The Job Is Online. Is the Job Open?

The sentence arrives after the work

The short answer is uncomfortable but useful: an online advert proves that a listing was visible. It does not, by itself, prove an approved budget, an unfilled post, an active search or an intention to appoint someone now. A role can be real but slow. A company can be collecting profiles for later. A listing can be duplicated, stale, cancelled or misleading. Those are different conditions, even when the applicant experiences them through the same silent inbox.

A posting is not the statistical vacancy

Eurostat defines a job vacancy as a paid post that is newly created, unoccupied or about to become vacant, where the employer is taking active steps to find an external candidate and intends to fill the post immediately or within a specific period. That is a statistical benchmark, not a universal certification attached to every advert.

The European Statistical System treats online job advertisements as a separate source. Its methodology says employers may use them to seek candidates, explore talent pools or meet other recruitment needs. The European Commission’s Joint Research Centre likewise distinguishes online advertisements from vacancies: vacancies are unfilled positions that may or may not be publicly advertised, while online adverts are collected continuously from internet sources.

That distinction matters in practice. A specialist search may take months. A requisition may pause while an approval is reconsidered. An employer may maintain an evergreen pool for recurring roles. A board may carry a record after the employer’s priorities have changed. These are analytical possibilities, not a finding that any particular employer acted properly or deceptively. The point is to avoid treating every unclear advert as the same kind of failure.

Why the ghost-job percentage keeps moving

There is no settled global percentage of ghost jobs because the term does not describe one settled population. Greenhouse’s report, published on December 10, 2024, says that three in five surveyed candidates suspected they had encountered a ghost job. It also says that 18% to 22% of jobs posted on the Greenhouse platform in any given quarter were classified by Greenhouse as ghost jobs. The page does not fully explain the classification rule or provide a representative sample of all employers and countries.

A working paper dated October 29, 2024 used Glassdoor data and an LLM-BERT classification method to estimate that up to 21% of job adverts might be ghost jobs. That is a model-based result from one dataset and definition. “Up to” should not be converted into a universal prevalence rate, and a classification model cannot directly observe what an employer intended when a listing was published.

MyPerfectResume’s August 5, 2024 survey of 753 U.S. recruiters reported that 81% said their employer had posted adverts for jobs that did not exist or were already filled. That is a self-published, non-probability survey and the wording combines two different conditions. Respondents also selected reasons including testing the market, maintaining a job-board presence, building a future talent pool and making the company look viable during a hiring freeze. Those answers describe reported practices; they do not measure the share of all adverts that are fictitious.

In this article, “ghost job” is a broad journalistic label, not a single scientific category. Research and surveys may include filled roles, adverts with no immediate intention to hire, future talent pools or other non-active postings. Duplicated feeds, stale records and paused budgets may be related warning signs, but they should not be silently folded into every study’s definition.

What official vacancy numbers can tell you

Official labour-market data answers a wider question. The U.S. Bureau of Labor Statistics’ Job Openings and Labor Turnover Survey defines an opening as a specific position with available work that could start within 30 days and for which the establishment is actively recruiting externally. This is a carefully defined establishment-level measure, not a database of authenticated internet adverts.

The August 2026 JOLTS release, published on September 29, 2026, reported 7.1 million U.S. job openings. BLS states that openings are measured on the last business day of the month, while hires and separations count payroll changes across the entire month. Those measures should not be presented as a count of unique online adverts or guaranteed hiring decisions.

Eurostat and BLS statistics can show labour-market demand under defined rules. They cannot authenticate the job ID in front of you, confirm that a particular budget remains approved or predict whether a recruiter will arrange an interview. That limitation follows from the aggregate and statistical design of the data, not from a failure of the statistics.

A verification routine with limits

Begin with the employer-controlled record. Compare the advert with the company’s own careers page: title, location, wording, requisition number, reporting line and application route. Record the posting date, closing date and any language saying the role is continuously open. These are practical heuristics, not a validated detection system. Employer systems can be stale, duplicated or poorly synchronised too.

If there is a human contact, ask a bounded question: “Is this role attached to a current requisition, and what is the intended hiring window?” You might also ask where the role sits in the organisation, whether interviews are already under way and whether the application is handled directly by the employer. The answer can help you decide how much preparation to invest. It cannot reveal every internal change, and a vague answer is not proof of bad faith.

Look for patterns rather than verdicts: repeated renewal, materially different descriptions across channels, an unexplained intermediary, no identifiable team or an application process that does not connect back to the employer. These signals may justify a lighter-touch application or a request for clarification. They cannot prove that a budget was cancelled, that a role was never intended to be filled or that anyone meant to mislead applicants.

The cost arrives before the rejection

Imagine spending an evening rebuilding your CV, studying the company’s strategy and preparing for a case exercise, only to discover later that the search had been paused. Or imagine arranging travel and references for a role that was technically online but no longer attached to an active decision. The emotional injury is not proof of deliberate deception. It is the predictable cost of asking candidates to invest before the vacancy’s status is clear.

That is the central dilemma. The information designed to make a role discoverable is often much stronger than the information showing whether someone is ready to hire. The fair response is neither blind trust nor universal suspicion. Check the employer-controlled record, ask about timing, calibrate the work you give the application to the evidence available and keep other options moving. An online job is an invitation to investigate—not confirmation that the employer is ready to make an offer.