ashby, or Workday page
3. **Match by title** - Locate the same role and read the ATS first-publish date
4. **Score the bracket** - Apply 0-7 / 8-29 / 30+ / 120+ to the true date

*The ATS first-publish date is the one timestamp a repost cannot reset.*

## The source-of-truth cross-check: careers page over aggregator

The employer's own careers page is the source of truth for whether a role is live, and after March 2026 it is the only reliable one. Applying through the company's official site confirms the role actually exists and reaches the hiring team without passing through questionable third parties.

The reason aggregator absence stopped being proof of a dead role is a distribution change. As of March 31, 2026, Indeed eliminated free organic visibility for jobs sent through single-source XML or API feeds that do not use an ATS integration with Indeed Apply. Workology confirms employers on a supported ATS can still receive some organic visibility, but free reach is no longer guaranteed for all methods. Earlier, in December 2025, Indeed capped free hosted posts at 3 per employer per month and cut organic visibility from 120 days to 30. So a role missing from Indeed organic may reflect a feed gap, not a closed seat.

That is why "stale on the aggregator" and "dead" diverged. The careers page, not the aggregator, is now the liveness test. Run the check both directions:

- **Live on careers page, absent from aggregator:** likely a distribution gap. Treat as live and apply through the company site.
- **Present on aggregator, absent from careers page:** proceed with caution. If a job is listed on Indeed or LinkedIn but does not appear on the company's official careers page, that is a real warning, not a feed artifact.

> **Watch out:** A major platform is not a legitimacy stamp
>
> A job on a major platform can still be fraudulent, and a role can be genuinely posted yet not correspond to an open seat. Presence on LinkedIn or Indeed proves distribution, not intent. Confirm on the employer's own site.

Record the exact requisition, confirm the employer-linked application page is open, and note the job ID and the date you checked, plus any repost or description change since. Doing this by hand for every posting on a volume search is slow. [Refolk](/candidates) writes your resume from your own history and tailors it to each posting you decide is worth applying to, so the minutes you save on drafting go into the verification the standard actually requires.

## Which listing fields carry signal, and which lie

No single field on a listing proves hiring intent; you can verify a role is open, compare signals, and decide how much application effort the evidence justifies. Weight the fields accordingly, and know what each looks like when it lies.

| Field | What it suggests | How it lies |
|---|---|---|
| Salary range present | Weak positive; ghost jobs rarely include salary | Transparency laws made ranges common, so ghost roles now carry them too |
| Title matches responsibilities | Positive; aligned scope reads as real | Aspirational titles like "Senior Manager" over entry-level duties |
| Single clean listing | Positive; real campaigns are streamlined | Near-identical duplicates collect resumes or inflate visibility |
| Named recruiter or hiring manager | Positive if verifiable | A plausible name may be an impersonation of real staff |

Salary presence is the field to weight lowest. It still holds directionally that ghost jobs rarely include salary, but pay-transparency laws made ranges widespread, so the signal now generates false negatives: a fully salaried ghost job. Treat a missing range or an astronomically broad one as a small nudge toward suspicion, never as a verdict.

Title-to-responsibility mismatch is a documented tell. If the title says "Senior Manager" but the responsibilities sound entry-level, the listing may be a template rather than a real seat. Read the requirements against the level named in the title and mark a fail when they do not track.

Duplicate listings mislead in the opposite direction. Many postings can look like aggressive hiring, but if the same job appears under slightly different titles or locations, it more likely reflects resume collection or inflated visibility than multiple open seats. Real hiring campaigns are typically streamlined.

#### Age against source-of-truth presence

Horizontal axis runs from Old on ATS to Fresh on ATS. Vertical axis runs from Absent from careers page to Live on careers page.

| Quadrant | What it means |
| --- | --- |
| Old and absent | Ghost - do not apply |
| Fresh but absent | Stale or feed gap - verify the careers page directly |
| Old but live | Aging - apply only if fit is strong, expect slow process |
| Fresh and live | Live - apply |

*The two highest-weight axes decide most verdicts before you reach the field checks.*

## Grade any posting in eight steps

Run the steps in order. The first four establish the evidence; the last four turn it into a verdict. The whole pass takes ten to fifteen minutes and produces a single defensible label with its evidence attached.

