Skip to main content
Nova evaluates candidates using a transparent, criteria-driven scoring system. Every score includes detailed reasoning and citations from the candidate’s resume.

​
What Powers the Score

Core Data Sources:
  1. Criteria – Your configured requirements (primary signal)
  2. Resume/CV – Main candidate information source
  3. Job Description – Context for seniority and domain
  4. Application Data – Cover letters, portfolios, form answers
Fallback Options:
  • LinkedIn URL if no resume attached
  • Additional portfolio documents
  • Custom application question responses

​
Score Interpretation

Score Ranges:
  • 9-10: Exceptional fit, strong match across all criteria
  • 7-8: Good fit, meets most requirements with some strengths
  • 5-6: Partial fit, meets some requirements, gaps in others
  • 3-4: Poor fit, significant gaps in key requirements
  • 1-2: Very poor fit, minimal qualification match
Reading Score Details:
  • Review criterion-by-criterion breakdown
  • Check resume citations for evidence
  • Look for reasoning behind each assessment
  • Consider overall pattern vs individual scores

​
Diagnosing Score Issues

SymptomProbable CauseFix
Everyone gets 4-6 scoresToo many Required criteriaConvert less-critical rules to Nice-to-have
Too many 9-10 scoresCriteria too loose/broadTighten experience ranges, add specific requirements
Great candidate scored lowCriteria too strict for their backgroundCheck if resume meets criteria, then broaden if needed
Inconsistent scoringVague or ambiguous criteriaMake criteria more specific and measurable

​
Continuous Improvement

Enhanced Optimization Features:
  • Real-time Re-scoring: Update all candidate scores instantly when criteria change
  • A/B Testing: Compare different criteria sets to find optimal configuration
  • Historical Analysis: Track scoring accuracy over time
  • Custom Feedback Loops: Train models based on your hiring decisions (optional)
Improvement Process:
  1. Review assessments for first 20-30 applicants
  2. Adjust criteria where mismatches appear
  3. Hit Re-score to update all candidates in seconds
  4. Monitor performance and iterate weekly
Nova’s scoring engine is audited quarterly for disparate impact across sex, race, age, disability, and intersectional groups. Full methodology and live metrics are available in our Bias Evaluation.Built-in Safeguards:
  • Regular bias audits across demographic groups
  • Transparent reasoning for all scores
  • Human override capabilities
  • Criteria-driven assessment (not pattern matching)

​
Frequently Asked Questions

Optionally—models can fine-tune when enabled (default off), but your configured criteria always outrank historical patterns. This ensures consistent scoring based on your explicit requirements.
Currently, Nova doesn’t support manual weighting of individual criteria. This is by design—traditional weighted scoring systems are difficult to maintain, configure accurately, and often introduce unconscious bias. Instead, Nova uses advanced AI to intelligently evaluate all criteria together.
Nova evaluates candidates based strictly on your written criteria. If scores don’t match your intuition, review whether your criteria fully capture what you’re looking for in candidates. Consider adding missing requirements or adjusting existing ones.
Criteria changes apply immediately to new candidates. For existing candidates, re-scoring (available on Enterprise plans) updates all scores within seconds of criteria modifications.
Nova can work with LinkedIn URLs as a fallback. If neither resume nor LinkedIn is available, scoring will be limited and the system will indicate insufficient data for reliable assessment.

Configuring Criteria

Create and refine your scoring criteria

Enabling Nova

Enable Nova scoring on your job postings