SEEK vs LinkedIn: When Is a Job Ad Too Old to Trust?

If you are applying for work, every listing carries an opportunity cost. A tailored resume and cover letter can take hours. Should you spend that time on a role advertised yesterday, one advertised six weeks ago, or one that appears to have been online for a year?

We compared the dates displayed in SEEK and LinkedIn’s default public search results for the same twelve occupation searches across Australia: 1,440 SEEK records and 1,480 LinkedIn records, collected on 26 August 2026.

The short answer is: prioritise recent listings. If an ad is more than 30 days old, verify that the employer is still recruiting before investing heavily. If you cannot verify an old listing, deprioritise it.

The practical takeaway

• Use LinkedIn’s date filter. Start with the past 24 hours or past week, then widen the search if your occupation or location produces too few results.
• Treat 30 days as a verification trigger, not an automatic rejection. Look for the same role on the employer’s careers site, a current requisition number or closing date, or confirmation from the recruiter.
• Avoid spending an hour on an old, unverified ad. If a listing is more than 30 days old, repeatedly advertised and absent from the employer’s site, and no contact can confirm it is active, put your effort into a more verifiable opportunity.
• Do not confuse a recent date with proof. Reposting, renewal and imports from recruitment systems can affect the date shown, and a recent-looking listing can still be inactive.

What 2,920 Australian search results showed

We used the same terms on both platforms: accountant, registered nurse, software engineer, project manager, marketing manager, warehouse, electrician, customer service, chef, administration assistant, sales representative and human resources. Searches were public, logged out and set to All Australia.

The pooled result was not being driven by one occupation. LinkedIn’s median displayed date was older than SEEK’s in all twelve searches. Across the individual searches, SEEK’s medians ranged from one to 6.5 days; LinkedIn’s ranged from eight to 16 days.

The SEEK collection contained 1,345 unique job IDs because the same ad could appear under more than one search term. LinkedIn’s 1,480 IDs were unique. After deduplicating SEEK, its median remained one day and its share older than 30 days remained 0.1%.

The finding is clear but narrow: in this one-day snapshot, a job seeker using the default results encountered substantially older displayed dates on LinkedIn than on SEEK.

Why the word “displayed” matters

A job card’s date is not necessarily the length of time a vacancy has been continuously open. An ad can be renewed or reposted. A platform can import a job from an employer’s applicant-tracking system. A role can cover continuous hiring or several vacancies. The platform may also retain a date from an external source.

This is particularly important on LinkedIn. Its current help page says a free job posted from a personal account remains active for up to 14 days and closes automatically after 30 days if it is not promoted. A promoted post can remain active until closed, but expires after six months. LinkedIn expressly says those rules do not cover enterprise jobs and job wrapping through Recruiter.

Our sample nevertheless contained 18 LinkedIn records with displayed dates more than six calendar months before collection, including six more than a year earlier. The oldest date was 574 days earlier.

Those records do not prove that a vacancy stayed open for 574 days. We could not observe the posting route or the source of the date. Enterprise products, imported metadata, renewal or a platform issue are all possible explanations. The defensible conclusion is that a LinkedIn date cannot safely be read as continuous time open or confirmation that applications are still being considered.

SEEK’s public terms describe a 30-day extension and treat reposting or refreshing as a new or additional ad. Its oldest displayed date in our sample was 31 days earlier, consistent with a roughly 30-day search surface once collection timing and date boundaries are allowed for. Our design cannot establish whether expiry, ranking, pricing, moderation or employer behaviour caused the platform difference.

Does an old ad mean a ghost job?

No. A ghost job is usually defined by employer intent: a role advertised without a current, genuine intention to hire. We did not observe intent. We also did not observe hiring outcomes, recruiter responses or whether a listing remained continuously available.

Age can still be a warning signal, but there is no validated Australian rule showing that a 31-day-old ad has a particular probability of being a ghost job.

Current Australian evidence also shows why a blanket cutoff would be misleading. In its June-quarter 2026 Recruitment Experiences and Outlook Survey found that 50% of employers recruiting Professionals and 53% recruiting Technicians and Trades Workers had been unable to fill all their vacancies within a month. The figure was between 27% and 45% for the other occupation groups reported.

