How Matching Works

Overview

When you create a profile, whether as a candidate or a company posting a job, our system compares your structured profile data against potential matches. Each match receives a score from 0 to 100, representing how well the two profiles align. The highest-scoring matches appear first, but all matches remain visible. The system ranks, it does not exclude.

How Scores Are Calculated

Each match score is built from five factors, each contributing a fixed percentage to the total score:

Skills Match - 35%

We compare the skills listed on your profile against the skills required for the job. Must-have skills are weighted most heavily. If a job requires Python, Node.js, and PostgreSQL, and you have two of the three, your skills score reflects that coverage. Nice-to-have skills provide a smaller boost.

Role Fit - 25%

This measures how well your overall profile aligns with the job description in a broader sense, beyond just individual skills. We use AI-generated embeddings (numerical representations of your profile and the job) to calculate semantic similarity. This captures things like whether your career trajectory, headline, and experience narrative align with what the role is looking for.

Experience Level - 15%

Your years of experience are compared to the level the job expects (entry, junior, mid, senior, lead, principal). Being slightly under-qualified results in a small penalty; being significantly over-qualified results in a smaller penalty. The system recognises that a senior developer might not want a junior role, and vice versa.

Location - 15%

If the job is fully remote, location scores 100% automatically. For on-site or hybrid roles, we compare your location to the job's location. An exact city match scores highest, same region scores well, same country scores moderately, and different countries score lower. If you've indicated willingness to relocate, that adjusts the score upward.

Salary - 10%

We compare your salary expectations with the job's offered range. When the ranges overlap, the score is high. When there's a gap, the score decreases proportionally. Neither party's exact salary figures are ever shared. Only the compatibility is calculated.

Your final score is the weighted sum of these five factors, scaled to 0-100. The same formula is used for every match. No one receives preferential treatment.

How AI Is Used

Matchboks uses AI in two specific ways:

Profile Embeddings

When you save your profile, we generate a numerical representation (called an embedding) of your skills, experience, and job preferences using OpenAI's text-embedding model. This embedding is used to calculate the Role Fit factor described above. We do not send your name, email, or contact details to OpenAI, only structured professional data like skills, job title, experience level, and location.

Match Insights

For each match, our system generates role-fit insights, a structured breakdown of strengths, considerations, and suggested questions based on how your profile aligns with the job. These insights are created by an AI model using your professional profile data and the other party's professional data. Each side sees only their own perspective. Your insights are never shared with the employer, and theirs are never shared with you. Insights are supplementary and don't affect your score.

What AI Does Not Do

• AI never automatically rejects candidates or removes them from consideration
• AI never makes hiring decisions. It provides recommendations that humans act on
• AI does not process personal identifiers like your name, email, phone number, or photo
• AI does not use age, gender, ethnicity, or any protected characteristics in scoring

What Data Feeds Into Your Score

Only structured, professional profile data is used in scoring:

From Candidate Profiles

• Skills (with proficiency levels)
• Years of experience and experience level
• Location and willingness to relocate
• Salary expectations (min–max range)
• Remote work preference
• Current and desired job title

From Job Listings

• Required and nice-to-have skills
• Required experience level
• Job location and remote work policy
• Salary range offered
• Job description (used for role fit embedding only)

You Are Always in Control

Matchboks is a recommendation tool, not an automated decision-maker.

• You decide which matches to pursue by swiping right or left
• Even low-scoring matches remain visible. The system ranks, it does not filter
• If you swipe right on a job, your match is sent to the company regardless of your score
• You can block specific companies from seeing your profile
• You control your profile visibility level (full, limited, or hidden)

Blocking and Privacy

Both candidates and companies can block each other. When you block a company, they cannot see your profile or match with you. Companies can also block candidates. Additionally, candidates can hide themselves from specific employers, which is useful if you don't want your current employer to see you're looking. These blocks are absolute and override any match score.

Your Rights Under GDPR

Under GDPR Article 22, you have the right not to be subject to decisions based solely on automated processing that significantly affect you. Matchboks is designed as a recommendation tool with human decision-making at every step. No automated decisions determine hiring outcomes. You can request an explanation of any specific match score by contacting us at hei@matchboks.no.

If you have questions about how matching works or how your data is used, contact us at hei@matchboks.no. hei@matchboks.no