How old a job already is when it reaches you, board by board

Is the date on the board real?

Mostly, yes. On most boards in our data, the board's date sits close to the day the posting first became visible to us. What the date leaves out is how long the posting sat somewhere else first.

Oink's data compares the date the board reports with the moment our database first stored the posting. I call the difference age at first receipt: how old a posting already is when a job seeker's tool first sees it. It does not measure when the employer opened the role, and it says nothing about when the role closes.

Without LinkedIn, across 15,850 dated postings, 66.1% arrived the same day, 25.6% within a week, 5.2% within a month and 3.1% later than that. On most boards in our data a daily sweep sees a posting on the day the board dates it, and the exceptions are specific boards.

Which boards show a job first?

Same window, 28 August to 27 September 2026, split by how old each posting was on the day Oink first stored it. Cells under 10 postings are withheld and shown as "·", boards under 20 are left out, and 1,284 rows were excluded because the stored text later replaced the original date, leaving 59,588 postings across 30 sources.

Nine smaller boards had under 100 postings each, too few to read much into. The eight with dates (Jobgether, SmartRecruiters, Fashionjobs, Dynamite Jobs, Ashby, Himalayas, Academic Positions and a university maths PhD page) arrived the same day or within a week; most of Devex's carried no date.

What should you do differently on each board?

Here is my reading of the table. Four patterns stand out.

Boards that publish on the day they date. Tecnoempleo, Just Join IT, ReliefWeb, Remote Rocketship, Hacker News, Wellfound and EURAXESS put 80% or more of their postings in the same-day column. A daily sweep there sees what is new. For the Hacker News thread there is a guide on searching it by hand.

Boards where a week's lag is normal. EURES puts 51% of its postings one to seven days behind and 18% more than a month behind. No Fluff Jobs, jobs.ac.uk and the shared feed of Remote OK, Remotive and We Work Remotely (76% one to seven days, 17% eight to thirty) have the same shape, and Remocate (24% over thirty days), Welcome to the Jungle (11% at eight to thirty, 4% over thirty) and Y Combinator (17% and 6%) carry a long tail. On the first four, a posting is often several days old by the board's own date when it first turns up. On all seven, go by the printed date and check how long it has been up before you apply.

Boards with no date in our data. Impactpool shows no date on 95% of its postings and Built In on 64%. Of the Devex postings, 41 reached Oink with no date. You have to open the posting for a closing date, and on Devex the full description stays behind its paid Career Account.

Open-web search finds old pages. Of the 142 postings our web search found, 41% were more than a month old. A page a search engine found is undated until the page says otherwise.

On LinkedIn, 72% same day still leaves 28% up to a week old when they reached us, and LinkedIn's own job alerts run daily or weekly, so if you rely on them, set them to daily. There's a guide on getting past the 100+ applicants wall.

The table measures when a posting started, not how long it stays open. Unlisted's Ghost Jobs Report for September 2026, computed on 30 September 2026 from 668,355 open postings across 15 applicant-tracking systems, found 28.0% of dated postings open for more than 90 days from the employer's own publication date, with a median of 35 days. It reports rates only and never labels a single posting a ghost job. The ghost-job checklist takes two minutes to check one posting.

What surprised me about job boards

I'm an ML engineer, and job hunting felt like a second job: you search for openings, sometimes while still going to your main job, then read each posting to work out whether you even fit. That's why Oink exists.

What surprised me most reading boards is that there really are boards tied to a place or field, especially for researchers and regions where a local board beats LinkedIn. When a new user joins, Oink analyses their case and prioritises the relevant boards, and I keep adding more.

What does Oink do with the date, and where does it fail?

Oink pulls jobs from several public sources, removes duplicates, screens out explicit conflicts with your location, work arrangement or seniority, and ranks the rest against your brief, and your resume if you add one, on a 0 to 5 scale. Which boards feed a search depends on what each supplies, how relevant it is, how often its jobs were delivered before, and its health, so no board has to show up every day (where Oink finds jobs). A posting takes the board's date when there is one. At send time, postings older than 21 days are dropped (7 days for open-web search results); undated postings pass through except from Devex, EURES and open-web search, where a missing date rejects it. Every card links the original posting.

In the 30 days to 27 September 2026, 1,760 matched postings hit the final check before sending. 18% were dropped as too old, a dead link or closed. Another 16% failed because a liveness check couldn't confirm the job was open, though some of those might have been fine. Oink can't guarantee a card is fresh or that the role is still open. Think of the card as a reason to open the original page today.