Skip to content

Discernment Matrix

The Discernment Matrix evaluates each persona across 6 subjective traits using a 7-dimension rating model. It captures the qualitative characteristics that shape how a persona exercises judgment, interacts with collaborators, and maintains professional standards.

Purpose

R.I.S.C.E.A.R. defines what a persona does. The dimension system defines who the persona is. The Discernment Matrix defines how well the persona exercises judgment. It answers questions like:

  • How humble is this persona when receiving feedback from the Critique phase?
  • How curious is this persona when exploring unfamiliar domains during the Find phase?
  • How inclusive is this persona when collaborating with personas from different categories?

These traits are not binary -- they exist on a spectrum, measured from multiple perspectives.

The 6 Traits

1. Humility

The persona's openness to feedback, willingness to revise its outputs, and ability to acknowledge limitations. A highly humble persona accepts critique from the Documentation Evangelist (DE) or Blueprint Validator (BV) without defensiveness and actively incorporates corrections.

Relevance: Essential during the Critique phase, where personas must accept and act on quality feedback. Champions with high humility orchestrate without dominating.

2. Professional Background

The depth and breadth of domain expertise the persona brings to its role. This trait measures not just knowledge but the ability to apply professional experience to novel situations within the FCC workflow.

Relevance: Directly impacts the quality of outputs in specialized roles like the Semantic Taxonomy Engineer (STE) or the Anti-fact Mitigation Specialist (AMS).

3. Curiosity

The persona's drive to explore beyond its immediate scope, ask clarifying questions, and seek deeper understanding. A highly curious persona proactively investigates edge cases, cross-references findings, and identifies connections that others might miss.

Relevance: Critical during the Find phase, where Research Crafter (RC) and Catalog Indexer Architect (CIA) must thoroughly explore the problem space.

4. Taste

The persona's ability to distinguish high-quality outputs from adequate ones. Taste encompasses aesthetic judgment in design (UI Mockup Crafter), editorial sensibility in writing (Documentation Evangelist), and architectural elegance in blueprints (Blueprint Crafter).

Relevance: Separates good documentation from great documentation. Personas with high taste push the team toward excellence rather than mere compliance.

5. Inclusivity

The persona's commitment to accessibility, diverse perspectives, and equitable representation in its outputs. A highly inclusive persona ensures documentation serves all audiences, not just the primary technical audience.

Relevance: Directly impacts the User Guide Crafter (UG) and Stakeholder Content Publisher (SCP), but affects every persona that produces human-facing artifacts.

6. Responsibility

The persona's accountability for its outputs, adherence to commitments, and ownership of quality. A highly responsible persona tracks its deliverables through completion, maintains audit trails, and proactively flags risks.

Relevance: Underpins the entire governance layer. The Governance Compliance Auditor (GCA) and Traceability Specialist (TS) require exceptionally high responsibility ratings.

The 7 Rating Dimensions

Each trait is assessed across 7 rating dimensions, providing a multi-perspective view that prevents single-source bias.

Dimension Field Description
Self Rating self_rating The persona's own assessment of its trait level. Useful for calibration against external perspectives.
Peer Rating peer_rating Assessment from personas that collaborate directly (upstream, downstream, or peer relationships).
Survey Rating survey_rating Aggregated rating from a broader set of stakeholders or users who interact with the persona's outputs.
Individual Weighted Rating individual_weighted_rating A calculated score that weights self, peer, and survey ratings according to a defined formula.
Organizational Rating org_rating Assessment from the organizational context -- how the persona's trait level serves broader team goals.
External Rating external_rating Assessment from outside the immediate workflow -- external reviewers, auditors, or consumer feedback.
Ranked Percentile Rating ranked_percentile_rating The persona's percentile rank for this trait across all 24 personas. Enables comparative analysis.

All ratings use a 0.0 to 1.0 scale, where 0.0 represents the lowest expression of the trait and 1.0 represents the highest.

Scoring Methodology

Individual Weighted Rating Calculation

The individual weighted rating is typically calculated as a weighted average:

individual_weighted = (self_weight * self_rating)
                    + (peer_weight * peer_rating)
                    + (survey_weight * survey_rating)

Default weights: self (0.2), peer (0.4), survey (0.4). These weights can be customized per deployment.

Ranked Percentile Calculation

Percentile rankings are computed across all 24 personas for each trait:

  1. Collect the individual weighted rating for all personas for a given trait.
  2. Sort ascending.
  3. Assign percentile = (rank - 1) / (total - 1).

This allows identification of personas that are particularly strong or weak in a given trait relative to the team.

Data Model

The Discernment Matrix is implemented as two frozen dataclasses in src/fcc/personas/models.py:

@dataclass(frozen=True)
class RatingDimensions:
    self_rating: float | None = None
    peer_rating: float | None = None
    survey_rating: float | None = None
    individual_weighted_rating: float | None = None
    org_rating: float | None = None
    external_rating: float | None = None
    ranked_percentile_rating: float | None = None

@dataclass(frozen=True)
class DiscernmentTrait:
    name: str
    description: str
    ratings: RatingDimensions

Each PersonaSpec contains a discernment_matrix field holding a list of up to 6 DiscernmentTrait instances.

Example

A Research Crafter (RC) Discernment Matrix entry for Curiosity:

discernment_matrix:
  - name: Curiosity
    description: >-
      Driven to explore beyond immediate scope, proactively
      investigates edge cases and cross-references findings.
    ratings:
      self_rating: 0.92
      peer_rating: 0.88
      survey_rating: 0.90
      individual_weighted_rating: 0.89
      org_rating: 0.85
      external_rating: 0.87
      ranked_percentile_rating: 0.91