How scoring works
Amazon Connect Talent is built to give you fair, evidence-based results that you can trust. This page explains the principles behind scoring, how scoring rubrics are generated, and what counts as evidence.
Scoring principles
Keep the following principles in mind as you read a candidate's scores.
-
Consistent and fair – Every candidate for the same role is evaluated on the same competencies and the same criteria. Results do not depend on who the candidate is.
-
Measured against a clear rubric – Amazon Connect Talent scores each competency against a rubric – a set of criteria that defines what a strong response and a weak response look like. Amazon Connect Talent evaluates the candidate's responses against that rubric and ties each conclusion to evidence from the interview. When there isn't enough evidence, Amazon Connect Talent reports that instead of guessing.
-
Based on evidence, not impressions – Scores reflect the behavior a candidate actually described – concrete past actions and results – not hypotheticals, self-ratings, or promises about what they could do.
-
What they said, not how they said it – Amazon Connect Talent evaluates the content of a candidate's responses. How a candidate speaks – accent, fluency, pace, pauses, or filler words – does not affect the score.
-
Personal circumstances are never evidence – Disclosures such as disability, health, or pregnancy are never used to score a candidate. Opt-out requests stay visible so you can consider them.
Important
Amazon Connect Talent informs; you decide. Amazon Connect Talent reports evidence-based competency scores and never recommends whether to hire. Hiring decisions are yours.
How scoring rubrics work
Amazon Connect Talent generates a scoring rubric for each competency before any candidate is interviewed. A rubric defines the specific, observable, job-related behaviors that distinguish a strong response from a weak one for that competency.
The rubric is derived from the job description and competencies, and it is fixed for the role – every candidate who interviews for the same job is evaluated against the same criteria. This ensures results are consistent, comparable, and free from variation between candidates.
Amazon Connect Talent applies rubrics automatically during scoring. This removes subjectivity from the evaluation process and ensures every candidate is measured the same way, every time.
Tip
Want to see the details behind a scoring rubric? Ask the Amazon Connect Talent Teammate about a specific evaluation. The Teammate can walk you through the behaviors the rubric looks for and how they map to competencies.
What counts as evidence
Amazon Connect Talent scores each competency by checking the candidate's responses against the rubric. Only demonstrated behavior counts – concrete actions and results the candidate describes from their past experience. The following do not affect a score:
-
Hypothetical statements or self-assessments ("I would..." or "I'm good at...")
-
How the candidate speaks – accent, fluency, pace, pauses, or filler words
-
Personal circumstances – disclosures such as disability, health, or pregnancy are never used in scoring
Each score is tied to specific evidence from the interview, so you can trace any result back to what the candidate said.
Interview scores
Amazon Connect Talent scores each competency on a 1–5 scale. A score of 5 means the candidate demonstrated the positive behaviors the rubric defines for that competency with no concerns. A score of 1 means the candidate did not demonstrate the expected behaviors, or demonstrated behaviors that indicate concern.
The overall interview score is the equal average of all competency scores, so every competency counts the same.
For more information about how interview scores appear on the candidate results page, see Review candidate results.
Assessment scores
Assessments use a separate scale from the interview. Assessment results appear on the four-level assessment scale. For more information, see How assessments work. These scores reflect how closely the candidate's responses match the behavioral patterns associated with success in that competency – they are not pass/fail judgments.
Amazon Connect Talent selects the most relevant assessments for the job's competencies from its assessment library and tailors each one to the role. Not every evaluation includes assessments – Amazon Connect Talent adds assessments only when the job's competencies match what an assessment measures. For more information, see How assessments work.