That survey measures recruitment duration, not how long an online ad remained visible, and its unit is the recruiting employer rather than the individual listing. It nevertheless shows that genuine recruitment commonly lasts beyond a month, especially for harder-to-fill work. The Reserve Bank of Australia likewise notes that vacancies can remain open longer when employers struggle to hire.

Two Australian experts quoted by ABC Radio National in 2024 proposed 30 days as a useful warning point. Ben Hamer described an ad older than 30 days as “potentially a ghost job”, while Sugumar Mariappanadar argued for disclosure or removal after 30 days. Those are expert heuristics, not prevalence data or a diagnostic test.

Both SEEK and LinkedIn require job posts to concern genuine opportunities for which the advertiser intends to recruit. Those rules do not prove compliance, but they reinforce why an old date alone is not a responsible basis for accusing an employer.

Fresh first. Verify after 30 days. Deprioritise an old ad when you cannot confirm that the employer is still recruiting.

What Australian and overseas research can and cannot tell us

We did not find a national or platform-specific Australian study with a disclosed method that estimates the prevalence of ghost jobs. Australian commentary contains estimates, but an expert estimate is not the same as a measured rate.

Overseas and non-Australia-specific evidence is useful only as context. Greenhouse reports that 18% to 22% of jobs on its platform are classified as ghost jobs in a typical quarter. Its public summary does not provide an Australian estimate or enough detail to reproduce that classification, so the percentage cannot be transferred to Australia, SEEK or LinkedIn.

A large study using the former global Jobs on Facebook product found that, within the same vacancy and after accounting for application order, applications sent to older vacancies were less likely to be viewed, contacted or recorded as leading to an interview. That supports applying earlier. It does not identify ghost jobs. The platform mainly served roles that did not require a university degree, and the authors did not report an Australian result.

These settings share some online-recruitment mechanics with Australia, but that does not make the markets statistically analogous. Labour-market conditions, occupations, platform products, samples and definitions differ. For that reason, no overseas ghost-job percentage is used as an Australian benchmark here.

A job seeker’s screening rule

An employer’s careers page can also be out of date, so a matching page is supporting evidence rather than proof. If the job-board card and employer site disagree, or an application link fails, seek confirmation before proceeding.

LinkedIn’s own search help says its default results can include jobs from any date and provides filters for the past 24 hours, week or month. In our sample, restricting LinkedIn results to the past week would have removed most of the older tail. That is the clearest immediate action from this research. For other checks, see our guide to five signs of a ghost job.

Method and limitations

Collection ran on 26 August 2026 using public, unauthenticated pages from an Australian browser session. No account or paid product was used. Ages are whole calendar days because LinkedIn publishes a date without a time of day; SEEK labels in minutes or hours were treated as day zero.

• Ranked results, not random inventory. Both platforms rank results and may personalise them. The study compares the default search experience, not the complete stock of ads on either platform.
• One day and twelve purposively selected searches. The results do not measure seasonal change, and the occupations were chosen to span different parts of the labour market rather than sampled at random.
• Displayed dates, not continuous duration. Reposting, renewal and imported metadata can affect the label.
• No employer intent or hiring outcome. The study cannot calculate a ghost-job rate or show that LinkedIn has more ghost jobs.
• Public, logged-out pages. Logged-in users may see different results or status information.
• Posting route unknown. We could not identify whether LinkedIn records were free, promoted, enterprise or wrapped jobs.
• Evidence review, not a systematic review. We searched Australian government, platform, academic and news sources available by 28 August 2026. Saying that we found no Australian prevalence study describes that search; it does not prove that none exists.
• No independent audit or platform response. The data and code are available for checking, but no third party has audited them and neither platform was contacted before publication.

Data and disclosure

The underlying dataset, collection code and analysis script are available to journalists and researchers on request via the contact page. They cover the 1,440 SEEK records, 1,480 LinkedIn records and the code used to reproduce the figures above.

Neither SEEK nor LinkedIn participated in, reviewed or funded the research. We have no commercial relationship with either company. We did not seek comment before publication and will publish any substantive response received.

This article may be quoted or republished with attribution and a link. Because the figures are a single-day snapshot, please retain the collection date and do not describe them as current platform-wide rates.