#### The Live Posting grading procedure

1. **Capture the requisition** - Record employer, exact title, location, job ID, and source URL so you hold a stable identifier, not just a board link.
2. **Read the age, then distrust it** - Note the "posted X ago" stamp and flag it as unverified, because it resets with every repost.
3. **Recover the true first-appearance date** - Find the employer's ATS board, match the role by title, and read the first-publish date, or document "not found on ATS."
4. **Cross-check the careers page** - Confirm the role is live on the employer's own site with an open application page, or confirm it is absent.
5. **Score the age bracket** - Apply the 0-7, 8-29, 30+, and 120+ day brackets to the true date and assign one tier.
6. **Check the fields** - Mark salary presence, title-to-responsibility match, and duplicate listings as pass or fail.
7. **Verify the named contact** - Confirm any named recruiter appears as a company employee and verify contact details against known-good channels.
8. **Render the verdict** - Combine age tier, careers-page presence, and field checks into live, stale, or ghost with the evidence attached.

The dossier's sources disagree on order, and it is worth naming. Age-first coaches lead with the posted date; verification-first sources lead with the careers-page requisition. This standard recovers the ATS date before the careers-page check because the ATS date is the harder fact to obtain and the one that most changes the verdict. If you are grading a long list under time pressure, you can flip steps three and four - confirm the role is live first, and only bother recovering the ATS date for postings that survive that filter.

## How the grading goes wrong: failure modes and false positives

Most misgrades come from trusting one signal in isolation. Each failure mode below pairs the false positive that produces it with the check that catches it. This is the part of the standard to internalize, because a confident wrong verdict costs more than an uncertain one.

- **Trusting the displayed date.** The false positive is a "3 days ago" role that is actually four months old. A 3-day-old posting that first showed up in February is a fundamentally different object than one genuinely new. Check by recovering the ATS first-publish date.
- **Reading a repost as rejection or a fresh search.** The false positive is assuming a repost means everyone was rejected, or that it signals a brand-new opening. A repost does not necessarily mean they rejected everyone: offers get declined, budgets wobble, and reposts happen while strong candidates are mid-process. Check by counting cycles; three or more unchanged cycles over months is the ghost pattern.
- **Treating aggregator absence as "closed."** The false positive is a role missing from Indeed organic after the March 2026 feed change that looks dead but is live on the ATS. Check by going to the careers page, not the aggregator.
- **Treating a major-platform listing as proof of legitimacy.** The false positive is "it is on LinkedIn, so it must be real." A job on a major platform can still be fraudulent. Check by confirming the role on the employer's own site.
- **Inferring intent from a single field.** The false positive is "salary is posted, so it must be real." Salary can appear on ghost roles too, and intent is unprovable from the listing. Check by triangulating age, source of truth, and contact rather than leaning on any one field.
- **A named recruiter who looks legit but isn't.** The false positive is a plausible recruiter name that lends false confidence. Scammers impersonate real staff. Check that the person appears under the company's LinkedIn employees, and if you must contact them, use a phone number you know to be legitimate, not one from the message.
- **Duplicate listings inflating apparent activity.** The false positive is reading many postings as aggressive hiring. Check for near-identical titles or locations, which suggest resume collection rather than multiple seats.

> **Tip:** Two failing signals beat one
>
> No single field decides a verdict. Require at least two independent failures before you grade a posting ghost: an old ATS date plus careers-page absence, or a repost pattern plus an unverifiable contact. One failing signal downgrades; two condemn.

Confirming named contacts is where a volume search stalls, because it means opening a company's LinkedIn and reading its employee list for each suspect posting. This is exactly the kind of narrow, verifiable lookup Refolk is built for. Instead of hunting one name at a time, you can ask for the people who would own a live requisition.

Ask me this: `Heads of talent acquisition at US companies using Greenhouse` - [run the search](https://www.refolk.ai/start?q=Heads%20of%20talent%20acquisition%20at%20US%20companies%20using%20Greenhouse).

*Returns named TA leaders you can cross-check against a posting's stated recruiter before you trust the contact.*

## The verification checklist

Run this before you commit application effort to any posting. Every item is a check you can mark done, not a topic to consider. If a posting fails an item, note it against the verdict rather than dropping the whole grade.

