Simi

Methodology

How a thirty-minute decision becomes a defensible score.

Simi is not a quiz. The score on a Skills Profile is a calibrated reading of how the candidate actually behaved — the emails they sent, the briefs they wrote, the questions they didn't ask. This page is the ledger behind every number.

The score ladder

Three bands. Defined by reveals and specificity, not vocabulary.

A candidate who asks one surface question per actor and accepts every opening answer should not earn the same band as someone who triggered a reveal and named a specific fix. The ladder is encoded directly in the assessor prompt, not left to taste.

01

Fundamental

Rule

Accepts opening answers. No probing. Generic or absent recommendation.

What it signals

Engaged, but stayed surface-level.

02

Diverse

Rule

Probes at least one actor beyond the opening. Recommendation names a problem category, but no specific fix, owner, or timing.

What it signals

Investigative instinct, no commitment.

03

Thorough

Rule

Triggers at least one reveal. Recommendation names a specific fix, owner, or evidence statistic. Triangulates across two or more actors.

What it signals

Investigation that leads to a real decision.

Memo quality alone cannot compensate for zero investigative depth. Any candidate with no actor reveals is capped at Diverse in the probe-dependent competencies, regardless of how well the artifacts read.

Evidence architecture

Two categories of evidence. Different scoring rules.

The most common failure mode in skills assessment is rewarding candidates for guessing the right vocabulary. Simi splits the evidence into two explicit categories — and freezes the classification before the run.

Required

Concepts that are discoverable through two or more independent paths. The rubric may require them.

Bonus

Concepts gated behind a specific probe. The rubric credits them when they appear, but never penalizes their absence.

Once release criteria are set, the assessor prompt names every concept explicitly as required or bonus. No implicit tests.

The twenty NSF skills

One rubric, twenty NSF readings.

The rubric draws from broadly accepted employability and competency frameworks. Every Skills Profile scores all twelve, and links each score to the evidence behind it.

Judgement

01

Decision quality under pressure

Communication

02

Clarity, tone, and audience fit

Critical thinking

03

Reasoning over the evidence available

Problem solving

04

Structuring options when the path is unclear

Digital literacy

05

Working with the tools and signals at hand

Numeracy

06

Reading and explaining the numbers

Self management

07

Holding yourself together when the case bites back

Adaptability

08

Updating your plan when the situation shifts

Teamwork

09

Reading and managing the room you're in

Cultural intelligence

10

Working across different working styles

Planning

11

Sequencing what you'll do, and what you'll skip

Ethical reasoning

12

How you handle the call without a clean answer

Guarantees

Four commitments we make on every score.

  • Every score links back to the email, brief, escalation, or risk log it cites.
  • The rubric is published before the run. No hidden vocabulary tests.
  • Examiner calibration notes are shared with cohort owners.
  • Cohort medians and p50 are visible to enterprise customers.

Read the rubric, then run a real one.