The Posting Odds Decoder, Signal by Signal Before You Apply

You are mid-scroll. A posting reads "215 applicants, posted 3 days ago, reposted," it is a role you could do, and you have to decide in ten seconds whether to spend an application on it. This guide is the row-by-row lookup for that moment: it decodes the per-posting competition and freshness signals a job seeker actually stares at, tells you what each one proves, how it misleads, and whether it raises or lowers your priority. It assumes you already know whether the posting is live and whether the employer is healthy; this is only about the numbers and labels on the posting itself.

What each visible signal actually proves

Every posting shows a handful of numbers and labels, and each answers a narrower question than it appears to. Below is the fast lookup: the signal, the one thing it genuinely proves, and the tell that it is lying to you.

Read down this table and you already avoid the two most expensive mistakes: trusting the applicant number and reading the date stamp as age. The rest of the guide is how to do each row properly.

Why the applicant count is a click meter, not a competition count

The applicant number is a tally of Apply-button clicks, not completed submissions, so its misleadingness scales with friction rather than with real competition. When a role uses Easy Apply, you finish on LinkedIn and it registers the submission, so the count is fairly accurate. When it uses the plain Apply button, LinkedIn hands you to the company's own site or applicant tracking system, loses sight of you, and counts the click because the click is the only thing it can see.

This is not a folk theory. LinkedIn's own Campaign Manager documentation keeps two separate metrics: Job Apply Clicks, meaning clicks on the Apply button, and Job Applications, meaning applications actually collected. A former recruiter put it plainly: the count is accurate only when the direct LinkedIn application option is used, because if it directs to the company page there is no way for LinkedIn to know whether anyone actually applied. The badge is also capped at "Over 100," so any large number is already a shrug.

Then screening compounds the illusion. Applicant tracking systems typically discard around 75 percent of applications, and a recruiter forum estimate holds that roughly 90 percent of applicants are wildly unqualified because many have not read the description, do not meet core requirements, or need sponsorship and are screened out immediately. Stack those and a "215 applicant" pool is realistically a low-double-digit qualified field.

The counterintuitive part: high counts inflate most on the roles with the longest external forms, because clicking costs seconds while finishing does not. So a huge number on an external-Apply role tells you about form friction more than about how many people you are actually competing against.

How to recover the true posting age

No job board volunteers the real first-seen date, because every repost resets the visible date to zero, so a "3-day-old" listing can hide a months-old search. The load-bearing signal is the posting's history, not its stamp, and you can reconstruct it in a few minutes from sources the board does not control.

The most defensible method is to cross-reference the dates the board hides. A posting carries more timestamps than the one it shows you: the original ATS posting date, the JSON-LD datePosted embedded in the page, the page last-modified header, the earliest Wayback Machine capture, and the sitemap lastmod value. A public first-seen checker surfaces all of these across more than 30 ATS platforms at once. If the Wayback Machine shows the URL existed months before the listed date posted, the job was likely reposted or refreshed. The company careers page date is usually more reliable than the board's, so following "Apply on company website" is often faster than any tool.

One caveat worth stating: LinkedIn blocks automated access to its job pages, so first-seen tools work best on the underlying ATS or careers URL rather than the LinkedIn listing. Get to the company page first, then check dates there.

Refresh or ghost: classifying a repost

A repost is not one thing, and the whole decision turns on which kind you are looking at. A huge amount of reposting is just a refresh: the old listing expires or goes stale and is posted again to bounce it back to the top of the board. That is benign, and a refreshed role can be genuinely live. A different pattern, a posting that keeps expiring and reappearing, or one sitting open for two months, is telling you it is not urgent hiring.

There is a real disagreement here worth naming. A Forbes contributor argues that a repost is often a reason to apply, because recruiters juggle 15 to 25 live roles at once, get fatigued, and repost out of genuine urgency. Ghost-job guides treat repeated reposts as a reason to deprioritize. Both are right about different cases. The tell is cadence: a single refresh on a roughly 30-day cycle points to Forbes's reading, while a recurring expire-and-reappear pattern points to the ghost reading.

Why does classification matter so much? Because the "never a hire" rate is not small, and it varies by how you measure it. Treat "roughly 1 in 5 to 1 in 3 postings never hire" as the honest band and you will not be surprised by any single method's number.

That spread is a methodology gap, not a contradiction. ATS outcome data, listing analysis, and employer self-report each measure a different question, which is exactly why the honest read is a band and not a point.

