What Candidate Scoring Is (and Is Not)
Candidate scoring is the systematic process of evaluating job applicants against defined criteria and expressing the result as a measurable score. It is not a black box that spits out verdicts — it is a structured framework that makes the criteria explicit, the evaluation consistent, and the output auditable.
Done correctly, scoring does not replace human judgment. It focuses human judgment on the candidates most likely to succeed.
Where Hiring Bias Actually Enters
- Resume review: unconscious preference for certain educational institutions or company names
- Shortlisting: recency bias (the last resume reviewed scores higher in memory)
- Interviews: affinity bias (interviewers favor candidates similar to themselves)
- Offer decisions: anchoring on first impressions formed before structured evaluation
Scoring reduces bias at the first two stages — resume review and shortlisting — which affect the largest number of candidates. If your shortlist is built on consistent criteria, bias has fewer entry points in later stages.
How TwynIt's Scoring Model Works
TwynIt evaluates every candidate across three weighted dimensions for each role: skills coverage (50%), experience fit (30%), and keyword relevance (20%). All weights are configurable at the organization level — a sales role might weight keywords higher; a technical role might weight skills at 70%.
- Skills coverage: uses a 4-layer matching system (exact → normalized → alias → substring) so 'JS' and 'JavaScript' are treated as the same skill
- Experience fit: years of relevant experience benchmarked against role requirements with non-linear penalties for gaps
- Keyword relevance: job-description-specific terminology matched against the full candidate profile
Score Thresholds and What They Mean
In TwynIt, scores above 60 indicate a strong match — these candidates are auto-flagged for shortlist consideration. Scores between 40–60 indicate a 'consider' range — worth reviewing with context. Below 40 indicates a weak match against your specific role requirements.
These thresholds reflect the position's requirements, not candidate quality in absolute terms. A candidate scoring 45 for one role might score 80 for a slightly different one.
The Audit Trail
Every score in TwynIt includes a breakdown: which skills matched, which were missing, how experience compared to requirements, and which keywords were found. This creates a defensible, documented record of why each candidate was shortlisted or not — useful for team alignment and important for compliance.
Common Objections to Candidate Scoring
- 'It misses intangibles like culture fit' — Correct, and that is by design. Scoring handles objective criteria. Culture fit is evaluated in interviews.
- 'It could reinforce existing bias if the criteria are biased' — This is why the criteria are configurable and visible. Review what you are scoring before you score.
- 'Candidates can game it' — Scores are computed on submitted resumes, not self-reported assessments. Gaming requires actually having the skills.
Getting Started
Define your evaluation criteria for one role. Upload the JD. Let TwynIt score your existing candidate pool. Review the top 20 percent and compare the output against your instinctive shortlist. Most teams find the AI and the experienced recruiter agree 80 percent of the time — and the disagreements are worth examining.