Science, Not Guesswork

How We Detect
Ghost Jobs

Our AI analyzes every job description across three research-backed dimensions and 15+ signals to calculate the probability of a posting being fake.

The Information Asymmetry Problem

When a company posts a job, they know if it's real. You don't. This is called information asymmetry — a concept that won George Akerlof the Nobel Prize in Economics (2001).

43% of companies admit to posting jobs they have no intention of filling (Clarify Capital, 2022). That means nearly half of your job applications might be going to ghost jobs — positions that exist only on paper.

GhostJob uses AI to reduce this asymmetry. We analyze the language, structure, and signals in every job description to estimate the probability of it being real.

WITHOUT GHOSTJOB
🏢 Company
Knows if real ✅
📋 Job Posting
Real or Ghost? ❓
👤 You
No idea ❌
WITH GHOSTJOB
📋 Job Posting
Real or Ghost? ❓
👻 GhostJob AI
Ghost Score: 73%
"Probably Ghost"
👤 You
Now you know! ✅

We Analyze Three Core Dimensions

Each dimension examines a different aspect of the job description to build a complete picture.

📐

Clarity — Is the job description specific?

Example Score90/100

The Clarity score measures how specific and well-defined the job description is.

  • Are responsibilities concrete?
    "Build data pipelines using AWS Glue" vs "Work on exciting projects"
  • Are requirements measurable?
    "3-5 years with Python" vs "Strong experience"
  • Does seniority match expectations?
    Junior role asking for 8 years = confused posting
  • Is language direct or buzzwords?
    "Leverage synergies" = says nothing
Based on: NLP readability analysis, Flesch-Kincaid Score, Plain Language principles
🎯

Realism — Are the requirements possible?

Example Score95/100

The Realism score checks whether what the company is asking for exists in the real world.

  • Years vs technology age
    "10 years SwiftUI" — SwiftUI released 2019
  • Skills vs role scope
    "Full-stack + DevOps + Design + PM + Data" = 5 people for 1
  • Salary vs market vs seniority
    "Senior Engineer, $35k, San Francisco"
  • Contradictory requirements
    "Entry-level, 5+ years required" — pick one
Based on: Labor market data, Credential Inflation research (Fuller & Raman, Harvard, 2017)
🔍

Transparency — What is the company hiding?

Example Score80/100

The Transparency score measures how open the company is about important information.

  • Is salary mentioned?
    "Competitive salary" is not a salary
  • Is the company named?
    "Confidential client" — what else?
  • Is team described?
    "Join our amazing team" — what team?
  • Work model clear?
    "Flexible" = probably 5 days office
Based on: Information Asymmetry Theory (Akerlof, 1970), Pay Transparency legislation

The Signals We Look For

Each job description is scanned for 15+ signals, each weighted by severity.

High Severity (+15-20 points each)

Impossible Requirements
More years than the tech exists
Chronic Reposting
Same job posted 6+ months
Contradictory Requirements
Entry-level + senior experience
Unicorn Job
One role, 5+ completely different fields

Medium Severity (+8-12 points each)

No Salary Info
Anonymous Employer
Vague Language
Copy-Paste Text
No Team Info
Evergreen Posting
Excessive Openings

Low Severity (+3-5 points each)

No Benefits
'Competitive Salary'
No Process Described
Buzzword Overload
Vague Company
No Location Clarity

What Makes a Job Look Legit

These positive signals reduce the Ghost Score.

Specific Salary Range
-15 pts
Detailed Tech Stack
-12 pts
Clear Responsibilities
-10 pts
Named Hiring Manager
-10 pts
Team Description
-8 pts
Realistic Requirements
-8 pts
Specific Benefits
-5 pts
Hiring Process Explained
-5 pts
Detailed Company Intro
-5 pts

Putting It All Together: The Ghost Score

0
✅ Legit
30
🤔 Sus
60
👻 Ghost
85
💀 Certified
100

✅ Looks Legit (0-30)

Few or no red flags. Apply with confidence, but always do your own research too.

🤔 Kinda Sus (31-60)

Some concerning signals. Proceed with caution. Verify the company through other channels.

👻 Probably Ghost (61-85)

Multiple red flags. Consider skipping this one and focusing on better opportunities.

💀 Certified Ghost (86-100)

Overwhelming evidence. Do not waste your time. Share it on the Ghost Wall to warn others.

"The Ghost Score is a probability, not a certainty. Like a spam filter, it catches patterns — but no algorithm is 100% accurate. We optimize for catching ghost jobs even at the cost of occasional false positives, because your time is too valuable to waste."

Standing on the Shoulders of Giants

GhostJob is built on established economic and sociological research.

📖

Signal Detection Theory

Green & Swets (1966)

Originally for radar operators detecting enemy aircraft. We apply the same framework: distinguishing real jobs (signal) from ghost jobs (noise).

📖

The Market for Lemons

Akerlof (1970, Nobel 2001)

When buyers can't assess quality, bad products drive out good ones. Ghost jobs pollute the market for real opportunities.

📖

Job Market Signaling

Spence (1973, Nobel 2001)

Spence showed participants send signals to convey quality. We read the employer's signals to assess posting legitimacy.

📖

Dismissed by Degrees

Fuller & Raman (Harvard, 2017)

This study showed employers routinely inflate requirements beyond what's needed. We use this to calibrate our Realism score.

The Data Behind Our Claims

ClaimSourceYear
43% of jobs are ghost jobsClarify Capital Survey2022
11 hrs/week job searchingBureau of Labor Statistics2023
2-3% application response rateJobvite Recruiter Nation2022
75% more likely to pass ATSTopResume Study2021
Companies post for 'optics'Resume Builder Survey2023
60%+ credential inflationHarvard Business School2017

Enough Theory. Let's Check Your Job.

Now that you know how it works, try it yourself and stop wasting time on fake postings.

Analyze a Job Posting — Free 👻

No signup required