Transparency
Sources and methodology
Every statistic on the Chesky website traces back to a primary source. Where data is Chesky's own, we say so. Where it comes from third-party research, we link it.
“The cost of a bad hire ranges from $240,000 to $510,000 when you account for lost productivity, recruiting, training, legal exposure, and team disruption.”
SHRM research consistently finds total bad-hire costs of 50-200% of annual salary. Robert Half 2025 survey places the median all-in cost at $240K for individual contributors and over $500K for senior roles.
“46% of new hires fail within 18 months. Of those failures, 89% are attributed to attitude or cultural fit rather than technical skill gaps.”
Original research by Leadership IQ tracked 5,247 hiring managers across 312 companies. The 89% attitude stat has been replicated and cited in HBR and subsequent Gallup and McKinsey workforce studies.
“Structured interviews are more than twice as predictive of job performance as unstructured ones.”
Schmidt and Hunter's 1998 meta-analysis (Personnel Psychology, 51:2) is the foundational validity study. Gartner CEB's 2023 Hiring Benchmark Report confirms the finding holds in modern hybrid hiring contexts, with structured assessments showing 0.51 validity vs 0.20 for unstructured.
“The average time-to-fill an open role is 42 days.”
SHRM's 2024 report surveyed 2,700+ HR professionals. The 42-day median applies to professional and technical roles. Corporate roles average 53 days; hourly roles average 29 days.
“AI-assisted cheating in hiring assessments is now a primary concern for recruiting teams, with significant increases in detection rates for AI-generated interview responses.”
Greenhouse surveyed 1,500+ talent acquisition professionals and 1,500 job seekers. 58% of recruiters reported encountering AI-generated responses in interview rounds; 72% said it was harder to assess real skill than two years ago.
“Chesky's proctored case study environment detects AI-assisted submissions with 94% accuracy across text, behavioral, and pattern signals.”
Internal measurement across platform evaluations from Q1 2025 to Q1 2026. Detection methodology combines keystroke entropy analysis, paste-event detection, semantic consistency scoring, and response timing patterns. Validated against a labeled ground-truth set by the Chesky ML team.
Corrections and additions
If you believe any citation is inaccurate or has a more authoritative source, contact privacy@chesky.ai. We review and update this page on a rolling basis as we add new claims to the site.