Timing: how fast your odds decay

The sharpest decay is front-loaded, so when you apply matters more than most candidates assume. A study linking roughly 80 million applications to about 10 million postings found that 41 percent of all applications arrive within the first 48 hours and 56 percent within the first 96 hours. Practitioner guidance holds that response odds start dropping noticeably after 7 days and fall sharply after 14 to 21 days, with early applicants reportedly up to 5x more likely to get a response than people who wait a week.

The mechanism explains the "be an early applicant" badge, which appears when you are among roughly the first 25 applicants and links that group to about 3x higher odds. With recruiters managing 15 to 25 reqs at once and more than half of applications landing inside four days, late arrivals compete for attention that is already spent. The 3x effect is a workload artifact, not a scoring rule the recruiter applies to you.

One number is worth being honest about: a precise "45-day odds drop" figure is not publicly established. The 45-day mark appears only as a staleness threshold inside practitioner tools, not as a measured decay point. Use it as a flag to verify age, not as a hard probability.

The step-by-step: from raw signals to apply, skip, or verify

Here is the whole procedure in order. Most of it takes under a minute; only the age check takes real time, and only when the stamp looks suspicious.

You can encode step six as a quick rubric and reuse it on every posting.

True first-seen date under 7 days: +1 / over 30 days: -1
Real qualified field looks small (Easy Apply, low count): +1 / large external count: -1
"Be an early applicant" badge present: +1 / absent: 0
Verified poster and salary shown: +1 / neither: -1
Repost pattern is a single refresh: +1 / recurring expire-reappear: -1
"Actively reviewing applicants" present: +1 / absent: 0

Once you know your real qualified field is small and the role is fresh, the remaining friction is the application itself. This is where Refolk earns its place: it writes your resume from your own history, tailors it to that specific posting, drafts the cover letter, and scores how well you actually fit, so being an early applicant does not mean sending a worse application.

How this goes wrong: the false positives

Every signal in this guide has a failure mode, and knowing the false positive is what separates reading the posting from being read by it. Work down this list before you commit to a decision.

• Trusting the applicant count. You skip a "215 applicants" role that is really about 20 qualified clicks. Check: is it external Apply? If so, treat the number as clicks and discount roughly 90 percent.
• Reading the date stamp as age. You apply to a "3 days ago" role that first appeared months earlier. Check: pull the ATS, Wayback, or JSON-LD first-seen date before deciding.
• Treating every repost as a bad sign. You skip a genuine refresh or a hard-to-fill role where you would be welcomed. Check: identical text on a roughly 30-day cycle is benign; recurring expire-and-reappear is the risk.
• Treating every repost as a good sign. You apply urgently to a pipeline-building ghost because a Forbes framing said reposts mean urgency. Check: confirm salary, a verified poster, and that the role has actually cleared before.
• Over-reading "actively reviewing applicants." You believe it guarantees your application is read or that you will be hired. Check: it only signals speed of processing, not outcome, and no LinkedIn filter surfaces it reliably.
• Assuming a count that dropped means rejection. You panic when the number falls. Check: the count is not a stable tally, since postings get refreshed, reposted, merged, or edited.
• Counting seat plurality as easier odds. You read "hiring 5" as five times your chance. Check: multiple openings can equally mean a high-turnover evergreen req; verify with posting history.

Why the same number means different competition in different markets

Two postings can show an identical "215 applicants" and represent wildly different real fields, because the supply pool behind a title depends on geography and seniority far more than on the badge. This is where a candidate should weight location and level over the raw count.

Refolk's index makes the size difference concrete.

In Refolk's index of professional profiles, the US Software Engineer pool is about 8.1x the size of the UK's. So the same applicant badge on a US role and a UK role implies very different real competition, and geography moves your odds more than any badge on the posting. Seniority tells a second story: Senior Software Engineer profiles are about 52 percent of the broad US Software Engineer pool, which means a "senior" applicant field is deep and the label alone is weak differentiation in the click pile.

The practical move is to size the pool you are actually competing in before you trust any per-posting number. If you want to gauge how fast a specific local senior market is really moving, ask directly.

Before you spend the application

Run this final check on any posting before you commit. If a line fails, drop back to the step it maps to rather than applying on faith.

Keeping this current

The signals in this guide are stable in shape but shifting in value, so re-check the mechanisms rather than memorizing numbers. Boards change how they display and reset dates, LinkedIn expands or retires labels like the responsiveness insights, and ghost-job rates move with the hiring cycle. Two habits keep you current. First, whenever a date stamp surprises you, pull the underlying ATS or archive date and confirm the board is still resetting on repost the way it does today. Second, treat the "1 in 5 to 1 in 3 never hire" band as a range to re-verify against the freshest listing-level and employer-survey figures, not a fixed constant. The method holds even when the specific numbers drift: decode the signal, name what it proves, name how it lies, then decide.