This comes from the 324,634 open jobs we track on the careers pages of 10,630 companies that use Greenhouse, Lever or Ashby, checked every day.
Of the 239,310 jobs posted in the last six months:
| Day | Share of new postings |
|---|---|
| Monday | 16.5% |
| Tuesday | 20.5% |
| Wednesday | 20.9% |
| Thursday | 20.1% |
| Friday | 19.1% |
| Saturday | 1.9% |
| Sunday | 1.0% |
Recruiters post in the middle of the week. Monday is quieter, probably because it goes to meetings and reviewing the applications that came in over the weekend. Weekends are almost empty: 2.9% of postings.
So if you check for new jobs once a week, Thursday or Friday catches most of that week’s postings while they are still fresh. If you check every day, you will see most new jobs on Tuesday, Wednesday and Thursday.
This is the number that matters more. We looked at every job that came down since 8 November and measured how long it had been up:
A posting coming down does not always mean the job was filled, but it does mean the company stopped taking applications for it. A third of jobs give you less than two weeks. Recruiters also tend to start screening the first applicants while the ad is still collecting more, so a strong application on day three is read with more attention than the same application on day twenty.
We don’t publish an hour-by-hour chart because the posting times we see are in UTC and the companies span many time zones, so any “post your application at 9:14 am” advice from this data would be made up. What we can say is that sending an application on a Sunday night is fine: it will be in the pile on Monday morning either way.
Big job boards pick up a company’s posting hours or days after it appears on the company’s own site, and often show it as “reposted” later. Going to the source is faster.
“Posted” is the date the employer published the job (Greenhouse’s first-published date, Lever’s creation date, Ashby’s published date). Weekdays are counted in UTC for jobs posted in the last 180 days. A job counts as closed when it disappears from its board; we have recorded closures since 2026-11-08. Numbers refresh monthly, last on 2026-11-15. A free monthly sample of the data is on Kaggle.
]]>We can put a number on it. Every day we check the public job boards of 10,630 companies that hire through Greenhouse, Lever and Ashby, the software behind the careers pages of most tech companies and funded startups. On 15 November 2026 those boards listed 324,634 open jobs.
| Posted more than | Share of open jobs |
|---|---|
| 60 days ago | 51.7% |
| 90 days ago | 40.9% |
| 180 days ago | 25.8% |
| a year ago | 14.0% |
The typical open posting went up 65 days ago, and 83,039 of them have been up for more than six months.
Now compare that with the jobs that actually came down. Postings that closed since 8 November had been up for a median of 33 days, and 48.5% of them closed within a month. A role that someone is working to fill rarely sits there for half a year.
Some roles stay open all year on purpose: sales reps, nurses, drivers, support agents, anything a company hires in batches. A posting from March for “Account Executive, Mid-Market” may well be real. What should make you suspicious is one very specific role (“Senior Data Engineer, Payments Platform”) that has been up for eight months.
Then there are the honest non-jobs. 2,178 of the open postings (0.7%) are “general application” or “future opportunities” ads that collect CVs for no particular role. They at least say so.
We built ATS Jobs Search for exactly this. It searches all 324,634 open jobs at those 10,630 companies, and every row comes with the employer’s posted date, the day we first saw the job, and how many new roles the company opened this week. Postings older than 180 days are left out unless you ask for them. The Closed Jobs Tracker shows which jobs came down and how long they were up.
If you only want the numbers, a free monthly sample is on Kaggle and Hugging Face.
Postings come from the public job board APIs that companies run for applicants. “Posted” is the date the employer published the job: Greenhouse’s first-published date, Lever’s creation date and Ashby’s published date. A job counts as closed when it disappears from its board; we have recorded closures since 2026-11-08. The numbers on this page refresh monthly, last on 2026-11-15.
]]>We track 324,634 open jobs on the careers pages of 10,630 companies (Greenhouse, Lever and Ashby). 62,540 of them are remote or remote-friendly, 19.4% of all postings, and 42.2% of those remote postings include a pay range.
These are mostly tech and tech-adjacent companies, so they pay above the national average. They are also the companies hiring remotely at scale.
