Skip to content

Methodology

How HirQ scores jobs and companies

Every number on HirQ comes from employers' own public job boards. This page shows exactly what we measure, how much each signal counts, and where the data falls short.

Scores recomputed 6m ago Methodology updated Nov 20, 2026By the HirQ Data team
Open jobs scored
679,099

1,246,911 tracked in total

Company job boards
26,390

22,257 with open roles

Median recheck
5.6 h

90% of boards at least daily

Closed in 30 days
365,780

Removed when the employer takes them down

Where the data comes from

HirQ reads job postings directly from the public job boards that companies run on their own applicant-tracking systems (Greenhouse, BambooHR, SmartRecruiters, Personio, Ashby, Recruitee, Lever and Workable). We check each board several times a day. That lets us see when a job first appears, when it disappears, and whether the same title comes back later.

We do not buy listings, we don't take paid or sponsored posts, and companies cannot pay to change their score. Every apply button goes to the employer's own site.

Boards we read, by hiring system

26,390 active boards
  • Greenhouse
    5,3894,929 with open jobs
  • BambooHR
    5,1342,640 with open jobs
  • SmartRecruiters
    3,8413,658 with open jobs
  • Personio
    3,7703,467 with open jobs
  • Ashby
    3,4993,066 with open jobs
  • Recruitee
    2,6942,471 with open jobs
  • Lever
    2,0531,842 with open jobs
  • Workable
    101 with open jobs

Definitions

The words we use on every job and company page, and exactly what they mean.

First seen
The first time our reader saw the posting on the employer's board.
True age
Days since the earlier of the employer's own posting date and the day we first saw the job. A refreshed “posted 2 days ago” date can't make an old job look new.
Repost
The company closed a posting and later opened a new one with the same title. Repeated reposts can mean a hard-to-fill role, a standing pipeline or a job that rarely gets filled. It is a signal, not proof.
Stale opening
A job that has been open for more than 90 days.
Closed
The posting disappeared from the employer's board. A close is not proof of a hire, which is why we look at reposts too.
Clean close
A close that is not followed by the same title reappearing soon after.
Follow-through
The share of a company's older roles that closed cleanly within the window. Our closest public signal for “this company fills its jobs”.
Ghost risk
Our estimate (low, medium or high) of how likely it is that a posting is not a live, specific opening. It is never a verdict.
Pay listed
The employer published pay in the posting. We never estimate pay that wasn't posted.
Data confidence
How much evidence we have about a company. Small companies give us little, so their scores are pulled toward the middle (below).

The Job HirQ Score

Every job gets a score from 0 to 100 that answers one question: is this posting worth applying to right now? It combines seven signals.

Current weights (admin-editable, applied at the next recompute)

  • Freshness

    25%

    Half of all new roles close within about three weeks (52.7% by day 21 in our data), so a job's freshness value halves every 21 days, with a floor of 10. If the employer does not publish a posting date, freshness is capped at 60.

  • Company hiring

    20%

    The company's Hiring Score (below): is it adding roles, and do its roles close cleanly?

  • Pay

    15%

    No pay listed scores 20. Listed pay scores 60 to 100 depending on where it sits against similar US postings: the same role in the same state first, then the same category in the state, the category nationally, then the seniority level. We need at least 10 comparable postings; otherwise listed pay scores 75. A very wide range (top more than twice the bottom) loses 10.

  • Ghost risk

    15%

    Low scores 100, medium 50, high 0. See ghost risk below.

  • Repost history

    10%

    How often the company closed and reposted this same title in the last six months: never 100, once 60, two or three times 35, four or more 10.

  • Description

    7.5%

    A real description of the work scores well; a very short one (under 500 characters) scores 20. If we could not read the description, this part is neutral (50).

  • Clarity

    7.5%

    Whether the posting states its remote policy, location and visa sponsorship.

Score bands

We always show the number and the word together.

Excellent80+
Strong65–79
Fair50–64
Weak35–49
Poorunder 35

How today's open jobs score

  • Excellent 1.8%12,490
  • Strong 14.5%98,784
  • Fair 31.0%210,462
  • Weak 30.1%204,581
  • Poor 22.5%152,658

Share of 678,975 open jobs. Excellent is rare by design: it takes a fresh posting with listed, competitive pay at a company that fills its roles.

The Company Hiring Score

Is this company actively and reliably hiring? Six signals, each from what we see on its own job board.

Current weights

  • Activity

    25%

    New roles in the last 30 days, both relative to how many roles the company has open and in absolute terms.

  • Follow-through

    20%

    Of the roles opened a while ago, the share that closed within the window without being reposted under the same title soon after. A close followed by a repost is churn, not a hire, so it doesn't count. The window is 30 days until we have 60 days of history (Dec 5, 2026), then 60 days.

  • Reposts

    15%

    Share of recent roles that are reposts of titles the company recently closed. Lower is better.

  • Freshness of openings

    15%

    Share of open roles that are not older than 90 days.

  • Pay transparency

    15%

    Share of open roles that list pay (US roles, when the company has at least 3).

  • Momentum

    10%

    New roles in the last 30 days compared with the 30 days before.

Data confidence

Small companies give us little evidence, so their score is pulled toward the middle: confidence = n / (n + 10), where n is the number of roles we have observed recently (open roles plus roles first seen in the last 90 days). Score = 50 + (raw − 50) × confidence.

A board that returns the maximum number of jobs its system allows (about 5,000) has its confidence halved and follow-through ignored, because jobs falling off the end of the list look like closes. Companies with confidence under 0.4 are shown as “Low data”.

Labels

The first rule that matches wins.

