AI Candidate Matching

📋 On This Page

• Overview
• How It Works
• Detailed Scoring
• Usage Guide
• Best Practices
• Limitations
• Related Features

---

Overview

The AI Candidate Matching feature uses an intelligent scoring algorithm to match candidates to job openings based on skills, experience, location, and compensation alignment. This helps recruiters quickly identify the best candidates for each role.

---

How It Works

Matching Algorithm

The system calculates a match score (0-100) based on four weighted factors:

• Skills Match (40% weight) — Overlap between candidate skills and job requirements
• Experience Match (25% weight) — Alignment of years of experience
• Location Match (20% weight) — Work arrangement and location compatibility
• Compensation Match (15% weight) — Salary/day rate alignment

Score Calculation

Overall Score:

Total Score = (Skills × 0.40) + (Experience × 0.25) + (Location × 0.20) + (Compensation × 0.15)

Score Range: 0-100 (higher is better)

---

Detailed Scoring

Skills Match (40% weight)

Required Skills:

• Up to 80 points for required skills match
• Percentage of required skills candidate has
• Formula: (matched_required / total_required) × 80

Nice-to-Have Skills:

• Up to 20 points for nice-to-have skills
• Percentage of nice-to-have skills candidate has
• Formula: (matched_nice_to_have / total_nice_to_have) × 20

Example:

• Job requires: PHP, Web Development, Database Management
• Candidate has: PHP, Web Development, JavaScript, Database Management
• Required match: 3/3 = 100% → 80 points
• Nice-to-have: JavaScript (bonus) → +5 points
• Total: 85 points

---

Experience Match (25% weight)

Within Range:

• Perfect match (100 points) if candidate experience within job range
• Example: Job requires 3-5 years, candidate has 4 years → 100 points

Below Minimum:

• Penalty: -20 points per year under minimum
• Example: Job requires 5 years, candidate has 3 years → 60 points

Above Maximum:

• Penalty: -10 points per year over maximum (less penalty)
• Example: Job requires 3-5 years, candidate has 8 years → 70 points

Not Specified:

• Neutral score (50 points) if experience not specified

---

Location Match (20% weight)

Work Arrangement (60% of location score):

• Remote jobs match anyone (60 points)
• Remote candidates match any job (60 points)
• Flexible matches everything (60 points)
• Exact match (hybrid=hybrid, office=office) → 60 points
• Mismatch → 20 points

Location (40% of location score):

• Exact location match → 40 points
• Partial match (contains) → 40 points
• No location specified → 20 points
• No match → 0 points

Total Location Score:

• Work arrangement score + Location score
• Maximum: 100 points

---

Compensation Match (15% weight)

For Permanent Roles:

• Compares candidate salary expectations to job salary range
• Checks for range overlap
• Calculates overlap percentage
• Penalizes gaps (20 points per £10k gap)

For Contract Roles:

• Compares candidate day rate expectations to job day rate range
• Checks for range overlap
• Perfect match if ranges overlap
• Penalizes gaps (20 points per £100 gap)

Not Specified:

• Neutral score (50 points) if compensation not specified

---

Usage Guide

Finding Matching Candidates

• Navigate to Jobs → Select job
• Click Find Matching Candidates
• System displays candidates ranked by match score
• Each candidate shows:
- Overall match score (0-100) - Score breakdown (skills, experience, location, compensation) - Candidate profile summary
• Review top matches
• Create submissions for best matches

---

Understanding Match Scores

Score Interpretation:

• 90-100 — Excellent match, strong candidate
• 75-89 — Good match, worth considering
• 60-74 — Moderate match, may need training
• Below 60 — Weak match, likely not suitable

Score Breakdown:

• View detailed breakdown to understand why score is high/low
• Identify areas where candidate excels
• Identify gaps that could be addressed

---

Filtering Results

Minimum Score Filter:

• Set minimum score threshold
• Only show candidates above threshold
• Default: 0 (show all)
• Recommended: 60+ for quality matches

---

Best Practices

• Set Minimum Score — Filter out weak matches (60+ recommended)
• Review Breakdown — Understand why candidates score high/low
• Consider Context — Score is a guide, not absolute
• Update Skills — Keep candidate skills current for better matching
• Refine Job Requirements — More specific requirements = better matches
• Manual Review — Always review top matches manually

---

Limitations

Current Limitations

• Only matches active candidates
• Skills matching is case-sensitive
• Location matching is basic (contains check)
• Compensation parsing may be imprecise
• No learning/improvement over time

Future Enhancements

• Machine learning for better matching
• Skill synonym matching
• Industry-specific matching
• Historical placement success weighting

---

🔗 Related Features

• Client & Job Management — Managing jobs
• Candidate Management — Managing candidates
• AI Relationship Graph — Relationship context for matches
• AI Job Parsing — Parsing job descriptions