> ## Documentation Index
> Fetch the complete documentation index at: https://docs.exterview.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Evaluation Design

> Design role-relevant Exterview evaluations by mapping competencies to rubric items, picking session types, and avoiding common assessment pitfalls.

## Purpose

Good evaluations start with role clarity, not tooling. This page shows how to design evaluations that measure what actually matters for a role.

## Start with role clarity

Before you configure anything, answer four questions:

<CardGroup cols={2}>
  <Card title="What does the role actually require?" icon="bullseye">
    The skills and responsibilities that matter, not a generic job description.
  </Card>

  <Card title="What does 'strong' look like?" icon="scale-balanced">
    The bar a candidate needs to clear, stated in terms you can score against.
  </Card>

  <Card title="Which stages add real signal?" icon="list-check">
    Every stage should justify itself, extra stages add drop-off, not insight.
  </Card>

  <Card title="Who owns the final call?" icon="user-check">
    The hiring manager owns what "good" looks like; talent owns consistency.
  </Card>
</CardGroup>

## What strong evaluation design looks like

<CardGroup cols={2}>
  <Card title="Role-first, not tool-first" icon="sitemap">
    Start from what the role needs, not from what the platform can run.
  </Card>

  <Card title="Only as many stages as needed" icon="filter">
    Fewer, higher-signal stages beat a long funnel that adds drop-off.
  </Card>

  <Card title="Traceable to the rubric" icon="code-branch">
    Every report ties back to a defined dimension, not a black-box score.
  </Card>

  <Card title="You define what matters" icon="user-gear">
    Exterview evaluates consistently against your criteria, it doesn't set them.
  </Card>
</CardGroup>

## Map the evaluation

A clean evaluation traces a straight line from the role to the report:

```mermaid theme={null}
%%{init: {'theme':'base','themeVariables':{'primaryColor':'#EEF2FF','primaryBorderColor':'#4B6CFE','primaryTextColor':'#0F172A','lineColor':'#64748B','clusterBkg':'#F8FAFC','clusterBorder':'#E2E8F0'}}}%%
flowchart LR
  A([Role]) --> B([Skills]) --> C([Rubric]) --> D([Session Type]) --> E([Report])
```

Each step should justify the next: skills come from the role, the rubric measures the skills, the session type collects the evidence, and the report reflects the rubric.

## FAQs

<AccordionGroup>
  <Accordion title="How many stages should an evaluation have?">
    Only as many as add signal. Extra stages increase drop-off without improving the decision.
  </Accordion>

  <Accordion title="Who should design the evaluation?">
    The hiring manager and talent lead together, the manager owns what "good"
    looks like; talent owns consistency.
  </Accordion>

  <Accordion title="Does Exterview decide the criteria for us?">
    No. You define what matters; Exterview evaluates against it consistently and traceably.
  </Accordion>
</AccordionGroup>

## Related

<CardGroup cols={2}>
  <Card title="Scoring Quality" icon="scale-balanced" href="/best-practices/scoring-quality">
    Keep scoring consistent once the evaluation is designed.
  </Card>

  <Card title="Operations Setup" icon="gears" href="/configure/operations">
    Where rubrics, pipelines, and gating are actually configured.
  </Card>
</CardGroup>