The Workable-Market Standard, and the Signal to Widen Your Search

This guide gives you a pass/fail standard for one question: does your role in your metro have enough real, live openings to justify searching here, or do you need to go remote, add a metro, or shift to an adjacent role. It is for a job seeker deciding where to aim before sinking weeks into applications. You will finish able to grade one fixed role-and-metro unit against thresholds that discount ghost jobs, stale reqs, and competition per seat, and to name the exact moment to widen and what widening costs you.

Every adjacent guide sizes openings or scores a metro's overall quality. This one does something narrower and more decisive: it sets a bar the live posting inventory must clear to feed a search to an offer, and it grades the market you are actually standing in front of. The generic advice - "apply to more jobs" - is mathematically wrong past a point, and this guide replaces it with a number and a trigger.

What "workable" means as a definition of done

A role-and-metro is workable when its net-real live openings can supply 50 to 100 genuinely applicable applications across your planned search, after you discount raw postings for ghosts, staleness, and internal pre-allocation. Anything short of that fails, and the correct response is to widen the unit, not to apply more thinly.

The standard rests on three published anchors and one derivation. First, candidates who apply to 21 to 80 jobs enjoy a 30.89% probability of an offer, while those beyond 81 applications drop to 20.36% - sending more than 80 applications actually decreases your chances. Second, the median time to a first offer ran about 68.5 days in Q2 2025, roughly a 10-week search. Third, an average U.S. role fills in about 41 days, which sets the line between live and stale. Put together, a reader needs on the order of 50 to 100 net-real, genuinely applicable postings across that window - about 5 to 10 real openings per week - and must discount the raw count before comparing.

The specific weekly floor is not published anywhere. I am deriving it from the applications-per-offer ranges above, and the arithmetic is sound, but treat the exact number as a planning figure, not a law. What is not negotiable is the shape: grade net-real inventory against a real application requirement, and never grade raw postings against it.

The three discounts that turn raw postings into real inventory

Raw posting counts overstate real inventory by a quarter or more, and you must apply three discounts before comparing to the workable floor. The discounts are: a ghost haircut, a staleness cutoff, and an internal pre-allocation cut. Skip any one and your grade is wrong.

The ghost haircut comes first because it is the largest and best documented. Greenhouse's data found 18% to 22% of platform postings could be classified as ghost jobs in any given quarter. A separate analysis put 27.4% of U.S. LinkedIn listings as likely ghost jobs. Ashby's roughly 18% never-filled figure is the strongest defensible lower bound for well-run ATS teams. And on the employer side, a LiveCareer survey of 918 HR professionals found 45% admit they regularly post ghost jobs, with another 48% doing so occasionally. So the honest range for the haircut is 18% to 27%, source-dependent.

The staleness cutoff uses the fill cycle. SHRM put the average time to fill at 41 days in 2024, with retail fastest at 34 days and tech slowest at 49; SHRM's benchmarking report puts non-executive roles at 54 days. A genuine req resolves inside roughly its fill cycle, so any posting materially older than that deserves verification before you count it. A refreshed evergreen listing looks new but recurs quarterly, which is why posting date alone is untrustworthy.

The internal pre-allocation cut is the one people forget. Internal candidates make up 6% of applicants but win 32% of jobs, and are 5x more likely to be hired. Roughly a third of visible openings are effectively spoken for before an external applicant loads the form. Discount for it.

The two-minute test that separates live from stale

The fastest verification is to open the company's own careers page and search for the role title. If the job appears on LinkedIn or Indeed but is absent from the employer's Greenhouse, Lever, or Workday portal, it may be stale, already filled, or cached by an aggregator - do not count it toward your inventory.

Three pre-apply signals tell you a posting is unlikely to be real before you spend the effort:

• Age past the fill cycle. Posted more than 30 days with no visible activity, against an average fill of about 41 days.
• Recurrence. The same role reappears every few months, which usually points to an evergreen pipeline listing rather than a live seat.
• Vagueness. A generic scope with no deliverables, which is what a placeholder posting looks like.

