Remote hiring data: three numbers we publish about this job market

Our ghost-posting rate is 38.3% as of 2026-10-07. Definition: the share of listings we still show as open whose age has already passed the median listing window that employers state themselves.

We compute it in two steps. First we take every active listing whose closing date genuinely came from the employer or the source board, never a date we inferred ourselves, and measure how many days sat between the posting date and that stated close. The median of those gaps becomes the reference window: how long a real listing on this corpus was actually meant to stay open.

Second, we measure the age of every active listing and count how many are already older than that reference window. That count divided by the whole active corpus is the figure above. The calculation only reads; it never closes or reopens a listing. A rising rate means more stale roles are being left visible, which is what job seekers experience as applying into silence.

Our salary disclosure rate is 20% as of 2026-10-07. Definition: the share of active listings whose pay range came from the employer, not from a model guessing at the description text.

Every listing in the corpus carries a provenance marker on its pay field. Only listings marked employer-stated count toward the numerator. Two groups are deliberately excluded: listings whose range was inferred by a model from description text, and older listings that carry a pay figure but no recoverable provenance. Both would inflate the figure, and a reader has no way to verify either. The denominator is every active listing on the same day.

The published figure is frozen to one resolved study rather than recomputed hourly, so anyone citing it gets a stable number with a stable date. The live corpus figure keeps moving as listings turn over, and the two are expected to stay close to each other.

Our USD premium index is 38x as of 2026-10-07. Definition: how many times higher median annual pay is on a USD-paying remote listing than on an Indonesian-market listing for the same broad role family.

Both sides use medians rather than averages, so a single outlier salary cannot move the result. On the remote side we take active listings priced in US dollars and annualize whichever pay period the employer quoted. On the Indonesian side we take locally advertised monthly salaries and annualize those, converting at the exchange rate recorded alongside the calculation. Implausible ranges are dropped on both sides before any median is taken.

A role family is published only when both sides carry at least thirty observations; thinner families are dropped entirely rather than shown with a footnote. The headline is the median across the families that cleared that bar, so it describes an ordinary role rather than the best-paid one.