How match scores work

A match score shows how well a candidate fits one job’s brief, on a scale from 0 to 100. Every score comes as a pair: the current score, based on the evidence Staffer has, and the potential score, what the candidate could score if the details their profile doesn’t mention turn out in their favor. Scores are per job. The same person can score 80 for one job and 35 for another, because each job has different criteria.

​

How a score is built

Staffer scores each criterion in the brief on its own, then combines them.

1. Staffer compares the criterion with everything it knows about the candidate.
2. An AI model reviews that evidence and sets the criterion’s score and potential score. It writes a one-sentence reason, cites up to three sources, and says how confident it is.
3. The overall score is a weighted average of the criteria. Required criteria count double compared with preferred ones.

If the model doesn’t answer every criterion, Staffer throws the attempt away and the candidate keeps the score they had. A criterion is never quietly scored as zero because an answer went missing.

​

Read the Match tab

Open a candidate and click the Match tab.

• Overall match shows the overall pair. Expand it for a short summary of the fit.
• Required criteria match and Preferred criteria match list one row per criterion, with its status and its pair of scores.

Expand a criterion to see the reason, the confidence (High confidence, Medium confidence or Low confidence) and the Sources the reason is based on: the profile, a message, a process step or a note. If a later rescore didn’t find new evidence for that criterion, you’ll see “Unchanged by the latest evidence — this criterion keeps its previous score.” Each criterion gets one of four statuses: Scores use the same three colors everywhere: green for 70 and up, teal for 40 to 69, and blue below 40. The scale deliberately avoids red and amber, so a low score reads as “weak fit for this brief”, not as a warning about the person.

​

Current score and potential score

The potential score is never lower than the current score. It’s higher when the evidence is silent on something the candidate plausibly has, given their seniority and roles. When the evidence is clear, the two numbers are the same. Hover over the two circles to see which is which: Current score and Potential score (or Potential not scored yet). Potential matters in two places you might not expect:

• Search results on the Scouting tab are ordered by potential score.
• The Match filter on candidate lists sorts people into Good fit (70+), Maybe (40–69), Weak (under 40) and Unscored using the potential score when there is one. A candidate showing a current score of 55 can appear under Good fit (70+).

​

What the score reads

• The candidate’s profile.
• Their six most recent messages with you.
• Completed and skipped process steps, including the candidate’s own answers, such as questionnaire responses and interview transcripts.
• Notes from your team. See Notes.

​

What the score ignores

• The candidate’s name.
• The evaluation an AI interview produces, and any score a reviewer gives in a review step. The score reads what the candidate said, not someone’s verdict on it.
• Notes left by people outside your workspace on a share link.

​

Where scores appear

• Scouting results. Hover over a score to see the per-criterion breakdown.
• The Match column and filter on a job’s Candidates tab and on the Applications page.
• The candidate header and the Match score section of the profile panel.
• The Match tab.
• Share links, if you include the match score when you create the link.

The talent pool doesn’t show scores, because a profile in the pool isn’t tied to a job.

​

What a score doesn’t do

A score is evidence for your decision, not the decision. Staffer doesn’t reject, advance or contact anyone because of a score on its own. A score also doesn’t update when you edit the brief. It changes when there’s new evidence about the candidate, or when you rescore them. See When scores change.

​

Common questions