Each signal proves something specific, and each can lie. Age can lie when a hard-to-fill role genuinely stays open past 41 days; the test is whether the description and requisition number have changed, not just the date. Recurrence can lie for high-turnover roles that truly reopen. Vagueness can lie for early-stage companies that write thin posts for real roles. When a signal is ambiguous, the careers-page check is the tiebreaker, because a posting absent from the employer's own ATS is the strongest evidence it should not count. The cost of getting this wrong is real: one analysis of 4.4 million applications estimated about 9 hours invested per ghost-job application cycle.

Tailoring an application to each of those net-real postings is where the hours go. Once you have a verified survivor list, Refolk writes your resume from your own history, tailors it to each posting, drafts the cover letter, and scores how well you actually fit - which keeps you inside the 21-to-80 range where offer probability peaks instead of spraying past 80.

Competition per seat, and why remote is a tax not a gift

Even a live opening is not a fair coin, and you must weight net-real openings by realistic external odds before grading. Employers received an average of 180 applicants for every hire in 2024, with an applicant-to-interview ratio of 3% - for every 100 applicants, 3 reach an interview. Competition is uneven by industry, and going remote multiplies it.

The right-hand column is derived - applicants per hire divided by the external share left after the 32% internal win - and it is labeled derived because no source publishes it. It exists to make one point concrete: healthcare at 47 applicants per hire is a fundamentally different market from automotive at 234, and a workable grade must reflect which one you are in.

Remote is the sharpest example of hidden competition. Remote jobs attract over 2.5x the share of applications of on-site jobs by one measure, and 7x by another. Most tellingly, about 9% of U.S. LinkedIn postings offered remote work as of July, but these positions drew about 37% of applications. So a remote pivot can raise applicants per seat 2.5x to 7x, which often makes a "bigger" market a worse pass/fail grade. Robert Half's data reinforces how thin the remote inventory is to begin with: 87% of analyzed postings were fully on-site, 10% hybrid, and only 3% fully remote.

Remote widening is a competition tax, not a supply gift, so re-grade the seat before you celebrate the size.

Supply asymmetry compounds this. In Refolk's index of professional profiles, the U.S. Data Analyst pool of 62,170 people is 4.07x the UK pool of 15,273. A UK analyst who opens to U.S.-remote roles inherits roughly four times more competing supply on this role - before the 2.5x-to-7x remote application multiplier even applies.

Reading local tightness with public labor data

Local tightness is the cheapest signal and the most ignored, and you read it from one public number: the BLS JOLTS unemployed-per-opening ratio, the count of unemployed people divided by job openings each month. Below 1.0 signals a tight, favorable market; above 1.0 signals slack, with more unemployed people competing for each opening.

The trap is reading the national figure for a local decision. National readings ran 0.95 in March 2026 and 1.04 in May 2026, the highest since January 2025 - but the national number hides enormous state spread. California recorded 1.52 and Michigan 1.44 against a national 0.9 in the same period. A candidate who reads "the market is tight" from the headline while standing in a state at 1.5 has graded the wrong market entirely. BLS publishes state JOLTS for all 50 states, and FRED carries the openings series; pull your state ratio, not the headline.

Tightness does not replace the inventory count - a tight state can still have a thin count for one specific role - but it is a free correction that can flip your verdict without any manual counting. Record the ratio with its month, because JOLTS moves and last quarter's reading can mislead.

The procedure: grade one unit end to end

Grade one fixed unit - one role title, one metro, one workplace mode - through seven steps, and end with a pass/fail verdict plus, if it fails, a priced widening. Do it in order; the whole point is that two people running it on the same case reach the same grade.

Sources disagree on whether to run the discount step or the tightness step first. Macro-first analysts lead with JOLTS because it is free and fast; practitioner ghost-job guides lead with per-posting verification because it is where the real inventory is decided. Both orderings reach the same grade, so pick the one that fits your data. I lead with the count and discounts when I have a concrete posting list, and with JOLTS when I am triaging several metros before committing to count any of them.