  1. 1

    Frozen7,347 companies

    No new roles in 90 days, or none in the last 60.

  2. 2

    Low follow-through2,385 companies

    At least 10 open roles and either 60%+ of them are older than 90 days, or half of recent roles are reposts (with 10+ new roles), or almost none closed cleanly while 30%+ are old. It describes the postings, not the company's intent.

  3. 3

    Hiring fast1,130 companies

    10+ new roles in 30 days, many relative to its size, not slowing, score 60+.

  4. 4

    Actively hiring4,692 companies

    3+ new roles in 30 days, a reasonable share of its size, score 50+.

  5. 5

    Slowing1,102 companies

    At least 5 new roles in the previous 30 days and 40%+ fewer now.

  6. 6

    Steady6,750 companies

    Everything else.

Companies by label today

Hiring fast: 1.1K (5%)Actively hiring: 4.7K (20%)Steady: 6.8K (29%)Slowing: 1.1K (5%)Frozen: 7.3K (31%)Low follow-through: 2.4K (10%)
23,406companies
  • Hiring fast1.1K · 5%
  • Actively hiring4.7K · 20%
  • Steady6.8K · 29%
  • Slowing1.1K · 5%
  • Frozen7.3K · 31%
  • Low follow-through2.4K · 10%

Ghost risk

We add up warning signs that a posting may not be a live, specific opening.

  • The title is a general talent pool or unsolicited application (“General Application”, “Talent Community”): +4.
  • Open a long time (only when the posting date is known): over a year +3, 181–365 days +2, 91–180 days +1.
  • The same title was closed and reposted: three or more times +2, once or twice +1.
  • The company has the Low follow-through label: +2 if the job is older than 30 days, otherwise +1. A Frozen company: +1.
  • No pay and a very short description: +1.
High at 4+ pointsMedium at 2–3Low at 0–1

Ghost risk across open jobs today

  • Low 41.1%279,153
  • Medium 29.4%199,941
  • High 29.4%199,881

Outside studies put ghost listings at 18–30% of postings. Our high-risk share is driven mostly by very old postings at frozen companies and old postings at low-follow-through companies.

A high ghost risk does not mean a posting is fake. It means the evidence suggests it may not be a live, specific opening, so it's worth checking with the company before investing time.

About these estimates

HirQ's ghost-risk estimate and company labels are automated opinions based on what we observe on employers' public job boards: how long a listing has been open, whether it was re-listed, wording in the description, and how often roles on that board close. They are not statements about an employer's intent, and a high-risk listing can still be a real opening. Data is refreshed several times a day; figures are as of the date shown. Employers can request a review.

How we compare pay

Pay comparisons only ever use pay that employers posted. We don't estimate or guess salaries.

  • Source: US jobs with a posted pay range in US dollars, first seen in the last 365 days or still open.
  • Each posting counts once at the midpoint of its range. Hourly pay is annualized at 2,080 hours (40 hours × 52 weeks).
  • Ranges wider than 4× (top divided by bottom) are dropped, and so are values outside $15,000–$1,000,000 a year.
  • Mass-posted requisitions count once: we de-duplicate by company, title, state and range.
  • For a job's pay signal we compare against the same role in the same state, then the category in the state, the category nationally, then the seniority level. We need at least 10 comparable postings.
  • Salary pages show the 25th, 50th and 75th percentiles and are published only when at least 10 postings back them.

See the results on salary pages, or check a number with the salary checker.

Limitations

What our data can't tell you, so you can weigh the scores properly.

  • We only see public employer job boards. Jobs posted only on LinkedIn, Indeed, agencies or internal boards aren't here.
  • Our close history starts in early October 2026. Until it covers 60 days (5 Dec 2026), follow-through uses a 30-day window.
  • Some boards don't publish a posting date. For those jobs age starts when we first saw them, and freshness is capped.
  • Some hiring systems rarely put pay in a structured field, so a job can show “Pay not listed” even when pay is written further down the description.
  • “Not mentioned” is not “no”: when a posting says nothing about visa sponsorship, we don't treat it as a refusal.
  • A close is not a confirmed hire, and a repost is not proof of a fake job. Both are signals.
  • Labels are recomputed every 6 hours, so a company close to a threshold can move between labels from one day to the next.
  • Boards that hit their system's listing cap can hide closes. We lower our confidence for those companies.

Changelog

Changes to weights, bands or definitions are listed here.

  1. Scheduled

    Follow-through window moves from 30 to 60 days

    By this date we will have 60 days of close history for every board, so follow-through will look at roles closed within 60 days instead of 30.

  2. HirQ Score v1

    Job HirQ Score (seven signals), Company Hiring Score (six signals with data-confidence shrinkage), points-based ghost risk and six company labels, all recomputed every 6 hours. Pay comparisons use employer-posted ranges only.

  3. Job categories and company sectors

    Every job gets one of 35 categories from title and department rules. Companies get one of 26 sectors from name rules, their mix of job categories, and, where rules could not decide, a language model that saw only the company name and its job titles.

Corrections and employer reviews

Employers: request a correction

If a label, sector or job looks wrong, tell us. We check it against your public job board, fix what we got wrong, and explain what we saw when the data is right.

Request a review

Job seekers: report wrong data

A closed job still showing, pay read wrong, a duplicate? Send us the link. Fixes to how we read data are added to the changelog above.

Report a problem

What we never do

  • Use a company's ads, payments or reviews in a score.
  • Guess salaries. Pay comparisons only use pay that employers posted.
  • Penalize a job for our own gaps: when a board doesn't publish something (a description, for example), that part of the score is neutral.

Figures on this page are live: 679,000+ open jobs, counts updated 1h ago.