#### Before you apply, confirm each of these

- [ ] Employer, exact title, location, job ID, and source URL are recorded as a stable identifier.
- [ ] The displayed "posted X ago" stamp is noted and flagged as unverified.
- [ ] The ATS first-publish date is recovered, or "not found on ATS" is documented.
- [ ] The role is confirmed live on the employer's own careers page, or confirmed absent.
- [ ] An age tier is assigned from the true date: Fresh 0-7, Aging 8-29, Stale 30+, or Dead 120+.
- [ ] Repost cycles are counted; three or more unchanged cycles over months is flagged.
- [ ] Salary presence, title-to-responsibility match, and duplicate listings are each marked pass or fail.
- [ ] Any named recruiter appears as a company employee and contact details check out against known-good channels.
- [ ] A single verdict - live, stale, or ghost - is written with the supporting evidence attached.

## Keeping the standard current

The mechanics that make this standard work are set by platform rules, and those rules change, so re-check them rather than trusting a fixed value. Two moving parts matter most: how boards distribute and age postings, and what fraud tooling employers deploy.

On distribution, the anchor to watch is Indeed's organic-visibility policy. It moved twice in quick succession - a December 2025 cap of three free hosted posts per employer per month and a cut from 120 to 30 days of organic visibility, then the March 31, 2026 removal of free organic reach for single-source feeds without ATS Indeed Apply support. Each change alters how much "aggregator absence means dead" you can assume. When you notice a posting live on a careers page but invisible on aggregators, re-read the current policy before you conclude anything.

On the employer side, ATS providers are adding verification layers. Greenhouse announced fraud and identity tooling under "Real Talent," and a verified-company jobs hub went live in 2025. As verified-company signals spread, a posting on a verified employer hub becomes a stronger positive than an unverified board listing. Fold that into your field checks as it becomes available for the employers on your list.

> **Note:** Re-check the mechanism, not the number
>
> Platform policies and fraud tooling shift often. The brackets and cross-checks in this standard stay valid, but the specific visibility rules and verification hubs change, so confirm the current policy before you rely on aggregator absence or a verified-company badge.

The standard itself is stable because it rests on facts a company cannot easily fake - the ATS first-publish date and the state of its own careers page - rather than on any one board's display conventions. Adopt it as team policy by fixing the age brackets, requiring the two-signal rule before any ghost verdict, and logging the evidence with each call. Two people running it on the same posting will reach the same answer, which is the whole point.

## Frequently asked questions

### How can I tell if a job posting is real before I spend time applying?

Grade three things and combine them. First, recover the true first-appearance date from the employer's ATS rather than trusting the displayed 'posted X ago' stamp, which resets on every repost. Second, confirm the role is open on the company's own careers page, not just an aggregator. Third, check the listing fields and any named recruiter. A posting that is fresh on the ATS, live on the careers page, and specific in its fields is worth your effort.

### Is this job still open if it only shows up on LinkedIn or Indeed and not the company site?

Proceed with caution, but do not assume it is closed. Since Indeed removed free organic visibility for single-source feeds on March 31, 2026, a live role can be absent from an aggregator because of a distribution gap. The reliable test is the employer's own careers page. If the role is open there, it is live regardless of the aggregator; if it is absent from the careers page while present on a third-party board, that is a real warning sign.

### Does a reposted job mean everyone was rejected?

No. A repost does not necessarily mean the company rejected all prior candidates. Offers get declined, budgets wobble, and reposts happen while strong candidates are still mid-process. The signal that matters is repetition over time: three or more unchanged cycles across several months on the same requisition is the ghost pattern. One repost by itself is normal hiring noise.

### What are the strongest ghost job red flags to weight most heavily?

Age against the true first-publish date is the primary tell: a role posted 48 hours ago is more likely to be actively hiring than one posted three months ago. Absence from the employer's own careers page is the second. Weight salary absence low, because pay-transparency laws made ranges common enough that some ghost roles now include them. Three or more unchanged repost cycles and near-identical duplicate listings round out the high-weight signals.

### How common are ghost jobs, really?

Scrape-based studies of live listings cluster between 14 and 27 percent: Clarify Capital's 2026 Indeed scrape found 1 in 7, and a LinkedIn analysis put it at 27.4 percent. Employer surveys report higher numbers, 39 to 81 percent, but those measure the share of companies that ever posted one ghost, not the share of postings that are ghosts. Keep the two apart; 40 percent of companies is not 40 percent of listings.

---

*From the Refolk guide library. I revise these guides rather than replacing them, so the current version is always at https://www.refolk.ai/candidates/guides/live-posting-standard-ghost-jobs*