Comparing within a stage
Candidates who sat the same stage of the same flow are directly comparable: the same skills, the same weights, the same difficulty and the same evaluation settings. The comparison view puts their per-skill scores side by side.
This is the strongest thing the platform offers, and it is stronger than comparing two human-led interviews, because two human interviewers never ask quite the same question.
Where scores are not comparable
- Across positions. Different skills and different weights. A 78 on one role and a 74 on another are not a ranking.
- Across leniency settings. If you changed evaluation leniency mid-round, candidates before and after are on different scales.
- Across stage types. An MCQ score and a video interview score measure different things.
The comparison view will not put non-comparable candidates on one axis, which sometimes means it shows you less than you wanted. That is the intended behaviour.
What to look at beyond the number
Open the per-skill breakdown. Two candidates on the same overall score are frequently strong in different places, and which of them you want depends on what the team already has — a question the score cannot answer.