The Apply-Window Score, Read From Posting Age and Fill Speed

You found a posting. The question is not whether it is a good job - it is whether the early-review window is still open, and how fast you have to move. This guide gives you a two-variable score you can run on one posting in under two minutes: its true age crossed against its fill-speed class, producing a concrete action - apply now, apply within a set window, or deprioritize. It is for people running a search across many companies at once, where you cannot afford to sprint at every listing or to write off every posting that looks a week old.

The blanket advice - "apply within 48 hours" - is an average that hides the variance. A warehouse posting and a director search decay at completely different rates. Applying the same rule to both means racing a role that has months of runway and skipping a role that was already dead on day three. The fix is to score the posting, not the calendar.

Why one "apply within 48 hours" rule fails

A single timing rule fails because postings decay at different speeds, and the number you can see is not the number you should act on. The headline figures are real but they describe different clocks, and confusing them is the root mistake.

Three separate things get called "how long a posting lasts," and they disagree:

• Employer time-to-fill. How long from posting to hire. One benchmark puts the cross-industry average at 42 days, up 24 percent from 2021.
• Posting duration. How long the listing stays live. A 125M-application dataset found a median of seven days, often two or three; a live-crawl index reports 20 to 30 days overall.
• Visible age. How old a posting looks when you find it. A live index put median visible age at about 9.6 days.

These are not contradictions. They are three points on the same lifecycle. The employer's clock runs longest; the posting comes down somewhere inside it; and by the time you see the listing, it is usually already partway through. The practical consequence is blunt: when you find a role, you are typically about a third of the way into the employer's process, not at the starting line.

So speed matters. But how much runway you have left depends on which of these clocks the role is running, and that is what the fill-speed class captures.

The two variables that decide the window

The Apply-Window Score reads exactly two variables: the posting's true age and its fill-speed class. Age tells you how much of the window is spent; class tells you how big the window was to begin with.

Age is not what the aggregator says. LinkedIn and other boards lag the source. Direct career-page monitoring runs roughly three days younger at the median than public boards, because the board only shows the role after it finishes indexing it. So "posted 2 days ago" on an aggregator can mean five days on the company site. Always timestamp from the ATS listing.

Fill-speed class sorts roles by how fast they close:

• Fast: hourly, warehouse, seasonal, admin. These fill in roughly 20 to 30 days on the employer clock, and the visible window is short.
• Medium: most professional individual-contributor roles, recruiting, mid-level.
• Slow: senior, technical, executive. Healthcare, technology, and financial services roles regularly take 45 to 65 days; mid-level, technical, and executive searches can run 60 to 90-plus days.

The dossier's two research camps disagree about which variable dominates. One camp says age matters most: across sectors, visible ages cluster in a tight 7-to-12-day band, with the fastest and slowest sectors only about 1.6x apart (7.1 versus 11.4 days). The other camp says class dominates: retail and hospitality at 20 to 30 days versus healthcare, tech, and finance at 45 to 65 days is a real gap on the employer clock. The right move is not to pick a side. Score both, and deprioritize if either one fails. A role that is old and fast-class is dead. A role that is young or slow-class is worth a look.

Median posting duration by role class

Fast-class roles come down first, so their apply windows close first. Here is the visible-duration data by class from a single live-crawl index, which you can use as your default when you have no other read on a role.

Read these as visibility estimates from one index, not employer publish-to-close records, and not gospel. They tell you the shape: hourly and admin roles vanish fast, retail and hospitality linger longer than their reputation suggests. The number to distrust most is the blanket "48 hours" - it fits the warehouse row and badly overstates urgency for a 33-day retail posting.

The channel gap most people miss

The channel you find a role on changes the window more than its industry does. The same job posted to two platforms shows completely different lifespans, because employers run shorter posting cycles on some boards and repost.

In one France study, computer and math roles showed a LinkedIn median of 17 days against an Indeed median of 3 days - the same category, a 5.7x visibility gap driven purely by channel. Transportation ran 21 days on LinkedIn versus 1 day on Indeed; education ran 26 versus 4. The lesson is that "I saw it three days ago on Indeed" and "I saw it three days ago on LinkedIn" are not the same read on remaining window. Faster-cycling channels reward faster action.

