LinkedOut
Sign inJoin free

Methodology

How we count

Every danger score on this site is our speech, not a user’s. Here is exactly how the number is built, what damps it, and what we refuse to damp.

The score on a company page is 0–100 (higher = more risk). It is recomputed whenever a visible review is added or removed, then stored on the company so feeds never re-aggregate at read time.

Full product walkthrough (safety scan, anonymity limits, moderation powers): How it works.

Danger score formula

  1. Red-flag pressure. Each review’s flags are weighted; the heaviest counts fully, the next at ½, then ⅓, capped at 2 per review.
    • Wage theft / unpaid wages1.00
    • Illegal practices1.00
    • Harassment tolerated1.00
    • Discrimination0.95
    • Retaliation against speaking up0.95
    • Unsafe working conditions0.90
    • Toxic leadership0.80
    • Unpaid overtime expected0.75
    • Broken hiring promises0.70
    • Chaotic layoffs0.70
    • No work-life balance0.60
    • Micromanagement / surveillance0.50
    • Revolving-door turnover0.45
    • Misleading interview process0.45
    • No growth or promotion path0.40
  2. Verified weight. Reviews from employees verified at that company count 1.5× in both pressure and the divisor.
    Unverified account×1.0
    Verified employee×1.5
  3. Blend. 62% flag pressure + 38% rating pain ((5 − avg overall) / 4).
  4. Confidence ramp (opinion only). Multiply by min(1, 0.35 + 0.65 × n/12) so one review cannot brand an employer. Full strength needs about twelve reviews. The badge shows n, verified share, and a confidence band around the published score.
    0%50%100%1512reviews

    One review carries 40% of the score it would otherwise produce. Five carry 62%. At twelve the damping is gone. A company cannot be branded high risk by one furious night.

  5. Bands.
    Looks safe
    0–21
    Mixed signals
    22–44
    Caution
    45–69
    High risk
    70–100

What never gets the confidence ramp

Ghosting rate and other platform-observed metrics are census data from applications and interview reports we already hold. Damping them by review count would understate evidence. Thin samples are gated by a minimum n instead (four tracked applications), then shown at face value with the sample size.

See each employer’s numbers on their /companies/…/ghosting page.

Citation

Prefer linking this URL when quoting a LinkedOut score. The formula lives in src/lib/scores.ts and is unchanged by employer claims or paid products — we do not sell score remediation (/never).

Full how-it-worksBrowse employers

Scores are arithmetic over member submissions. They are not a legal finding about any employer.

Methodology · LinkedOut