Ghost Jobs in India: We Measured the Age of 29,372 Live Job Postings

Every job seeker in India eventually asks the same question: is anyone actually reading this? You apply to forty roles, tailor each one, and hear nothing. The standard advice blames your resume. Sometimes that is fair. But there is a second explanation nobody measures, because measuring it requires watching job boards over time rather than surveying people about them.

So we measured it. JobTether runs a scraper that reads job postings directly from company career boards — the actual applicant tracking systems companies host their openings on, not aggregator listings. On 17 October 2026 we took a snapshot of the entire pool: 29,372 live postings across 1,057 companies hiring in India, and checked how long each one had been sitting there.

Of those with a genuine publish date, about one in four was more than 90 days old. One has been open since July 2015.

We first published this on 1 August 2026 (6,342 postings, 33.0% over 90 days) and re-ran it on 23 September (6,266 postings, 33.6%). On 24 September the pool nearly tripled, to 17,730 postings, and the headline moved to postings with a genuine publish date (29.5% over 90 days). By 26 September it had grown again, to 29,815 postings from 1,057 boards, almost all of the growth from large employers on Workday, whose postings run younger than the rest. That change in the sample is why the numbers below moved, not a change in the market. On Lever and Workable, the two sources whose board lists barely changed, 43.5% of postings are still older than 90 days. We re-ran it on 17 October on a sample of almost the same size (29,372 postings, same 1,057 boards), so this is the first run that compares like with like: the share of genuinely dated postings older than 90 days moved from 24.0% to 25.0%. Three weeks on, the backlog of stale listings has not shrunk.

What a “ghost job” actually means

A ghost job is a listing that is publicly visible with no active hiring behind it. It is not always a lie. The category covers several very different situations:

• The role was filled, and nobody closed the posting.
• The req was frozen mid-process during a budget review.
• The recruiter who owned it left, and the posting outlived them.
• It is an evergreen ad for a high-churn role, kept open permanently to collect resumes.
• It is deliberate — pipeline building, or signalling growth to the market.

From your side of the screen, all five look identical. That is the whole problem. The best-known benchmark comes from Greenhouse, which found that 18–22% of postings on its own platform were ghost jobs, and that roughly 70% of companies using it had posted at least one in a single quarter. Other estimates run as high as one in three.

Those figures are all from the US market. We could not find equivalent data for India, so we generated it.

The headline numbers

Here is the age distribution of the 24,034 postings whose boards report a genuine first-published date: Lever, Workable, Ashby and Workday, 81.8% of the pool. Every one was returned as an open, applyable role by the company's own career board on the day we checked.

We leave two sources out of this headline. Until 16 October our scraper read Greenhouse's last-edited timestamp rather than its first-published date (more on that below), so earlier runs could not use it, and we keep it out here so the runs stay comparable. SmartRecruiters ages are cut off at a year by our own scraper. Counted across all 29,372 postings, with Greenhouse now on its real publish dates, the over-90-days share is 26.1%.

The median posting in that group was 32 days old. The mean was 98 days, pulled up by a long tail of listings that have been open for years. Put the other way round: only 47.7% of these postings went up in the last 30 days, and only 16.2% in the last week.

This reframes the “I applied to 50 jobs and heard nothing” experience. If a quarter of what you applied to was already stale, a meaningful chunk of that silence was never about you. It is still worth fixing the things you control, but the denominator matters.

The applicant tracking system predicts freshness

This was the most useful thing we found, and the most actionable. Posting age varies dramatically depending on which applicant tracking system hosts the board — and you can see which one a company uses just by looking at the apply link.

* Greenhouse ages are on its first-published date from 16 October 2026; earlier runs of this study used a last-edited timestamp and showed Greenhouse as the freshest system, which was an artefact. A few boards had not been re-read yet, so if anything these figures understate Greenhouse's age. † Our scraper drops SmartRecruiters postings older than 365 days, so its “over a year” figure is zero by construction, not by reality. More on both in the methodology section.

Even allowing for those caveats, the spread is real. The four systems that report a genuine first-published date, Lever, Workable, Ashby and Workday, together 24,034 postings, 81.8% of the pool:

• Median age: 32 days
• Older than 90 days: 25.0%
• Older than 180 days: 13.3%
• Older than one year: 6.3%
• Older than two years: 2.0%

More than one posting in eight that is still standing has been standing for six months. On Lever and Workable alone it is one in four.

How to read an apply URL

You do not need our scraper to use this. The ATS is visible in the link behind any “Apply” button on a company careers page:

Each linked name above goes to that system’s own public API documentation — the interfaces this study reads. Workday publishes no such documentation, but its career sites serve the same data to any browser. Everything here is reproducible from them without our scraper.

This is a prior, not a verdict. A Lever posting from last week is fresh; a Workday posting can still be a frozen req. But when you are triaging thirty tabs on a Sunday evening, a prior is worth having.

The postings that have been open for years

The tail is where this stops being an abstraction. The single oldest live posting in our snapshot is a Software Engineer role at Navaratan Technologies Private Limited, published 24 July 2015 and still returned as an open, applyable job — 4,103 dayslater. It is the only posting on that board. Anyone can verify it: the company's public Workable feed still lists it today.

More instructive is Tech Firefly. In September its public board mixed fresh roles with postings from 2019 and 2020. By 17 October the fresh roles were gone and the old ones were not: all 11 of its live postings are more than four years old. Among them:

A role advertised as Immediate has been immediate for nearly seven years. This is the clearest illustration we found of why urgency language in a job title tells you nothing at all.