The applicant-count badge and what it actually proves

The "Over 100 applicants" badge proves the click-flood has passed - not that the shortlist is set, and not your odds. It is engineered to mislead in both directions, so treat it as one weak signal, never a gate.

Here is what the visible signals prove and how they lie:

• "Over 100 applicants" counts apply-button clicks, not finished or qualified applications. It is a display cap introduced at the end of 2023, so a role at 101 clicks and one at 4,000 look identical. It proves the early click-flood is over; it lies when you read it as your competition, because 20 to 25 percent of clickers never finish registration and 40 to 60 percent of applications are rejected for not meeting requirements. The real field is a fraction of the badge.
• "Be an early applicant" appears when you would be among the first ~25 applicants. It proves you are plausibly in the first-read wave. It rarely lies, but it is fleeting - it is the strongest apply-now signal you can see for free.

Score the window: the procedure

Run this in order on the posting in front of you. The whole pass is under two minutes; only the tailoring step in the middle takes real time, and you do that once you have decided the role is worth it.

The scoring step is a lookup, not a judgement. Use this grid.

Steps four and six are where most of the value sits: the score turns a vague "should I rush?" into a named action, and the tailoring step protects the timing edge you just earned. Being first in with a generic resume still loses.

Where this goes wrong

The score fails in predictable ways, almost all of them caused by trusting a single number that was never built to be trusted. Here are the failure modes and how to check each one against a real posting.

Two more traps do not fit the table cleanly. First, speed without tailoring: first-in with a generic resume loses to a slower, keyword-matched application, so never trade the tailoring pass for a few minutes of speed. Second, stale time-of-day advice: the widely cited "apply around 10 a.m., Monday beats Saturday by 46 percent" figures trace to a study from around 2019 with unverified recency. Treat time-of-day as directional at best - submit during business hours if you can, but do not hold a finished, tailored application overnight to hit a "peak" the data may no longer support.

Why niche-skill roles widen the window

A niche-skill role justifies a wider apply window than any headline rule suggests, because it draws from a far smaller candidate pool and therefore fills slowly. This is the one case where a visibly old posting is often still worth a full-effort application.

The size of the effect is large. In Refolk's index, 347,096 US profiles hold a "Software Engineer" title, but only 3,465 list Rust as a skill - roughly one-hundredth of the broad pool.

A role that can only be filled from a pool a hundred times smaller cannot close in a week. Fewer qualified applicants arrive per day, the shortlist takes longer to assemble, and a strong late applicant is still a live candidate. When you classify a posting, treat a genuinely rare skill requirement as a slow-fill signal in its own right, even if the title reads generic. The pool-size gap also travels across borders: in Refolk's index, the same Software Engineer title covers 347,096 US profiles against 21,889 in Germany, a 15.9x difference, so the same role fills at different speeds in different markets.

If you want to see how thin a specialist pool is before you decide how hard to push on a posting, Refolk can size it from its index of professional profiles in one query.

Keep the score current and beat the crawl lag

The score is only as fresh as the postings you feed it, and default alerts feed you stale ones. Alert infrastructure structurally removes the very speed advantage the data rewards: index-then-batch pipelines add 18 to 48 hours before you are notified, yet nearly half of applications land in the first 48 hours.

One test measured LinkedIn averaging 31 hours of lag between a career page posting and the in-app push. Against a 48-hour application flood, that means default-alert users arrive after the peak has passed. LinkedIn is also capped at 20 saved alerts and defaults to daily or weekly email, which compounds the delay.

The fix is to skip the index for your highest-priority targets:

1. List your top target companies and bookmark each one's careers/jobs page directly.
2. Check those pages on your own cadence (daily for fast-fill targets, twice a week for slow).
3. Set daily - not weekly - LinkedIn and Indeed alerts for your role and location as a backstop.
4. For each hit, timestamp from the career page, not the alert email.
5. When you find a fresh fast-class role, run the score and apply the same day.

Before you call any single application done, run this check.

The score does not promise you the job. It promises you spend your limited hours on the postings where timing still helps, and stop sprinting at roles where the window closed before you arrived. Run it on every posting, trust the ATS date over the badge, and widen the window for anything that draws from a genuinely small pool.