Spoke

performance-calibration

Persisted pre/post calibration pairs — cycle rollups, ELT-aware leader ladders, trajectory windows.

Character

You own calibration rituals but finance insists on reproducible tuples proving distributions moved—not anecdotal facilitator notes.

Problem

External. Decks summarise “movement” qualitatively; auditors cannot reconcile leader-level deltas to captured employee-cycle facts.

Internal. Spreadsheets rot after HRIS mergers duplicate identifiers mid-cycle.

Philosophical. Calibration is operational measurement—paired ratings deserve immutable ledgers analogous to ledger postings.

Guide

PAT-159 landed durable calibration_pairs keyed by tenant, employee, calibration cycle—with shared parseRating coercion (stub evolving PAT-152) classifying completeness before rollups. Summaries deliver percent unchanged plus directional deltas; hierarchical leader vistas align to ELT / L3-L5 expectations; trajectory routes dedupe the freshest capture per cycle for longitudinal storytelling. performance-calibration primes performance-validity Q5 analytics without illicit SQL cross-talk.

Abstract

Background. Calibration consumes expensive facilitator cycles—without longitudinal pair persistence there is zero falsifiable evidence of behavioural impact.

Methodology. Ingestion paths omit incomplete parses before aggregating per-cycle KPIs; rollups summarise calibration accuracy akin to facilitator scorecards plus hierarchical overlays; trajectory APIs order historical captures respecting dedupe semantics for audit reproducibility.

Scope. Stores calibration artefacts exclusively—interpretive psychometrics reside in performance-validity; deterministic merge conflicts route through worker-resolution upstream.

Contribution. Lightweight GET summaries feed performance-validity Q5 calculators via HTTP respecting PAT cross-spoke bans on DB imports—clean contract boundary hygiene.

Evidence / Provenance. PAT-159 README references shared rating-codec stewardship across spokes.

Plan

01
Ingest pairs
POST pair writers referencing tenant identifiers + calibration cycles under PAT-11 service keys.
02
Pull cycle summaries
GET /summary?tenantId=… optionally narrowing cycleIds for targeted facilitator retrospectives.
03
Replay trajectories
Examine chronological arcs for focal employees versus managers after dedupe aligns freshest captures.
04
Wire diagnostics
Pass structured rollups downstream to performance-validity Q5 calculators for quantitative ROI overlays.

Call to Action

Direct. Populate demo tuples then rehearse summary GET curls with your sandbox tenant slug.

Transitional. Coordinate with PAT-159 session dossier for DDL + dedupe nuances.

Spoke I/O (visual language v1)

Every toolbox spoke shares the same abstract choreography: typed inputs on the left, distilled verbs in the center, typed outputs on the right, and (when relevant) cross-spoke HTTP composition along the bottom rail. Source package: @people-analytics-toolbox/spoke-illustrations.

Try it now

Copy this curl. Paste in any terminal. Public read — no auth needed.

performance-calibration.cycles.summary

GET

Needs persisted rows—swap tenant id after seeding; returns empty scaffolding when unseen.

curl -sS "https://people-analytics-toolbox.vercel.app/api/spokes/performance-calibration/summary?tenantId=demo"

Vendor the contract

The Zod contract is the source of truth. Vendor a copy into your consumer app — you keep it; we don't break it underneath you. Re-vendor when the version bumps.

// Vendor canonical types:
// src/spokes/performance-calibration/contracts/types.ts

Source path: src/spokes/performance-calibration/contracts/types.ts · GitHub

Try in Dev Console Read the explainer Back to all spokes

Failure

Calibration theatre continues—finance cannot correlate promotion dollars or merit pools to evidenced distribution shifts keyed per employee-cycle.

Success

Immutable paired tuples aggregated by leader layer land in board-ready explainable tables ahead of fiduciary reviews.