Zooming out, we counted boards with at least 10 open roles and a median posting age above 180 days. There were 35 of them, holding 1,616 jobs, 5.5% of the entire pool. (On 26 September there were 28, holding 1,269 jobs. Part of the increase is Greenhouse boards whose real dates we can now read, so not all of it is new decay. In August, on a pool a fifth this size, 24 such boards held 893 jobs.) These are not scams. They are mostly real companies whose career pages have quietly stopped being maintained. Applying into them is not dangerous; it is just time you will not get back.

The part that matters most if you are a fresher

We also looked at what these 29,372 roles are actually for. Classifying by title keywords:

• 55.0% carried a senior signal — senior, lead, principal, staff, head, director, manager, architect.
• 7.9% carried an entry signal — intern, trainee, fresher, graduate, junior, associate, apprentice.
• That is a ratio of 7.0 senior-sounding roles for every entry-level one.

Now combine the two findings. Entry-level titles that were also posted in the last 30 days came to 1,134 postings, 3.9% of the whole pool. That is the realistic live target list for a fresher applying through company career pages on any given month.

If you are a fresher sending 100 applications into that market and hearing nothing, the arithmetic, not your ability, is doing most of the work. The response is not to apply harder into the same 3.9%. It is to widen the surface: referrals, the roles that never get posted, and campus and cohort channels that never enter a public ATS in the first place.

Where the jobs actually are

Since we had the snapshot, the geographic split is worth recording. Of postings we could resolve to a metro:

Bengaluru alone is well over a third of everything. Remote-eligible roles were 4.6% of the pool — a real share, but far below what the discourse around remote work in India would suggest.

A five-minute check before you apply

None of this means you should stop applying. It means you should triage. Before spending forty minutes tailoring for a role:

1. Find the date on the company's own careers page, not the aggregator. Naukri and LinkedIn re-stamp dates when a listing is refreshed; the ATS usually does not. A “2 days ago” on an aggregator is frequently a six-month-old req.
2. Look at what else the board has. If every other posting is also months old, the whole page is unmaintained. If there are fresh roles alongside, the company is hiring and this specific req may just be slow.
3. Check for repeats. The same title reposted every few weeks usually means the req is not converting — either the bar is unrealistic or the pay is not competitive.
4. Ignore urgency words.“Immediate joiner”, “urgent requirement” and “hiring now” carry no information. Our oldest “Immediate” posting is from 2019. The same goes for “competitive salary” — 6.9% of postings say it and it predicts nothing.
5. Spend the saved time on fewer, better applications. A tailored application into a two-week-old posting beats ten into six-month-old ones. If you are tracking properly — and you should be keeping a tracker — log the posting date alongside the application date. After a month you will see the correlation in your own data.

Methodology, and what this data cannot tell you

We are publishing the limits alongside the numbers, because a statistic without its caveats is just a headline.

What we did.On 17 October 2026 we took a single snapshot of JobTether's scraped job pool: 29,372 postings across 1,057 company career boards, all filtered to roles based in India. Postings come from six applicant tracking systems — Greenhouse, Lever, Ashby, Workable, SmartRecruiters and Workday — read from each company's public board API. Age is the difference between the posting date the board reports and the snapshot date.

The caveats, in order of how much they matter:

• Age is evidence, not proof. An old posting is not automatically a ghost job. Some roles genuinely take six months to fill, and some companies run legitimate evergreen postings. What age gives you is a probability, not a verdict.
• We corrected how we read Greenhouse dates.Greenhouse's public feed carries both a last-updated timestamp and a first-published date. Until 16 October 2026 our scraper used the last-updated one, which made Greenhouse look like the freshest system (a 10-day median in September). On its real publish dates its median is 53 days and 32.3% of postings are over 90 days old. We keep Greenhouse out of the headline so this run stays comparable with the earlier ones.
• The SmartRecruiters figures are censored by our own scraper. SmartRecruiters boards do not remove closed postings and expose no status field — we verified a posting released in 2015 that still reports itself as active and public. To avoid importing archives, our scraper drops SmartRecruiters postings older than 365 days. So its “over a year” rate reads as 0% because we cut it there, not because such postings do not exist. Its true stale rate is higher than the table shows.
• This is not the whole Indian job market. It is company career boards on six modern ATS platforms — which skews toward tech, startups, GCCs and mid-to-large organised employers. It excludes Naukri-native listings, walk-ins, consultancies and the substantial share of Indian hiring that never touches a public board at all.
• Title-keyword seniority classification is crude.“Associate” means junior at one company and mid-level at another. Treat the 7.0:1 ratio as a strong directional signal rather than a precise measurement.
• The sample is chosen, not random. We add company boards ourselves. The 245 Workday employers, now 65% of the pool, were added in hand-picked batches and carry 16 to 442 India postings each, and we left out Accenture, whose roughly 33,600 India postings would have been most of the pool. Because the sample changes, comparing one run to the next measures our coverage as much as the market.
• Five snapshots, not a time series. We measured the standing stock of postings, not flow. A longitudinal version — watching how long postings survive before removal — would be the stronger study, and it is what we plan to publish next.

The honest conclusion

The Indian job market is not fake. Roughly half of what you can apply to right now went up in the last month, and those roles are real, staffed and moving. But the board you are looking at does not distinguish between those and the quarter that have been sitting there for more than three months, and neither does your application.

The practical shift is small: stop treating “listed” as “open”. Check the date, check the board around it, and spend your effort where the evidence says someone is actually on the other end. Then make the applications you do send genuinely tailored — because on a fresh posting with a real recruiter behind it, that is the part that decides the outcome.

The tables behind this study are free to download and cite (CC BY 4.0) on our research and data page.