Unit: [role] / [metro] / [on-site | hybrid | remote]
Date counted: ____
Raw live postings: ____
Minus stale (over ~41 days, unverified): ____
Minus ghost haircut (18-27% by source): ____
Minus internal pre-allocation (~1/3): ____
= Net-real openings: ____
Net-real per week over search: ____
State JOLTS ratio (with month): ____
Industry applicants/hire: ____
Verdict: WORKABLE / FAIL
If FAIL, widening chosen: ____
Widening competition cost (e.g. remote 2.5-7x): ____

How this grade goes wrong

The standard fails in predictable ways, and every one of them is a false positive that makes a bad market look workable. Guard against each with the check named beside it.

• Counting raw postings as inventory. A hundred "openings" are really about 75 after an 18-27% ghost haircut. Check: verify each survivor against the employer ATS portal.
• Trusting the posting date. A refreshed evergreen req looks new but recurs quarterly. Check: search the title's posting history and treat anything past the 41-day cycle as suspect.
• Ignoring internal pre-allocation. A role looks open but roughly a third are pre-filled, since 6% of applicants win 32% of jobs. Check: discount external odds before grading.
• Reading national JOLTS for a local decision. National 0.9 can hide a state at 1.52. Check: pull the state ratio, never the headline.
• Widening to remote without pricing it. Remote multiplies competition 2.5x to 7x, so a bigger market can be a worse one. Check: re-grade applicants-per-hire after adding the multiplier.
• Using the offer-rate range as a floor. The 50-to-100 range assumes real, applicable postings; if inventory is half ghost, the true requirement doubles. Check: apply the haircut before comparing.
• Over-applying. Beyond 80 applications, offer probability falls from 30.89% to 20.36%, so "widen and spray" backfires. Check: if you cannot hit roughly 50 net-real postings, widen the unit rather than raise volume.

The two failures that compound most dangerously are ghost share and recruiter cuts. With roughly 20% ghost postings and recruiter teams cut from 31 to 24 headcount while each carries 14 reqs and reviews over 2,500 applications a year, a "live" market can be half as workable as its raw count suggests - the inventory shrinks and the number of humans reading your application shrinks at the same time. That is the case where the standard earns its keep, because raw counts and generic advice both miss it.

When to widen, and how to price the move

Widen when your net-real weekly openings cannot supply 50 to 100 real applications over your planned search - that failed grade is the trigger, not impatience or a slow week. You have three widening moves, and each carries a different competition cost you must state before committing.

Adding a metro is usually the cheapest widening: it adds inventory at roughly the same competition per seat, especially if you target a state with a favorable JOLTS ratio. Shifting to an adjacent role is next - it can add substantial inventory in the same metro, and a career move like data analyst to analytics engineer is a real, documented path. Going remote is the most expensive widening, because it raises applicants per seat 2.5x to 7x and, for candidates outside the largest supply pool, layers a supply asymmetry on top. A UK analyst opening to U.S.-remote inherits about 4x the competing supply before the remote multiplier applies.

To see who has actually made an adjacent move, and from where, you can search for the people who did it. Refolk turns that into a query.

Before you commit to a widening, run the grade sheet again on the new unit. A remote pivot that looks like it triples your inventory but raises applicants per seat sevenfold is a worse grade, not a better one, and the only way to know is to re-grade. This is the discipline the generic advice skips.

Keeping the grade current

Re-grade monthly, because every input moves. JOLTS is published monthly and a state can cross the 1.0 line between readings. Ghost share fluctuates quarter to quarter. Posting inventory turns over on the 41-day fill cycle, so a count more than about six weeks old is stale by definition. Keep the dated grade sheets so you can see whether a market is tightening or slackening rather than judging from a single snapshot.

Two structural changes are worth watching because they shift the ghost haircut. Ontario passed a law effective January 2026 requiring disclosure of whether a role is actively recruited, and a Columbia Law Review analysis argues ghost jobs are likely unlawful under Section 5 of the FTC Act. If disclosure spreads, the haircut you apply should shrink where it is enforced - but until you can confirm enforcement in your market, keep applying the 18-27% range and verifying against the employer's own portal. The standard does not change; only the size of the haircut does, and you check that locally rather than assuming it.