Ghost-job risk methodology

kery.io estimates ghost-job risk from public listing history it can verify: how long an ad has been up, distinct reposts, closed status, related postings, and evidence quality. Thin history produces a cautious score. The method cannot prove a job is fake or genuine.

Signals evaluated

When you save a job with a public link, the dashboard can look at:

  1. How long the listing appears to have been up (first seen date, where verified).
  2. Whether there are distinct repost cycles, not just a refreshed LinkedIn date stamp.
  3. Whether the listing is closed or not currently accepting applications.
  4. Whether the same company and title appear as other specific job URLs in the same country.
  5. How much verified posting history was available (evidence quality).

kery.io does not read employer inboxes, applicant volumes, or internal ATS stages. If those sources are missing, it does not invent them.

What the scores mean

  • Low: little or no verified repost history, and the ad does not look long-stale on what we can see.
  • Medium: the role has been open a long time, history is limited, or there are weaker reuse signals without a clear repost pattern.
  • High: several verified reposts, or a long-open role that has also been put back up.

Known limits: LinkedIn can reset the visible posted date on the same job ID. Other job boards can show older copies of the same title. Scores can be wrong. An interview last week can still sit on a Medium or High listing if the ad is old. Product page: ghost job checker.

Questions people ask

What does evidence quality mean?

How much verified posting history was available. If sources are thin, the score stays careful instead of inventing missing dates or applicant counts.

Do LinkedIn date refreshes count as a new posting?

Not automatically. LinkedIn can reset the visible posted date on the same job ID. kery.io tries not to count that as an extra ad for this listing.

When was this method last described?

This page was last updated on 16 October 2026. The live dashboard uses the current analysis version deployed with the product.