Middle of the posted range, for role types with at least 30 remote postings that list a USD salary:
| Role | Typical pay | Postings |
|---|---|---|
| Engineering - Management | $238,000 | 274 |
| Engineering - Security | $195,000 | 481 |
| Engineering - Infrastructure | $188,000 | 472 |
| Product Management | $186,000 | 639 |
| Engineering - Frontend | $186,000 | 76 |
| Engineering - ML/AI | $184,000 | 1,427 |
| Engineering - Backend/Fullstack | $177,000 | 2,757 |
| Engineering - Mobile | $170,000 | 304 |
| Engineering - Data | $165,000 | 198 |
| Design | $162,000 | 251 |
| Finance/Legal | $155,000 | 950 |
| Sales | $154,000 | 2,928 |
| Marketing | $151,000 | 922 |
| People/Recruiting | $142,000 | 278 |
| Healthcare/Clinical | $138,000 | 1,160 |
Engineering management, security and infrastructure pay the most. Sales sits lower on base pay, but those ranges usually leave out commission, so the real number is higher.
The top of the posted range, one row per title:
Very wide ranges usually span several levels, or include equity or commission. The bottom of the range is a better guide to the offer you would get.
Some companies pay the same for a role wherever you live; others set bands by location, and remote roles get the band of where you are. Postings rarely say which. If the range is listed “for US-based candidates” or has a city next to it, ask during the first call whether the number changes with your location.
Remote means the employer marked the job as remote or put “remote” in the location (this includes remote-friendly roles). Salaries are posted ranges in US dollars per year; typos and figures outside $20,000 to $1,500,000 are left out. Role types are worked out from job titles. Talent-pool ads are excluded. Numbers refresh monthly, last on 2026-11-15. A free monthly sample is on Kaggle.
]]>The people getting interviews use AI for the slow parts of a job search (finding the right openings, judging fit, tailoring a CV) and keep the judgment for themselves. Here is a workflow that does that, with prompts you can copy.
Timing beats volume. Of the jobs that came down from company career pages recently, 33.0% were gone within two weeks of going up, and half within 33 days. An application in the first few days is read by someone who is still looking; one in week four often lands after the shortlist is made.
Job boards copy postings from company career pages, sometimes days later. Going to the source is faster:
You now have a list of 20 to 50 fresh roles instead of a feed of 2,000.
Paste your CV and the list into ChatGPT, Claude or Gemini. Keep the job rows short (title, company, location, salary, link) so the model can compare them.
Here is my CV, then a list of job openings. Score each opening from 1 to 10 for how well my experience matches it, based only on what is in my CV. For each, give one line on the strongest match and one on the biggest gap. Sort by score. Do not reward jobs just for being senior or well paid.
Then open the top ten yourself and read them. The model is good at spotting obvious mismatches; it is bad at knowing which team you would actually enjoy.
For each job you pick, give the model the full posting and your CV:
Rewrite the bullet points in my CV so the most relevant experience for this job comes first and uses the same words the posting uses where they are true for me. Do not add skills, tools or results I don’t have. Keep my numbers exactly as they are. Show me what you changed.
The “show me what you changed” line matters. Read every change. If the model turned “helped with” into “led”, put it back. Interviewers ask about the things on your CV.
Many hiring managers skim cover letters. A short note does more:
Write a 3-sentence note to the hiring manager for this role. Sentence 1: the specific thing in the posting that made me apply. Sentence 2: the one piece of my experience that proves I can do it, with the number. Sentence 3: a plain ask for a conversation. No flattery, no “I am excited to”.
Edit it so it sounds like you. If you would not say it out loud, cut it.
Use the “Apply” button on the company’s own careers page rather than a quick-apply button on a job board. Your application goes straight into the system the recruiter works in, with your tailored CV, instead of a generic profile.
A simple sheet with company, role, date applied, link and status is enough. If you run the search from step 1 on a schedule, the new jobs can land in that sheet every morning on their own.
They exist: browser agents and auto-apply services that fill in forms for you. They save time per application but send the same generic CV everywhere, answer screening questions with guesses, and can apply you to jobs you would never take. Recruiters also notice when twenty applications in a week read the same. If you use one, point it only at a short list you have already checked, and read what it submits.
Fewer, earlier, better-matched applications beat a hundred automated ones. Use AI to find new roles at the source, rank them, and tailor your CV without inventing anything. Then apply yourself in the first week.
Timing figures come from our daily check of 10,630 company job boards on Greenhouse, Lever and Ashby (324,634 open jobs; closures recorded since 2026-11-08). They refresh monthly, last on 2026-11-15. More numbers: the best time to apply for jobs and ghost jobs in 2026.
]]>