Most hiring tools give you a single score. Resume match: 78%. That number is meaningless without context. How confident is the system? What’s the score based on? A resume alone, or a full evaluation?
Progressive scoring is our answer. The score starts low and increases as we gather more signal. A CV-only score of 45 means something very different from a post-interview score of 85. Both are useful, but only if you understand the confidence level behind them.
This post explains the mechanics of progressive scoring: how each stage works, what the numbers mean, and why we believe it’s a fundamentally better way to evaluate candidates than the single-score model.
The problem with single scores
A single score is an average that hides everything interesting. A candidate who scores 70% on a resume screen could be a perfect fit who has an unconventional background, or a mediocre fit who happens to have the right keywords. You can’t tell which. The number is identical, but the underlying reality is completely different.
Single scores also create anchoring bias. Once a hiring manager sees “78%,” that number shapes every subsequent interaction with the candidate. If the resume score is high, the interviewer looks for confirmation. If it’s low, they look for problems. The first number you see becomes the frame through which you evaluate everything else.
Progressive scoring avoids both problems by being explicit about what the score is based on and how confident the system is at each stage.
The four scoring stages
Stage one: CV analysis. Maximum score of 50. This is a rough screen based on background, experience, and inferred skills. It’s enough to decide whether to continue, not enough to make a hiring decision.
Stage two: Q&A responses. Maximum score of 65. The candidate answers role-specific questions, and we evaluate depth of thinking, communication clarity, and domain knowledge. These questions are generated by Rob based on the hiring brief, so they’re tailored to the role rather than generic.
Stage three: skills assessment. Maximum score of 80. This is a practical evaluation — a work sample, a technical exercise, or a case study, depending on the role. It tests whether the candidate can do the actual work, not just talk about it.
Stage four: AI-assisted interview. Maximum score of 95+. Rob conducts a structured interview covering the areas where the earlier stages left uncertainty. The conversation is adaptive — it probes the specific gaps and concerns surfaced by previous stages.
How the stages interact
Each stage builds on the previous ones. The Q&A questions are informed by what the CV analysis revealed. The skills assessment is calibrated to test the specific competencies the Q&A flagged as uncertain. The interview focuses on the remaining open questions.
This means the evaluation gets progressively more targeted. By stage four, Rob isn’t asking generic interview questions. It’s asking the specific questions that will resolve the specific uncertainties about this specific candidate for this specific role.
Candidates also experience the process as fair and transparent. Each stage gives them a chance to demonstrate something new. A candidate who has a weak resume but strong responses in the Q&A will see their score climb. The system adapts to the candidate rather than filtering them through a static rubric.
Why confidence matters more than the number
A score of 70 at stage four is more meaningful than a score of 70 at stage one. The number is the same, but the confidence is completely different. Stage one is a guess. Stage four is a conviction.
We display the confidence level alongside the score so hiring managers can calibrate their decisions appropriately. Don’t reject a candidate at 45 — that’s a stage-one score. Keep going.
Confidence is expressed as a percentage. A stage-one score of 45 might have 30% confidence. That means: “based on limited information, this is our best estimate, but we’re not very sure.” A stage-four score of 45 with 85% confidence means: “we’ve evaluated this candidate thoroughly, and they’re not a strong fit.” Same score, very different implications.
What this looks like in practice
A typical candidate journey through progressive scoring: CV analysis returns a score of 42 with 28% confidence. The hiring manager might hesitate, but Rob recommends proceeding because the brief emphasizes a skill set that’s hard to assess from a resume alone.
After the Q&A, the score jumps to 58 with 52% confidence. The candidate’s thinking is strong. The skills assessment pushes it to 74 with 71% confidence. The work sample is excellent. The final interview confirms the fit: 88 with 89% confidence.
Without progressive scoring, this candidate might have been filtered out at stage one. The resume didn’t look like a match. But the progressive approach gave them four chances to demonstrate their abilities, and each chance added real signal.
This is the core insight: a single score at any point in the process is less valuable than a trajectory of scores across multiple evaluations. The trajectory tells you not just where the candidate is, but where they’re going.